University of Southern California Dissertations and Theses

A Study Of Classical Conditioning And The Eyeblink Response In Rabbit: Issues Of Learning Vs. Performance

(USC Thesis Other)

A Study Of Classical Conditioning And The Eyeblink Response In Rabbit: Issues Of Learning Vs. Performance

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note w ill indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to tight in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. M l 48106*1346 USA 313'761-4700 800.K1-0600 A STUDY OF CLASSICAL CONDITIONING AND THE EYEBLINK RESPONSE IN RABBIT: ISSUES OF LEARNING VS. PERFORMANCE by Dragana Ivkovich A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Behavioral Neuroscience) May 1995 Copyright 1995 Dragana Ivkovich UMI N u m b e r: 9 6 0 0 9 9 4 UMI Microform 9600994 Copyright 1995, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES. CALIFORNIA 90007 This dissertation, written by ............................................ under the direction of h&K. Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School in partial fulfillment of re quirements for the degree of DOCTOR OF PHILOSOPHY c . Dean of Graduate Studies D ate...... DISSERTATION COMMITTEE Chairperson DEDICATION To my parents for devoting their lives to my sister and It for showing us the opportunities of the world and for guiding and supporting us from the heart. PUNO VAS VOLIM ACKNOWLEDGMENTS I would tike to express my deepest thanks and admiration to my advisor, Dr. Richard F. Thompson, for providing a truly supportive research environment in which an eclectic group of young scientists and visiting researchers have gathered together to study neural bases of learning and memory. I owe special thanks to Dr. David Lavond for his hands-on assistance on innumerable occasions, for the stimulating and challenging (though sometimes frustrating) exchange of ideas, and for his friendship. Many thanks also to Dr. Joseph Hcllige for all his advice, encouragement, and laughs, to Dr. Larry Swanson for his insights, excitement, and expertise, and to Dr. Albert Hcrrara for his advice, time, and support. I am deeply indebted to my colleague and friend Dr. Christine Logan for taking me under her wing and training me in the ways of the Thompson lab. To my classmates JoAnnc Tracy, Bob Clark, and Alan Nordholm: For all those early mornings, days, and late nights of partnership and commradcry, I toast you! My gratitude goes out to Dr. David Krupa, Dr. Rodney Swain, Martha Berg, Jeff Grethe, Judith Thompson, AJ Annala and Aunc Moro on whom I could always rely for advice and a helping hand. Thanks to Jon Lockard, Neil Chatteijce, Ulin Sargeant, and Keona Jones for assistance with some of the work presented here. I wish also to relay my genuine appreciation to my fellow lab mates, past and present, for creating a stimulating research environment at USC. Last, but not least, to my best friends and fellow graduate students Elizabeth Cowin and Jovanka Nikolic my love and best wishes. Throughout, you've been my companions, my sounding boards, my fan club, and, at times, my sanity. iii TABLE OF CONTENTS DEDICATION.............................................................................................................. ii ACKNOWLEDGMENTS...........................................................................................iii TABLE OF CONTENTS............................................................................................ iv LIST OF TABLES.................................................................................................... vii LIST OF FIGURES................................................................................................... viii ABSTRACT................................................................................................................. xii CHAPTER 1: General Introduction............................................................................1 1.1 The Emergence of Classical Conditioning as a Method of Studying Learning and Memory.............................................. .................................... 1 1.2 The Cerebellar Circuit........................................................................................3 1.2.1 The Unconditioned Stimulus and Response Pathways 1.2.2 The Conditioned Stimulus and Response Pathway 1.2.3 Convergence and Association 1.3 Issues of Learning vs. Performance.................................................................11 1.4 Some Modulatory Influences on Eyeblink Behavior...................................... 13 1.4.1 Motor Cortex 1.4.2 Habituation 1.4.3 Behavior-Evoking Stimulation 1.5 Summary of Experiments..................................................................... 18 CHAPTER 2: Studies Addressing the Learning vs. Performance Debate............ 20 2 .1 An Input/Output Study of Performance Measures Over the Course of Classical Conditioning...................................................................................... 20 2.1.1 Methods 2.1.2 Results 2.1.3 Discussion 2.2 Training with Low Intensity Unconditioned Stimuli Does Not Unmask a Performance Deficit Following Interpositus Lesions .......................... 42 2.2.1 Methods 2.2.2 Results 2.2.3 Discussion 2.3 Do Lesions of Eyeblink Motomeurons Equally Impair the CR and the UR?...................................................................................................................... 50 2.3.1 Methods 2.3.2 Results 2.3.3 Discussion CHAPTER 3: Does Motor Cortex Modulate the Cerebellar Circuit for Classical Conditioning?.....................................................................................................59 3.1 Motor Cortex Lesions Do Not Affect Acquisition or Retention of the Delay Conditioned Eyeblink in Rabbits............................................................ 59 3.1.1 Methods 3.1.2 Results 3.1.3 Discussion 3.2 Effects of Motor Cortex Lesions on Acquisition of Trace Eyeblink Conditioning........................................................................................................81 3.2.1 Methods 3.2.2 Resutts 3.2.3 Discussion v CHAPTER 4: Factors Influencing Unconditioned Reflex Performance in Rabbit............................................................................................................ ....98 4.1 A Systematic Study of Habituation to Peripheral Airpuff Presented at Different Inter-Stimulus Intervals...................................................................99 4.1.1 Methods 4.1.2 Results 4.1.3 Discussion 4.2 Parametric Study of Direct Stimulation of the Cerebellum and Elicited Behaviors ...................................... 123 4.2.1 Methods 4.2.2 Results 4.2.3 Discussion 4.3 Interference in the Classically Conditioned Eyeblink Response in Rabbits by Direct Stimulation of the Cerebellum during the CS-US Interval.............. 142 4.3.1 Methods 4.3.2 Results and Discussion 4.3.3 Summary CHAPTERS: Conclusion.........................................................................................162 LIST OF TABLES 2.1 Summary of training protocol for paired group.................................................. 23 2.2 Summary o f training for accessory abducens group...........................................52 vii LIST OF FIGURES 1.1 Cerebellar Circuit for Eyeblink Conditioning....................................... 5 1.2 Simple Nictitating Membrane Reflex Circuit.......................................................7 1.3 Motor Cortical Connections to Cerebellar Eyeblink Circuit............................. 13 2.1 Paired Group Conditioning:....................................... 30 (A) % Conditioned Responses (B) Conditioned Response Amplitude 2.2 Paired Group I/O Tests:..................................................................................... 32 (A) UR Amplitude (B) UR Area (C) % URs (D) UR Risetime (E) UR Peak Latency 2.3 Unpaired Group:....................... 36 (A) % Conditioned Responses (B) Conditioned Response Amplitude (C) I/O UR Amplitudes 2.4 Intcrpositus Lesion Reconstruction.................................................................... 38 2.5 Low US Training - Response Amplitudes.........................................................47 2.6 Interpositus Lesion Reconstruction................... 48 2.7 % Pre-Op Amplitude ............ 55 2.8 Accessory Abducens Lesion Reconstruction.......................... 57 3.1 Motor Cortex Surgical Opening........................................................................ 64 3.2 Maps of Cortical Surface for Rabbit and Rat........................................ 65 3.3 Percent Conditioned Responses - Delay Conditioning ............................67 3.4 (A) Response Amplitudes * Acquisition Group............................................ 68 (B) Response Amplitudes - Retention Group 3.5 (A) Onset Latencies - CS-AIone Trials............................................................70 (B) Onset Latencies - Paired Tone-Airpuff Trials 3.6 (A) I/O Onset Latencies * Acquisition Group.................................................72 (B) I/O Onset Latencies - Retention Group 3.7 (A) Surface Lesion Reconstruction....................................................................73 (B) Sample Photos of Motor Cortex Lesion 3.8 (A) Coronal Lesion Reconstruction - Acquisition Group............................... 76 (B) Coronal Lesion Reconstruction - Retention Group................................... 77 (C) Normal Template with Labels....................................................................78 3.9 Percent Conditioned Responses - Trace Conditioning....................................84 (A) CS-AIone Trials (B) Paired Trials 3.10 (A) CR Amplitudes - CS-Alonc T rials............................................................86 (B) CR Amplitudes - Paired Click-Airpuff Trials 3.11 (A) I/O Response Amplitudes - Lesion Group............................................... 87 (B) Paired Training UR amplitudes - Both Groups 3.12 (A) Onset Latencies - CS-AIone Trials............................................................ 88 (B) Onset Latencies • Paired Trials 3.13 I/O Onset Latencies................................................................ 89 3.14 Coronal Lesion Reconstruction...........................................................................92 3 .15 Anatomical Links from Hippocampus to the Cerebellar Circuit.....................95 3.16 Individual Lesion Reconstructions of Retrosplenial Area...................... 97 4 .1 Absolute Response Amplitudes........................................................................... 105 4.2 Absolute Response Amplitudes: by group........................................................106 4.3 Scaled Response Amplitudes.............................................................................108 4.4 Scaled Response Amplitudes: group x segment........................................ ....109 4.5 Scaled Amplitude by group ................................................................110 4.6 Peak Latency: group x block............................................................................112 4.7 Measure of Response Area........................................................ 113 4.8 UR Area: psi x block..........................................................................................115 4.9 UR Area: order x psi x day................................................................................ 116 ' 4.10 UR Area by group and order............................................................................117 4.11 Map of Cerebellar Electrode Placements and Stimulation-Evoked Behavior 126 4.12 Varying Stimulation Duration: Percent Amplitude Change......................... 128 4.13 Sample Data: Animal #94-482........................................................................130 4.14 Sample Traces: Animal #94-450 and # 9 3 -3 9 9 ......................... 131 4.15 Varying Stimulation Intensity: Percent Amplitude Change and Latencies.. 132 4.16 Sample Traces for Increasing Stimulation Intensity: #93-399..................... 134 4.17 Varying Stimulation Frequency: Percent Amplitude Change........................135 4.18 Varying Stimulation Frequency: Peak and Onset Latency...........................136 4.19 Sample Traces : Animal #94-448....................................................................137 4.20 Varying Pulse Duration :Percent Amplitude Change and Latencies........... 138 ■ • 4.21 Sample Traces for Range of Pulse Widths: #94-450.................. 139 4.22 Cerebellar Stimulation During Conditioning: Protocol Diagram.................145 x 4.23 Map of Cerebellar Electrode Placements....................................................... 147 4.24 Percent Conditioned Responses................................................ 148 4.25 Sample Traces (Group 1) Without and With Stimulation.............................. ISO 4.26 Sample Traces (Animal #94-193) Varying Stimulation Duration................... 151 4.27 Sample Traces (Group 1) Varying Stimulation Duration...............................152 4.28 Sample Traces (Group 1) Separating Tone and Stimulation..................153 4.29 Sample Traces (Animal #94-193) Varying Stimulation................................155 4.30 Sample Traces (Group 2) Varying Stimulation Duration........................156 4.31 Sample Traces (Group 2) Simultaneous Tone and Stimulation Onset...158 xi ABSTRACT A Study of Classical Conditioning and the Eyeblink Response in Rabbit: Issues of Learning vs. Performance by Dragana Ivkovich Classical conditioning of the rabbit eyeblink response has become a model preparation for the study of basic associative learning and memory. Research aimed at identifying the engram for skilled motor movements must separately manipulate the learned response without affecting the animal's ability to perform the natural reflexive response. The following scries of experiments addresses issues of learning vs. performance. Lesions of the intcrpositus nucleus of the cerebellum selectively abolish the conditioned eyeblink response. Reflex responsiveness testing at 4 different unconditioned stimulus intensities over the course of training indicate that interpositus lesions have no effect on performance under normal training conditions or with training utilizing lower intensity unconditioned stimuli. Lesions of the accessory abducens motor nucleus severely impaired alt responses initially, but with continued training there was faster recovery of the conditioned than the unconditioned response. Though generally accepted as a contributor to motor behavior, motor cortex is not necessary for optimal performance of reflexive or delay conditioned xii responses. Lesioned animals do, however, appear to exhibit slightly poorer performance overall, though not statistically significant, with trace conditioning. Focusing on the unconditioned eyeblink reflex, habituation of the rabbit eyeblink is consistent with the properties of habituation outlined by Thompson and Spencer (1966) including: exponential curves, spontaneous recovery, better habituation at higher stimulation rales, and stronger habituation to lower intensity stimuli. Direct stimulation of interpositus nucleus which evokes eyeblink behavior is influenced by similar manipulations. Frequency and pulse duration determine whether a response is initiated. Stimulus intensity determines the size of the movement. Peak latency consistently follows stimulation offset, while onset latency is unchanged by different stimulation durations. However, onset latency is reduced with longer pulse durations. When classically conditioned animals receive stimulation of cerebellar dentate nucleus, conditioned response behavior is significantly delayed until the offset of stimulation. If the interpositus nucleus is stimulated, the response is potentiated. Overall, the results of these experiments provide an extensive characterization of the eyeblink response in rabbits under both associative and non- associative conditions. The findings reconfirm that learning and performance deficits can be distinguished and show that the eyeblink reflex changes with experience and stimulus parameters. CHAPTER 1 General Introduction 1.1 The Emergence of Classical Conditioning as a Method of Studying Learning and Memory The importance of learning and memory in our society can be traced back to the early days of lyric poetry, before the printed word, when knowledge and experience were transferred to younger generations though oration. Philosophers studied the techniques of the greatest orators and memory as an art form was cultivated and perfected (Boorstin, 1983). The study of memory may be said to have begun as early as a few hundred years B.C. (c. 300 B.C.) when mnemonic strategies were being developed and taught by the Greek lyric poet Simonides of Ceos. Cicero praised Simonides as the inventor of the mnemonic device by which menial images of items to be remembered were created and stored in select locations. These Associated" locations served to organize the memories and aid later recall (Boorstin, 1983). Work on the skill or “art" of memory flourished over the centuries and helped shed light on conscious learning processes. While such conscious techniques of acquiring and retrieving information are quite useful, powerful associations can be formed when one is unaware. In a review of memory literature, Hermann and Chaffin (1988) quote the Greek philosopher Heraclitus as saying “...some men are as ignorant of what they do when awake as they are forgetful of what they do when they are asleep.” What conscious and unconscious memory formation have in common is the importance of association processes. As early as the fourth century B.C. Aristotle suggested that learning could be accounted for by frequency of associations. Associationism persisted for centuries and in the 1700’s, Hume suggested three basic principles governing the association of ideas: resemblance, contiguity in time or place, and cause or effect (Hermann and Chaffin, 1988; Klein, 1991). Both of these men were ahead of the times for it wasn’t until the 20th century that Pavlov was able to show these principles to be part of learning by classical conditioning. While philosophers contemplated the abstract nature of human behaviors and phenomena such as learning and memory, a parallel path was being forged by early anatomists who demonstrated that the brain could be a physical substrate of behavior (Dudai, 1989). In the 1600’s Descartes defined the simple reflex and suggested the basis for all action to be a mechanical response to sensory stimulation. This view of behavior shifted the notion of central control to that of external controlling factors (Jenkins, 1979) and lead to a new era of psychological rescarch-'bchaviorism. « The purely mechanical theory of behavior, without consideration for experience and change, was not adequate, however. Clearly, living organisms are in constant contact with their environment and continually receive information about that environment and their own relation to it. Survival is facilitated by the ability to amass this information and use experience to adapt behavior, whether consciously or unconsciously. It was in the late 1800's that Sechenov first suggested that the ideas of associationism might be combined with the mechanical view of behavior. He proposed that psychological reactions might be the result of associations modifying reflexive behavior (Gleitman, 1983). 2 It was Ivan Pavlov, in the early 1900's, who elaborated on Sechenov's theory and determined that definite rules governed which behaviors were learned (Klein, 1991). He demonstrated that learning occurred by repeated exposure to stimuli having contingent relationships where one of the stimuli evokes a reflexive response. This form of learning came to be called Pavlovian or classical conditioning. Further, he proposed that for an association to occur, the conditioned stimulus (CS) and unconditioned stimulus (US) centers must somehow be connected (Jenkins, 1979). At the same time Karl Lashley promoted the lesion technique as a means of localizing certain behavioral functions to specific brain regions. In particular, he wanted to find a brain region where learning processes could take place, associations would be formed, and possibly even stored. This place in the brain where memories might be located has been called the engram. So, rather than a connection between two separate centers, there may be a third center where CS and US information converge and a memory trace is established. As Pavlov suggested, in order to address the question of mechanisms underlying association during classical conditioning and to search for a location in the nervous system where an engram might be located, the basic neural circuits, by which sensory information from the CS and US reaches motor outputs producing the behavioral responses, had to be identified. 1.2 The Cerebellar Circuit Classical conditioning of the rabbit nictitating membrane (NM) response has become a classic model preparation for the study of learning and memory (See 3 Gormezano et ah, 1983 and Thompson, 1990 for reviews of the field). During conditioning, a neutral stimulus (tone) becomes associated with another stimulus (airpuff) which unconditionally or involuntarily elicits a behavioral response (extension of the NM). As a result, the previously neutral stimulus, now a conditioned stimulus (CS), will evoke a behavioral response identical to, or closely resembling, the unconditioned response (UR). This is the conditioned or learned response (CR). The general anatomy of the neural circuitry underlying the classical conditioning of the rabbit eyeblink reflex has been identified at the simplest level to involve structures of the brainstem and cerebellum (see Thompson, 1989, and Lavond et ah, 1993, for reviews). The ultimate goal of this tracing work has been to provide a foundation for the investigation of putative loci of the memory trace by studying sites where US and CS information converge and an association may occur. Various manipulations have been made which affect acquisition or retention of the learned response (CR), independent of the unconditioned reflex response, and have enabled detailed characterization of the neural substrates of teaming and memory for skilled motor movements. (Refer to Figure 1.1 for a diagram of the critical neural circuitry described below.) 1.2.1 The Unconditioned Stimulus and Response Pathways. An airpuff unconditioned stimulus directed at the cornea also has a diffuse effect on the external eyelids. Somatosensory fibers carry information via both the ophthalmic and maxillary branches of the trigeminal nerve (V) to the sensory trigeminal nuclei. 4 I f. Parallel fiber Climbing fiber Mossy fiber Cerebellum Behavior: UR, CR ^ ( c o rn e a ) CR ^ R e d N . Reflex path US ^ N. VIA V II N. V (Sp) Purklnje cell Interpositus N. Pontine N. I.O. (DAO) (to n e ) V. Coch. N. Figure 1.1 Diagram of the cerebellar circuit underlying classical conditioning of the eyeblink response in rabbit. Abbreviations: DAO-dorsal accessory olive, 1 0= inferior olive, N V(sp)= spinal Vth cranial nucleus (trigeminal), N VI= sixth cranial nucleus (abducens), N VII = seventh cranial nucleus (facial), V Coch N = ventral cochlear nucleus, (from Thompson, 1986) 5 The majority of these projections terminate on the rostral portion of the pars oralis of the spinal Vth nucleus and the caudal portion of the principal Vth sensory nucleus. The ophthalmic fibers are distributed ventrally within these nuclei and the maxillary fibers terminate just dorsal to the ophthalmic projections. HRP injections from the cornea also indicate that some sensory terminal fields extend to areas o f reticular formation, facial nucleus, and solitary nucleus (Cegavske et al., 1987). Secondary projections from the trigeminal nuclei follow two major routes, both having a common output pathway to motor nuclei which produce the behavioral response. The first route is that of the simple reflex arc involving just the ipsilateral brainstem motor nuclei and reticular formation. The major motor nuclei involved in producing the simple reflexive eyeblink UR and which receive trigeminal projections are the accessory abduccns (aVI), abducens (VI), oculomotor (III), and facial (VII) nuclei (Cegavske et al., 1987). These nuclei send projections to various muscles controlling eyeball movement and eyelid closure. Movement of the external eyelids is controlled by the obicularis oculi muscle which is innervated by fibers from the motor region of the facial (VHth) nucleus (Cegavske et al., 1976; Inagaki et al., 1989). Extension of the nictitating membrane is itself a passive movement occurring when the eyeball is retracted into the orbit. This causes displacement of Harder's gland which, in turn, pushes the cartilage of the NM across the cornea (Berthier et al., 1987; Cegavske et al., 1987; Prince, 1964). The retractor bulbus muscle is responsible for the actual eyeball retraction and is innervated by the Vlth cranial nerve (abducens; see Figure 1.2). 6 Figure 1.2 Diagram of the simple reflex circuit for the rabbit nictitating membrane response. Sensory information from the cornea (C) is relayed through the Vth cranial nerve and nucleus (trigeminal) to theaccessory Vlth cranial nucleus (abducens) which, in turn, innervates the retractor bulbus (RB) muscle controlling eyeball refaction. Retraction of the eyeball displaces Harder's £land (HG) and pushes the nictitating membrane (NM) across the eye. Sensory information is also projected contralaterally. Conditioned response (CR) information from the cerebellum is relayed to the motor nuclei to influence behavior, (figure from Berthier et al., 1987) 7 Several reports agree that the primary source of motor neurons to the retractor bulbus (RB) is actually the accessory abducens nucleus (Cegavske et al., 1987; Disterhoft et al., 1985; Marek, 1984). All motor neurons in the accessory abducens are labeled following HRP injections to the RB muscle. However, there is some evidence of labeling in parts of the abducens and oculomotor nuclei. Stimulation of the abducens and accessory abducens nuclei results in eyeball retraction and NM extension whereas stimulation of the oculomotor nucleus results in eyeball rotation and NM retraction (Cegavske et a)., 1976). There is evidence that the cxtraocular eye muscles controlled by the oculomotor nerve are a secondary mechanism for the defensive eyeball retraction when necessary (for instance, following lesions of the accessory abducens nucleus or cuts of the Vlth nerve; Berthier & Moore, 1980; Disterhoft et al., 1985). In addition, crescent shaped cells in the dorsal part of the oculomotor nucleus have specifically been shown to innervate the levator palpebrae superioris (LPS) muscle which is responsible for actively retracting the NM into the nasal canlhus of the eye following passive extension (Cegavske et al., 1987). In contrast, the second route for US information passing through the trigeminal is that of the US training or reinforcement pathway through the cerebellum which ultimately reaches the motor nuclei. Sensory information proceeds from the trigeminal to the contralateral dorsal accessory olive (DAO) which, in turn, sends decussating climbing Tiber projections to the Purkinje cells of cerebellar cortex and neurons of the interpositus nucleus (IP) of the cerebellum. The IP sends excitatory projections, via the superior cerebellar peduncle (SCP), to the contralateral red nucleus (Weiss et al., 1985) from which fibers extend through the ventral tegmental bundle and rubrobulbar tract to act on the motor nuclei (e.g., 8 Thompson et al., 1986; Schreurs, 1988). Stimulation of the DAO can produce an eyeblink and serve as an effective US leading to successful conditioning (Mauk et al., 1986). There is also evidence of IP inhibitory feedback projections controlling activity in the IO (Krupa et al., 1991; Sears & Steinmetz, 1991). 1.2.2 The Conditioned Stimulus and Response Pathway. In the case of an auditory CS, information about a tone, for example, is relayed to the ventral cochlear nucleus, crosses the midline, and continues to the pons (see Thompson et al., 1986 for review). Mossy fibers from the lateral pons then project to back across the midline to cerebellar cortex and the interpositus nucleus of the cerebellum via the middle cerebellar peduncle (Steinmetz & Sengelaub, 1988). Evidence for this comes from studies in which presentation of a tone produces increases in evoked activity of Purkinje cells in areas HVI, Crus I and Crus II of cerebellar cortex (Donegan et al., 1985; Foy & Thompson, 1986). In addition, auditory responsive cells have been found in IP (Tracy et al., 1991). Functionally, stimulation of the lateral pons (Steinmetz et al., 1985; 1990) is an effective substitute for an auditory CS during conditioning. At this point, the efferent pathways, from the interpositus, of both the US and CS which produce the behavioral responses, the UR and CR, are presumed to be the same. Fibers exit the interpositus through the superior cerebellar peduncle on the side of training, cross to the contralateral red nucleus, and descend the rubral pathway crossing again to act on ipsilateral motor neurons (see previous section for details). 1.2.2 Conwrgence and Association. Electrical brain stimulation has been used to reduce the classical conditioning preparation and to begin approximating the most 9 essential elements of the neural circuitry. Stimulation of mossy fiber pontine projections are an effective CS when stimulation of climbing fiber inferior olive projections serve as the US (Steinmetz et al., 1989). More recently, stimulation of parallel fibers, which relay mossy fiber signals to Purkinje cells in cortex, has been paired with stimulation of the underlying white matter just above the interpositus nucleus (Shinkman et al., 1992) resulting in successful conditioning. Afferent CS and US information converge in cerebellar cortex and deep nuclei and follow a common efferent motor pathway. Lesions of as little as 1 cubic mm of tissue at the interpositus/ cerebellar cortical junction completely abolish conditioned responding and the ability to relearn (Lavond et al., 1984) without affecting unconditioned response performance. Meanwhile, lesions of the red nucleus also abolish CRs without affecting the UR. Newer reversible inactivation techniques (Clark & Lavond, 1993; Krupa, et al., 1993) have demonstrated that when the red nucleus is inactivated during classical conditioning there is no performance of conditioned responses, but removal of that inactivation results in immediate asymptotic conditioned responding. In contrast, inactivation of motor nuclei by infusion of muscimol (Thompson et at., 1993) or cooling (Zhang and Lavond, 1991) prevented expression of both URs and CRs. Once the inactivation was reversed these animals, too, showed immediate asymptotic levels of conditioned responses indicating that learning had occurred and the motor nuclei do not contribute to memory trace formation. Afferent to the interpositus, inactivation of the dorsal anterior interpositus and overlying white matter by infusions of either muscimol (Krupa et al., 1993) or lidocaine (Nordholm et al., 1993), as well as with cooling (Clark and Lavond, 1993) consistently resulted in prevention of learning. As in to the previously mentioned inactivation studies, these animals did not show any conditioned responses during inactivation. Unconditioned responses were unaffected. Upon reactivation, animals did not show immediate conditioned responding, but rather learned acquired the response at the same rate as normal control animals. Together these results have yielded the current hypothesis that memory trace formation involves the interpositus and occurs before the level of the red nucleus. 1.3 Issues of Learning vs. Performance The usefulness of classical conditioning as a tool in the search for a memory trace is due to its unique inherent element of scientific control which allows assessment of the unconditioned reflex performance separately from conditioned response performance both at behavioral and now also neuroanatomical levels. This enables us to study critical brain regions by administering drugs, applying stimulation, recording and lesion techniques, or manipulating training conditions and to distinguish effects on the learning process and subsequent memory storage from motivational or physical constraints on response production. For example, there are two important criteria to consider a neural structure as the site where associations take place which can be nicely demonstrated with classical conditioning. First, electrophysiological recordings must show a change in the pattern of cell firing uniquely associated with the learned response. Second, lesioning this structure should prevent acquisition of a conditioned response and result in the complete abolition of a previously learned response and its re acquisition with continued training. The important control here is that the animal it must still be able to produce a reflexive response to an unconditioned stimulus, showing the distinction between a learning and a performance deficit. As outlined in the previous section, extensive research using classical conditioning of the eyeblink response in rabbit has been successful in identifying just such a region where learning can be affected independently of performance, the interpositus nucleus of the cerebellum. Recently, this result has been challenged by Welsh and Harvey (1989) who claim that lesions of the interpositus do affect performance and, therefore, the cerebellum could not be involved in learning and memory processes associated with the learned response, but only with its production. Welsh and Harvey (1989) present two points of contention. The first point is that interpositus lesions do not completely abolish conditioned responses, but rather leave behind a residual low amplitude, long latency, less frequent conditioned response. This result had already been reported by Clark et al. (1984) when destruction of the critical area of the interpositus was incomplete. In addition, Steinmctz et al. (1992) reanalyzed data from animals with effective lesions using less stringent criteria for identifying CRs. The percentage of learned responses was still well below spontaneous response rates. The second point Welsh and Harvey make is that interpositus lesions do have a small effect on unconditioned response performance at low US intensities, and that the stimulus intensity typically used for training is severe enough to mask this deficit. The three experiments in Chapter 1 of this dissertation provide an extensive study of UR topography during paired and unpaired training, before and after lesions of interpositus, at a variety of stimulus intensities, and in relation to CR following a true performance deficit. 12 1,4 Some Modulatory Influences on Eyeblink Behavior One interesting result from the study of UR topography over the course of training was the finding that there was a significant enhancement of UR amplitude following both associative and non-associative training conditions. The eyeblink UR, which has been taken for granted to be a rather stable and replicable motor behavior, was clearly being modulated. Several factors may be involved in modulating eyeblink behavior. This section serves to introduce the few which are to be discussed further in this paper. More specific information as related to the individual experiments will appear immediately preceding each experiment. 1.4.1 Motor Cortex: One major brain area implicated in the modulation of conditioning of discrete motor movements is the motor cortex (Houk, 1989; Woody, 1984). Though decorticate animals are able to acquire the conditioned NM response (Norman et al., 1977; Oakley and Russell, 1977), it is suggested that the motor cortex may have some modulatory role in the behavioral responses. Bures and Bracha (1990) suggest that motor cortex may be involved in Elaboration of discrete voluntary movements, the acquisition of new motor skills, and in the instrumentation of already existing unconditioned reflexes." The neuroanatomical connections between cerebral motor cortex and the cerebellar eyeblink circuit are numerous, though the primary cerebellar influence on neocortex is via red nuclear projections through the ventrolateral nucleus of the 1 3 thalamus (VL). Four motor cortical efferent structures which have the potential to modulate conditioned eyeblink behavior are schematically represented in Figure 1.3. First, motor cortex projects heavily to the red nucleus (Allen and Tsukahara, 1974; Brodal, 1981; Ito, 1984) which may either effect behavior directly via the motor nuclei or indirectly by inhibitory feedback to the trigeminal (Davis and Dostrovsky, 1986) and inferior olive (Weiss et al., 1985), components of the US pathway. The red nucleus has been implicated in adaptive timing of the conditioned response (Krupa et al., 1994), Second, motor cortex projects to pontine nuclei (Allen and Tsukahara, 1974; Wicscndangcr, 1986) which carry CS information to cerebellar cortex and deep nuclei. Third, motor cortex projects to the lateral reticular nucleus (LRN) which also receives projections from red nucleus (Brodal, 1981; Ito, 1984; Wiescndangcr, 1990). The LRN itself projects back heavily to deep nuclei of the cerebellum as well as to cerebellar cortex. Fourth, direct projections from motor cortex to the trigeminal nucleus have been identified in primates and may also exist in lower mammals (Wiesendanger, 1990). Based on this intercommunication, it seemed appropriate to study the effects of motor cortical lesions on classical conditioning with a closer emphasis on response topography. The studies in Chapter 3 address the role of motor cortex in rabbit eyeblink conditioning using two different behavioral paradigms, delay and trace conditioning. 14 C E R EB E L LA R C O R T E X AIRPUFF M otor N u d o i: III. VI. »VI. VII EYEBLINK Figure 1.3 Diagram of motor cortical efferents which provide links to the cerebellar learning circuit and have potential to modify behavioral output. Structures of the cerebellar circuit (see also Figure 1.1) include the trigeminal (V) and ventral cochlear (VCN) sensory nuclei, inferior olive (IO), pontine nuclei (PN), cerebellar cortex, including Purkinje cells (shaded oval) and parallel fibers (curved line receiving PN projection, interpositus (IP) and the red nucleus (RN). Motor cortex forms a cerebral loop with basal ganglia (BG ) and the ventro-lateral thalamus (VL), but also projects down to RN, PN, lateral reticular nucleus (LRN), and V nucleus. All paths ultimately loop through either RN or V to influence motor nuclei III (oculomotor), VI (abducens), aVI (accessory abducens), and VII (facial). Filled circles indicate known inhibitory connections. Numbers apply to text description. 1.4.2 Habituation: i t In addition to learning related changes which lead to the development and maintenance of conditioned responding, there are also changes occurring within the unconditioned stimulus reflex and reinforcement pathways which may or may not be associative in nature (Ison & Leonard, 1971; Steinmetz et al., 1992; Thompson & Spencer, 1966; Weisz & Mclnemey, 1990). Particular to the studies in this paper is a general enhancement of unconditioned response amplitude observed over the course of experience with both associative and non-associative stimulus presentation. With the wide use of the rabbit eyeblink as a model for classical conditioning in the study of learning and memory in recent years, the necessity for a more extensive characterization of the reflexive or unconditioned eyeblink and the factors influencing it has increased. The most extensively characterized behavioral change to repeated stimulus presentations is habituation. Habituation is the process by which a stimulus- evoked (unconditioned) response decreases in strength with repeated presentations of that stimulus. Thompson and Spencer (1966) outline 9 basic points or “rules" of habituation among which are that higher frequency stimulus presentations support greater degrees and more rapid rates of habituation. Mild habituation of an unconditioned eyeblink to an airpuff stimulus is sometimes seen within the first few trials of a behavioral training session at which time responses tend to level off and stabilize. Recently, however, extreme degrees of habituation producing response decrements of over 90% within a single training session (108 trials) have been observed in a preliminary study with just a few animals receiving trigeminal stimulation as a US (Berg, Lockard, and Thompson, personal communication). 1 6 Such extreme habituation is not usually obseived during normal conditioning sessions. This observation has spurred an interest and uncovered the need for a study in which peripheral US (e.g., airpuff) stimulation parameters, such as inter- stimulus interval and intensity, are systematically manipulated to study potential habituating effects and to characterize the changes in unconditioned responding. This would allow future comparisons of how habituating to a response-eliciting stimulus under particular conditioning parameters might influence conditioned response development to a neutral stimulus. 1.4.3 Behavior- Evoking Stimulation: As an extension to the parametric analysis of the effects of external airpuff stimulation on the eyeblink response, the final experiments presented here involve a more direct path of evoking movement— stimulation of the interpositus nucleus of the cerebellum. The interpositus nucleus of the cerebellum is the direct output path for projections from cerebellar cortex. The interpositus projects to motor nuclei via the red nucleus and when stimulated will elicit movement. Though direct interpositus stimulation is ineffective as a US and can interfere with the expression of a previously learned response, there is significant transfer from tone-stimutation training to tone-airpuff training (Chapman et al, 1988; Thompson and Krupa, 1994), An evaluation of stimulation parameters which effectively evoke movement in rabbits and the behavior modulation resulting from manipulation of these parameters is the subject of the latter portion of Chapter 4, 17 1.5 Summary of Experiments The first two experiments in Chapter 2 provide evidence that lesions of the interpositus nucleus (IP) of the cerebellum selectively abolish the conditioned eyeblink response while the ability to perform a response remains intact. Input/ output performance measures taken over the course of normal acquisition* IP lesion* retention training, as well as measures taking during similar training using lower intensity unconditioned stimuli, indicate no deficit in performance associated with loss of conditioned responding. Interestingly, the third experiment in Chapter 2 shows that when a performance deficit is produced by lesioning the motor output ( accessory abducens nucleus), conditioned response recovery is faster than that of the unconditioned response. Classical conditioning produces a robust learned response which is not easily susceptible to performance deficits. While the severest performance deficits are produced by damaging motor nuclei, performance may also be modulated by other higher brain structures, especially motor cortex. Chapter 3 describes the results that large bilateral lesions of frontal cortex have on different types of conditioning. Neither acquisition nor retention of delay classical conditioning was affected by lesions of motor cortical areas. Acquisition of trace conditioning was mostly unaffected by similar lesions. Lesioned animals were, however, significantly delayed (longer onset latency) in responding on this more complex task. Though performance deficits were rarely observed in the previous experiments, the unconditioned response did consistently exhibit amplitude enhancement following repeated stimulus exposure, whether associative or non*associative. Chapter 4 provides an extensive study of habituation of the reflexive eyeblink to peripheral 1 8 airpuff stimulation. The results are discussed in terms of the 9 properties of habituation described by Thompson and Spencer (1966). Data are also presented for behaviors evoked by varying parameters of intra-cranial stimulation of the IP. In the final experiment, stimulation of the IP region during the delay interval of conditioning is shown to affect response execution. Together these data present an extensive study of the eyeblink response under both associative and non-associative conditions. They suggest that performance and learning deficits can be distinguished and that the conditioned eyeblink is a robust phenomenon. The data also indicate that higher cortical modulation of classical conditioning is minimal and that it is the interpositus nucleus of the cerebellum that is essential for producing the learned response. 1 9 CHAPTER 2 Studies Addressing the Learning vs. Performance Debate Presented here are three experiments that provide additional strong evidence for a critical role of the cerebellum in the learning and memory of the Pavlovian conditioned response. These experiments include: (a) failure to find systematic or persisting decrements in unconditioned response amplitude (i.e., the eye blink reflex) after appropriate interpositus lesion, (b) failure to unmask a performance deficit following effective interpositus lesion in animals trained with lower than normal unconditioned stimulus intensities, and observations of differential effects on conditioned and unconditioned response recovery after lesions in the region of the accessory abducen motor nucleus responsible for executing the eyeblink response. These data are discussed in the context of performance versus learning issues; evidence presented here rules out the possibility that interpositus lesion abolition of the eyeblink CR is simply due to lesion effects on performance. 2.1 An Input/ Output Study of Performance Measures Over the Course of Classical Conditioning Welsh and Harvey (1989) make much of their reported finding that their “effective" interpositus lesion causes persisting changes in the UR. The actual change they report is extremely smalt, involving only modest changes in the percent, peak latency and rise-time of the UR to very low intensities of US. They report no lesion effect on reflex amplitude or onset latency, even at the lowest US 2 0 intensities. Furthermore, they did not determine the effect of the lesion on performance of the UR in the same animals, i.e. they did not determine the effect of the lesion on the UR compared to the same UR prior to lesion. This experiment evaluates the effect of interpositus lesions that completely abolish the eyeblink CR on properties of the UR in the same group of animals, comparing pre- and post lesion URs. A control group of animals given the same amount of experience with stimuli but with random unpaired presentations of the stimuli is used for comparison. 2.1.1 Methods: Subjects. The subjects for this experiment were 19 adult male New Zealand White rabbits (Oryaotagus cunicutus). All procedures and animal care were in accordance with NIH guidelines. Rabbits were housed individually in a room with a regulated 12 hr light/ 12 hr dark cycle. All animals were maintained on an ad lib food and water diet. Animal care was provided by the experimenter, staff and veterinarians of the University of Southern California. Protocol. The amplitude of each rabbit's reflexive reaction to different air puff intensities was measured at various points during the course of behavioral training. Each input/output (I/O) testing measured the amplitude, latency, amplitude-time area and rise-time of the reflexive blinks. There were two groups of animals. Eight animals participated in classical conditioning using a standard delay conditioning paradigm pairing tone CS and airpuff US (Paired Group). Another 11 animals received random unpaired presentations of tone and airpuff to determine if any changes that might occur in I/O response topography as a result of training are due to non-associative processes (Unpaired Group). Refer to Table i i. 2.1 for a summary of the training protocol described below. On the first day, the left eye of each rabbit was anesthetized with ophthalmic lidocaine and a loop of 6*0 nylon suture was placed into the temporal margin of the nictitating membrane. Later, a minitorque potentiometer would be attached to this loop to record nictitating membrane (NM) movement. The following day each rabbit was tested for reaction to air puff. I/O test 1 was used to assess reactions to air puff intensities with each rabbit having as little experience with the experimental condition as possible (I/O tests are described below). For this and all testing and training to follow, animals were restrained, placed in a sound attenuating chamber, and their external eyelids were retracted to prevent interference with NM response measurements. A headpiece holding the potentiometer and airpuff nozzle was placed on the animal's head. A piece of suture thread extended from the arm of the potentiometer, and the needle at its end was modified into a light hook which was placed through the suture loop in the NM. Movements of the membrane cause rotation of the arm and were transduced into signals for data collection. A 1 mm displacement was calibrated to represent 10 units of conversion on an 8-bit a/d converter. Both stimulus presentation and data collection were achieved on-line using interface (Lavond & Steinmetz, 1989) and software for the IBM-PC/XT, Bach rabbit was then adapted for three days to the conditioning apparatus. Adaptation consisted of placing each rabbit in the restrainer and animal chamber for a period of 1 hour during which all conditions matched those of subsequent training sessions with the exception that no stimuli were presented and no data 2 2 Table 2.1 Summary of Training Protocol Day Paired Croup Unpaired Croup 0 Nictitating membrane suture 1 I/O Test 1 2-4 Adaptation (3 days) 5 I/O Test 2 6-10 Paired Training (5days) Unpaired Training (S days) U I/O Test 3 I/O Test 3 12 Overtraining (1 day) I/O Test 4 13-20 Interpositus Lesion & Recovery (7 days) 21 I/O Test 4 22-26 Retention Training 27 I/O Test 5 28 I/O Test 6 collected. Following adaptation, each rabbit was tested for reflexes to the four air puff intensities as before. This I/O test 2 was used to assess the effects of adaptation on reflexive responding. Over the following 5 days rabbits received either paired or unpaired training as described below and were again tested for reflexes to the four air puff intensities (I/O test 3). This tested for the influence of training on reflexes. At this point animals in the Unpaired group received I/O test 4 to test for consistency and were finished. On the other hand, each rabbit in the Paired group was given one additional day of training. This additional over-training was to guard against the possibility that I/O test 3 interfered with performance. Then, the left interpositus nucleus of these 8 rabbits was destroyed by electrolytic lesion. After 7 days of recovery from the surgery each rabbit was again tested for reflexes to the four intensities of air puff (I/O test 4). This tested for the effects of the lesion before retention/relearning began. Next began S days of retention/relearning on classical conditioning using the same protocol as before the lesion. Training occurred on the same side as the original training and therefore on the same side as the interpositus lesion. Finally, reflexes to the four air puff intensities were tested for two days at the end of training. I/O test 5 tested for reflexive responses immediately following retraining. I/O test 6 on the subsequent day tested for consistency of responding. 24 Behavioral Training Paradigms Training Stimuli. The standard classical conditioning stimuli used were the tone and airpuff. The CS was a 1 kHz, 85 dB SPL, 348 ms tone presented through speakers located about 15 cm above and in front of the animal’s ears. The US was a 3 psi (2.1 N/cm^), 98 ms airpuff to the left cornea. Normal Paired Training Sessions. In general, the term “paired training” refers to sessions in which the tone CS was presented 250 ms prior to the airpuff US and both continue another 100 ms and co-terminate. This is the delay paradigm in classical conditioning. Paired training sessions consisted of 108 trials, 12 blocks of 9 trials, where the first trial of each block was a tone-alone test trial and was followed by 8 tone-air paired trials. The intertrial interval varied between 20 and 40 seconds (30 s average). Pseudo-Random Unpaired Training Sessions. These sessions consisted of 204 trials (12 blocks of 17 trials) in which each block contained 8 US-alone and 9 CS-alone trials, corresponding to the number of stimuli presented in paired training sessions. The ITI varied pseudo-randomly, from 0 to 33 seconds, to approximate the truly random procedure suggested by Rescorla (1988). The shortest possible time interval between stimuli was actually 500 ms when accounting for data collection time. This interval is not likely to produce associations of two consecutive stimuli (Nordholm et al., 1990). The trial sequence was specified in order to ensure all animals received the same experience (a=air, t=tone), as follows: a, t, t, a, a, t, a, t, t, a, a, t, a, t, t, a, t. One complete session lasted about 50 minutes. 25 R d k x Responsiveness Testinz (UQ tests). Reflex responses were measured on separate days over the course of training. These input/output (I/O) tests consisted of 40 US-alone trials, 10 consecutive trials at each airpuff intensity (1, 2, 3, and 4 psi). The normal training intensity is 3 psi. Based on earlier test cases (personal communication, D. Lavond and R.F. Thompson), randomly varying the intensity of presentations does not qualitatively differ from sequential presentations. In the following study, either an ascending or descending order was maintained per animal. Each I/O trial consisted of a 100 ms corneal airpuff presentation with the intertrial interval (ITI) varying from 20-40 seconds (average 30 s). Between each block of 10 trials there was a pause while intensity was manually adjusted. The entire I/O session lasted about 20 minutes. Data Analysis. Presentation of stimuli and collection of data were made on-line by an interface (Lavond & Stcinmctz, 1989) and software (8088 machine language and Forth) for the 1BM-PC/XT and compatibles. Each trial began with a 248 msec baseline period, followed by a 248 msec CS period (the period starting with tone onset), and ended with a 248 msec US period (the period starting with air puff onset). During unpaired training sessions trial length was only 396 ms to allow for shorter inter-trial intervals. On CS-alone trials the tone was presented at 48 ms for 348 ms, and on US-alone trials the airpuff was presented at 296 ms for 100 ms, ending the trial. The computer measured amplitude, latency (onset and peak) and amplitude time areas during the CS and the US periods, and determined trials in which a conditioned response (CR) occurred and when criterion for learning was reached. A CR was any movement after CS onset that equaled or exceeded 0.5 mm of 26 nictitating membrane extension. Criterion for learning consisted of the first time that a CR occurred in 8 out of 9 consecutive trials. Percentage of CRs was reported for each block and for the entire training session. Trials were not counted toward criterion if movement in the baseline period exceeded 0.7 mm or if a movement of 0.7 mm occurred in the CS period within 25 msec after CS onset. Typically about 5% of trials are rejected from analysis by these criteria (e.g., Gormezano et al., 1983; Lavond et al., 1987). Sureerv. Each rabbit was anesthetized with xylazine (8 mg/kg) and ketamine (60 mg/kg) and maintained on 1-2% halothane. Under aseptic conditions, the midline was incised and the dura exposed. The head was positioned in the stereotaxic plane of McBride and Klemm (1968) in which bregma is placed 1.5 mm dorsal to lambda. Two stainless steel insect pins (size 00), insulated with Epoxylite except for 0.5 mm exposed tips, were inserted deep into the cerebellum. The coordinates for the anterior electrode were AP +1.5, ML +5.0 and DV -13.5 mm from the bony landmark lambda (+ represents anterior, lateral and dorsal). The coordinates for the posterior electrode were AP +0.5, ML +5.0 and DV -14.5 mm. The final vertical coordinate was actually determined by the distance from lambda of the posterior electrode (-14.5 mm), by the distance from the first recorded cerebellar cortical cells (-7.5 mm), by the recording of unit activity near the target location (0 mm), and by ascending from a recording of auditory evoked responses from the dorsal cochlear nucleus (approximately 2 mm). At the targeted location 2 mA of anodal direct current was passed for 150 seconds through each electrode referenced to the skin. The skin was sutured and wound wiped with the antiseptic microbicide Betadine ointment. 27 Each rabbit was monitored for postoperative recovery from anesthesia and allowed at least 7 days to recover. Histology. The rabbits were overdosed with sodium pentobarbital and perfused intra-aortically with saline followed by formalin. The brains were extracted and postfixed in formalin for at least one week. Next, the brains were embedded in a gelatin-albumin matrix and fixed with paraformaldehyde in preparation for normal histology. The brains were then frozen and sectioned at 80 mm. The slices were mounted onto chrome-alum subbed slides and dried. The sections were then stained with crcsyl violet for cell bodies and counterstained with potassium ferrocyanidc (Prussian Blue) for the iron deposited by the lesions. A coyer slip was cemented over the section and allowed to dry. Serial reconstructions were made using light microscopy and a projected image through a photographic enlarger. 2.1.2 Results: Paircd.Group- Conditioning Performance on classical conditioning was analyzed with separate repeated measures ANOVA for percent of conditioned responses, amplitude of conditioned responses, and amplitude of unconditioned responses. A two-factor within group design compared the first 5 days of acquisition with the S days of retention/relearning (before/after lesion) and each day of training (days 1 through S). All F tests reported are significant at p < .05 unless stated otherwise (non significant F values are not reported). Only a few additional comparisons were made to keep the experiment-wise error low. 2 8 Percent Conditioned Responses. There were significant main effects before/after the lesion (£(1,7) = 70.0), over days (E(4,28) = 35.0) and of their interaction (£(4,28) * = 30.0; Figure 2.1 A). Hie first day of acquisition was significantly different than the 5th day of acquisition (£(1,7) - 253.0), indicating that the animals learned. The 5th day of acquisition was significantly different than the 1M day of retention/relearning (E.(l,7) * = 333.0), indicating that the lesion abolished the previously learned response. The 1* day of acquisition did not significantly differ from any day of retention/relearning, indicating that there was no reacquisition after the lesion with repeated training. Conditioned Response Amplitude. There were significant main effects before/after the lesion (E(l,7) « = 15.0), over days (E(4,28) « 8.0) and of their interaction (£ (4,28) * * 8.0; Figure 2. IB). The first day of acquisition was significantly different than the 50 1 day of acquisition (£ (1,7) = 16), indicating that the animals learned. The 5th day of acquisition was not significantly different than the 1H day of rctcntion/rcleaming indicating that there was no reacquisition after the lesion with repeated training. Paired Group - Input/ Output Tests Reflexes to 1, 2, 3 or 4 psi at various points during training (see Table 1.1) were analyzed by repeated measures ANOVA for unconditioned response amplitude, amplitude'time area, frequency, rise-time and peak latency. A three- factor within group design compared reflexes for acquisition with retention/relearning (before/after lesion effect), each with three tests days (day) and each of these with four intensities (1 through 4 psi). As above, all F tests 2 9 A 100 80 3 60 0 1 40 £ 20 0 3 1 2 4 S B Acquisition 2 3 4 Reacqulsltlon “i 5 1* 3 E C C 1 O Acquisition e o 1 2 3 4 5 ~i-------- 1 ---- 1 ~ 2 3 4 Reacaulsltlon Figure 2.1 Percent CRs (A) and CR amplitudes (B) recorded before and after left interpositus nucleus lesions in animals given paired classical conditioning training. 3 0 reported are significant at q < .05 unless stated otherwise, and only a few additional comparisons were made to keep the experiment-wise error low. Amplitude. There was no significant main effect on reflex amplitude before/after the lesion, indicating that the size o f the reflexes were the same before and after Lhe lesion (Figure 2.2A). There were significant effects of days (£(1,14) = 3.9), o f psi (E (3,21) = 19) and of the interaction lesion x I/O test (£ (2,14) * * 6.4), indicating that the reflexes generally changed size over days and that there were differential effects on the reflexes depending upon the lesion and day observed. No other main effects were significant. Overall, neither reflex amplitudes to 1 psi nor 2 psi changed, while reflex amplitudes elicited by 3 psi and 4 psi did change significantly (£ (5,35) * = 3.4 and 3.4 respectively). Furthermore, there were no effects at any pressure level on reflexes before versus after the lesion. Together, these results indicate no effect of lesion on reflexive amplitudes at any psi value. However, 3 and 4 psi values did change across days. To further explore this finding, we compared I/O tests at 3 and 4 psi only. Reflexive amplitudes grew significantly larger when comparing I/O 1 and 2 combined (pre training) versus I/O 3 (post-training), E (1,7) * = 8.9 and 5.8, respectively for 3 and 4 psi. Paired t-tests were then used, showing that the 4 psi values did not change from I/O 3 to I/O 4, from I/O 3 to I/O 5 and from I/O 3 to I/O 6. That is, reflexive responding to 4 psi remained high regacdiess of the \es\on. On the other hand, 3 psi did significantly decrease following the lesion (I/O 3 versus I/O 4,1(7) = 2.9). However, there was no difference at 3 psi between I/O 3 and either I/O 5 or I/O 6, indicating that reflexive responses to 3 psi returned to pre-lesion levels after retention training. 31 I I I B , ------1 ----- 1 J----- 1 -----. 100 i r M OO 1 0 0 _ _ 00 1 I V O Tailtaaalon i — i r ~ VO Tati ■•MlOA 1 1 0 - 100 . I f | H 5 0 . i t a 4 VO T a il I u i Im Palrad Training M m La don (M -0 ) ■— ■ 4pat a a a pal *— * > P * < a a »P*» Figure 2.2 UR amplitudes (A), UR amplitude/time areas (B), percent URs (C), UR rise times (D) and UR peak, latencies (E) recorded during input/output test sessions in rabbits given paired training. The input/output functions were established with 4 levels of air puff pressure. 32 Amplitude-time Area. The main effects for amplitude-time are the same as for the amplitude data above, except that there was also a significant 3-way interaction for amplitude-time (E (6,42) = 2.7; Figure 2.2B). There was no significant main effect for amplitude-time before versus after the lesion, but there were significant day (E (2,14) = 5.3) and pressure (E (3,21) = 34) main effects and a lesion by day interaction (E (3,21) = 3.5). As with amplitude data above, there were no changes in reflexive areas for 1 and 2 psi, but the reflexes for 3 and 4 psi did change (E (5,35)3.5 and 4.5). Both psi 3 and 4 areas increased from I/O test 1 and 2 combined to I/O test 3 (E (1.7) = 8.7 and 9.7, respectively). Paired t-tests were then used, showing no change al 4 psi from I/O 3 to I/O 4, from I/O 3 to I/O 5, and I/O 3 to I/O 6. On the other hand, at 3 psi, I/O 3 differed from I/O 4 Q (7) « 3.2) and from I/O 5 ft (7) = 2.5) but not from I/O 6. These results arc very similar to the analysis of amplitude data above. Unconditioned Response Frequency. There were no significant main effects before versus after the lesion or over days, but there was a significant pressure effect (E (3,21) = 17.0; Figure 2.2C). No interactions were significant. There was a significant difference between 1 and 2 psi (E (1,7) = 34) and between 2 and 3 psi (E (1,7) = 5.6), but not between 3 and 4 psi. These results indicate that there were fewer unconditioned responses with decreasing airpuff intensity, but that there were no lesion effects on percentage of unconditioned responses. Unconditioned Response Rise-time. Welsh and Harvey (1989) used an unusual method of scoring UR onset in their rise-time analysis: they measured backward from the 0.5 mm UR amplitude to an amplitude of 60 mm and defined this as UR onset. The current data were analyzed for rise-time using a comparable procedure; measuring backward from the 0.5 mm UR criterion to an amplitude of 33 100 mm and defining this as the UR onset. The rise-time was then calculated as the difference between peak latency and onset latency. Analysis of variance indicated no significant effects, although pressure level approached significance (£(3,21) = 3.0, n.s; Figure 2.2D). Nevertheless, limited comparisons were performed to examine these data. There was no significant difference at 3 psi comparing both I/O 1 and 2 combined with I/O 3. At 4 psi, there was a significant difference comparing both I/O 1 and 2 combined with I/O 3 (E (1,7) « = 7.7). In general, there was no effect on rise-time of the unconditioned reflexes to the differing psi intensities, and there was no lesion effect. Unconditioned Response Peak Latency. There were no significant main effects for before versus after the lesion, over days or of intensity (Figure 2.2E). None of the interactions were significant. Summary. The amplitudes and amplitude-time areas of unconditioned reflexes to 1, 2, 3 and 4 psi before any training were small. The reflexes to I and 2 psi did not change throughout all I/O tests. After training, reflexes to 3 and 4 psi were larger than before training. It may be significant that the intensity used for training was 3 psi. After the lesion, there was a significant reduction in the reflexes to 3 psi (I/O 3 versus 4) before training resumed. However, the responses to 3 psi returned to pre-lesion levels (I/O 3) by the end of retention/ relearning (I/O 6). Control Group - Unpaired Training Performance on the unpaired training condition was analyzed with separate repeated measures ANOVA for percent of pseudo-conditioned responses (pseudo- CRs) and amplitude of such responses. 34 Percent Pseudo-Conditioned Responses. The mean percent pseudo-CRs to tone alone presentations averaged over the five days of pseudorandom stimulation was 2.4 (Figure 2.3A). There was no significant change in mean percent pseudo- CRs over five days of training. Pseudo-Conditioned Response Amplitude. The mean amplitude of pseudo- CRs to tone alone trials averaged over the 5 days of pseudorandom stimulation was 0.06 mm (Figure 2.3B). There was no significant change in this measure over the five days of training. Control Group - Input/ Output Tests Reflexes to 1, 2, 3 or 4 psi at various points of training were analyzed by repeated measures ANOVA for unconditioned response amplitude (Figure 2.3C). A two-factor within group design compared reflexes over four test days (day), each of which included testing at 4 different aiipuff intensities (1, 2, 3, and 4 psi). As above, all F tests reported are significant at p < .05 unless stated otherwise, and only a few additional comparisons were made to keep the experiment-wise error low. There were significant main effects for day (E (3,30) = 4.3) and for pressure (E (3,30) = 4.9) but not for their interaction. There was no change in reflexes for 1 psi but there were significant increases in the reflexes for 2, 3 and 4 psi (E (3,30) = 4.1, 3.5 and 3.6, respectively). I/O tests 1 and 2 did not significantly differ from one another as was the case with I/O tests 3 and 4. When I/O 1 and I/O 2 (before and after adaptation) were combined and compared with I/O 3 (following unpaired training) there was a significant difference (E (1,10) = 6.8), indicating that reflexes over all intensities increased as a function of random exposure to the training stimuli. 35 B 3 0 20 10 0 2 3 4 1 5 5 2 . 2 1 3 4 5 Daya of Unpaired Training Daya of Unpaliad Training 8 7 6 5 4 3 2 1 0 1 2 3 4 Unpaired Training ( N s 11) 4 pal 3 pal 2 pal 1 pal I/O Teat Seaalon Figure 2.3 Percent CRs (A) and CR amplitudes (B) recorded during unpaired training sessions. Also shown are UR amplitudes for four US intensity levels that were recorded during input/output test sessions presented to the unpaired group (C). 36 In summary, these results indicate that the size of the unconditioned reflex is not an unvarying parameter. Reflex amplitudes increase over days of training whether the animals are given random unpaired tones and air puffs, or given paired training, as reported above. Histology. Figure 2.4 shows reconstructions of the lesion for these 8 experimental animals. All rabbits had damage to the anterior 2/3 of the deep nuclei, encompassing the lateral border of the intcrpositus and the medial border of the dentate. The lateral extent of the lesions included most of the dentate nucleus in all but two animals (126 and 128), and the medial extent included all or most of the interpositus. The lesions, therefore, included the beginnings of the superior cerebellar peduncle which originates in the hilus of the dentate nucleus and divides the interpositus into dorsal and ventral halves. The critical region of the intcrpositus, the dorsal aspect of the anterolateral intcrpositus, is destroyed in all animals. In addition, at least 3 (005, 006 and 126) and possibly 4 (005) rabbits have complete interpositus damage in the anterior-posterior plane. The electrode track shows some incomplete cortical damage to anterior cerebellar lobules (anterior to HVI). HVI is largely damaged in two rabbits (125 and 128) with some damage in two others (004 and 126) but not damaged in the remaining 3 rabbits (005, 006 and 127). In summary, the electrolytic lesions have destroyed the pari of the interpositus nucleus which we have previously identified as critical for classical conditioning in all 8 animals. 37 Smallest Largest 0.0 + 0.5 + 1.0 + 1.5 V C * Figure 2.4 Smallest and largest extent of the lesions for rabbits given paired training before and after an inteipositus lesion. The electrolytic lesions include the anterior and middle extent of the interpositus nucleus as well as damage to the efferent superior cerebellar peduncle. In addition, there is some damage to the dorsal cochlear nucleus below the cerebellum. Refer to boxed figure for structure identification. Numbers in left column are mm relative to lambda. Labels: HVI= hemispheric lobule VI, ANS =ansiform lobe, PF=parafloculus, ANT = anterior lobe, D E = dentate, IN = interpositus, f= fibers, FA= fastigial, VCN= ventral cochlear nucleus, icp= inferior cerebellar peduncle, VN= vestibular nucleus, IO = inferior olive. 38 2.1.3 Discussion: As previously reported in a number of studies, rabbits conditioned with tone and airpuff acquired the conditioned eyeblink response and showed no evidence of conditioned responses following lesion of the inteipositus nucleus of the cerebellum. In addition, this experiment demonstrated no persisting effects of the interpositus lesion (that abolished the CR) on the UR during paired training or for I/O tests to US alone presentations at four different US intensities (1, 2, 3, 4 psi) on any measure of UR performance: amplitude, amplitude-time area, frequency, latency or rise-time. The lowest intensity used (1 psi = 0.68 kg/cm^) is a value at which Welsh and Harvey reported a substantial lesion effect on UR frequency, peak latency and rise-time. While the present study reveals a transient decrease in the mean UR amplitude and amplitude-time area to the training intensity US (3 psi) immediately following lesion, this decrease quickly recovered to pre-lesion levels. No latency, frequency, or rise-time deficit was observed. In comparing these results to those of Welsh and Harvey, it is important to emphasize that they did not in fact measure the effects of interpositus lesions on reflex responses. Instead, all their analyses were done after lesion by separating the lesion animals, post hoc, into different groups. In contrast to Welsh and Harvey, the present study measured the effects of interpositus lesions on URs in the same animals before and after lesion. Because many factors can influence the UR and because of substantial individual differences in URs, it is essential to compare URs in the same animals before and after lesion. 39 In addition, the interpositus lesions of Welsh and Harvey often included varying amounts of cerebellar cortex (Welsh, 1987, unpublished master's thesis). There is now evidence that lesions of the cerebellar cortex that markedly impaired or abolished the eyeblink CR caused a significant increase in the amplitude of the UR in the same animals (Logan, 1991, unpubtished doctoral dissertation). It is thus very possible that the differences Welsh and Harvey report between their post lesion groups are due to relative changes in their "ineffective" lesion groups, due in turn to cerebellar cortical damage in the absence of damage to the critical region of the interpositus nucleus. In the present study, lcsioned animals showed a significant but very transient decrease (I/O UR at 3 psi), not increase, in amplitude and amplitude-time area immediately following lesion, and completely recovered to preoperative levels. However, latency, frequency and rise-time measures did not show this effect, i.e. they were not sensitive measures as Welsh and Harvey have argued. US intensity is reliably the most powerful determinant of the UR. Yet Welsh and Harvey suggest that UR frequency, peak latency and rise-time do not even vary with US intensity except at the lowest intensities, intensities that are ineffective as USs. On the other hand, their data suggest that UR amplitude does vary systematically with US intensity. Importantly, as noted, they report no differences between "effective” and "ineffective" lesion groups in UR amplitude as a function of US intensity. The present study indicates that UR rise-time and latency do not vary systematically or significantly with air pressure, in marked contrast to amplitude and amplitude-time area measures, which increase in a highly significant manner with increasing air pressure (see Figures 2.2 and 2.3C). This argues that UR rise-time and latency are not meaningful measures of behavioral 40 performance in this behavioral learning paradigm. It was observed, however, that UR frequency increased significantly with US intensity. The interpositus lesion effective in abolishing the CR had no effect on UR frequency at any US intensity. A striking and somewhat unexpected finding in the present experiment is the marked overall increase in UR amplitude over US alone I/O tests that resulted from experience (Figures 2.2A and 2.3C). This increase does not appear to be associative in nature; the same overall increase occurs after training in the control group given pseudorandom unpaired training. Further, the interpositus lesion that abolishes the CR has no effect on this experience-dependent increase in UR amplitude. This finding indicates that the UR itself is “plastic” and can change as a result of experience with the situation and stimuli, independent of increases in associative strength of the CR. There is a literature indicating that adaptation of the cyebiink reflex occurs. Specifically, in both rabbits and humans, increased or decreased loading of the external eyelid can increase or decrease the gain of the reflex and this gain persists immediately following removal of the load (Evinger & Manning, 1988). Further, the reflex can be modulated by presentations of “neutral” stimuli prior to learning, the phenomenon of reflex facilitation (e.g., Weisz & LoTurco, 1988). The important point of these observations and our results is that the amplitude and other properties of the eyeblink UR can change considerably quite independently of the development of conditioned eyeblink responses to a CS. The key findings in this experiment are: 1) interpositus lesions that abolish the eyeblink CR have no persisting effects on any property of the UR to US alone tests, even at low intensities, and 2) experience in the situation results in a marked increase in UR amplitudes to US alone tests, independent of the development of CRs. 41 2.2 Training with Low Intensity. Unconditioned Stimuli Does Not Unmask a Performance Deficit Following Interpositus Lesions In an attempt to support their claim of a performance deficit following interpositus lesion, Welsh and Harvey (1989) specified the conditions under which they observed such a deficit. They claimed that lesions of the interpositus effective in abolishing the eyeblink CR did have effects on the UR on US alone trials at low US intensities. However, their analysis was entirely post hoc: They selected only some of the interpositus lesion animals where the CR was abolished and compared them, in terms of UR measures, with some of the animals where the lesion did not abolish the CR. All comparisons were done post-lesion, no pre-lesion UR data were given and the criteria for selecting the animals used for the UR comparisons were not given. The previous experiment showed that if the URs on US alone trials are compared in the same animals before and after lesion, interpositus lesions effective in completely abolishing the eyeblink CR had no persisting effect on any property of the UR over a wide range of US intensities. Welsh and Harvey (1989) argued that “when one attempts to equate the CS and UCS as response-eliciting stimuli, the [lesion] deficits in the CR and the UCR become more alike” (p. 309). However, when low intensity USs were used which elicit URs on US-alone trials equivalent to CRs prior to lesion, the lesion abolished the CR and has no effect at all on the equated UR (I/O results of previous study). It is true that the US intensity used for training was well above UR threshold and elicited vigorous reflex blinks, typically considerably larger in amplitude than the asymptotic CR. It may be conceivable that use of the relatively high intensity US in training may have resulted in some as-yet-undeflned effect that could influence 42 subsequent UR performance on US-alone trials at low US intensities, i.e. could somehow mask lesion effects on the UR. The eyeblink UR itself does exhibit a degree of plasticity, increasing in amplitude with repeated US presentations independent of associative learning (Steinmetz et al., 1992), or as a result of reflex facilitation (Weisz & LoTurco, 1988), or following eyelid loading (Evinger & Manning, 1988). To evaluate this possibility reported here are the effects of interpositus lesions on performance in animals trained with very low US intensities. 2.2,1 Procedures: Protocol. Twenty New Zealand White rabbits were trained using standard procedures, measuring extension of the left nictitating membrane (NM), with a 1 kHz, 85 dB 348 msec tone CS and a 98 msec corneal airpuff at one of two intensities (0.13 or 0.50 psi) as the US, CS and US coterminating (Clark et al., 1984; Lavond et al., 1990; Steinmetz et al., 1992).1 Intertrial interval (ITI) varied randomly from 20-40 sec (average 30 sec). All animals were given 5-10 days of acquisition training, (108 trials per day, every ninth trial CS-alone). Animals that learned to criterion (8 out of 9 consecutive CRs) and exhibited consistent CR 1 The actual US source air pressures used by Welsh and Harvey (1989) are puzzling. They state (p.300; Figures 8-10; Table 2) that their training US source pressure was 2.2 kg/cm^. This equals 31.79 psi, a pressure that could damage the cornea. The training air pressure standard in the field is 3-4 psi (Gormezano, 1966) and in our laboratory is 3 psi, equal to 0.21 kg/cm^, which equals 2.1 Nfcm^ (see, e.g., Steinmetz et al,, 1992). The lowest pressure used by Welsh and Harvey in their post-lesion US-alone tests was 0.09 kg/cm 2, equal to 1.29 psi. The two training air pressures we use here, 0.50 and 0.13 psi, equal 0.035 and 0.009 kg/cm2, well below this pressure. 43 performance (seven of the ten 0.5 psi and one of the ten 0.13 psi animals) were given electrolytic lesions of the left anterior interpositus nucleus, using standard electrode placements (Clark et al., 1984; Steinmetz et al., 1992) and a 2 mA anodal direct current for about 150 sec. Lesioned animals were also given performance tests at the end of prelesion training, one week post-lesion and at the end of post-lesion training, consisting of 40 US alone presentations of the same intensity US used during training for that animal, m varying from 20-40 sec (average 30 sec). Previous studies have shown that amplitude is the most sensitive and reliable measure of performance o f the NM response. It (and the highly correlated measure of area under the response envelope) correlates highest (0.95) with amount of unit activity in motor nuclei (Ccgavske, Patterson, & Thompson, 1979; McCormick, Lavond, & Thompson, 1982). Further, on US- alone trials response amplitude shows consistent and systematic variation with US intensity whereas other measures do not (Steinmetz et a l., 1992). Since later components of the CR are masked by the UR and UR amplitude is confounded by the CR on paired CS- US trials, analyzed here are CRs on CS alone trials and URs on US alone trials. Surgery. Each rabbit was anesthetized with a mixture of ketamine (.08 ml/kg) and xylazine (.6 ml/kg) and placed in a stereotaxic headholder. Anesthesia was maintained during surgery using 1-2% halolhane. The head was positioned in the stereotaxic plane of McBride and Klemm (1968) in which bregma is leveled 1.5 mm dorsal to lambda. Stainless steel electrodes (00 insect pins) insulated with epoxylite except for 150-200 mm at the tip were implanted stereotaxically in the left interpositus nucleus of the cerebellum at approximately + 1 .0 mm AP, + 5.0 mm ML, and -14.0 mm DV relative to 44 lambda. Precise DV placement was assisted during surgery electrophysiological L. recording. Subjects were allowed at least one week post-operative recovery before training continued. Histology. Following completion of the experiment, all lesioned animals were overdosed with 5 cc sodium pentobarbital injected intravenously and perfused with physiological saline followed by 10% formalin. Brains were removed, stored in formalin and embedded in an albumen-gelatin mixture. Alternate 80 ft coronal sections were stained with cresyl violet for cell bodies and counterstained with potassium ferrocyanide (Prussian Blue) for the iron deposited by the lesion. 2.2.2 Results: Of the 10 animals trained with the 0.5 psi US, seven animals reached learning criterion in an average of 395 ± . 109 (s.e.m.) trials and received a lesion of the left interpositus. Five of the seven reached criterion within the first 5 days o f acquisition training. The other two animals did not reach criterion within S days but indicated increased conditioned responding and were given an additional 5 days of training during which criterion was achieved. As for the ten animals trained with the 0.13 psi US, only two animals reached criterion and only one exhibited sufficiently reliable performance ( consistently above 50% conditioned responding from the first day criterion was reached) to merit lesion. Thus, the two US intensities used appear to have bracketed the US threshold intensity effective for learning. A between group repeated measures ANOVA was performed for CR amplitudes on CS-alone test trials comparing first (Acql) and last (LastAcq) days 45 of acquisition for both US intensity groups pre-lesion. The main effects of group and training session were significant (E(l, 18) = 17.3 and EO* 18) = 26.9, respectively) as was the group by session interaction (E(l, 18) = 18.7). CR amplitudes of the 0.5 psi group were generally greater than the 0.13 psi group, and there was a highly significant increase in CR amplitudes for animals in the 0.5 psi group from Acql to LastAcq (E(l, 18) = 45.3) but no significant increase for the 0.13 psi group. A within group repeated measures ANOVA including the seven lesioned animals of the 0.5 psi group was performed for CR amplitudes (CS alone-trials) across acquisition and retention training sessions. There was a significant effect of lesion on CR amplitudes (E(l> 6) = 43.9), a significant effect of training session (E(l, 6) = 54.7), and a significant interaction (E(l, 6) = 61.8). CR amplitudes increased significantly (E(l, 6) = 58.4) from 0.1 ± 0.1 to 1.7 ± 0.2 mm during acquisition, dropped off to 0.0 ± 0.0 (E(l. 6) = 53.0) following the lesion and never recovered (see Figure 1). The one 0.13 psi animal lesioned exhibited similar results. All analyses of variance were repeated including this animal with identical results. A one way repeated measures ANOVA showed that reflex response amplitudes to US-alone trials for the 0.5 psi group pre-lesion (1.3 ± 0.3 mm), post-lesion (0.8 ± , 0.3), and following post-lesion training (1.7 ±, 0.8) did not differ significantly (see Figure 2.5). There was, thus, no effect of lesion on performance of the UR (see Figure 2.5). 4 6 E < D X3 Q . E' < 5 CRs (CS-alone) URs (US-alone) 4 tn 3 t/> co 2 a . c r o S’ us u Acquisition Retention Figure 2.5 Response amplitudes for 7 animals in the O.S psi group which learned and received unilateral interpositus lesions are presented here. Triangles represent CR amplitudes on tone-alone trials for the first and last days of acquisition (Acql and LastAcq) and retention (Retl and LastRet). UR amplitudes on airpuff-alone test trials are represented by squares. Error bars indicate s.e.m. 47 Figure 2.6 Coronal brain sections at 0.0, 0.5, and 1.0 mm anterior to lambda showing the extent of the largest and smallest lesions in the vicinity of the interpositus nucleus. Forward slashes represent the largest lesion and reverse slashes represent the smallest lesion. (Abbreviations: ANS, ansiform lobule; ANT, anterior lobe; DE dentate; F, fibers; FA, fostigial nucleus; HVI, hemispheric lobule VI; icp, inferior cerebellar peduncle; IN, interpositus nucleus; IQ inferior olive; PF, paraflocculus; VCN, ventral cochlear nucleus; VN, vestibular nucleus.) 48 Lesion reconstructions are presented in Figure 2.6. Lesions included dorsolateral portions of the anterior interpositus nucleus and sometimes portions of the dentate nucleus and dorsal white matter. 2.2.3 Discussion: Results of prelesion training indicate that the two US intensities used (0.13 and 0.S psi) approximately bracketed the threshold intensity effective as a US; two of ten animals learned at 0.13 psi and seven of ten learned at 0.5 psi. In terms of group performance, the 0.5 psi animals showed significant learning but the 0.13 animals did not. It is always possible that with extended training more of the 0.13 psi trained animals would leam. However, previous experience would suggest that if animals show no signs of CRs in the first five days of training, they are unlikely to leam with more extended training. At the end of prcoperative training the CR and the UR (US alone trials) were essentially identical: mean amplitudes: CR = 1.7 ± 0.1 mm; UR « 1.3 ± 0.9 mm (see Figure 2.5). The CS and the US were thus equated as response eliciting stimuli (actually, the CR was numerically, but not statistically, a larger response). Effect of the interpositus lesion was clear and consistent. The CR was completely abolished in all lesion animals ( see Figure 2.5). In marked contrast, the lesion had no significant effect at all on UR amplitude. There was a numerical, but not significant, decrease in UR amplitude immediately post-lesion that recovered fully to the prelesion level by the end of post-lesion training. These results exactly replicate results we reported earlier, the effects of appropriate 4 9 interpositus lesions on the CR and the UR when the training US intensity was well above threshold, as is standard in the field (Gormezano, 1966). In sum, when the CS and the training US are equated in terms of amplitude as response eliciting stimuli, effective lesions of the interpositus nucleus completely abolish the CR and have no effect at all on the pre*lesion-equated UR. This result, in conjunction with all of the other studies in which higher intensity USs were used (see introduction), decisively disproves Welsh and Harvey’s (1989) "performance” argument, i.e. the argument that interpositus lesion abolition of the eyeblink CR is somehow due to lesion effects on the UR. 2.3 Do lesions of eyeblink motorneurons equally impair the CR and the UR? Welsh and Harvey (1989) argue that abolition of the eyeblink CR following intcrpositus lesion is merely a "performance" effect—that the lesion alters the properties of the UR and this will have disproportionate effects on the CR because the CR is somehow more "fragile" than the UR. Welsh and Harvey do not define "performance" other than in terms of the properties of the UR. The experiment reported here was designed to test their assertion that procedures that alter performance of the UR will have disproportionate effects on the CR. In an important series of studies, Disterhoft and associates (1985) made lesions of the accessory abducens nucleus in rabbits previously trained in eyeblink conditioning (they used a white noise CS, comeal air puff US and measured NM extension). Results were striking: There were significant reductions in amplitudes of the CR and the UR immediately after lesion. However, in all cases they report varying degrees of recovery of the CR and the UR. Indeed, in the one example 5 0 shown (Figure 5, pg. 947), the CR recovered to a greater extent than the UR. If anything, the CR appeared to be less fragile, less influenced by variables that act on performance, than the UR itself. Because this finding has fundamentally important implications for the “performance" argument these observations were replicated and extended in the present study using both measures of URs to US alone trials and CRs to paired CS-US trials and CS-alone trials. 1.3.1 Methods: Subjects. Seventeen New Zealand White rabbits were classically conditioned and underwent surgical procedures. Results are reported for nine surviving animals: 3 effective lesions, 3 non-effective control lesions, and 3 sham operated controls for anesthetic effect one day after surgery. Housing and care were in accordance with N1H guidelines as described in the previous experiments. Protocol. All animals were trained and tested using the same general schedule as in the first experiment in this chapter, with simplification. Animals were habituated to the apparatus for two days. On Day 3 they were given the I/O tests at 1, 2, 3 and 4 psi, using the same procedure described earlier. They were then trained for five days, using the standard conditioning procedure followed by another day of I/O testing. Lesions of the accessory abducens region were then made, and animals were given their first post-lesion I/O test the very next day. This was followed by five days of continued training, another I/O test day, and additional training and testing sessions (see Table 2.2). 5 1 Table 2.2 Summary of Training Protocol for Accessory Abducens Study Day Type of Session I Nictitating membrane suture and Adaptation day 1 2 Adaptation day 2 3 I/O Test 1 4-8 Paired acquisition training (5 days) 9 I/O Test 2 and Surgery (lesion or anesthetic only) 10 I/O Test 3 IM S Paired retention training (5 days) 16 I/O Test 4 (“non-effective" group finished) 17-21 Paired retention training (5 days) 22 I/O Test 5 23-27 Paired retention training (5 days) 28 I/O Test 4 (“effective" group finished) Animals showing a significant performance impairment on the first day of retention following the accessory abducens lesion made up the “effective lesion" group. Those not showing significant impairment made up the "non-effective " group. Another group of three rabbits served as controls for possible lingering effects of surgical anesthetic 24 hours after surgery. These animals received the same pre-lesion training described above. On the day of surgery they were placed in the stereotaxic headholder and exposed to the same anesthetics as all other animals for a comparable period of time without any invasive surgery. After 24 hours they were given an I/O test which demonstrated that there were no anesthetic effects on UR performance. Data Analysis. Absolute response measures for unconditioned and conditioned performance post-lesion were transformed to the percentage of preoperative levels and compared across post-lesion time points using repeated measures ANOVA between effective and non-effective groups. Surgery. Each rabbit was anesthetized with a mixture of ketamine (.08 ml/kg) and xylazine (.6 ml/kg) and placed in a stereotaxic headholder. Anesthesia was maintained during surgery using halothane. The lesion electrode (00 insulated insect pins, 150-200 mm exposed tip) was lowered into the vicinity of the accessory abducens nucleus ipsilateral to the trained eye. Stimulus trains were given (100 msec train duration, 0.1 msec pulse duration, 10-15 mA, 400 Hz, cathodal stimulation) and the electrode was positioned so that the threshold response to stimulation was "pure" eyeball retraction and NM extension. The coordinates for the accessory abducens nucleus were those reported by Gray and colleagues (1981): AP obtained from regression equation X = 0.69y + 1.0 mm where y = distance between lambda and bregma; ML = 2 mm from midline; DV 53 = 20.9 mm ventral to bregma (however, our best DV was 1 to 1.5 mm lower than this value). The lesion current was 1 mA for 30-45 sec. Following one day of recovery, animals were given the I/O tests and training as indicated above. Three sham lesion animals were also prepared, using the same schedule and same duration of anesthesia as the lesion animals, to control for possible long-lasting effects of the anesthetic. In addition, three animals sustained lesions dorsal to the region of the accessory abducens nucleus and thus served as lesioned controls. Histoloev. Upon completion, animals were overdosed with 5 cc sodium pentobarbital and the brain tissue was collected and prepared as in previous experiments. 2.3.2 Results: Figure 2.7 shows results for three animals, all of whom sustained very large lesions in the region of the accessory abducens nucleus ipsilateral to the trained eye, plotting mean amplitude (as percent of pre-lesion amplitude) of the UR to US alone stimuli of the training intensity (3 psi) and mean amplitude (as percent of pre-lesion amplitude) of the CR to CS-US and CS alone trials at successive 5-6 day periods over the course of postoperative training and testing. The mean pre lesion amplitudes were: UR = 13.3 mm; CR = 6.7 mm. Note that the mean amplitude of the UR to US alone stimulation at the end of preoperative training was twice as large as the mean CR amplitude. Following destruction of the 5 4 70 60 a > 50 .3 40 CL O & 30 a. 20 10 0 ■ URS n c r s m n 6 12 Post-Op Day 18 Figure 2.7 UR amplitudes (dark bars) and CR amplitudes (open bars) recorded 1, 6, 12, and 18 days after abducens lesions were given. The amplitudes are expressed as percents of pre-surgery levels. 55 accessory abducens nucleus both the UR and the CR were virtually abolished. However, over the course of postoperative training and testing the CR amplitude showed pronounced recovery whereas the UR showed relatively little recovery. These data were analyzed with a 2 X 4 between groups analysis of variance. Response type (UR versus CR) was not significant; session was highly significant (E(3,12) = 17.9, p < .001) and the interaction was significant (H(3,12) = 4.2, p < .05). Planned comparisons indicated no significant change in the percent UR measure over postoperative sessions but a significant increase in the percent CR measure over postoperative sessions (E(l .2) = 50.74, ji< 05). In short, the CR amplitude recovered significantly over the postoperative training sessions but the UR did not exhibit significant recovery over postoperative training sessions. The sham lesion animals showed no changes in cither CR or UR amplitudes when tested at the same time after anesthesia as the lesion animals. Further, a group of three animals served as lesion controls; in these animals the lesion caused little or no damage to the accessory abducens nucleus. These animals showed no decrement in CR or UR performance following lesion. The effective accessory abducens lesions in the three animals are reconstructed in Figure 2.8. The largest lesion completely destroyed the accessory abducens and invaded regions surrounding the nucleus. The smallest lesion invaded a portion of the accessory abducens and included some adjacent tissue. The animal with largest lesion (91- 058) was continued for an additional 10 days of training and testing. The UR to US alone showed no further recovery beyond that seen at 15 days but the CR amplitude showed further recovery. 56 Lesion Croup Anterior Middle Posterior 90-292 91-058 91-146 IV vet* I kcabO htvut*. nVL f t m o . aecABO Control Group * m . D $ 90-282 90-283 90-284 Anterior IV vent % utvut ■ ccA BO Middle m vn. trap* a. Pyr.tr. Posterior IV V tM . Dip.n. ltmn.lL p-M cr I metVn. Figure 2.8 Coronal brain sections showing the extent of the lesion in the vicinity of the accessory abducens nucleus for the Lesion and Control groups. Abbreviations: IV vent, fourth ventricle; SCP, superior cerebellar peduncle; MCP, middle cerebellar peduncle; ICP, inferior cerebellar peduncle; accABD, accessory abducens nucleus; VIII, eighth cranial nerve; VII, seventh cranial nerve; mV n., motor 57 2.3.3 Discussion: In sum, lesions of the accessory abducens nucleus that massively and permanently impair performance of the UR (recovery not statistically significant and to only 23% of pre-lesion amplitude) result in much less impairment of the CR (recovery statistically significant and to 56% of pre-lesion amplitudes), even though the UR amplitude was substantially higher than the CR amplitude before lesion. This finding replicates and extends the earlier report by Disterhoft and associates (1985) and demonstrates clearly that direct and substantial impairment in performance of the behavioral response, as measured by the UR amplitude in response to US alone stimuli, causes much less impairment in the CR amplitude in response to the CS. In this instance the CR recovery was likely due to compensation via increased activation of other extra-ocular muscles and perhaps from abducens motor neurons following damage to the accessory abducens nucleus. The animals were still able to achieve the necessary adaptive learned response. This stands in marked contrast to the effect of the interpositus lesions in the present experiments, which caused complete and permanent abolition of the CR but had no effects on the UR. Welsh and Harvey (1989) defined “performance” as performance of the UR and argue that very small effects of lesions on the UR will cause vastly greater effects on the CR. The present results do not support this and are, in fact, the opposite. Although the UR exhibits some plasticity following lesion impairment, the CR exhibits much greater plasticity in recovery following lesion impairment of the UR than does the UR itself. 58 CHAPTER 3 Does Motor Cortex Modulate the Cerebellar Circuit for Classsical Conditioning ? This chapter addresses the functional role of motor cortex in classical conditioning, as it is a purported modulator of both learning (see Houk (1989) and Woody (1984) for reviews) and response performance (Bures and Bracha, 1990). While cortical decerebration studies have demonstrated that rabbits are able to leam the conditioned eyeblink (Norman et al.t 1977; Mauk and Thompson, 1987), the present experiments take a closer look at response topography following bilateral motor cortex lesions for animals trained with both the delay and trace conditioning paradigms. 3.1 Motor Cortex Lesions Do Not Affect Acquisition or Retention of the Delay Conditioned Eyeblink in Rabbits The following study will focus on potential changes in conditioned and unconditioned response topography during acquisition and retention phases of delay conditioning in rabbits when frontal cortex, including motor cortex, has been lesioned bilaterally. Because of its interconnections with the cerebellar learning circuit (discussed in Chapter 1), motor cortex is a reasonable candidate as modulator of behavior during conditioning. Lesions (Woody et al., 1974) and cortical spreading depression (Megirian and Bures, 1970), both of which disable motor cortex, have been reported to prevent the performance of the conditioned responses without 5 9 affecting unconditioned responses. Similarly, rabbits conditioned to a shock CS and an airpuff US and then subjected to cortical application of potassium chloride to produce spreading cortical depression, were transiently unable to produce the learned response while the UR remained intact (Megirian and Bures, 1970). These profound effects on memory for the learned response are inconsistent with decerebration studies in rabbits (Mauk and Thompson, 1987) which suggest classical conditioning is possible. It is true that the primary efferents of motor cortex which are connected with the cerebellar circuit for learning synapse on structures most closely related to the conditioned response (RN and PN) These structures can, however, also act as intermediaries connecting motor cortex to the motor nuclei for the behavior and affect both conditioned and unconditioned response performance. In fact, the connection through red nucleus has some interesting implications. The red nucleus sends inhibitory projections to the trigeminal nucleus (Davis and Dostrovsky, 1986) and may modify the initial sensory component of response production. The red nucleus also sends inhibitory projections to the inferior olive where it might prevent the transmission of US information. Finally, the red nucleus has been implicated in adaptive timing of the conditioned response (Krupa et at., 1994). In the present study, acquisition and retention of delay conditioning were compared between lesioned and non-Iesioned animals and basic CR and UR topographic measures, such as amplitude, onset latency and area, are reported. The results clearly indicate that large bilateral frontal cortical lesions have no effect on either acquisition or retention of delay conditioning, nor do they affect any of the response characteristics measured. 6 0 3.1.1 M eth o d s: Subjects. Fourteen New Zealand White rabbits (Oryctolagus cuniculus) were used in this study. All procedures and animal care were in accordance with NIH guidelines. Rabbits were housed individually in a room with a regulated 12 hr light/ 12 hr dark cycle. All animals were maintained on an ad lib food and water diet. Animal care was provided by the experimenter, staff and veterinarians of the University of Southern California. Protocol. Seven animals received bilateral ablations of frontal cortex, including motor areas, prior to any training (Acquisition Group) and were then given 5 days of normal paired training. Another 7 rabbits were trained for 5 days, lesioned and allowed 7-10 days recovery, and then given 5 more days of training (Retention Group). Each rabbit's reflexive responses to airpuff intensities of 1, 2, 3, and 4 psi were measured on separate test days (I/O tests as described in Chapter 2) before and after acquisition for both groups and before and after post-lesion training for the Retention Group. Delay conditioning sessions consisted of 108 trials, 12 blocks of 9 trials, where the first trial of each block was a tone-alone test trial followed by 8 tone-air paired trials. On paired trials the tone CS (1 kHz, 85 dB SPL, 350 ms) was presented 250 ms prior to the airpuff US (3 psi at the source) and both continued together for 100 ms to co-lerminate. The intertrial interval varied between 20 and 40 seconds (30 s average). Basic Data Analysis. On-line data collection sampled signals of NM movement every 4 ms during the 744 ms epoch comprising each trial. The first 248 ms made up a baseline period during which no stimuli were presented. 61 Following the baseline period the 248 ms conditioned stimulus was presented, if i , applicable. This was followed 248 ms later by presentation of the US (98 ms). The 248 ms period between CS onset and US onset is referred to as the CS-period. Data collection continued for another 248 ms following US presentation (referred to as the US-period). Bad trials were identified by movements exceeding 0.7 mm in the baseline period or 0.S mm within the first 25 ms of the CS period. These trials were excluded from analyses to prevent contamination of conditioned response recordings by spontaneous movement artifact. The percentage of conditioned responses was monitored as were measurements of CR and UR amplitude and peak and onset latencies for each block of training and I/O testing. Conditioned responses were defined as deflections exceeding 0.5 mm within the CS period, excluding bad trials. Onset latency was defined as the first time within a trial that NM movement exceeded 0.5 mm. Statistical analyses were performed using the CSS/Statistica software package (Stalsoft, 1990) and a significance criterion of p < .05. Data for each response measure collected during acquisition for both groups were analyzed using a between groups ANOVA with repeated measures (5 days of training). Another analysis was run for the Retention group using a within subjects repeated measures ANOVA where pre/post lesion (2 levets) and session (5 levels) were the factors. I/O amplitudes and latency data were analyzed separately for each group as 2 factor repeated measures ANOVAs looking at I/O test (2 levels for Acquisition group and 4 levels for Retention group) x intensity (4 air pressure levels). 6 2 Surgery. Each rabbit was anesthetized with a mixture o f ketamine (.08 ml/kg) and xylazine (.6 ml/kg) and placed in a stereotaxic headholder. Anesthesia was maintained during surgery using halothane. The surface of the brain was exposed by making an opening in the skull, bilaterally, which extended 12-14 mm anterior to bregma and 3 mm posterior to bregma. In the medial-lateral direction, the opening began 3 mm lateral to the midline, so as to leave a protective bridge over the saggital sinus, and extended to the orbital arch at the anterior end, following the arch postero-laterally to reach an extent of 7-8 mm lateral to the midline along the posterior edge. The opening, therefore, was about 15x5 mm (AP x ML; see Figure 3.1). Large lesions of motor cortex, including some somatosensory regions, were made by aspiration. Lesion dimensions were based on stimulation studies in rabbit (Brooks & Wootsey, 1940; Woolsey, 1958) with some reference to the motor cortex lesions of Hobbclcn and van Hoof (1986) and detailed maps of rat cortex (Neafscy, 1990; see Figure 3.2). The eye region of motor cortex in rabbit is believed to be located along the dorsal surface near the midline and the medial surface of the hemispheres. For this reason, aspirations extended further medially than the bone opening. Since the AP borders o f motor cortex are rather vague, lesions also extended beyond the bone opening as far as possible anteriorly to include much of the anterior forebrain, and about 2 mm posteriorly. Because the amount o f tissue removed was extensive, the cavity was loosely packed with sterile absorptive gelatinous sponge (Gelfoam, The Upjohn Company) and bonewax sealed the skull opening. Hfsrology. Upon completion, animals were overdosed with 5 cc sodium pentobarbital, injected intravenously, and perfused with physiological saline 63 Figure 3,1 Tbp view of a rabbit skull with right half removed to expose underlying neural tissue. The area of skull removed for motor cortex suigery is outlined with a dashed line on the left surface of the skull. This same opening was made bilaterally for the actual suigeiy. Each opening was about 15x5 mm (AP x ML) (Adapted from Shenk et al., 1986.) 64 SM I SM II V is primary visual V IIs secondary visual Aud= Auditory M 1= primary motor M 1 1 “ supplementary motor SM I “ primary sensory mo tor SM II" seondary sensorymotor B— Bregma AGM “ agranular medial AGL“ agranular lateral RAG — agranular relrosplenial SI “ primary comatosensory Sll“ secondary somatosensory OC“ occipital T E - auditory INS« insular PIR— piriform RF“ rhinal fissure Figure 3.2 Schematic maps of rabbit (A) and rat (B) cortex taken from Woolsey (1958) and Neafsey (1990), respectively. Note that eye related movements are elicited by stimulation of medial M I of the rabbit and AGM of the rat. Motor cortex is typically described as agranular, hence the alternative nomenclature. In addition, Neafsey reports eyelid movement resulting from stimulation of the eye region in lateral SI. Damage did not extend quite to this region in the present studies. 65 followed by 10% formalin. The brains were removed and stored in 10% formalin for at least a week. Drawings and measurements of the extents of the lesions were made upon visual inspection of the fixed tissue. The brains were embedded in an albumin-gelatin mixture and ready for sectioning on a freezing microtome one week later. Every fifth 80 pm coronal section was collected and stained with cresyl violet for further lesion reconstruction and identification of damage. 3.1.2 Results: Both lesioned (Acquisition Group) and unlesioncd (Retention group) animals acquired the conditioned cycblink without difficulty. The average trials- to-critcrion for the Acquisition group was 190 ± 35.6 S.E.M. and 185.7 41.0 for the Retention group. Percent Conditioned Responses. There was no difference between groups in percent conditioned responses over the course of 5 days of acquisition training, indicating that lesions of motor cortex had no effect on learning rate (see Figure 3.3). There was a significant effect of training session (£(4,48)=34.4) reflecting the increase in conditioned responses observed with acquisition. The Retention group showed no impairment following bilateral motor cortex lesion. Condilioned Response Amplitude. Conditioned response amplitudes on CS- alone test trials increased with acquisition in both groups of animals (E(4,48) = 14) and were within the same range (see Figure 3.4). CR amplitudes were unimpaired by subsequent motor cortical lesions for the Retention group. Post-lesion amplitudes were, however, significantly higher than pre-lesion (£(1,6)=15.4). 66 Percent CRs on CS-alone trials & c n r# 4ft 20 1 2 3 4 5 1 2 3 4 5 Flel&nlion (H-/> Acquisition (H 7) Acquisition Ratantian Figure 3.3 Bilateral motor cortex lesions did not affect the acquisition or retention of the classically conditioned eyeblink response. Percent conditioned responding did not differ between lesioned and unlesioned animals during acquisition. In animals lesioned after learning, there was no effect on subsequent retention of CRs. 67 E JE < D T D E < 12 E E o a> ■a o o <c M 2 S a s i £ A pll £ 1 ptd — A — 2 P#1 — + — t |>si —D — C R s <08* alone} Acquisition 12 o a > t i A 0 1 2 3 A 5 * 1 2 3 4 5 ” r> —■ — 4 pai 3 p si — A — 2 p si — ♦ — 1 p si C R s (CS-alone) A cquisition R e te n tio n Figure 3.4 (A) Response amplitudes of animals which received motor cortex lesions prior to acquisition show characteristic rises in amplitude across training and are within the same range as unlesioned animals (compare with Figure 3.3B, acquisition phase). (B) Response amplitudes before and alter lesion do not show any deficit in performance of either CRs on CS-alone trials or URs on US-alone trials (I/O tests) following bilateral motor cortex lesion. 68 Unconditioned Reflex Responses. A within-subjects repeated measures ANOVA performed on the two I/O tests for the Acquisition group indicated that the difference between pre and post acquisition I/O tests approached significance (E(l,6)=5.6, p=.055). There was a significant effect for US intensity (E(3,18)=6.9) which was due to the much lower response amplitudes at 1 psi as compared to 2, 3, and 4 psi combined (contrast comparison E(l,6)=9.8; see Figure 3.4A). For the Retention group, analyzing 2 lesion levels (pre/post) x 2 I/O tests x 4 US intensities, there was no difference across I/O test overall; however, I/O 1 and I/O 4 were significantly different (contrast comparison £(1,6)=6.7). There was no main effect for intensity, but there was a significant interaction between lesion and intensity (E(3,18)=4.9) which was due to a significant response increase to 4 psi post-lesion (£(1,6)=7.4; sec Figure 3.4B). Conditioned Response Onset Latencies. Response onset latencies on both CS-alone trials and paired trials are presented in Figures 3.SA and 3.SB. Onset latencies dropped to adaptive levels with acquisition in both lesioned and unlesioned animals and were maintained post-lesion for the Retention group. Similar trends were observed for both CS-alone and paired trials. On CS-alone trials there was a significant effect for the Retention group between pre and post lesion latencies (£(1,6)=61.3) probably reflecting the adaptive learning process taking place pre-lesion and finer-tuned response in the overtrained animal post lesion. This is also reflected in the significant effect of training session (E(4,24) = 13.6) and interaction between lesion and session (E(4,24) = 13.1). 6 9 (/) E E IT c= a> co 500 450 400 350 300 250 200 150 1 0 0 50 0 CS-alone Trials L * a l o n « d Post-Acq L M l o n a d Prs-Acq 1 2 3 4 5 A cq u isition 1 2 3 4 5 R e te n tio n B 350 300 250 200 tu 150 .3 100 50 O Paired Trials L » i o n « d P i»n I Ai M — L a o i o n a u I 'r a - A c q 1 2 3 4 5 A c q u i s i t i o n 1 2 3 4 5 Ratantlan Figure 3.5 Motor Cortex lesions do not impair adaptation of response latency with acquisition or its maintenance during post-lesion retention training. (A) Onset Latencies for both Acquisition and Retention groups on CS-alone test trials. (B) Onset Latencies for both Acquisition and Retention groups on paired training trials. 70 Unconditioned Reflex Response Latencies. Response onset latencies on I/O tests changed over the course of training. As amplitudes increased (see Figure 3.4 A and B), it seems response were initiated earlier (see Figures 3.6 A and B). In terms of onset latency for the Acquisition group, there was a significant difference between I/O tests (E(l,6)=12.1). There was also a significant effect of US intensity (£(3,18)=4.9) reflecting the much slower response onset on both I/O tests for 1 psi relative to other intensities (see Figure 3.6A). An interesting pattern emerged in the Retention group (see Figure 3.6B). There was no significant difference between pre and post-lesion I/O tests overall, but there was a significant effect of I/O test (E (l,6 )= 17.2) representing the decrease from I/O test 1 to 2 (before and after acquisition) and from I/O test 3 to 4 (before and after retention). This reflects the amplitude enhancement occurring with continued training both before and after the lesion. As with the Acquisition group, there was a significant intensity effect (E(3,18)=7.5). Responses to 1 psi were significantly slower overall. There was no interaction between I/O test and intensity pre or post-lesion, but there was a significant interaction between US intensity and pre or post-lesion testing (E(3,18)— 4.7) which probably relates to the small decrease in onset latency at 1 psi with retention training as compared to previous acquisition tests. Histology. Surface measurements and sketches of the frontal cortical ablations were made and are reconstructed in Figure 3.7A. Even the severest lesions spared 5-7 mm of the most anterior frontal pole. The smallest AP extent was 13 mm and the largest was about 17.5 mm. The smallest ML aspect of the ablation spared some tissue along the midline surface, while the largest did not, and both smallest and largest lesions extended laterally 8-9 mm. 71 A 2 0 0 W E. u e < D 3 B 150 100 50 I/O 1 I/O 2 a pci 3 pci 2 pci 1 p& i 200 150 in E 100 I/O 1 I/O 2 I/O 3 I/O 4 4 pCi 3 pci 2 pci 1 pel Figure 3.6 Reflex response onset latencies on US-alone I/O tests at 4 different US intensities for Acquisition group (A) and Retention ^roup (B). Unconditioned reflex onset latencies decrease with training, and a rebound is observed in the retention group following the lesion and a no training recovery period. 72 A A _ 6 - 7 mm apared — r 1 3 mm (vm tlestl ♦ A S rrm Carpegl) Figure 3.7 (A) Surface reconstruction of the smallest (filled) and laigest (hatched) extent of lesion for both groups combined. (B) Actual photos of respresentative lesions in two animals, top and side views (black bar= 5 mm). 73 Cresyl stained coronal sections indicate the dorsal hippocampus remained intact in all cases, although swelling in the cavity of some animals may have exerted pressure causing some displacement or distortion. Reconstructions of minimal and maximal damage are presented for Acquisition and Retention groups separately in Figures 3.8 A and B, respectively. 3.1.3 Discussion: As had been expected, motor cortex lesions did not effect the acquisition or retention of the conditioned cycblink when animals were trained under standard delay conditions. There was an expectation, however, that there might be some changes in response topography. Neocortex is often described as having an inhibitory controlling role over the more primitive brain regions associated with instinctual and survival behaviors. As such, it might be expected that significantly larger response amplitudes would result from cortical lesions. This idea of an inhibitory role for motor cortex, specifically during learning, is supported by recordings from neurons during discrimination reversal training (Disterhoft and Segal, 1978). Rats were observed for anticipatory movements toward a pellet dispenser in response to a tone. One tonal frequency was rewarded (CS4-) while another was not (CS-). When the reward contingency was reversed, there was increased activity in motor cortex neurons corresponding to the decrease in movement associated with the tone switching from CS+ to CS-. As has already been described, the motor cortex can exert this inhibitory influence on the cerebellar learning circuit via its connections to the red nucleus, 74 Figure 3.8 On the following three pages are serial reconstructions of the common and largest extent of damage presented in the coronal plane from cresyl stained sections at 5 mm increments (labels relative to bregma) for Acquisition group (A) and Retention group (B). Normal brain sections with labels based on Paxinos and Watson (1982) and Shcnk ct al.(l986) arc available for comparison (C). Abbreviations: FrCtx«= Frontal Cortex, Mtr= Motor Cortex, SS= Somatosensory Cortex, Rh F>= Rhinal Fissure, LOT = Lateral Olfactory Tract, aCing= anterior Cingulatc Cortex, pCing= posterior Cingulate Cortex, d Hip and v Hip= dorsal and ventral Hippocampus, respectively, Iat vent= lateral ventricle, sub= subiculum, and hip= hippocampus. 75 -13 -8 * 3 +2 +7 76 ^ ^ ^ ^ F r C t x pontine nuclei, trigeminal nucleus, and lateral reticular nucleus (see Chapter 1). With all this intricate anatomical connectivity, it is interesting to note that in the present study there were no differences between animals with or without motor cortex during acquisition training on performance measures for either conditioned or reflexive responses. All I/O test results confirmed previous findings of amplitude enhancement with experience. For animals lesioned after initial acquisition training, a significant increase in overall post-lesion CR amplitude was observed, but this result is influenced by no or low CR amplitudes on initial training days. The only other disinhibitory type of effect observed for amplitude was an intensity specific increase on post-lesion I/O tests. Retention animals showed a significantly higher amplitude response to a 4 psi airpuff after lesions to motor cortex. Though this may just be a coincidence, it is also possible to speculate, that the disinhibitory influence of motor cortex might be more noticeable for larger motor movements involving cortico-spinal connections. As airpuff intensity is increased and becomes more aversive, some discomfort is evident in animals at the 4 psi intensity and evasive tactics evoke larger body movements. On the other hand, in a study of neural discharge patterns in motor cortex and red nucleus associated with skilled motor movements in monkeys, Martin and Ghez (1988) found both cell types to be related to initiation of movement, but motor cortical cells coded for direction of movement toward a target (see also Georgopolis, 1992) while red nucleus cells did not. The authors speculate that the red nucleus may be involved in the adaptive modification of motor patterns with changes in load or task requirements. The spatial nature of this type of motor task may be a determining factor of motor cortical modulation of movement and may explain why broader effects were not observed in the current experiments which did not incorporate a spatial component in the required motor response. The overall behavior of these animals was quite normal for the most part despite massive tissue damage. The only anomaly was what seemed like an inability to cease running once initiated. When an animal is retrieved from its home cage its initial response often, though not always, is an escape attempt which makes the animals hyperactive. In the lesioned rabbits this hyper-response was more extreme, though it never appeared aggressive or directed toward the researcher, and it took animals longer to calm down to normal. This, too, is consistent with motor cortex inhibition of larger body movements, but was specifically and only evident in the fight-or-flight situation. Though there was no direct septal damage in any animal which would facilitate aggressive emotional behavior, the destruction of anterior collosal fibers was extensive in many cases and may have disrupted limbic function. Additionally, other research (Buchanan et at., 198S) indicates that stimulation of medial frontal cortex produced autonomic cardiovascular responses in rabbit, a disruption of which may lead to the behavior observed in the present study. In sum, while the fine skilled eyeblink movement produced by cranial motor neurons is unaffected by motor cortex lesions, the inhibitory influence of this neocortical region may preferentially affect spinal motor movements or target- oriented responses. 8 0 3.2 Effects of Motor Cortex Lesions on Acquisition of Trace Eyeblink Conditioning Several of the studies which have suggested involvement of motor cortex in classical conditioning have used the trace paradigm (Woody & Yarowsky, 1972; Megirian, 1970), in which there is an intervening delay between the CS and US during which no stimuli are presented. This is consistent with the theory that more difficult, taxing, or complex learning paradigms involve structures above the level of the cerebellum. Specifically, the involvement of motor cortex in classical conditioning of a discrete motor movement has been demonstrated using the trace paradigm. Woody and Yarowsky (1972) have successfully trace conditioned cats using stimulation of motor cortex as a CS and a glabella tap as the US. In addition, Woody and Engel (1972) have identified changes in threshold and activity levels of neurons in motor cortex with conditioning. This study is designed to assess the effects of motor cortex lesions on eyeblink conditioning using a paradigm simitar to Woody et al. (1974) by using a 1 ms auditory click, instead of a tone, for the CS and by introducing a 500 ms trace interval between presentations of the CS and US. The only significant difference was in the onset latency of conditioned responses, though all performance measure appeared to be slightly impaired for the lesion group. 3.2.1 Methods: Subjects. Thirteen New Zealand White rabbits (Oryctolagus cuniculus) were included in this study. All procedures and animal care were in accordance 81 with NIH guidelines. Rabbits were housed individually in a room with a regulated 12 hr light/ 12 hr dark cycle. All animals were maintained on an ad lib food and water diet. Animal care was provided by the experimenter, staff and veterinarians of the University of Southern California. Protocol. Seven animals received bilateral ablations of frontal cortex, including motor areas, prior to any training (Lesion Group). Following one week of recovery, each animal received training using the trace paradigm with a click CS and airpuff US until they achieved 2 consecutive sessions at or above criterion performance (8 out of 9 consecutive CRs) with a maximum of 25 days of training. Another 6 rabbits served as untesioncd controls receiving the same behavioral training (Control Group). The Lesion group underwent reflex responsiveness testing. For each rabbit in the Lesion group, reflexive responses to airpuff intensities of 1, 2, 3, and 4 psi were measured on separate test days (I/O tests as described in Chapter 2) before and after acquisition. Trace conditioning sessions involved presentations of a 1 ms click CS followed by a 500 ms trace interval and a 100 ms airpuff US. Each daily session consisted of 108 trials, (12 blocks of 9 trials), just as with delay training in the previous experiment. Data Analysis. In the case of trace conditioning, on-line data was sampled for a 1250 ms trial length. The baseline period was 250 ms, CS-period 500 ms, and US-period 500 ms. Conditioned responses and bad trials were detected as described earlier. Lesion and control animals were compared on the basis of trials-to criterion using an independent t-test. Also, measures such as percent CRs, UR and CR amplitude, and onset latency were compared across first, middle and criterion days of training and between groups using a between groups repeated measured 82 ANOVA. These points of training were chosen to standardize comparisons across animals with widely varying learning rates. Surecrv. Each rabbit in the lesion group underwent bilateral motor cortical ablation as described previously. Histology. Histological procedures were identical to those for frontal cortex lesioncd animals in the delay conditioning experiment. 3.2.2 Results: Both lesioncd (Lesion Group) and unlcsioned (Control group) animals acquired the trace conditioned eyeblink with some difficulty. One animal in the Lesion group never reached the criterion of 8 out 9 consecutive CRs and was dropped after 25 days of training. One animal in each group reached criterion once, but not twice, and training was discontinued after 20 days. All others reached criterion at least twice and training was terminated at that point. The average trials*to-first criterion for the Lesion group was 1023 d t 270 and 900 ± . 252 for the Control group. Though the learning rates were not significantly different for the two groups, the Lesioned animals tended to be slower learners. This non-significant trend toward poorer performance for the Lesion group was observed consistently on all measures to follow. Percent Conditioned Responses. Control animals achieved 59.1 ± . 6.0 percent conditioned responses on the first criterion day, whereas animals in the Lesion group were only at 40.4 ± . 9.8 percent (see Figure 3.9). The groups were not, however, significantly different. There was a significant effect for training 83 % CRs 70 60 50 40 30 20 10 0 A C Q 1 MID CRIT - CONTROL <N-S) ~ LESION (N-7J Training Day Figure 3.9 Percent conditioned responses on CS-alone test trials. Though lesion group seems impaired, the difference is not statistically significant. 8 4 day (£(2,20)=41.6) indicating a considerable increase in conditioned responses, but there was no interaction between group and training day. Conditioned Response Amplitude. Conditioned response amplitudes on both CS-alone and paired trials increased significantly during the course of training (E(2,20) *=9.9 and 9.8, respectively). There was no significant difference between groups, though the lesioned animals performed slightly worse (see Figures 3.10 A and B). There was also no interaction effect. Unconditioned Reflex Responses. I/O tests were performed before and after training for the Lesion group and the characteristic increase with experience was observed. There was a significant increase for all US intensities post-training (E(I,6)=9.2). There was no difference in response amplitude to any particular airpuff intensity and no interaction effect (sec Figure 3.11 A). Response amplitudes during the US-period following airpuff presentation are shown for both groups of animals in Figure 3.1 IB for group comparison. Lesioned animals were not significantly impaired. There was a significant effect of training day (£(2,20) = 10.7) as response amplitudes increased with training. Conditioned Response Onset Latencies. These data are presented in Figure 3 .12A. Onset latency on CS-alone trials during trace conditioning was the only measure significantly affected by motor cortex lesions (E (l,10)= 6.5). Lesioned animals showed longer latency responses than controls. Onset latency on paired trials was not significantly different between Lesion and Control groups (see Figure 3.12B). Unconditioned Reflex Response Latencies. Response onset latencies on I/O tests changed over the course of training. As amplitudes increased (see Figure 3.11 A), responses were initiated earlier (see Figure 3.13). I/O 1 and I/O 2 were 85 B e E, ® ■ a 3 2= a. E < E E a T J a. E < 5 4 3 — CONTROL (N-6) LESION (N-7) 2 1 0 ACQ1 MID CRIT Training Day 5 4 3 2 1 0 ACQ1 M ID CRIT CONTROL < M « ) LeaiON w-tt T raining Day Figure 3.10 Conditioned response amplitudes on (A) CS-alone test trials and (B) on paired click-airpuff trials. The main effect for group difference was not statistically significant. 86 E £ £ 3 "5 . E < 12 8 4 0 I/O 1 I/O 2 ■ 4 psl • 3 psi A 2 psi ♦ 1 psl I/O T e st B E E a ■a 3 a. E < 12 10 8 6 4 2 0 ACQ1 MID CRIT — CONTROL (M A) teaioN (N -/J Training Day Figure 3.11 (A) Unconditioned response amplitudes to 4 different airpuff intensities for the lesion group on separate I/O test days before and after classical conditioning. (B) Response amplitudes for lesion and control groups measured during the trial period following US presentation on paired trials of trace conditioning training (Note absolute response amplitude is influenced by CR amplitude, if such a responsewas taking place.) 87 to E. >. o c © a B n E, >i u c u < * - * J 3 800 700 600 500 400 * slgnilicanl m.«. of group and ol day 300 200 100 CRIT MID ACQ1 C ON TRO L C N -W LESION (N -7 ) Training Day 800 700 no tlgnificint difference 600 500 400 300 200 100 ACQ1 CRIT CON TRO L (N « ) LESION <rwj Training Day Figure 3.12 Onset latencies for the first movement to exceed 0.S mm on (A) CS-alone test trials and (B) paired click-aiipuff trials. The averaged onset latencies reported above include non-CR trials (or < .05 mm) where the maximum trial length was assigned as the onset value. 88 E E, © T D Q. E < 200 160 120 8 0 40 0 I/O 1 I/O 2 4 psi 3 psi 2 psl 1 psi I/O Test Figure 3.13 Lesion group onset latencies for unconditioned reflex responses at 4 different airpuff intensities and on separate I/O test days before and after trace conditioning. 89 significantly different (£(1,6) =9.0). A significant main effect of intensity (£(3,18) = 11.6) was observed reflecting slower response latencies for 1 psi as compared to 2, 3, and 4 psi combined (£(1,6) — 19.9). Histology. Lesions were comparable to those made in the previous experiment on delay conditioning. Cresyl stained coronal sections indicate the dorsal hippocampus remained intact in all cases, although swelling in the cavity of some animals may have exerted pressure causing some displacement or distortion. Reconstructions of minimal and maximal damage are presented in Figure 3.14. 3.2.3 Discussion: Although deccrcbration studies (Mauk and Thompson, 1987; Norman et al., 1977) have demonstrated that structures above the level of the cerebellum are not critical or essential for the acquisition of standard classical conditioning, other studies have indicated a modulatory role for these structures. In some cases, the involvement of higher brain structures is necessary for optimal learning and performance. Neurons in auditory regions of thalamus shift their optimal response to match the frequency of a auditory conditioning stimulus (Bakin and Weinberger, 1990 and Diamond and Weinberger, 1986 for example). Neural models of conditioned behaviors (eyeblink, jaw movement, hindlimb flexion) have been recorded from hippocampus (Berger and Thompson, 1978, Berry and Oliver, 1982, and Thompson et al., 1980, respectively) and stimulation of hippocampus facilitates eyeblink conditioning (Prokasy et al., 1983). Further, the hippocampus has been specifically implicated in complex discrimination reversal tasks. Animals with damage to dorsal hippocampus bilaterally were impaired in reversal learning 90 Figure 3.14 Coronal sections reconstructing the common (solid) and largest extent (shaded) of cortical lesions in 7 animals. Refer to Figure 3.8C for corresponding normal sections. 9 1 (Berger and Orr, 1983). In addition, dorsal hippocampal lesions altered performance for trace classical conditioning. Animals were massively impaired in conditioned responding following tone and airpuff pairings separated by as much as a 750 ms silent period. When the hippocampectomized animals did respond to the CS, the distinctive feature of these responses was a short onset occurring 600 ms prior to presentation of the US (Solomon et al., 1986; Port et al., 1986). These results have led to a theory that the hippocampus has a role in timing conditioned behavior. Another more recent study (Kim et al., in press) has shown that hippocampectomy impairs the memory of recently, but not remotely, acquired trace eyeblink conditioned responses. Animals trained with a 500 ms trace interval and lesioned 1 day after reaching criterion did not retain conditioned responding when tested post-lesion. However, animals lesioned 1 month following criterion showed perfect retention. There is still some disagreement in the literature as to whether the role of the hippocampus is modulatory or essential to bridging the temporal gap between the CS and US during acquisition (Akase et at., 1989; Moyer et al, 1990). Distcrhoft and associates (Moyer et al., 1990) suggest that complete lesions of both dorsal and ventral hippocampus are necessary to observe an acquisition deficit with a 500 ms trace period. On the other hand, lesioned animals can condition with a 300 ms trace, but have difficulty extinguishing the behavior with extinction training. This effect resembles the deficit Berger and Orr (1983) observed in reversal learning which required a change in conditioned behavior from one CS to another, previously unreinforced stimulus. In fact, damage to retrosplenial cortex, which links the hippocampus to both cerebellum and motor cortex (Bassett, 1987), was sufficient to cause similar impairment (Berger et al., 1986). Hippocampal information is transmitted to the 9 3 retrosplenial cortex via the subiculum and is linked to cerebellum by way of projections to the ventral pontine nuclei (see Figure 3.15). Other projections from retrosplenial cortex synapse in the ventrolateral thalamus which feeds into motor cortex and can also influence cerebellar circuits (see Chapter 1.4.1 : Motor Cortex). This connection is particularly interesting because, following lesions of motor cortex. Woody and associates find deficits in cats trained to a tone CS and glabella tap US with a 500 ms trace interval between (Woody et al., 1974; Woody, 1984). The animals arc unable to acquire a conditioned response identified as a short-Iatcncy (20 ms) electromyographic (EMG) response in eyelid muscles to the tone CS. Woody and Yarowsky (1972), further, have shown that stimulation of motor cortex can act as a conditioned stimulus and recordings from neurons in the cortex exhibit Icaming>rclatcd activity changes. These data have lead Woody to suggest that motor cortex is essential for classical conditioning. As the experiments in this chapter have demonstrated, motor cortex is not essential for conditioning of the adaptive long-latency eyeblink response in rabbits trained either with a delay paradigm or using a 500 ms trace procedure. Bilateral frontal lesions did slightly, but not significantly, impair overall performance on the trace conditioning task. Onset latencies were the only measure for which a statistically significant difference was observed. The delay in response onset for lesioned animals may have been an artifact of lower conditioned response rates and, therefore, the assignment of a maximum latency value. However, performance of lesioned animals was consistently impaired relative to normal controls. There were no dramatic impairments in the timing of conditioned behavior, such as those associated with hippocampal damage or disruption of 9 4 Globus Pallidus C audate Nucleus Cerebellum Red Nucleus Motor Cortex Subiculum Hippocampus Motor Neurons Retrosplenial Cortex Ventrolateral Thalamic Nucleus Ventral Pontine Nuclei Figure 3.15 Anatomical efferents from hippocampus. Shaded areas are links to eyeblink response behavior (adapted from Bassett, 1987). Two of seven lesion animals in this study sustained minimal damage to retrospenial cortex. hippocampal efferents. Of the seven animals in the trace conditioning lesion group, four animals did sustain minimal damage to anterior cingulate cortex, 2 unilateral (92-285 and 92-304) and 2 bilateral (92-231 and 92-286). One of these animals also received partial damage to dorsal anterior retrospenial cortex bilaterally (92-231) while one other animal had unilateral damage to anterior retrosplenium sparing a significant portion of cingulate cortex (see Figure 3.16). Unconditioned response performance on I/O tests before and after trace conditioning followed the same pattern of enhancement over the course of training as observed in the preceding experiments. Unfortunately, I/O data were not collected for normal subjects in this experiment, but a comparison to normal subjects in the delay conditioning task (see Figure 3.4) suggests that there is a possibility that motor cortex lesions facilitated UR enhancement from I/O 1 to I/O 2. These data are interesting on at least two counts. One, animals with motor cortex lesions in the previous experiment exhibited an overall increase in response amplitude following frontal lesions. Two, if this effect were to be significant, it would provide further evidence for cortical modulation of behavior. The present study does not support motor cortex as a critical structure for acquisition of the conditioned eyeblink in rabbits. As for its role as a modulator of motor behavior, reflexive behavior of lesioned animals was consistent with previous patterns over the course of training (though comparable data was not available for control animals trained with trace conditioning). Conditioned motor behaviors exhibited slight, but highly consistent, impairments which may warrant further study before a modulatory role for motor cortex in classical conditioning is ruled out completely. 96 Cortex _ Corpus Callosum Hippocampus Figure 3.16 (A and B) The area of the cerebral cortex called the retrosplenial cortex, after its location dorsal to the splenium, which is believed to extend posteriorly about 9 mm (from Bassett, 1987). PCG = precentral gyrus, ACG = anterior cingulate gyrus, Cng— cingulate cortex, Rsp— retrosplenial cortex, Ps=postsubiculum. (C) Lesion reconstructions at + 2 mm relative to bragma for the animals in this study which sustained some damage in the cingulate/retrospenial region. No animal damage was observed at the next posterior section (+7). 97 CHAPTER 4 Factors Influencing Unconditioned Reflex Performance in Rabbit Reflexive behavior is often thought of as instinctual in nature, a “necessary reaction to some external stimulus” (Woody, 1982), and, therefore, critical in some way to the survival or function of an organism. For this reason, changes in reflex behavior are viewed as significant and believed to be adaptive for the organism, otherwise change is presumably unwarranted. The adaptation of reflexes can be divided into three categories according to Woody (1982): “those that are induced by simple stimulus presentation, those induced by repetitive stimulus presentation, and those induced associativcly by the presentation of different stimuli...in temporal relationship to each other." The changes that occur may be either facilitory or inhibitory on the behavior. First, the simple adaptive reflex is a transient change to a single stimulus which alters immediately subsequent perception or response to a stimulus, but is not maintained beyond a test session. At the other end of the spectrum, associative adaptive reflexes are long lasting and involve repeated pairings of different stimuli such as with classical conditioning (adaptive “reflex" being a slight misnomer in this case since reflexive and conditioned behaviors can be different behaviors as well as separate processes, as discussed in earlier chapters). Finally, adaptation to repeated presentations of the same stimulus have been labeled sensitization (facilitory) and habituation (inhibitory). While these processes don’t require association, in the classic sense, 9 8 they are long-lasting and there is speculation that they may form the building blocks of classical conditioning (Kimmel, 1973; Hawkins and Kandel, 1984). The series of experiments in this chapter generally address the properties of and changes occurring in the rabbit's unconditioned reflexive eyeblink response over time under specific conditions of training. As reported in previous studies, input/ output tests of reflex responsiveness to 4 different airpuff intensities reveal that, in addition to the conditioned response, the unconditioned response undergoes changes over the course of training and repeated stimulus presentations. While a facilitory change in eyeblink amplitude is observed on separate test sessions across training, inhibitory changes can be observed within a test session (Kimble and Ost, 1961; Weiss, personal communication). The first of the following experiments addresses stimulus conditions which.result in habituation to a peripheral airpuff stimulus. In order to better understand the nature of reflexive changes during eyeblink conditioning and the conditions which facilitate them, an additional study of electrical brain stimulation parameters effecting evoked behavior was undertaken in the second experiment. Finally, data are presented from a preliminary study of cerebellar stimulation during conditioned response performance. Together, these studies characterize situational changes which influence reflex behavior. 4.1 A Systematic Study of Habitation to a Peripheral Airpuff Presented at Different Inter-Stimulus Intervals Classical conditioning of both animals (see Gormezano et al., 1983; Thompson, 1990) and humans (Hilgard, 1931; Solomon, 1989) commonly uses 9 9 peripheral airpuff stimulation as an unconditioned stimulus which reliably evokes an eyeblink reflex. Sensory information regarding the airpuff is received by the trigeminal nucleus and then transmitted directly to motor nuclei as well as indirectly, via cerebellar structures, for possible integration with other sensory information. Since the airpuff-evoked response changes with experience (see Chapter 2), there must be alterations in the sensory or motor processes preceding behavioral output. In order to approach an understanding of the neural mechanism(s) mediating changes in the airpuff-evoked eyeblink, the parameters necessary for inducing modulation of that response must be elaborated. Recent preliminary developments in our laboratory suggested that animals receiving stimulation of the pars oralis region of the trigeminal nucleus as a US during conditioning show a profound reduction (>90% ) in UR amplitude to repeated stimulation at a single intensity by the end of one training session. Such a response decrement following repeated stimulation is known as habituation. Rate and degree of habituation are positively related to frequency of stimulation (see Thompson & Spencer (1966) for review). In contrast to the trigeminal stimulation study, habituation is not generally observed to such an extent under normal training conditions with a peripheral airpuff US. Information about peripheral stimulus frequencies which produce effective habituation of the rabbit eyeblink is not available. The following experiment involved systematic manipulation of interstimulus intervals between airpuff presentations to assess the rate and degree of habituation of the rabbit eyeblink response to an airpuff intensity used in standard classical conditioning sessions. The standard training airpuff intensity (3 psi) which reliably evokes an eyeblink was used. In addition, there is evidence in the habituation too literature that weaker repeated stimulus presentations produce more rapid habituation than stronger intensity stimuli (see Thompson & Spencer (1966) for review). A comparison, therefore, was included between the training intensity airpuff at 3 psi and a much weaker, but still reliable (see Chapter 2.2), 1 psi intensity. 4.1.1 Methods: Subjects. The subjects in this study were SO (10 per group) New Zealand White rabbits (Oryctolagus cuniculus). All procedures and animal care were in accordance with NIH guidelines. Rabbits were housed individually in a room with a regulated 12 hr light/ 12 hr dark cycle. All animals were maintained on an ad lib food and water diet. Animal care was provided by the experimenter, staff and veterinarians of the University of Southern California. Protocol. Animals were divided into 5 groups of 10 animals. Each group received airpuff alone trials presented at different fixed inter-trial intervals (ITI; 60, 30, 10, 5, and 1 sec). Under normal training conditions ITI is varied randomly between 20 and 40 s, averaging 30 s. A blank trial was presented at the beginning of each 9-trial block in order to test whether the animals learned to predict the occurrence of the next stimulus. Each animal participated in 4 habituation sessions on separate days each consisting of 96 airpuff-alone trials and 12 blank trials (12 blocks of 9 trials for a total of 108 trials). This is the same number of airpuffs presented during a normal conditioning session. Individual trial length was 1000 ms with a 98 ms airpuff presented at 500 ms. The longest session, therefore, lasted 108 minutes and the shortest session was 108 seconds. tot Each animal was habituated for two consecutive days at each of two different airpuff intensities. There was a one week break between the different intensities. One pair of habituation sessions was at 3 psi and the other at 1 psi. Within each ITI group, half of the animals were habituated to 3 psi first and the other half to 1 psi first. On-line Data Analysis. On-line data collection sampled NM movement every 2 ms during the 1000 ms epoch comprising each trial. The first 250 ms made up a baseline period during which no stimuli were presented. Following the baseline period was another stimulus-free 248 ms epoch referred to as the CS period. A response within this period would be predictive of the subsequent airpuff US. Data collection continued for another 500 ms following US presentation (referred to as the US-period). Bad trials were identified by movements exceeding 0.7 mm in the baseline period or 0.5 mm within the CS period. In addition, non-responses were defined as bad trials and excluded from subsequent data summaries. Onset latency was defined as the first time within a trial that NM movement exceeded 0.5 mm. Statistical Analyses. Statistical analyses were performed using the CSS/Statistica software package and a significance criterion of p < .05 (non significant results are not reported). The data were arranged so that both order of US intensity received (order) and inter-stimulus interval group (group) were treated as independent variables. Response frequency, amplitude, peak latency and time- amplitude area were dependent repeated measures. The five ITI groups were compared across stimulus intensity, habituation days, and 12 blocks of habituating trials using a 2 (psi) x 2 (day) x 12 (block) between groups repeated measures ANOVA. 102 For the ease of presenting data effectively, block data were combined into two 6-block segments per day for several of the measures. Because these coarser segments can be less sensitive to change over time (see Petrinovich and Widaman, 1984), it was then necessary to confirm these data reflected statistical significance o f the 12-block data. A 2 (psi) x 2 (day) x 2 (segment) between groups ANOVA was used. Since the number of statistical tests performed on a single data set can affect experimentwise error it is important to note that the primary analysis reported is that using the more sensitive 12-block division and that the second analysis was only performed to confirm that data presented in coarser segments still reflected the significant changes being discussed. Petrinovich and Widaman (1984) also report that the large degrees of freedom resulting from the use of many small blocks of data may produce a large number of statistical differences which actually account for very little of the variance in the data. The differences between the two analyses and their relevance are discussed, as necessary. Response amplitude was also converted to a percentage of the amplitude in the first block at a given US intensity where block 1 was set to equal 100% (scaled response amplitude). This was a more accurate comparison of degree of habituation between animals compensating for individual differences in absolute response amplitude. 4.1.2 Results: Response Frequency. A significant difference was observed in the number of responses made to 1 versus 3 psi stimulus intensity (£(1,40) = 18.6). Fortunately, response rates for both intensities were high enough to provide 1 0 3 sufficient data for comparisons on other measures. At an airpuff intensity of 1 psi, rabbits responded 90.2 ±. 1.7 % of the time as compared to 96.8 ± , .6 % for an airpuff at 3 psi. The percentage of responses timed to precede normal airpuff presentation was recorded on test trials in which no stimulus was present. No more than four animals per groups showed any such responses and for those which did the percentage never exceeded 16.7% (2 out of 12 test trials). The fixed inter-trial interval did not seem to facilitate anticipatory responding. The incidence of anticipatory responding was higher at 3 psi (17/50 animals) than at 1 psi (6/50) and suggests a higher spontaneous blink rale at the higher intensity. Absolute Response Amplitude. There was no significant effect of order, group or habituation day for absolute eyeblink amplitude. There were, however, significant main effects of intensity (psi; £(1,40)=39.2) and block (E(l 1,440)= 18.4). Responses to 3 psi were larger (4.4 ±. .4 mm) on average than responses to I psi (2.3 ±. .3 mm). The block effect reflects the drop in UR amplitudes across a one-day session. These data are presented as 6-block (half session) averages in Figure 4.1. Another effect which comes to light in this figure but is also reconstructed in Figure 4.2 is that there is not only spontaneous recovery at the beginning of day two, but response amplitudes are even higher than the original response. This leads to greater habituation on day 2 than on day 1 (day x block: E(11,440)=4.1) of habituation. These results were all consistent between the 12-block and 2-segment analyses. There were several other significant interactions in the 12-block analysis, all involving a change across block: (a) order x block, (b) psi x block, group x psi x block, and (d) order x psi x day x block. Only the 4-way interaction was 1 0 4 Absolute U R Amplitude (mm) 6 5 4 Absolute R esponse Amplitudes B D1 D2 1 p s i D1 D2 3 p si p si B lock s 1*6 EZZl B locks 7*12 [ = □ m e a n 1 p si m e a n 3 p si Figure 4.1 (A) Absolute response amplitudes by intensity and habituation day. (B) Average amplitude overall for each intensity. 105 Absolute U R Amplitude (mm) Absolute R esponse Amplitudes day x block interaction 6 5 4 3 2 1 0 b1*b6 b 7 -b 1 2 b 1 -b 6 b 7 -b 1 2 1 p ci 3 P 6 * Figure 4.2 Absolute response amplitudes comparing day 1 and day 2 of habituation between the two airpuff intensities. replicated in the 2-segment analysis and it seemed more appropriate to represent these results by percent change in amplitude as reported in the following section rather than by raw absolute amplitude. Scaled Response Amplitude. All significant effects observed with the 12- block analysis were replicated in the subsequent 2-segment analysis. A significant main effect was observed for the order of US intensity (E( 1,40)=5.3). Rabbits exposed to 1 psi first showed more often showed increases in response amplitude across a session (sensitization) whereas animals presented with 3 psi first seemed to show greater overall habituation (see Figure 4.3). A significant main effect of day (E (M 0)«5.2) reflected the general increase in response amplitude from day 1 to day 2 at cither training intensity (Figure 4.3). A significant effect of block (E( 11 v 440)=6.1) represents changes across a session without specifics to direction of change, which varies. This variation is picked up by a highly significant day x block interaction (£(11,440) =7.4). Figure 4.3 shows that in 3 out of 4 cases on day 1 there was an increase in response amplitude whereas day 2 in all cases exhibited a decrease in response amplitude. The only animals to show habituation on day 1 rather than sensitization were those initially exposed to 3 psi on the day when they were tested at 1 psi (order x psi x day x block: E O 1.440) »2.1; see also time-amplitude area). Though there were no significant group differences with intensity (psi) or day, there was a significant effect for group x block (E(44,440)= 1.6). As seen in Figure 4.4, this would appear to be a result of the greater degree of overall habituation observed in the 1 s.ITI group as compared to others which were much more variable as to whether there was habituation or sensitization within any given session, if any change at all (see Figure 4.5 for individual group performances). 1 0 7 Scaled Response Amplitudes < D •o 3 "5. E < D 1 D2 D 1 02 1 psi 3 psi 1 -> 3 p si 01 D2 D 1 d: 1 psi 3 psi 3 ->1 p s i Blocks 1*6 E Z Z 3 Blocks 7-12 Figure 4.3 Scaled response amplitudes as a percentage of first block at a given intensity. A significant order effect is clearly evident. Effects of day, and session segment are also represented. 108 3 % Amplitude Scaled Response Amplitudes group x segm ent interaction -A- - o - 1 6 ITI 56 ITI 106 ITI 306 ITI 6 0 s ITI bi -b6 b7-b12 Figure 4.4 Scaled response amplitudes combined across day and intensity to examine group by sessionsegment interaction. 1 0 9 (uiui) epmi|duiy (tutu) epniijdury (tutu) epnjijdwy (uui) •patgduiv (uiut) epnpiduiy Day 1 blocks 1*12 Day 2 blocks 13*24 1 * ITI e i p » < a p n ftla th 1 B • ITI t pal S pal 1 10 a ITI B • • 10 IB IB 10 IT IB SI BS i pal a pal i ao a ITI 1 pal a pal 1 1 pal a pal Figure 4.5 Absolute response amplitudes by ITI group and airpuff intensity further describe the group x block effect on percent amplitudes. Another group difference illustrated in Figure 4.4 is that animals exposed to stimuli separated by 1, 30 and 60 s actually sensitized within the first blocks (% amplitude greater than 100%) while the S and 10 s ITI groups remained rather steady. Peak Latencies. The latency at which animals' eyeblink responses peaked were significantly affected by airpuff intensity (E (l,40)= 12.8). The peak latency at 1 psi was 147.6 ± . 4.1 ms, but occurred a little sooner, at 133.7 ;fc 3.5 ms, for 3 psi. In addition, there was a significant group by block effect (£(44,440) = 1.5) which was lost when data were combined into 6-block segments. Since this effect excludes any influence of order, intensity, or day, and no other higher order interaction were significant, data were combined across these conditions for each of the 5 groups and are presented in Figure 4.6. For the 1 s ITI group there was a trend toward increasing latencies in the middle blocks corresponding to the general sensitization observed in Figure 4.4 which is followed by a decrease as habituation processes set in. Peak latencies for 5, 10, and 60 s ITI groups were rather stable across session. In contrast, the 30s ITI group seemed to start out with much longer latencies and stabilize by block 6. A slight non-significant trend of increasing latency with increasing ITI may be noticeable. Time-Amplimde Area. This measure is heavily dependent on response amplitude and many of the effects parallel those discussed earlier. Significant main effects o f US intensity and block were obtained (£(1,40)<=35.5 and E(11,440) = 10.5, respectively). The area of the UR to 3 psi was much greater than that produced by 1 psi (see Figure 4.7). On most days there was a dramatic reduction in response area across a session, but there was very little change observed for day 1 at 3 psi (significant day x block: E(11,44)=3.9). Because of 111 Peak Latency (ms) Peak Latency group x block interaction 170 160 150 140 130 120 1 2 3 4 5 6 7 8 0 10 11 12 Block Figure 4.6 Peak latencies averages across intensity and day in order to address significant group x block interaction. ■ — A— 1* ITI — 5s ITI —O - 10s ITI —V— 30s ITI — 60s ITI 112 Measure of Response Area 0 0 k- < 0 0 E i~ O C O SCO Blocks 1*6 Blocks 7*12 1psl* moon 3psi* mean D 1 D2 D 1 D2 1 psi 3 psi Figure 4.7 (A) Measure of response area averaged across all ITI groups. Eyeblink time-amplitude area was significantly larger at 3 psi as compared to 1 psi and reflected habituation within a a single day's session (D1 = day 1, D2*= day 2). (B) Mean area for all habituation sessions averaged to single out the significant intensity effect. 113 this, overall habituation to 1 psi was greater than to 3 psi as represented by a i . significant psi x block interaction reconstructed slightly differently in Figure 4.8. Significant higher order interactions included order x psi x day (E(l,40) = 5.8) and order x psi x day x block (E(l 1,440)=1.9) represented in Figure 4.9. Basically, rabbits exposed to 1 psi first exhibited significant sensitization to that intensity on day two and particularly in the first half of the session, but less so on day two of 3 psi habituation. The opposite held for rabbits exposed to 3 psi first. They showed profound sensitization on day two at that intensity also in the first half, but less on day two of 1 psi training. Interestingly, even when scaled response amplitudes (Figure 4.3) indicated sensitization, time-amplitude area decreased slightly. This suggests that sensitization causes larger but quicker responses. An extremely complex order x group x day x block interaction was obtained (£(44,440)»1.4) indicative of the many factors influencing response production. An attempt to make the data accessible for interpretation is presented in Figure 4.10. 4.1,3 Discussion: Thorough investigation of habituation of the spinal reflex in cat has led to a widely accepted 9 point operational definition of the habituation process (Thompson and Spencer, 1966; see also Thompson and Glanzman, 1976 for more recent review) based on changes in response size or amplitude. The following is a brief discussion of the data from the present study in the context of these points closing with a summary of the additional topographic measures reported above. 114 UR Area psi x block interaction a © * • » < © ■ © Z 3 o L E < © £ P cc 3 4 50 425 4 00 375 350 325 300 275 250 225 200 175 150 125 100 1 10 1psl 3 psi Block Figure 4.8 Response time-amplitude area exhibited a decrease paralling that of response amplitude over the course of habituation. The weaker 1 psi stimulus yielded greater habituation as reflected by response area. 115 UR Area order x psi x day x block interaction (9 O < a > o E 1 = [ 01 D2 D1 D2 1 p s i 3 p s i 1 - > 3 p s i D2 01 1 p s i 3 p si 3 -> 1 p s i ■ ■ B lo c k s 1*6 T 7 Z X B lo c k s 7 -1 2 Figure 4.9 Time-amplitude area's dependence on absolute amplitude is evindent by the similarities and diferences between this graph and Figure 4.3. 1 1 6 3 Day 1 blocks 1*12 Day 2 blocks 13-24 2 = a . e « E • o •) * E E u a E 4 u 1 ■ I u E Is III 2 4 0 0 1 0 1 2 1 3 1 5 1 7 1 9 2 1 2 3 2 4 O 0 1 0 1 2 1 3 1 5 1 7 1 9 2 1 2 3 1»>3 psi J -» 1 p s i S« ITI 1 - * - J p s i a*»i pw ID s T Q 1 0 1 2 1 3 1 5 1 7 1 9 2 1 2 3 1O O 0 3 0 s ITI 1 .-J p»| a-fi psi 1 - > - 4 p s i 3>:<l psi 2 4 6 0 1 0 1 2 1 3 1 6 1 7 1 9 2 1 2 3 1000 6 0 s ITI 1-<'J psi a-:-1 pM 2 4 0 0 1 0 1 2 1 3 1 5 1 7 1 9 2 1 2 3 Figure 4.10 Time-amplitude area combined across airpuff intensities but displayed by ITI group and order of airpuff intensity presentation. The first three characteristics of habituation can be addressed together. First, Thompson and Spencer (1966) list the finding that a stimulus which elicits a behavioral response will lead to an exponential decrease of the response with repeated presentations. The second point is that when the habituating stimulus is withheld for a period of time, the response will show spontaneous recovery. Third, if a series of habituation training sessions with intervening spontaneous recovery are given, subsequent habituation sessions are increasingly more rapid (potentiation of habituation). Reviewing the results of the current study, a smooth and steady decline was not always observed within a single habituation session, but a general overall decline from the beginning to end of a habituation session was observed when looking at absolute response measures (sec Figure 4.2) supporting point one. However, habituation on day two was always more robust and representative of Thompson and Spencer’s (1966) third point. In addition, spontaneous recovery (point #2) between sessions was clearly evident. In fact, not only did response amplitude recover between the first and second habituation sessions, but responses were even bigger at the beginning of day two. This was especially significant for training to the stronger 3 psi intensity. This result will be discussed further in relation to point five. The fourth point on the list explains that stimulation frequency and degree of habituation are positively correlated (see also Groves et al., 1969). The higher the frequency of stimulation, the greater the degree of habituation. Note, since the studies on which this premise is based involved electrical stimulation, the corresponding manipulation for airpuff frequency is actually the rate of airpuff presentations. Data representative of stimulus frequency and habituation rate, its regardless of order are best depicted in Figure 4.5. The greatest change in response amplitude over 12 block habituation sessions was indeed observed in the highest frequency group, animals presented with airpuffs 1 s apart. A smaller decrease was observed across blocks for animals trained at 5, 10 and 30 s ITIs, though they did not differ from each other. The lowest frequency group (60 s ITI) did not show significant habituation, but rather exhibited sensitization. A closer look at the data (Figure 4.5 and Figure 4.10) indicates that there was a small degree of habituation on day two for this group, but it was overshadowed by no change on day one. In Woody's (1982) review of the literature on adaptive reflexes, he describes a facilitory phenomena following repetitive stimulation which may apply to this situation. The observations, first labeled by Ukhtomsky as "reflex dominance" and also known as "latent facilitation", involved an increase in motor response with the slow repetition of a strong stimulus (see day 1 at 3 psi vs. 1 psi for the 60s ITI group in Figure 4.7). Reflex dominance was named such because Ukhtomsky viewed the facilitory response as becoming generally dominant so that it can be elicited by a variety of stimuli not normally effective. According to Woody (1982), such latent facilitation is observed following glabella tap stimulation. The fifth general observation of habituation is that weaker stimuli yield stronger habituation, whereas strong stimuli may not lead to any habituation. In the present study, how much habituation was observed at a given intensity was influenced by the order in which the two intensities were presented and day of habituation, but combining the data across these categories indicates that both stimuli resulted in habituation and, true to the principle being discussed, habituation to the weaker stimulus was greater. This result is reflected in the time- 119 amplitude measure presented across a session’s 12 training blocks for each stimulus intensity. Sixth, habituation training may proceed beyond a zero response level. This point refers to training continued even when a response has disappeared or stabilized. In such cases, Thompson and Spencer (1966) point out, recovery is slower. This point cannot be confirmed with the current data set. Significant recovery was reliably observed, and the persistent response enhancement observed across days (consistent with the enhancement observed on input/output tests over the course of training in previous experiments) never allowed for this type of maximal habituation. However, unpublished observations (Nordholm) of animal responsiveness given massed trials of airpuffs at 9 s ITIs was severely reduced in number, and some animals were observed to stop blinking completely. Clearly, this was not the case in the present study where response rate was at least 90%. Seventh, stimulus generalization may be observed in habituation. Though this phenomena often relates to stimuli affecting different modalities, the importance is that the stimulus be perceptively different. This could certainly be true for the 1 vs. 3 psi stimuli used in the current study. Actually, stimulus generalization appeared more effective in animals first exposed to the higher intensity airpuff (3 psi). They showed better habituation to 1 psi than animals first exposed to 1 psi. On the other hand, animals first exposed to 1 psi did not readily transfer habituation to 3 psi. Adaptively this would seem to make sense. If an organisms has adjusted to a strong aversive stimulus such as the airpuff, it is unlikely that any weaker stimulus coutd bring harm, and it becomes less significant. In fact, the eighth characteristic listed by Thompson and Spencer (1966) states that the presentation of another stimulus (usually strong) results in 120 dishabituation, or response recovery from habituation. This usually relates to a strong stimulus presented unexpectedly within a habituation session; However, this phenomena certainly could be applied to the lack of transfer of habituation observed between 1 and 3 psi in animals habituated to the weaker stimulus first. Finally, point nine states that repeatedly presenting a dishabituating stimulus will result in habituation to that stimulus. Though half of the animals in the current experiment experienced a shift from 1 to 3 psi US intensity, this did not occur within a single habituating session and, as explained above does not specifically relate to the present point. However, given that the dishabituation phenomena might be applicable to these animals even across sessions, the subsequent, though latent, habituation observed on day two at 3 psi would be consistent with this final criteria. That both habituation and sensitization processes are observed with the extensive manipulations performed in the current study is not unusual. In fact, Thompson el al. (1973) described habituation as a dual process involving a balance between the independent influences of habituation and sensitization. The application of this theory to the phenomena of dishabituation is an illustrative example. The authors suggest that dishabituation is not a disruption of habituation, but rather a superimposition of sensitization or facilitation (see also Thompson and Glanzman, 1976, Figure 7). So, what they were actually describing was a complex process, both facilitory and inhibitory to varying degrees, which regulates response change with experience. Though the term sensitization has been used to describe the response facilitation observed, the term itself relates not to the stimulus or response, but to the perceptive state of the organism being observed. Groves et al., (1969) 121 described two types of sensitization, one short-term, the other long-term. Short term sensitization was observed primarily as higher stimulation frequencies (shorter ITI equivalent), whereas long-term sensitization was observed at all frequencies. The intensity of stimulation interacted with frequency to influence the outcome of habituation. At high frequency and high intensity, the sensitization process prevailed and resulted in response facilitation (Note, large time blocks of averaged responses may block out the observance of sensitization, as seen in the difference between block and segment presentations of data). Thompson et al. (1973) studied the interaction of frequency and intensity a bit further and suggested that frequency has a stronger effect than intensity. This, too, is reflected in the current data set when comparing rates of habituation between intensities (Figure 4.3) and groups (Figure 4.6). Rate of habituation was only slightly faster at 3 psi compared to 1 psi. The difference in rates between high, intermediate, and low frequency groups (short to long ITI) shows a more dramatic effect as discussed earlier. In conclusion, to the degree that conditions of the present study were comparable to electrical stimulation, the results are consistent on all 9 points with classic habituation research. In particular, the presence of both sensitization and habituation processes under repetitive stimulation conditions was clearly evident and provide further support for the dual-process theory. 122 4.2 Parametric Study or Direct Stimulation of the Cerebellum and Elicited Behaviors As evidence has accumulated'for the role of the cerebellum in associative learning, studies have turned toward the underlying mechanisms of learning and the output signaling the behavioral response. Systematic studies of electrical stimulation of cerebellar regions could potentially provide insight into the characteristics of underlying cell populations. In conditioning experiments where electrical stimulation of cerebellar white matter was subsequently used as an unconditioned stimulus, Swain (1992) included a pre-training session of stimulation-only testing in which stimulus train duration was varied. Swain noted that behavioral response onset latencies remained stable across varying stimulus train durations. In contrast, the latency to peak varied with duration and occurred about SO- 100 ms following termination of electrical stimulation. The rabbits in that study also exhibited an increase in response amplitude as stimulus duration increased. The behaviors elicited by stimulation in these animals were lip and head movements. The present study extends these observations to electrical stimulation of the interpositus which produced eyeblink behaviors and includes variations of train duration, stimulus frequency, intensity, and pulse duration. 4.2.1 Methods: Subjects. The subjects for this experiment were 37 surgically naive New Zealand White rabbits (Oryctolagus cuniculus) weighing 2.0 to 3.2 kg. Several of 1 2 3 the rabbits had previously been subjects in Experiment 4.1 of this chapter (Habituation). All procedures and animal care were in accordance with NIH guidelines. Rabbits were housed individually in a room with a regulated 12 hr light/ 12 hr dark cycle. All animals were maintained on an ad lib food and water diet. Animal care was provided by the experimenter, staff and veterinarians of the University of Southern California. Surgery. Each rabbit was anesthetized with a mixture of ketamine (.08 ml/kg) and xylazine (.6 ml/kg) and placed in a stereotaxic headholder. Anesthesia was maintained during surgery using 1-2% halothane. The head was positioned in the stereotaxic plane of McBride and Klemm (1968) in which bregma is leveled l.S mm dorsal to lambda. Stainless steel electrodes (00 insect pins) insulated with cpoxylite except for 150-200 mm at the tip were implanted stcrcotaxically in the left intcrpositus nucleus of the cerebellum at approximately + 1.0 mm AP, + 5.0 mm ML, and -14.0 mm DV relative to lambda. Precise DV placement was determined during surgery by a combination of recording and stimulation results. Electrodes were implanted at the position which when stimulated best elicited eye-related movement. The DV range of implantation varied between - 12.0 and -15.0 mm . Stimulation during surgery consisted of 200 ms trains of 0.1 ms pulses at 350 Hz, ranging from 20-200 mA, as required to elicit movement. In 10 of the animals electrodes were implanted at the ideal coordinates where cellular activity was observed even though movement was not elicited by stimulation. Subjects were allowed at least one week post operative recovery before parametric testing began. Protocol. Following recovery, each animal exhibiting a distinct and measurable stimulation- evoked response underwent individual testing in which 124 stimulation parameters were manipulated. Amplitude, latency (onset and peak), and time-area characteristics of the evoked behavior were measured. Data were collected in blocks of 5 trials with a maximum of 200 trials on a given day and at most 5 days of input/output testing, Testing began with 200 ms trains of stimulation at 350 Hz with a pulse width of 0.1 ms at an initial intensity of 60 mA, First, data were collected for a range of behavior eliciting current intensities. These ranged from the smallest intensity which elicited a measurable response to that which cither produced a ceiling effect on response amplitude or appeared to cause discomfort. From these data, a reliable stimulation intensity was selected and held steady as a range of frequencies (50 - 500 Hz) and durations (50-850 ms) were applied and evoked behaviors recorded. 4.2.3 Results: Of the 37 animals which underwent surgery, one died and 6 were sacrificed during surgery because of difficulty with electrophysiological recordings and evoking movement. Of the remaining 30 rabbits with electrodes implanted in the interpositus region of the cerebellum, 10 placements were based on coordinates alone. Two of these were successful and underwent further testing. Of the 20 rabbits which showed some evidence of stimulation- evoked movement during surgery, 15 exhibited complex, unmeasurable, unreliable, or no movements following surgical recovery. Ten of these provided enough information to be suitable for mapping the interpositus nucleus and were combined with 7 successfully studied animals to produce Figure 4.11. ^ ru n n in g ♦ Eycblink + head A Head turn+cyeblink ■ Head ahnigfrcrunch Head nod Head lum * W hitkcn (4- head nod lateral) • Borepaw(left) HV1 • Nothing J * Eycblink + body v Eyelid + head ■ Head ahruc/scrunch Head nod Head turn • W hitkcrt + head nod • Nothing T Eyelid A Eye opening ■ Head movement * W hiikert/head ihrug Figure 4.11 Reconstruction of electrode placements whcih yielded an identifiable behavioral response upon stimulation. Labels in section 1.0 represent: hemispheric lobule VI (HV1), dentate nucleus (D). interpositus nucleus (IP), fastifial nucleus (F), and the inferior cerebellar peduncle (icp). Of all marked electrode placements, three stimulation* evoked behaviors were clear eyeblinks. Some behaviors were more complex, but included eyelid movement or eyeball rotation (not measurable). The prevalent movement in these cases was most often head movement such as a nod, turn or lift. These electrode placements were classified with other head related movements. Seven animals participated in parametric manipulations. Due to the exploratory nature of this study not all manipulations were common across all animals and are best presented case by case. In some cases, within animal replication of certain series manipulations arc available. Stimulus train duration and stimulus intensity were most commonly varied for all animals, and some between animal comparisons can be made for these measures. Duration. In four animals, a range of 6 train durations (50, 100, 150, 200, 250, 300) were studied keeping stimulation constant at 350 Hz and an intensity which produced a reliable behavior at a train duration of 200 ms. The percent change in amplitude for three of these subjects is presented in Figure 4 .12A . One animal (94-022) showed a large increase in amplitude at short train durations. The other two animals showed little change until a train duration of 200 or 250 ms. It is possible that some unaccounted for factor caused the dip observed at 200 ms for 94-022, in what otherwise seemed like a fairly steady amplitude increase coupled with increasing train durations, Individual differences are especially evident in the amount of change at higher train durations. The fourth animal (94-482) underwent 4 replications of the duration manipulation over a two day period (see Figure 4 .12B). The data include two separate series of eyeblinks at 350 Hz on day 1 and day 2 of testing, a series at 350 Hz in which head turn prevailed, and a series of eyeblink measurements at 200 Hz. 127 1000 900 © o > c aoo 700 © x: o © " O 600 S O O 3 400 Q . E < 300 200 100 100 150 200 250 300 50 © O ) c © J Z O © ■ o 1600 1300 1000 = 700 E < ^ 400 100 94-450 94-448 - + - 04-022 Stimulus Duration (ms) d2-350Hz (ayabllnk) dl-350Hz (ayabllnk) - A - dl-350Hz (haad turn) - o - 200Hz (oyabllnk) 50 100 150 200 250 300 Stimulus Duration (ms) Figure 4.12 Amplitude expressed in terms of percent change relative to shortest duration (SO ms)- Data in A are from different animals. Data in B represent the results of one animal (#94-482) for repeated duration series. 128 A considerable increase in response amplitude was observed on most counts between 50 and 250 ms train duration. There was a sharp reduction and stabilizing of the head turn at 200 ms which may be an artifact of restraint limitations and never quite returning the head to the original start position. In addition, the eyeblink on day 1 showed little change in amplitude at durations greater than 100 ms and a big drop at 300 ms. What did change in this case, however, was the length o f the response (see Figure 4 .13A) best described by time-amplitude area and risetime (see Figure 4.13B and C , respectively). Because onset latency fluctuated in some cases and not others, a consistent trend was not visible with increasing stimulus durations, but latency to peak did increase considerably with prolonged stimulation, as might be expected (see Figure 4 .13C). Two nice examples of the change in response behavior with increasing train duration are presented in Figure 4.14). While initial onset of the response varied little, note that the extension of eye closure with prolonged train duration seemed to be accomplished in smaller steps. The response peak always seemed to occur just beyond stimulus offset. Intensity. Eight series of increasing stimulation intensity were performed in 5 animals. Stimulation intensities ranged from 20 to 130 mA with a maximum of 5 steps for any given animal. Whengrouped together, the percent change in amplitude relative to the amplitude at the lowest intensity was calculated in 10 mA step increments (see Figure 4 .15A). A steady increase in amplitude with increasing stimulus intensity was observed. At the same time there appeared to be a decrease in onset latencies and small increase in peak latencies resulting in a general increase in risetime (peak- onset; see Figure 4 .15B). A typical example of actual response trials is presented 1 2 9 SOffifcec r a S • X J £ a . E < B E SOD 4 0 0 1O 0 S O 200 2 B O 300 S tim u lu s D uration (ms) too 800 so aao SOD P m t t i U m » I R l N < ' l i f l " ' yimului Uut Klton {mat Figure 4.13 Sample data animal #94-482. (A) Single traces over the range of stimulus durations increasing top to bottom. (B) Time-amplitude Area and (C) latency measures for the same data. 1 3 0 I— H —•f" * 1 1 ■- » I i>" 1 f I | »■ — ■ 1 ■ . . __ _ * — • _ _ — 1 —1 —1 —1 — V I I ■' 1 I I !■ I ■ I 1 I 1 ■ I Figure 4.14 Sample traces with varying stimulation duration. (A) Animal #94-450 : vertical dashed lines represent 50, 100, 150, 200, 250, and 300 ms electrical stimulation. (B) Animal #93-399 responses to 50,250, and 350 ms stimulation. 1 3 1 1 2 3 4 6 Stim ulation Intensity (10 uA increm ents) « E, >. u c ffl a 6 0 0 SOO 400 300 200 100 0 100 -B — Rlsctim* Peak f — O o tst 1 2 3 4 5 Stimulation Intensity (10uA increments) Figure 4.15 Combined data for comparison across a series of 10 pA increments of stimulation intensity (A) percent amplitude change (B) response latency measure. 132 in Figure 4.16. The decrease in onset reflect the faster rise toward peak which at lower intensities seemed more precisely tuned to stimulus offset. At higher intensities the peak occurred a soon after stimulus offset. Frequency. Two out of three animals exhibited a reliable increase in amplitude with increasing frequency of stimulation (94-450 and 94-448; see Figure 4.17). The other seemed to vary considerably (94-482). Interestingly, there was a steady decrease in latency to peak (see Figure 4 .18A) for all 3 animals. An even greater decrease in onset latency was observed for the same three animals (see Figure 4 .18B). It seems increasing stimulation frequency produced a sharper response with a faster risetime. See sample traces in Figure 4.19 and note the initial slope of the response. Pulse Duration. In one animal (94-450) with a solid evoked eycblink behavior a small range of pulse widths was explored. Figure 4.20A shows the difference in amplitude for 4 pulse durations (.05, .1, .2, .3 ms) while stimulation was held at 450 Hz and 40 mA for 150 ms. While a pulse duration of .05 ms produced no behavioral response, an cyeblink was elicited with a . 1 ms pulse duration and grew bigger at .2 ms, but began to drop off again at .3 ms. The latency to peak remained stable at about 173 ms, just beyond the offset of the 150 ms train. However, onset latency changed with amplitude (see Figure 4.20B). Sample traces are presented in Figure 4.21. 133 I 50msec Figure 4.16 Sample traces for animal #93-399 with increasing intensities (60, 70, 75, 80 uA) every 5 trials from top to bottom. 134 % Amplitude Change 4 0 0 0 3 0 0 0 2000 1000 6 0 0 « ’ 5 0 0 4 0 0 4 3 0 0 200 100 1 0 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 50 94*462 - A — 9 4 -4 5 0 9 4 -4 4 8 Frequency (Hz) Figure 4.17 Varying stimulation frequency: percent change amplitude. 1 3 5 Onset Latency (m s) P,ak Latenc» 1"“ ) 600 600 400 300 200 100 0 100 200 250 300 360 400 450 F r e q u e n c y (H z) 300 200 100 0 100 200 250 300 350 400 450 Frequency (Hz) Figure 4.18 Vary stimulation frequency: peak and onset latencies. H— 04-462 04-460 04-446 STIM O N r i I 200 Hz 250 Hz 300 Hz 350 Hz 400 Hz 450 Hz 50msec Figure 4.19 Animal # 94-448: sample traces of eyeblink behavior resulting from increasing stimulation frequencies. 137 1 0 E E tj a. E < .O S .1 .2 P o l i o D u r a tio n (m s ) 200 - + 150 E E © 100 C L E < V ----------V .05 .3 .1 .2 —s — nisotlmo - + - P a a k — *■ - O n s e t P u ls e D uration (m s) Figure 4.20 Varying Pulse Duration: percent amplitude change and response latencies. Note, these data are only Tor one animal. 138 SUM O H t i ) * i i — i— i— i— 1 — i— i— i— i— i— i— i— r**—i— i— i— i— i— i— 50msec Figure 4.21 Sample (races of for rabit #94-450 behavior when stimulation pulse width was changed. Pulse width is expressed in ms. 4.2.3 Discussion: Consistent with earlier findings in cerebellar white matter (Swain, 1992), varying stimulus train duration did not effect response onset, but peak response latency was adjusted to occur shortly after stimulation was terminated. This also affected time-amplitude area and risetime accordingly. As for changes in amplitude paralleling the change in duration, the results are less clear. The trend appears to be an increase in amplitude with longer stimulus trains. The variability arises at the longer durations (250 and 300 ms). It is possible that for a given stimulation frequency and intensity there are individual differences in maximal recruitment of cells regulating the size of the behavior. Beyond this point, the response expand in time, not amplitude. In contrast, when stimulation intensity was increased there was a very clear and steady increase in response amplitude and a slight delay in peak latency. Interestingly, there was generally a decrease in onset latency. Onset seemed to « accommodate the longer risetime required to reach a larger peak amplitude so that peak latency is only slightly affected. But it also seems that intensity is the key factor in achieving maximum amplitude. Response peaks were more closely tuned to stimulus offset at lower intensities than at higher intensities. At lower intensities the response was more easily reversed when stimulation was terminated because the maximum amplitude had been reached. At high intensities that peak amplitude took a little longer to reach, but this delay could also be due to rebound excitation of interpositus (Eccles et al., 1967). Stimulation in the area of interpositus activates cells of interpositus as well as mossy fibers projecting to cerebellar cortex and Purkinje cells. At lower intensities stimulation is more localized, reaching 140 fewer mossy fibers, and there is less Purkinje cell inhibition and less rebound. At higher intensities, Purkinje cell inhibition prevents maximal cell activity and removal of this inhibition causes a rebound which manifests itself as a prolonged increase in behavior response amplitude. Similar changes in amplitude and latency were observed with increases in stimulation frequency. Responses were larger, initiated sooner and peaked faster (note slope to peak in Figure 4.21). This gave the appearance of a sharper and more vigorous response. A wide range of frequencies could rarely be studied because higher frequencies often seemed to cause discomfort to the animals. In addition to stimulation intensity and frequency, pulse duration also affected onset latency, but not peak. While increases in intensity yielded increasingly larger responses, pulse width and frequency appeared to have a small range of effective values including an optimal value for best response. In short, stimulation frequency and pulse duration seemed to determine whether a response would be initiated. Stimulus intensity then determined the size of the stimulation- elicited movement and stimulus train duration determined the risetime to peak. Taking this one step further the data are consistent with well established findings that intensity determines the number of cells recruited or activated by stimulation. Frequency and pulse width may be the code for the optimal conditions of activation for a particular groups of cells, and train duration maintains this activation. That movement persisted only as long as electrical stimulation, suggests that the motor output cells of interpositus are phasic in nature. This is consistent with casual observations (Tracy, personal communication) of single cells in the interpositus identifying phasic cells primarily in the dorsal area of interpositus. 1 4 1 Stimulation in this region best elicits specific movements and cells are believed to project to the red nucleus motor pathway. On the other hand, tonic cells are located more in the ventral region (Tracy, personal communication) and are thought to project to the inferior olive (McCrea et al., 1978). 4.3 Interference in the Classically Conditioned Eycblink Response in Rabbits by Direct Stimulation of the Cerebellum during the CS-US Interval Eight of the animals used in the previous experiment were subsequently taken and classically conditioned using tone and airpuff to confirm that stimulation had not damaged the area of the interpositus critical for learning. Once they had acquired the conditioned eyeblink response they received stimulation during the CS-US interval to interfere with conditioned response production and examine the effect(s) on response initiation. 4.3.1 Methods: Subjects. Eight New Zealand White rabbits (Oryctolagus cuniculus) with multi-unit recording and stimulating electrodes in the region of the left interpositus nucleus of the cerebellum were used in this study (surgical procedure described in previous experiment). All animals had been previously tested for stimulus-evoked movements. Two of the animals (94-448 and 94-450) reliably produced eye movements and had undergone a series of stimulus manipulations in the previous experiment. The others responded to electrical stimulation either with a complex head or facial movement or not at all. 1 4 2 Training Protocol Acquisition Phase. Rabbits received 4 days of delay classical conditioning. Each 108 trial training session consisted of 18 blocks of one tone-alone (T) and 5 paired tone-airpuff trials (T&A) in the sequence T, T&A, T&A, T&A, T&A, T&A. On paired trials a 654 ms tone (1 kHz, 85 dB) co-tcrminated with a 102 ms airpuff (see Figure 4.22 -Training Phase). Stimulus presentation was controlled on-line by the Forth Runtime/Summary Program and interface (Lavond & Steinmetz, 1989) described earlier . Testing Phase d- Following acquisition, training continued as in the acquisition phase but with the addition of a trigger pulse following tone onset by 96 ms. This trigger pulse was delivered to a Grass 8800 stimulator and initiated stimulation trains to the electrodes implanted in the interpositus region of the cerebellum (sec Figure 4.22 - Test A). The modified training sequence was then TS, TS&A, TS&A, TS&A, TS&A, TS&A. The stimulator could be turned on and off as needed throughout the session without programming changes. Stimulation parameters (frequency, duration, and intensity) were manipulated during the session by adjusting settings on the stimulator manually. The primary manipulation was stimulation duration within and beyond the CS-US interval. Frequency and intensity were set to produce a reliable and measurable (just above threshold) movement. Pulse duration was typically 0.1 ms unless otherwise noted. Testing Phase Q. To better distinguish stimulation effects on the CR vs. UR components of behavior, in this phase a trace interval of 96 ms was introduced between the tone and airpuff (see Figure 4.22 - Test B). Air-alone (A) trials were now added and stimulation still followed tone onset by 96 ms. The new sessions 143 now contained 100 trials, 10 blocks of 10 trials in the sequence T, TS&A, A, TS&A, TS&A, T, TS&A, A, TS&A, T. As before, a variety of stimulation parameters were tested. Testine Phase Q. To assess whether the initial 96 ms of tone only presentation arc critical to the behavioral response produced, in this phase stimulation and tone onset were simultaneous (see Figure 4.22 - Test C). All other details were the same as in Testing Phase B. Data Analysis. Data were analyzed on-line for amplitude, latency (onset and peak) and amplitude-time areas during the CS and the US periods. A CR was any movement after CS onset that equaled or exceeded 0.5 mm of nictitating membrane extension. Trials were not counted if movement in the baseline period exceeded 0.7 mm or if a movement of 0.7 mm occurred in the CS period within 25 msec after CS onset. Due to the exploratory nature of this study, the changes in response topography resulting from manipulating electrical stimulation are largely descriptive. Hisrolopv. At the completion of Testing Phase C, marking lesions (80 mA, 10 sec) were made through the stimulating electrodes. Animals were perfused with saline, followed by a 50/50 mixture of potassium ferrocyanide and 10% formalin, and then with 10% formalin. The brain was excised and stored in 10% formalin for a few days. It was then embedded in an albumin-gel matrix and stored in 10% formalin for at least one week. The tissue was sliced on a freezing microtome in 80 mm sections, mounted, and stained with cresyl violet. 144 6 5 4 m s T o n e L 1 0 2 m s Air 6 5 4 m s T o n s v a ria b le S tim 96 m tf~ ~ “ I 1 0 2 m s Air 1 1 4 5 6 m s T o n e 1 0 2 m s Air J [0 8 M 1 v a ria b le S tim 9 6 m s [ I 456 m s T o n e 1 0 2 m s Air r |B6 m s | | • • r - v a ria b le Stim n 1 0 0 m s Figure 4*22 Training and testing protocol. 1 4 5 4.3.2 Results and Discussion: On-line data collection hardware allowed for the recording of 4 channels of data simultaneously. The animals in this study were run four at a time and data will be similarly grouped. Group 1 subjects were animals 94-193, 94-194, 94-195, and 94-196. Group 2 involved 94-446, 94-448, 94-449, and 94-450. Histology. Placement of electrodes stimulated during the CS-US interval are presented for each animal in Figure 4.23. Group 1 placements were all a little lateral to the regions of cerebellum specifically implicated in the learning circuit. Two placements were in the molecular layer of cerebellar cortex lateral to HVI and above the dentate (94-193 and 94-195). The other two placements were more ventrally located in the dorso-anterior dentate nucleus (94-194 and 94-196). Group 2 placements were more medial and in regions of cortex and deep nuclei associated with conditioning. As with Group 1, the placements were in one dorso-vcntral column and two placements were more cortical while the other two were in the deep nucleus. The cortical placements were at and near the base of cerebellar lobule HVI ( 94-449 and 94-446, respectively) which is the primary cortical input to the interpositus. The other two animals had electrodes directly in the interpositus nucleus (94-448 and 94-450). As was demonstrated in the previous study o f stimulation- evoked behaviors, the interpositus placements resulted in eyeblinks. Stimulation at the base of HVI produced whisker movement, and the more lateral cortical and dentate placements all involved some form of head movement. Acquisition. After 4 days of tone-aiTpuff training all animals achieved at least 65 % conditioned responses on tone-alone test trials. Five animals were 1 4 6 • 94-193 • 94-446 A 94-194 T 94-448 m 94-195 • 94-449 * 94-196 ■ 94-450 Figure 4.23 Histological reconstruction of electrode placements where stimulation was applied in well trained animals during the CS-US period of conditioning. 1 4 7 100 80 60 to C C O 40 N«8 20 1 2 3 4 Acquisition Day Figure 4.24 Percent conditioned responses on normal acquisition traingin days to show the critical brain lesions had not sustained damage during participation in the previous stimulation study. 148 responding at 100 % and as a whole the group (N =8) averaged 91.2 4.7 percent on day 4 (see Figure 4.24). Group k Testing Phase d.. The first group (94-193, 94-194, 94-195, 94- 196) underwent Testing Phase A, while the second (94-446, 94-448, 94-449, 94- 450) proceeded directly to Testing Phase B. Figure 4.25 demonstrates the effect a 300 Hz, 350 ms train of stimulation presented 96 ms after tone onset had on the conditioned response in the first group o f 4 animals. The onset of the CR was delayed until stimulation offset. In one animal (94-195) movement seemed to begin slightly in response to the tone but was further prevented until stimulation offset. The delayed CR was most clearly demonstrated by animal 94-193 (see Figure 4.26) with increasing stimulus durations (electrode in lateral cortex; see Figure 4.23). Group k Testing Phase £ . When this same group was tested in Phase B where the tone-stimulation relationship was identical, but a delay was inserted before the airpuff, the results of Phase A held up for the two animals with cortical electrode placements (94-193 and 94-195; see Figure 4.27). CR onset was slightly delayed with increasing stimulus durations. Animal 94-194, with an electrode in dentate nucleus very dorsal and medial, also exhibited a some delay of behavior onset when stimulation was present. Animal 94-196 showed a potentiated conditioned response amplitude in the presence of electrical brain stimulation. None of the animals appeared to condition to the electrical stimulation. Figure 4.28 presents a comparison between paired tone-stimulation-airpuff trials (A) and stimulation-airpuff trials (B). While presentation of the tone initiated some behavioral change in the CS-period in most animals despite the addition of 1 4 9 Tone Airpuff 94-193 94-194 94-195 94-196 Stim Tone Airpuff Figure 4.25 Sample responses for Group 1 animals during test phase A on a normal paired trial (A) and with stimulation added 96 ms after tone onset. (B). \ / x / V / ’ ■ N / \ • • •s 4 - 4 Stim Ton© Airpuff Figure 4.26 Sample traces (94-193) varying stimulation offset: 100, 200, 650, 850 ms, top to bottom. 94-193 94-194 94-195 94-198 94-193 94-194 94-195 94-196 94-193 94-194 94-195 94-196 Figure 4.27 Individual traces for animals in Group 1 during phase B testing with 0, 200, and 360 ms stimulation durations. The solid bar indicate the duration of the tone. The shaded bar represents stimulation. The vertical line mark the onset of the airpuff 96 ms following tone offset. 1 5 2 94-193 94-194 94-195 94-196 94-193 94-194 94-195 94-196 Figure 4.28 Individual traces for Group 1. Stimulation-alone (shaded bar) did not evoke eyeblink behavior in these animals (bottom), nor did animals condition to it after more than 100 trials where stimulation was added to the normal tone-airpuf pairing in the CS-period (compare bottom and top traces). 1 5 3 stimulation, stimulation alone did not. Also of interest is that stimulation in the lateral cerebellar regions had no effect on the unconditioned reflexive eyeblink. Group h Testing Phase Q. To address the issue of whether or not behavior was initiated before stimulation began in Phases A and B, it was necessary to compare responses on those days to Phase C responses where tone and stimulation onset were simultaneous. By far, the most interesting changes in Group 1 on this point and in general were observed for animal 94-193. Figure 4.29 demonstrates how stimulation following tone onset by 96 ms (Phase B) temporarily “interrupts” conditioned behavior on test trials until stimulation is removed. When stimulation is simultaneous with tone onset (Phase C), the initial movement observed to the tone when there was a delay is completely abolished. In addition, the discontinuation of stimulation consistently results in a rebound effect on behavior. This is presumably due to the release of cortical inhibition to deep nuclei involved in motor behavior. Group 2z Testing Phase £. An interesting distinction arises in Group 2 based on electrode placements. The animal with the cortical placement slightly lateral to the base of HVI (94-446) showed slight delays in CR onset to 100 and 200 ms stimulation, similar to Group I animals with lateral placements. As durations continued to increase there were no further delays and CR amplitude increased (see Figure 4.30 top trace). The other three animals exhibited stimulation-related increases in CR amplitude. Almost immediately after recovery from stimulation there was a second wave of movement peaking just around airpuff onset presumably a response to the ongoing tone. As stimulus duration increased these responses blurred, but were consistently followed by a secondary response in the US period. When 154 A Figure 4.29 Sample traces on paired trials (tone, stimulation, and airpuff) for animal #94-193 under a variety of stimulation conditions. Black bar represents tone, shaded bar represents duration of stimulation (A *0, B«200, C=360, D — 0, E=650 ms). 1 5 5 94-440 94-448 94-449 94-450 Figure 4.30 Sample tracs for Group 2 varying stimulation duration. First black bar represents tone CS, second black bar marks airpuff US presentation, and the shaded bar represents stimultiton. A: no stim, B: 100 ms, C: 200 ms. D: 360, E: 450, F: 550, G: 650. 1 5 6 stimulation was extended beyond the airpuff and there could be no reflexive eyebtink to the airpuff, this response remained. It would appear to be a movement produced by rebound excitation of interpositus following removal of cortical inhibition, especially in the case of 94*449, where the electrode was in the white matter of HVI. Since the behavior of 94-450 parallels this, the same explanation could apply (see discussion of previous experiment). Group 2z Testing Phase £ . Figure 4.31 demonstrates the effects of cerebellar stimulation beginning at tone onset and extending for various durations. Stimulation did not seem to effect conditioned response performance for animal 94- 446 (top trace) at any duration, but did potentiate the unconditioned component of the response (distinct peak following airpuff onset). As for the other three subjects, the shortest stimulus duration (100 ms) produced a distinct peak within the stimulation period (Figure 4.3IB) which occurred earlier than the first changes observed in baseline recordings (Figure 4.31A). For the remaining cortical animal (94-449), this initial response to stimulation was consistent and was followed by a potentiated response in the CS- period causing some reduction in the UR because the eye is already partially closed. No further correspondence was observed between stimulation and behavior for this animal. Both animals with electrodes in interpositus showed behavior which clearly reflected stimulation train duration. There were distinct dips in the eyeblink trace corresponding to the offset of stimulation at each duration. When the tone was disconnected and only a 350 ms stimulus train was paired with an airpuff, it seemed the initial response slope was more gradual (this is especially evident for 1 5 7 A D Figure 4.31 Sample traces (Group2): simultaneous tine and stimulation onset but varying stimulation offset. Black bars represent tone and aiipuff stimuli while the shaded bar represents electrical stimulation. A:not stimulation, B: 100 ms, C: 250 ms, D: no tone, 250 ms stimulation test trial, E: 350 ms tone and stimulation overlap completely, F: 650 ms stimulation overlaps tone and airpuff. 1 5 8 94-450) suggesting that the combined effected of tone and stimulation processes on response initiation differed from that of either alone (see Figure 4.31 A and D). Also these same animals continually showed a delayed response within the US period. Because of the delay, it is not clear that the response is reflexive to the airpuff. When stimulation duration was 650 ms and overlapped the airpuff completely, only the two animals with electrodes in cortex (94-446 and 94-449) showed a distinct unconditioned response to the airpuff. 4.3.3 Summary: Group I . Stimulation in areas of cortex and deep nuclei lateral to those involved in eyeblink classical conditioning (Group 1) delayed conditioned response performance for the duration of the stimulation. When there was some movement initiated to the brief tone, it was abolished by overlapping stimulation. Presumably, stimulation must have activated inhibitory mechanisms acting on response production. Inhibition on the interpositus could easily have occurred via stimulation of cortical parallel fibers to Purkinje cells or stimulus spread to neighboring white matter. Indeed Yang and Weisz (1992) recorded from neurons in dentate and anterior interpositus during conditioning. Unconditioned stimulus elicited single unit activity of cells in dentate was depressed by a 500 ms auditory CS, but a 30 ms did nothing. In contrast, they found that US-elicited activity in inteipositus neurons was facilitated by an auditory tone CS (see next section). The deeper placements in dentate may have disrupted the normal activity of cells in the area which are responsive to continuing tone stimulation. 1 5 9 Group 2. Stimulation in these animals evoked more eye-related behaviors and, therefore, contributed to conditioned response amplitude in most cases (See Figure 4.30). The animals with cortical electrodes were more responsive to the tone during the CR period, as is evident by comparing all traces in Figures 4.31 to part D. A more robust conditioned response was observed to tone-alone (A) and in the presence of tone (B, C, H, and F) than to stimulation alone (D). Stimulation neither interfered with perception of the tone nor with the behavioral response to the tone. Interesting effects on the unconditioned component of the eyeblink response were observed. Animals with cortical electrode placements exhibited potentiated URs probably resulting from rebound activation of interpositus neurons following prolonged inhibition from overlying cortex. When stimulation duration was 6S0 ms and overlapped the airpuff completely, only the two animals with electrodes in cortex (94-446 and 94-449) showed a distinct unconditioned response to the airpuff. In contrast, animals with electrodes in interpositus initially showed a normally timed UR (see Figure 4.30A) but later showed a delayed response in the US-period (see Figure 4.30D-F and Figure 4,31). While the explanation seems to be recovery from stimulation in some cases, this argument is not confirmed consistently. It is unusual for the airpuff not to elicit an immediate response. Possibly the response consisted of eye opening before closure. 1 6 0 Overall reliable effects extracted from these data are consistent with current understanding of cerebellar anatomy. In particular, the modulatory role of cortical inhibition in both conditioned and unconditioned response production and timing. Also, the independence of the unconditioned reflex mechanism is once again demonstrated. 1 6 1 CHAPTERS Conclusion For years, researchers interested in the plasticity of learning and memory have strived toward the ultimate goal of identifying a specific locus where associations take place and memories can be stored. Classical eyeblink conditioning in rabbit has all the advantages of a model system suited for just such an endeavor. Classical conditioning itself is a simple, yet elegant paradigm, which can be manipulated in order to address a variety of complexities of learning and memory, which arc identified by adaptive changes in behavior which are maintained for a period of time. Particularly useful, and essential for the studies presented in this paper, is its “built-in" control system which permits the distinction between the process of acquiring or remembering an association, as measured by a behavioral response, and the physical ability of the animal to produce the same or a similar response. This is essential because, what appears to be a learning deficit may reflect an inability to express the required behavior due to damaged motor systems. Similarly, there may be an impairment of sensory or perceptive processes, thereby altering the significance of a stimulus and not meriting a response. There may also be a motivational deficits preventing the initiation of a response. For all these reasons, classical conditioning is ideal. It allows the researcher to control for these alternatives by testing reflexive responses periodically using only the natural response eliciting stimulus (US) to assess performance ability. 162 The first part of this paper emphasizes the use of input/ output tests of reflex response performance at different points of time over the course of training experience. These tests were used to address the arguments of Welsh and Harvey (1989) who challenged the involvement of cerebellar structures in learning and memory. Welsh and Harvey (1989) reported that lesions of the interpositus nucleus of the cerebellum did not affect an animal's ability to acquire classical conditioning, but rather it's ability to show that it had learned. They went on to stipulate that these performance effects were only noticable at low US intensities because higher intensities (presumably such as those standardly used in the field), were strong enough to necccsitate defensive activation of associated muscle groups to compensate for the deficit. Their results were contrary to numerous other studies (see introduction for review) which have implicated the cerebellar cortex and deep nuclei to be essential to learning and memory of the classically conditioned eyeblink. The studies in Chapter 1 directly investigated the points Welsh and Harvey (1989) made and, in conjunction with the work of many others, firmly established that animals with lesions of the interpositus which effectively abolished all conditioned responding did not exhibit any permanent performance deficits which would prevent them from exhibiting CRs. Reflex responsiveness tests covered a significant range of US intensities above and below that normally used for training (3 psi). Only a transient response decrement was observed at 3 psi immediately following lesion of interpositus nucleus, but returned to normal after S more days of conditioning training. In fact, all unconditioned reflexes were enhanced by conditioning training and remained at or returned to those enhanced levels following additional training. In addition, animals trained at the low intensities 163 reported by Welsh and Harvey exhibited marginal rates of learning. In a group of ten animals trained using 0.13 psi, only 1 animal learned, as compared to seven in the 0.5 psi group. This study demonstrated that even animals whose entire training experience consisted of low intensity airpuffs, did not exhibit performance deficits following interpositus lesions which effectively abolished all conditioned responding. When a post- learning performance deficit was specifically produced by targeting the motor nucleus responsible for eyeball retraction and nictitating membrane movement (accessory abducens), a profound performance deficit was produced. As expected, this also significantly affected conditioned response expression. Response recovery was monitored for several weeks post-lesion and and a very interesting result emerged. The conditioned response recovered more quickly and to a greater extent than the unconditioned response. Though it is difficult to say much about the exact mechanism supporting this recovery, it is clear that the conditioned response is not as easily affected by performance deficits as Welsh and Harvey had suggested. Another structure strongly implicated in control of motor behavior, and which has several anatomical links to the cerebellar circuit, is the motor cortex (see introduction for review of interconnections). As a result of the previous studies mentioned in this paper, performance measures and more detailed analyses of changes in response topography became an issue of interest in relation to motor learning. Earlier work by Woody and colleagues (see Woody, 1984 for review) showed that motor cortex stimulation could serve as a CS in the classical conditioning of cats, that there were learning-related changes in firing patterns of cells in motor cortex, and that lesions of motor cortex results in a profound 164 reduction of CRs. This was inconsistent with earlier findings in the rabbit literature (Norman et al., 1977; Mauk and Thompson, 1987) that decerebrated animals can learn. Similarly, results reported in the present paper suggested that motor cortex was not necessary for either acquisition or subsequent retention of delay classical conditioning. However, when a paradigm similar to Woody's (click CS and glabella tap US separated by a 500 ms trace interval) was used with rabbits, some impairments began to emerge. Though not statistically significant, the impariments were very consistent across all measures and may suggest that the measures used were not sensitive enough. The only significant effect observed was for timing of the onset latency for the response. The data suggest that lesioned animals were slower to initiate an eyeblink, though this may be the result of fewer overall responses. This is opposite to studies where hippocampus was lesioned and produced short latency maladaptive cycblinks. Hippocampus was spared in all animals and only slight damage to anterior rctrosplcnium was sustained in a few animals. Another interesting outcome from these studies was the finding that unconditioned reflexive cyeblinks are modified by exposure to stimuli, whether or not there is an associative component to training. The long-term nature o f these changes observed in I/O performance may be either explained by a slow accumulative facititory process in the brainstem reflex arc or possible by higher level processing processing. The most extensively studied effect o f repeated stimulus exposure has been habituation. In order to further study and characterize the topography of the UR and conditions which influence it, a systematic study of peripheral airpuff stimulation was undertaken varying airpuff intensity and frequency of occurrence. The results were consistent with classic habituation 1 6 5 studies of cat spinal reflex. The major findings included : (1) Response amplitude i. which exhibited a characteristic decline across blocks of repeated stimulation, (2) exhibited spontaneous recovery, and (3) faster rate of habituation on subsequent sessions; (4) lower intensity stimulation yielded more profound habituation, and (S) with higher frequency stimulation responses to the stronger aipuff habituated belter than at lower frequencies where there was little change. Eyeblinks and other motor behaviors produced by direct stimulation of the cerebellum also exhibited specific characteric changes dependent on stimulation parameters. A narrow range of intensities, frequencies, and pulse widths were effective in producing a response. Stimulus intensity seemed to follow a U-shape function reaching an optimal value and falling off. Frequency and pulse width seemed to be critical parameters for initiating a response and may be a type a of coding mechanism for distinctccll populations. By far, the most consistent trend observed in this limited sample was that of increasing stimulus durations. As stimulus duration changed, peak latency followed closely peaking just beyond stimulus offset. The extended rise in amplitude at longer stimulus durations contributed to larger peak amplitudes. Onset latency, however, did not change, as was reported earlier by Swain (1992). Animals trained on delay classical conditioning who then received stimulation in lateral regions of the cerebellum, especially dentate nucleus, showed behavior consistent with earlier recording studies where and auditory CS depressed US-elicited activity in the dentate nucleus (Yang and Weisz, 1992). In conclusion, the present group of studies have addressed a broad number of issues relating to learning and performance of the eyeblink during classical conditioning related training. Most importantly they have confirmed that the 166 production of conditioned and unconditioned motor behaviors are independent processes. The unconditioned response is plastic in its own right. It exhibits long term faciltation over both associative and non-associative training experience and demonstrates all the laws of habituation following massed stimulus presentation. Despite intricate cerebroccrebellar neuroanatomtcal connections, lesions of cerebral cortex, including motor regions, did not significantly effect any performance measures. Some evidence was obtained for cerebral cortical involvement in modulating motor behavior during more complex tasks than standard delay conditioning; however, results obtained herein are not significant enough upon which to base such a claim. 167 REFERENCES Akase, E., Alkon, D.L., Disterhoft, J.F. (1989). Hippocampal lesions impair memory of short-dclay conditioned eye blink in rabbits. Behavioral Neuroscience 103: 935*943. Allen, G.I. and Tsukahara, N. (1974). Cerebroccrebellar communication systems. Physiological Review 54: 957*1006. Andersson, G., Gorwicz, M, and Hesslow, G. (1987). Effects of bicuculline on cerebellar inhibition of the inferior olive. Abstracts o f the Second World Congress on Neuroscience, p.5631. Bakin, J.S. and Weinberger, N.M. (1990). Classical conditioning induces CS- specific receptive field plasticity in the auditory cortex of guinea pig. Brain Research 536: 271-286. Bassett, J.L. (1987). Identification of a multisynaptic hippocampal output pathway through the posterior cingulale gyrus: anatomical and electrophysiological studies. Doctoral Dissertation, University of Pittsburgh. Berger, T.W. and Orr, W.B. (1983). Hippocampcctomy selectively disrupts discrimination reversal conditioning of the rabbit nictitating membrane response. Behavioral Brain Research 201:411 -417. Berger, T.W. and Thompson, R.F. (1978). Neuronal plasticity in the limbic system during classical conditioning of the rabbit nictitating membrane response. I. The hippocampus. Brain Research 145:323-346. Berger, T.W ., Wcikcrt, C.L., Bassett, J.L., Orr, W.B. (1986). Lesions of the retrosplenial cortex produce deficits in reversal learning of the rabbit nictitating membrane response: Implications for potential interactions between hippocampal and cerebellar brain systems. Behavioral Neuroscience 100:802*809. Berry, S.D. and Oliver, C.G. (1982). Hippocampal activity during appetitive conditioning in rabbits. Society fo r Neuroscience Abstracts. 8: 315. Berthier, N.E., Desmond, J.E., and Moore, J.W. (1987). Brain stem control of the nictitating membrane response. In I. Gormezano, W.F. Prokasy, and R.F. Thompson (Eds.) Classical Conditioning, 3rd Edition. Hillsdale: Lawrence Erlbaum Associates, (pp. 275-286) Berthier, N.E. and Moore, J.W. (1980). Role of extraocular muscles in the rabbit (Oryctolagus cuniculus) nictitating membrane response. Physiology and Behavior 24: 931-937. 168 Boorstin, D.J. (1983) V ie Discoverers. New York: Raven Press. Brodal, A. (1981). Neurotogial Anatomy. New York: Oxford University Press. Brooks, C. McC. and Woolsey, C.N. (1940). Placing and hopping reactions in relation to the electrically excitable “motor” areas of the cerebral cortex of the rabbit. Bulletin o f the Johns Hopkins Hospital 67: 41*60. Buchanan, S.L., Valentine, J., and Powell, D.A. (1985). Autonomic responses are elicited by electrical stimulation of media, but not lateral frontal cortex in rabbits. Behavioral Brain Research 18: 51*62. Bures, J. and Bracha, V. (1990). The control of movements by motor cortex. In B. Kolb and R.C. Toss (Eds.) V ie Cerebral Cortex o f the Rat. Cambridge: The MIT Press. Cegavske, C .F ., Harrison, T.A ., and Torigoe, Y. (1987). Identification of the substrates of the unconditioned response in the classically conditioned rabbit, nictitating*membrane preparation. In I. Gormezano, W .F. Prokasy, and R.F. Thompson (Eds.) Classical Conditioning, 3rd Edition. Hillsdale: Lawrence Erlbaum Associates, (pp. 65*92) Cegavske, C .F ., Patterson, M .M ., Thompson, R.F. (1979) Neuronal unit activity in the abducens nucleus during classical conditioning of the nictitating membrane response in the rabbit (Oryctolagus cuniculus). Journal o f Comparative Physiological Psychology 93:595-609. Cegavske, C .F., Thompson, R.F., Patterson, M ., and Gormezano, I. (1976). Mechanisms of efferent control of the reflex nictitating membrane response in rabbit (Oryctolagus cuniculus). Journal o f Comparative and Physiological Psychology 90: 411 -423. Chapman, P.F., Steinmetz, J.E ., and Thompson, R.F. (1988). Classical conditioning docs not occur when direct stimulation of the red nucleus or cerebellar nuclei is the unconditioned stimulus. Brain Research 442: 97- 104. Clark, G .A ., McCormick, D.A., Lavond, D.G., and Thompson, R.F. (1984). Effects of lesions of cerebellar nuclei on conditioned behavioral and hippocampal neuron responses. Brain Research 291: 125-136. Clark, R.E. and Lavond, D.G. (1993). Reversible lesions of the red nucleus during acquisition and retention of a classically conditioned behavior in rabbits. Behavioral Neuroscience 107: 264-270. Davis, K.D. and Dostrovsky, J.O. (1986). Modulatory influences of red nucleus stimulation on the somatosensory responses of cat trigeminal subnucleus oralis neurons. Experimental Neurology 91: 80-101. 169 Diamond, D.M. and Weinberger, N.M. (1986). Classical conditioning rapidly induces specific changes in frequency receptive fields of single neurons in secondary and ventral ectosylvian auditory cortical fields. Brain Research 372: 357-360. Disterhoft, J.F. and Segal, M. (1978). Neuron activity in rat hippocampus and motor cortex during discrimination reversal. Brain Research Bulletin 3: 583-588. Disterhoft, J.F ., Quinn, K.J., Weiss, C., and Shipley, M.T. (1985). Accessory abducens nucleus and conditioned eye retraction/ nictitating membrane extension in rabbit. The Journal o f Neuroscience 5: 941-950. Donegan, N.H., Foy, M.R. and Thompson, R.F. (1985). Neuronal responses of the rabbit cerebellar cortex during performance of the classically conditioned eyelid response. Society fo r Neuroscience Abstracts 11:835. Dudai, Y. (1989). The Ncurobiology o f Memory. Oxford: Oxford University Press. Ecclcs, J.C. Ito, M., and Szcntagolhai, J. (1967). The cerebellum as a neuronal machine. New York: Springer-Verlag. Evinger, C. and Manning, K.A. (1988) A model system for motor learning: adaptive gain control of the blink reflex. Experimental Brain Research 70:527-538. Foy, M.R. and Thompson, R.F. (1986). Single unit analysis of Purkinje cell discharge in classically conditioned and untrained rabbits. Society fo r Neuroscience Abstracts 12:518. Gleitman, H. (1983). Basic Psychology. New York: W.W. Norton & Company, (p. 71) Georgopoulos, A.P., Ashe, J., Smymis, N., and Taira, M. (1992). The motor cortex and the coding of force. Science 256: 1692-1695. Gormezano, 1. (1966). Classical conditioning. In J.B. Sidowski (Ed.) Experimental Methods and Instrumentation in Psychology. New York: McGraw Hill. Gormezano, I., Kehoe, E.J., and Marshall, B.S. (1983). Twenty Years of classical conditioning research with the rabbit. Progress in Psychobiology and Physiological Psychology 10: 197-275. Gray, T.S., McMaster, S.E., Harvey, J.A ., Gormezano, I. (1981) Localization of retractor bulbi motoneurons in the rabbit. Brain Research 226:93-106. 1 7 0 Groves, P.M ., Lee, D. and Thompson, R.F. (1969). Effects of stimulus frequency and intensity on habituation and sensitization in acute spinal cat. Physiology and Behavior 4: 383*388. Hawkins, R.D. and Kandel, E.R. (1984). Is there a cell*biological alphabet for simple forms of learning? Psychological Review 91:375-391. Hermann, D.J. and Chaffin, R. (Eds). (1988). Memory in Historical Perspective. New York: Springer-Verlag. Hobbelen, J.F. and van Hof, M.W. (1986). The hopping reaction in the rabbit after early and late removal of motor cortex. Behavioral Brain Research 21: 73-77. Houk, J. (1989). In R.M.J. Cotterill (Ed.) Models o f Brain Function. Inagaki, M ., Takeshita, K., Nakao, S., Shiraishi, Y., and Oikawa, T. (1989). An electrophysiologically defined trigcmino-reticuto-facial pathway related to the blink reflex in the cat. Neuroscience Letters 96: 64-69. Ison, J.R. and Leonard, D.W. (1971). Effects of auditory stimuli on the amplitude of the nictitating membrane reflex of the rabbit (Oryctolagus cuniculus). Journal o f Comparative and Physiological Psychology 75: 157-164. Ito, M. (1984). The cerebellum and motor control. New York: Raven Press. James, W.A. (1890). The Principles o f Psychology. New York: Holt (p. 107) Jenkins, H.M. (1979). Animat learning and behavior theory. In The First Century o f Experimental Psychology. E. Hcarst (Ed.) Hillsdale: Lawrence Erlbaum Associates. Kim, J.J., Clark, R.E., and Thompson, R.F. (in press). Hippocampectomy impairs the memory of recently, but not remotely, acquired trace eyeblink responses. Behavioral Neuroscience, Kimble, G.A. and Ost, J.W.P. (1V£1) A conditioned inhibitory process in eyelid conditioning. Journal o f Experimental Psychology 61: 150-156. Kimmel, H.D. (1973). Habituation, habituality, and conditioning. In H.V.S. Peeke and M.J. Herz (Eds.) Habituation Volume J: Behavioral Studies. New York: Academic Press, pp. 219-238. Klein, S.B. (1991). Learning: Principles and Applications (2n d Ed). New York: McGraw-Hill, Inc. Krupa, D.J., Weng, J. and Thompson, R.F. (1994), Involvement of the red nucleus in accurate timing of the rabbit’s classically conditioned eyeblink response. Society fo r Neuroscience Abstracts 20: 1010. 1 7 1 Krupa, D.J., Weiss, C., and Thompson, R.F. (1991). A irpuff evoked Purkinje cell complex spike activity is diminished during conditioned responses in eye blink conditioned rabbits. Society fo r Neuroscience Abstracts 17:322. Krupa, D.J., Thompson, J.K., and Thompson, R.F. (1993). Localization of a memory trace in the mammalian brain. Science. 260: 989*991. Lavond, D.G., Kim, J.J., and Thompson, R.F. (1993). Mammalian substrates of aversive classical conditioning. Annual Review o f Neuroscience 44: 317* 342. Lavond, D.G., McCormick, D.A., and Thompson, R.F. (1984). A nonrccoverable learning deficit. Physiological Psychology 12: 103*110. Lavond, D.G. and Steinmctz, I.E. (1989). An inexpensive interface for the IBM* PC/XT and compatibles. Behavioral Research Methods: Instruments and Computers 21: 435*440. Lavond, D.G., Steinmctz, J.E., Yokaitis, M.H., Thompson, R.F. (1987) Rcacquisition of classical conditioning after removal of cerebellar cortex. Experimental Brain Research 67:569-593. Lavond, D.G., Logan, C.G., Sohn, J.H ., Gamer, W.D.A., and Kanzawa, S.A. (1990). Lesions of the cerebellar interpositus nucleus abolish both nictitating membrane and cyelod EMG conditioned response. Brain Research 514: 138-248 Marck, G.J., McMastcr, S.E., and Harvey, J.A. (1984). The rote of the accessory abduccns nucleus in the rabbit nictitating membrane response. Brain Research 299: 215- 229. Martin, J.H. and Ghez, C. (1988). Red nucleus and motor cortex: parallel motor systems for the initiation and control of skilled movement. Behavioral Brain Research 28: 217-223. Mauk, M.D. and Thompson, R.F. (1987). Retention of classically conditioned eyelid responses following acute decerebration. Brain Research 493: 89-95. Mauk, M.D., Steinmetz, J.E., and Thompson, R.F. (1986). Classical conditioning using stimulation off the inferior olive as the unconditioned stimulus. Proceedings o f the National Academy o f Sciences 83:5349-53. McBride, R.L. and Klemm, W.R. (1968). Stereotaxic atlas of the rabbit brain, based on the rapid method of photography of frozen unstained sections. Communications in Behavioral Biology Part A 2: 179-215. 172 McCormick, D.A., Lavond, D.G., Clark, G.A., Kettner, R.E., Rising, C.E., and Thompson, R.F. (1981). The engram found? Role of cerebellum in classical conditioning of nictitating membrane and eyelid responses. Bulletin o f the Psychonomic Society 18: 103-105. McCrea, R.A., Bishop, G.A., and Kitai, S.T. (1978). Morphological and electrophysiological characteristics of projections neurons in the nucleus interpositus of the cat cerebellum. Journal o f Comparative Neurology 181: 397-420. Megirian, D. and Bures, J. (1970). Unilateral cortical spreading depression and conditioned eyeblink responses in the rabbit. Experimental Neurology 27: 34-45. Moyer, J.R., Deyo, R.A. and Disterhoft, J.F. (1990). Hippocampectomy disrupts trace eye-blink conditioning in rabbits. Behavioral Neuroscience 104: 243- 252. Neafsey, E.J. (1990). The complete ratunculus: Output organization of layer V of cerebral cortex. In B. Kolb & R.C. Tees (Eds.) The Cerebral Cortex o f the Rat. Cambridge, MA: The MIT Press. Neafsey, E.J., Bold, E.L., Haas, G., Hurley-Gius, K.M., Quirk, G., Sievert, C.F., and Terrcberry, R.R. (1986). The organization of the rat motor cortex: a microstimulation mapping study. Brain Research Reviews 11: 77- 96. Nelson, B. and Mugnoini, E. (1987). GABAergic innervation of the inferior olive complex and experimental evidence for its origins. In: The Olivocerebellar System in Motor Control. New York: Springer-Veriag. Nordholm, A.F., Swain, R.A., and Thompson, R.F. (1990). CS/US pre-exposure retards acquisition of discrete motor responses using HVI stimulation. Society fo r Neuroscience Abstracts 16: 270. Nordholm, A.F., Thompson, J.K., Dersarkissian, C,, and Thompson, R.F. (1993). Lidocaine infusion in a critical region of cerebellum completely prevents learning of the conditioned eyeblink response, Behavioral Neuroscience 107: 882-886. Norman, R.J., Buchwald, J.S., and Villablanca, J.R. (1977). Classical conditioning with auditory discrimination of the eyeblink in decerebrate cats. Science 196: 551-553. Nudo, R.J., Jenkins, W.M., and Merzenich, M.M. (1990). Repetitive microstimulation alters the cortical representation of movements in adult rats. Somatosensory and Motor Research 7: 463-483. 1 7 3 Oakley, D.A. and Russell, I.S. (1977). Subcortical storage of Pavlovian conditioning in the rabbit. Physiology and Behavior 18: 931*937. Paxinos, G. and Watson, C. (1982). The Rat Brain in Stereotaxic Coordinates. New York: Academic Press. Petrinovich, L. and Widaman, K.F. (1984). An evaluation of statistical strategies to analyze repeated-measures data. In H.V.S. Peeke & L. Petrinovich (Eds.), Habituation, Sensitization, and Behavior {pp. 155-201). Orlando: Academic Press, Inc. Port, R.L., Romano, A.G., Steinmctz, J.E., Mikhail, A.A., and Patterson, M.M. (1986). Retention and acquisition of classical conditioned responses by rabbits with hippocampal lesions. Behavioral Neuroscience 100: 745-752. Prince, J.H. (1964). The rabbit in eye research. Springfield, IL: Charles C. Thomas. Prokasy, W.F., Kesner, R.P., and Caldcr, L.D. (1983). Posttrial electrical stimulation of the dorsal hippocampus facilitates acquisition of the nictitating membrane response. Behavioral Neuroscience 97: 890-896. Rescorla, R.A. (1988). Behavioral studies of Pavlovian conditioning. Annual Review o f Neuroscience II: 329-352. Rescorla, R.A. and Wagner, A.R. (1972). A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and non-reinforcement. In A.H. Black & W.F. Prokasy (Eds.) Classical Conditioning II: Current Research and Theory. New York: Appleton, pp. 64-99. Schreurs, B.G. (1987). Parameters and sites of brainstem stimulation capable of eliciting the rabbit nictitating membrane response. Behavioral Brain Research 25: 155-160. Schreurs, B.G. (1988). Stimulation of the spinal trigeminal nucleus supports classical conditioning of the rabbit's nictitating membrane response. Behavioral Neuroscience 102:163-172. Sears, L.L. and Steinmetz, J.E. (1991). Dorsal accessory inferior olive activity diminishes during acquisition of the rabbit classically conditioned eyelid response. Brain Research 545:114-122. Shenk, J.W ., Wen, G.Y. and Wisniewski, H.M. (1986) Atlas o f the Rabbit Brain and Spinal Cord. New York: Karger. Shinkman, P.G., Swain, R.A., and Thompson, R.F. (1992). Intracerebellar electrical stimulation as both CS and US in classical conditioning of discrete motor responses. Society fo r Neuroscience Abstracts 18: 1562. 174 Solomon, P.R ., Stowe, G .T., Pendleberry, W.W. (1989). Disrupted eyelid conditioning in a patient with damage to cerebellar afferents. Behavioral Neuroscience 103:898*902. Solomon, P.R ., Vander Schaaf, H.R., Thompson, R.P., and Weisz, D.J. (1986). Hippocampus and trace conditioning of the rabbit’s classically conditioned nictitating membrane response. Behavioral Neuroscience 100: 729-744. Steinmetz, J.E. (1990). Neuronal activity in the rabbit interpositus nucleus during classical NM-conditioning with pontine-nuclcus stimulation CS. Psychological Science 1:378-82. Steinmetz, J.E ., Lavond, D.G., and Thompson, R.F. (1985). Classical conditioning of the rabbit eyelid response with mossy fiber stimulation as the conditioned stimulus. Bulletin o f the Psychonomic Society 23:245-48. Steinmetz, J.E ., Lavond, D.G., and Thompson, R.F. (1989). Classical conditioning in rabbits using pontine nucleus stimulation as a conditioned stimulus and inferior olive stimulation as an unconditioned stimulus. Synapse 3: 225-233. Steinmetz, J.E ., Lavond, D .G ., Ivkovich, D., Logan, C .G ., and Thompson, R.F. (1992). Disruption of classical eyelid conditioning after cerebellar lesions: Damage to a memory trace system or a simple performance deficit? The Journal o f Neuroscience 12: 4403-4426. Steinmetz, J.E. and Scngclaub, D.R. (1988). Direct projections from the lateral pontine nucleus to the anterior interpositus nucleus: a potential CS pathway for classical conditioning. Society fo r Neuroscience Abstracts 14:782. Swain, R.A. (1992) Intraccrebcllar stimulation and classical conditioning of discrete motor movements. Doctoral Dissertation, University of Southern California. Thompson, J.K ., Krupa, D .J., Weng, J., and Thompson, R.F. (1993). Inactivation of motor nuclei blocks expression but not acquiition of rabbit's classically conditioned eyeblink response. Society fo r Neuroscience Abstracts 19: 999. Thompson, J.K ., Lavond, D.G., and Thompson, R.F. (1986). Preliminary evidence for a projection from the cochlear nucleus to the pontine nuclear region. Society fo r Neuroscience Abstracts 12: 754. Thompson, R.F. (1989). Neural circuit for classical conditioning of the eyelid closure response. In J.H . Byrne & W.O. Berry (Eds.) Neural M odels o f Plasticity. New York: Academic Press, pp. 160-177. 175 Thompson, R.F. (1990). Neural mechanisms o f classical conditioning in mammals. Philosophical Transactions o f the Royal Society o f London 329: 161-170. Thompson, R.F., Berger, T.W ., Berry, S.D ., Hoehler, F.K ., Kettner, R .J., and Weisz, D.J. (1980). Hippocampal substrate o f classical conditioning. Physiological Psychology. 8: 262-279. Thompson, R.F. and Glanzman, D.L. (1976). Neural and behavioral mechanisms of habituation and sensitization. In T.J. Tighe adn R.N. Leaton (Eds.) Habituation: Perspectives fro m Child Development, Anim al Behavior, and Neurophysiology. New York: Lawrence Erlbaum Associates, pp. 49-93. Thompson, R.F. and Krupa, D. J. (1994). Organization o f memory traces in the mammalian brain. Annual Review o f Neuroscience 17: 519-549. Thompson, R.F., McCormick, D.A., Lavond, D.G ., Clark, G.A., Kettner, R.E., and Mauk, M.D. (1983). Theengram found? Initial localization o f the memory trace for a basic form of associative learning. Progress in Psychobiology and Physiological Psychology 10: 167-196. Thompson, R.F., Groves, P.M ., Teyler, T .J., and Roemcr, R.A . (1973). A dual-proccess theory of habituation: Theory and behavior. In H. V.S. Peeke and M.J. Herz (Eds.) Habituation Volume I: Behavioral Studies. New York: Academic Press, pp. 239-271. Thompson, R.F. and Spencer, W .A. (1966). Habituation: A model phenomena for the study of neuronal substrates of behavior. Psychological Review 73: 16- 43. Tracy, J ., Weiss, C ., and Thompson, R.F. (1991). Single unit recordings of somatosensory and auditory evoked responses in the anterior inteipositus nucleus in the naive rabbit. Society fo r Neuroscience Abstracts 17:322. Tsukahara, N., Oda, T ., Notsu, T. (1981) Classical conditioning mediated by the red nucleus in the cat. The Journal o f Neuroscience 1:72-79. Weiss, C .G ., McCurdy, M .L., Houk, J.C ., and Gibson, A.R. (1985). Anatomy and physiology of dorsal column afferents to forelimb dorsal accessory olive. Society fo r Neuroscience Abstracts 11: 182. Weisz, D .J., Harden, D .G ., Xiang, Z. (1992). Effects of amygdala lesions on reflex facilitation and conditioned response acquisition during nictitating membrane response conditioning in rabbit. Behavioral Neuroscience 2: 262- 273. 1 7 6 Weisz, D.J., LoTurco, J.J. (1988) Reflex facilitation of the nictitating membrane response remains after cerebellar lesions. Behavioral Neuroscience 102:203*209. Weisz, D.J. and Mclnemey, J. (1990). An associative process maintains reflex facilitation of the unconditioned nictitating membrane response during the early stages of training. Behavioral Neuroscience 104: 21-27. Welsh, J.P., Harvey, J.A. (1989) Cerebellar lesions and the nictitating membrane reflex: Performance deficits of the conditioned and unconditioned response. The Journal o f Neuroscience 9:299-311. Welsh, J.P., Harvey, J.A. (1989) Intra-cerebellar lidocaine: dissociation of learning from performance. Society fo r Neuroscience Abstracts 15:639. Whalen, P.l. and Kapp, B.S. (1991). Contributions of the amygdaloid central nucleus to the modulation of the nictitating membrane reflex in the rabbit. Behavioral Neuroscience 105: 141-153. Wiesendangcr, M. (1986). Redistributive function of the motor cortex. Trends in Neuroscience Woody, C.D. and Yarowsky, P.J. (1972). Conditioned eye blink using electrical stimulation of coronal-precruciate cortex of the cat. Journal o f Ncurophysiology 35: 242-252. Woody, C.D. (1982) Memory, Learning, and Higher Function. New York: Springcr-Vcrlag. Woody, C.D. (1984). In G. Lynch, J.L. McGaugh, and N.M. Weinberger (Eds) Neurobiology o f Learning and Memory. Woody, C.D. and Brozek, G. (1969). Changes in evoked responses from facial nucleus of cat with conditioning and extinction of an eye blink. Journal o f Neurophysiology 32: 717-726. Woody, C.D. and Engel, J. Jr. (1972) Changes in unit activity and thresholds to electrical microstimulation at coronal-pericruciate cortex of cat with classical conditioning of different facial movements. Journal o f Neurophysiology 35: 230-241. Woody, C.D., Yarowsky, P., Owens, J., Black-Cleworth, P., and Crow, T. (1974). Effect of lesions of cortical motor areas on aquisition of conditioned eye blink in the cat. Journal o f Neurophysiology 37: 385-394. 1 7 7 Woolsey, C .N ., Settlage, P.H ., Meyer, D.R., Spencer, W ., Hamuy, T .P., and Travis, A.M. (1952). Patterns of localization in precentral and “supplementary” motor areas and their relation, to the concept of a premotor area. In P. Bard (Ed.) Patterns o f Organization in the Central Nervous System (Proceedings o f the Association fo r Research in Nervous and M ental Diseases). Baltimore: The Williams & Wilkins Company, pp. 238*264. Woolsey, C.N. (1958). Organization of somatic and motor areas of the cerebral cortex. In H.F. Harlow & C.N. Woolsey (Eds.) Biological and biochemical bases o f behavior. Madison, WI: University of Wisconsin Press. Yang, B.Y. and Weisz, D.J. (1992). An auditory conditioned stimulus modulates unconditioned stimulus-elicited neuronal activity in the cerebellar anterior interpositus and dentate nuclei during nictitating membrane response conditioning in rabbits. Behavioral Neuroscience 106:889-899. Zhang, A. A., and Lavond, D.G. (1991). Effects of reversible lesions of reticular or facial neurons during eyeblink conditioning. Society fo r Neuroscience Abstracts 17: 869. 178

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Inactivation of the interpositus nucleus prevents transfer of the rabbit's classically conditioned eyeblink response from a light to a tone-conditioned stimulus

PDF

The Role Of Cerebellar Cortex In Classical Conditioning Of Discrete Motor Movements: Intracranial Electrical Stimulation Studies

PDF

Genetic dissections of the cerebellar circuitry involved in classical eyeblink conditioning

PDF

The role of brain -derived neurotrophic factor (BDNF) in fear conditioning

PDF

Dopamine related temporal processing errors in prenatally stressed rats

PDF

The exploration of the neural mechanism of conditioned taste aversions with reversible blockade.

PDF

Effects of muscimol injections on conditioned eyeblink versus limb flexion responses in the rabbit cerebellar interpositus nucleus

PDF

The Regulation Of Neurotrophins And Their Receptors: Studies On The Roles Of Bdnf Andtrkb During Development And Neurodegeneration In The Rat

PDF

Hemispheric asymmetries in maintaining vigilance for younger and older adults

PDF

Cortical spreading depression in classical conditioning: Issues in associated cerebral and cerebellar eyeblink circuitry.

PDF

Simulation and modeling of neurobiological networks with applications to cerebellar networks implicated in classical conditioning of the eyeblink response

PDF

Hemispheric Differences In Visual Search Of Tachistoscopically Presented Nonunitizeable And Unitizeable Verbal Material

PDF

Effects of a corneal anesthetic on the extinction of the classically conditioned response in the rabbit

PDF

Studies on the organization of learning and memory in a classical conditioning model system: Participation of nuclearis reticularis tegmenti pontis (NRTP) and lateral pons in rabbit eye blink con...

PDF

Contributions of the cerebellum and neural pathways in the extinction of discrete motor responses

PDF

Neuroinformatics and the cerebellum: Towards an understanding of the cerebellar microzone and its contribution to the well-timed classically conditioned eyeblink response

PDF

Perceptual integration and noise exclusion in the etiology of developmental dyslexia

PDF

Is Ideology A Determinant Of Local Government Fiscal Policy? A Case Study Of Los Angeles County Board Of Supervisors

PDF

Disorientation Behavior In Person - Computer Interaction (Cognitive Models, Menu-Based)

PDF

Cerebellar cortical-nuclear projections and classical eyeblink conditioning

Asset Metadata

Creator Ivkovich, Dragana (author)

Core Title A Study Of Classical Conditioning And The Eyeblink Response In Rabbit: Issues Of Learning Vs. Performance

Degree Doctor of Philosophy

Degree Program Behavioral Neuroscience

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag biology, neuroscience,OAI-PMH Harvest,psychology, experimental,psychology, psychobiology

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c20-574018

Unique identifier UC11226378

Identifier 9600994.pdf (filename),usctheses-c20-574018 (legacy record id)

Legacy Identifier 9600994.pdf

Dmrecord 574018

Document Type Dissertation

Rights Ivkovich, Dragana

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA