University of Southern California Dissertations and Theses

Biofeedback Control Of The Eeg Alpha Rhythm And Its Effect On Reaction Time In The Young And Old

(USC Thesis Other)

Biofeedback Control Of The Eeg Alpha Rhythm And Its Effect On Reaction Time In The Young And Old

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content BIOFEEDBACK CONTROL OF THE EEG ALPHA RHYTHM AND ITS EFFECT ON REACTION TIME IN THE YOUNG AND OLD by Diana Stenen Woodruff 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 (Psychology) June 1972 INFORMATION TO USERS This dissertation was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted. The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction. 1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity. 2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame. 3. When a map, drawing or chart, etc., was part of the material being p h o to g rap h ed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete. 4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced. University Microfilms 300 North Zoob Road Ann Arbor, Michigan 4*100 A Xaron Education Company 72-26,067 WOODRUFF, Diana Stenen, 1946- BIOFEEDRACK CONTROL OF THE EEG ALPHA RHYTWt AND ITS EFFECT ON REACTION TIME IN THE YOUNG AND OLD. University of Southern California, Ph.D., 1972 Psychology, experimental University Microfilms, A Company, Ann Arbor, Michigan THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED UNIVERSITY O F S O U T H E R N CALIFORNIA T H I ORADUATKSCHOOL UNIVERSITY PARK LOS A N O S L U . C A LIFO R N IA * 0 0 0 7 This dissertation, written by Diana Sjten eno;od ru f f........ under the direction of A®£... Dissertation Com mittee, and approved by all its members, has been presented to and accepted by The Gradu ate School, in partial fulfillment of require ments of the degree of D O CTOR OF P H IL O S O P H Y Dole Ju. n® . i . 9. ? 2 SERTATION COMMITTEE I . VJ I -Jt_ PLEASE NOTE: Some p ag es may have In d istin ct p rin t. Filmed as r e c e iv e d . University Microfilms, A Xerox Education Company ACKNOWLEDGMENT A dissertation is the product of input and contri butions from a number of individuals in the training and development of the doctoral candidate. In this sense, | dissertation research represents the work of more than i one individual, and at this point I want to acknowledge the many people who have shared in this project with me. | Equipment and technical assistance were made available to me from several sources. Garst Reese pro- I vided me with the knowledge and apparatus for biofeedback conditioning, and George Seacat, Dick Burton, and George Rhodes, of the Long Beach Veterans Administration Hospital invited me to use the laboratory at that facility. Dr. Seacat and Dr. Burton helped me to recruit subjects, and Dr. Rhodes assisted me in many ways throughout my stay at the Veterans Administration J Hospital. Most of all, I appreciated Dr. Rhodes' willing ness to assist me when there were apparatus failures, no matter when they occurred. He bailed me out of tight | situations a number of times. I As a N. I. C. H. D. trainee at the Gerontology | Center, I profited in many ways. In addition to the | identity and support provided by the Center, I had the opportunity to interact with so many competent individuals 11 that they are too numerous to name. I would like to single out one, however, who had an important role in this dissertation. As Dr. Birren's assistant, Eleanor James j helped me to work with him, and she also gave me her I personal support and attention. Her contribution was to make some difficult aspects of this project more manageable I and enjoyable. | A primary source of support for this project has I come from the members of my dissertation committee. I feel unusually lucky for having worked with some of the most competent and supportive professors a candidate could have. Each of these professors made a unique contribution to my development. Although he was the out* side member of the committee, David Lindsley's input to | my training and research was substantial, and I expect I i | that his Influence on my interests will become even more apparent in future projects which I undertake. Gary i Galbraith is in many ways responsible for this research, for, without his knowledge of electrophysiology and his technical assistance in the laboratory, 1 would have been unable to complete this dissertation. Jacek Szafran I made a tremendous contribution to my training in the j psychology of aging and in my growth as a professional. I 1 was imprinted early in my training at USC on Dr. Szafran's challenging style and high standards, and it ill is to my profound regret that he died three days before my oral examination. It is my hope that his ideas will | live on through me and his other colleagues and students, j I am grateful to James Walker for being so kind as to J represent Dr. Szafran on my committee. | The chairman of my committee, James Birren, has my total admiration and devotion. His supportive style, | his awesome capacity, his complete commitment and the I many other superlatives which he personifies have been and will continue to be my greatest inspiration. It is i I with tremendous pride that I be known as his student. i i Other people also helped to make this task a pleasant experience. The subjects who participated in the experiment were extremely generous with their time, | and even more generous have been my family, and i especially my husband. Encouragement and tolerance have j consistently come from my family. It is my husband, Rob, i | though, who deserves the most credit for supporting me in this endeavor. In many respects this final product 1 represents his hard work as much as it represents mine. In no way could I have done it without him. iv TABLE OF CONTENTS ACKNOWLEDGEMENT INTRODUCTION CHAPTER [ I BACKGROUND ............................... Age related changes in the speed and | timing of behavior i Age changes in the electroencephalogram (EEG) j The relationship between EEG and reaction | time I Biofeedback manipulation of the EEG I I Reaction time-aging-EEG relationships and biofeedback manipulation II METHOD Subjects Apparatus Procedure Analyses III RESULTS ................................... EEG frequency-reaction time relationships Biofeedback conditioning i ! IV DISCUSSION ............................... I i Interpretations > Implications TABLE OF CONTENTS (CONT’D) I CHAPTER SUMMARY AND CONCLUSIONS 125 REFERENCES 131 vl LIST OF TABLES Effect of variation in number of alpha cycles within a RT interval on RT variability (hypothetical data) ........ Effect of variation in number of alpha cycles within a RT interval on RT variability ............................. Summary table for analysis of variance of the effect of age, brain wave fre quency, and subject on simple reaction t i m e ................................... Correlations between EEG period and reaction time in the baseline condition for young and old experimental and control subjects ....................... Correlations between EEG period and reaction time in the mean, fast, mean, slow, and mean conditions for young and old experimental subjects .......... Correlations between EEG period and reaction time for all conditions for young and old experimental and control subjects................................. Summary table for analysis of variance of the effect of age, feedback, and practice on learning to increase per cent time in the mean brain wave frequency . . . . Summary table for analysis of variance of the effect of age, order, and frequency on trials to criterion in a biofeedback task ..... ......................... LIST OF FIGURES FIGURE PAGE 1. Block Diagram of biofeedback conditioning apparatus.................... 65 2. Experimental design.................... 71 3. The effect of brain wave frequency on mean reaction time, reaction time variability, and EEG frequency for young and old sub jects in three biofeedback conditions . . 75 4. The effect of brain wave frequency on reaction time— a comparison of selected and total reaction time data for young and old subjects............................. 81 5. The effect of biofeedback on EEG frequency-- a comparison of selected and total EEG data for young and old subjects........ 82 6. Acquisition curves for biofeedback condition ing at mean frequency for young and old experimental subjects ............ 93 7. Acquisition curves for biofeedback conditioning at mean frequency for matched experimental and control subjects................................. 94 8. Acquisition curves for biofeedback con ditioning in the speed condition for young ana old subjects.................. 96 9. Acquisition curves for biofeedback con ditioning in the slow condition for young ana old subjects.................. 97 | 1 viii I I I INTRODUCTION i The purpose of this research was to provide an | experimental test of the hypothesis that the frequency 1 of the electroencephalogram (EEG) is the master timing mechanism for behavior. Slowing of reaction time with age [ Is a we11-documented phenomenon, and an age change which is almost equally well established is the slowing of the EEG alpha rhythm with age. Descriptive research undertaken by Surwillo (1960, 1961, 1963, 1964a, 1968, 1971) established a correlational relationship between age, brain wave frequency, and reaction time. The advent of the biofeedback technique which provides a method for the operant con ditioning of brain wave frequency presented a means for experimentally manipulating this age- and behavior-related variable. Hence, the biofeedback technique was used to investigate the causal nature of the relationship between age, brain wave frequency, and reaction time. For the past several decades investigators ,interested in the psychology of aging have undertaken and !produced a large body of descriptive research literature covering age changes and age differences in a wide range of I behaviors. These investigations have typically involved i comparisons of individuals of different age groups on a 1 2 1 given behavior (cross-sectional design), or they have I Involved the repeated measurement o£ selected behaviors in i the same group of individuals over a span of time j | (longitudinal design). Descriptive research is useful i inasmuch as it can lead to the identification of those | behaviors which differ between age groups and which may change over time. Hence, descriptive studies are useful : for the identification of age-functional relationships. Descriptive research is useful, but it represents only the first step of the scientific endeavor which involves the description, explanation, and modification of natural ; phenomena. Once age-related changes have been identified, the next logical step is to determine the causes of such ; changes and to try to bring them under experimental control j To describe behavior as a function of age and to identify i those behaviors which change with age does not lead to the understanding of ontogenetic change. Time itself is not a < causative factor, and as Flavell (1963) stated, "Age is a jvehicle rather than a cause in itself" (p. 36). Hence, the i task of experimentally-minded developmental researchers is | to explain the age variable--to identify the antecedents j 1 for the observed changes which occur with age. | While developmental psychologists have identified i i I numerous difficulties with descriptive research and have 3 'refined research strategies and designs aimed at more ’accurate description of age functions (e.g. Baltes, 1968; jBaltes & Nesselroade, 1970; Birren, 1959; Birren, Woodruff j j& Bergman, 1972; Schaie, 1965; Wohlwill, 1970), there has been a lag in the refinement of methodology and design | which would lead to the explanation of the observed age changes in behavior (Baltes & Goulet, 1971). To understand and develop empirical laws for behavior over the life span, an important thrust must be to develop and implement ex perimental research strategies for the major variables underlying development. A major drawback to the implementation of experi- | mental research strategies in developmental psychology is j the fact that it is impossible to manipulate chronological age. Chronological age is an organismic, assigned variable (Kerlinger, 1964). It cannot be varied, replicated or arbitrarily assigned, and it is therefore not an experi mental variable. In an attempt to devise indices with I greater explanatory power than chronological age, Birren I 1 (1959) suggested the use of indices of functional age in j terms of biological, psychological, and social parameters, j The substitution of an index of functional age for ! chronological age eliminates some of the drawbacks of chronological age as an experimental variable as in some 4 cases it Is possible to manipulate such an index, i In the present experiment, brain wave frequency was jused as an index of age with old age simulated by slow 'brain wave frequencies and young adulthood simulated by i ifaster frequencies. Since brain wave frequency could be |manipulated (Bundzen, 1966; Brown, 1968, 1969; Dewan, 1964, |1966; Kamiya, 1968, 1969; Green, Green & Walters, 1969, ,1970, 1971; Hart, 1967; Woodruff, 1971), subjects could be randomly assigned to a given age category, their "age" (in terms of brain wave frequency) could be varied, and differing "ages" could be replicated as subjects learned to shift from one frequency to another and back again. If brain wave frequency is an antecedent for some of the con- i sequent age-related changes in behavior, then the experi- ] I mental manipulation of brain wave frequency should lead to jalterations in behavior which are similar to alterations I which normally occur developmentally. Thus the attempt to manipulate brain wave frequency in the present experiment represented an attempt to experimentally manipulate one variable associated with age. There is a wide range of dependent behavioral variables which have been previously identified as changing with age, and a number of these variables have been related i :to parameters of the EEG. Hence, the number of dependent 5 variables which could have been used in this experiment is I large. Cognitive and intellectual variables known to i | change with age have been related to ongoing EEG (e.g. Bunkier, 1967; Obrlat et al.. 1962; Stoller, 1949; ; Wang et al., 1970), and EEG changes with age also have implications for age changes in memory (e.g. Hoagland, 1954 | Green, Green & Walters, 1970), attention and arousal ; (e.g. Schenkenberg, 1970; Tecce, 1971), personality (e.g. Davis, 1942; Hurst et al., 1954), and speed and timing of behavior (e.g. Obrist, 1965; Surwillo, 1968). There may be scientific merit in observing the effect of brain wave manipulation on all of the behavioral parameters . which have been related to age and EEG changes. In all of j the developmental literature, however, the most impressive | EEG-behavior correlations and the most specifically stated | hypotheses about age, behavior, and EEG have been 1 presented by Surwillo (1960, 1961, 1963a, 1963b, 1964a, 1964b, 1966, 1968, 1971) in the area of speed and timing. Since changes in the speed and timing of behavior are such a primary concomitant of age and since Surwillo presented i clear evidence of an EEG-age-reaction time relationship, ' reaction time was selected as the major dependent variable ’ in this investigation. The theoretical background and rationale for this r 6 experiment involve four fairly distinct areas of research. i i 1.) The literature on age changes in the speed and timing i | of behavior serves as a background for the choice of the | dependent variable. 2.) Discussion of the independent variable involves the body of research dealing with age- i 1 related changes in the EEG. 3.) Descriptive research ; relating parameters of the EEG to the speed and timing of behavior and to age changes in speed and timing provides the rationale for the present experiment. 4.) The back ground material for the methodology used to manipulate EEG parameters is provided by the literature describing the biofeedback technique. Since each of the four areas has evolved independently, separate sections will be devoted to a summary of the literature in each area. Subsequently, : the synthesis of the material along with the hypotheses i method, results, discussion, and conclusion for this dissertation research will be presented. CHAPTER I BACKGROUND . Age-Related Changes in the Speed and Timing ) or Behavior The slowing of response speed with age is a j reliable and we11-documented phenomenon which probably occurs in every member of the human species if he lives long enough (Birren, 1964, 1965). Psychomotor slowing does not appear to be dependent on age-related changes in health status as slowing occurs in even the healthiest of individuals (Birren, et al., 1963; Szafran, 1966, 1968). Age changes in speed and timing are general inasmuch as they are not related to changes in specific sensory ' modalities (e.g. Koga & Morant, 1923), and because the i ' slowing is not specific to certain tasks (Birren, Riegel, j & Morrison, 1962; Chown, 1961). Psychomotor slowing with i age also generalizes to other species as slowing has been observed in animals as well as in humans (Birren, 1955; Birren & Kay, 1958; Brody, 1940). Thus data from a number of different lines of research support the notion ' that the slowing of psychomotor speed with age is an in- i variant biological transformation. Because slowing is a I | general phenomenon and occurs irregardless of health i status, it has been considered as a primary aging factor 2______________________ 8 ; (Birren, 1965). j Reaction time is typically measured in a paradigm i in which a stimulus (or several stimuli) is (are) presented j ! to a subject who is required to make a response (such as i pushing a button) within the shortest possible delay. The time elapsing between the onset of the stimulus and the ! response of the subject is operationally defined as the | I reaction time. Simple reaction time involves the presentation of one stimulus to the subject while complex (also called disjunctive and choice) reaction time involves i the presentation of two or more stimuli and the opportunity for two or more means of response. , Simple Reaction Time Some of the first reaction time data was collected in the nineteenth century in Galton's Anthropometric \ Laboratory. Part of these extensive data from more than 9000 male and female subjects, ranging in age from young children to the aged, were analyzed by Koga and Morant i , (1923). These data indicated that simple reaction time increased with age to both visual and auditory stimuli, : and the slowing was independent of age differences in | acuity in a given sensory modality. The fastest reaction i i times were recorded for subjects aged eighteen to twenty when mean auditory reaction time was 154 msec, and mean 9 visual reaction time was 182 msec. Subjects in the J seventies had mean reaction times of 174 msec, and 205 i msec, for auditory and visual reaction time, respectively. i 1 Hence, both visual and auditory reaction time increased by | about 20 msec, between the ages of 18 and 70. Subsequent i | studies confirmed these results for visual and auditory I i ' reaction time (e.g., Beilis, 1933; Birren & Botwinick, 1955; | Miles, 1931; Obrist, 1953). Additional confirmation of the i finding that psychomotor slowing occurs independently of ! sense modality has been provided by studies showing age ; changes in simple reaction time to tactile stimuli : (Hugin, Norris & Shock, 1960; Magladery, Teasdall & Norris, 1955). Whenever speed of reaction has been compared in I 1 adult groups of different ages, older individuals have i | been found to be slower than younger individuals. Hicks and Birren (1970) summarized studies relating age and reaction time and found that the data demonstrated that simple reaction time slowed anywhere from 11% (tactile stimulus; Magladery, et al., 1955) to 102% (visual stimulus; Beilis, 1933). All of the reaction time studies which have been | described have been of the cross-sectional design involving I comparisons of different age groups on reaction time. Since age differences in cross-sectional studies can 10 result from cohort differences as well as from age changes, 1 additional data are required to establish the fact that ' the difference in reaction time between young and old I 1 groups is indeed an age change. i i The evidence which establishes psychomotor slowing | as an age change comes from several sources. First, | reaction time data have been collected for almost 100 years, Galton's data, analyzed by Koga and Morant (1923) was collected at the International Health Exhibition in London in 1884. The cohort bom in 1864 which was twenty years old and at peak reaction time as measured by Galton in 1884 would be an aged cohort measured around 1934. Reaction time studies undertaken around 1934 demonstrated 1 age differences in reaction time (e.g., Beilis, 1933; I Miles, 1931), hence suggesting that age related slowing j occurred in that cohort. Galton's data are consistent with : data collected in subsequent decades on different cohorts of the same age groups. Taken as a whole, the cross- i sectional data on reaction time indicate that the slowing in reaction time with age is consistent over a number of cohorts and is indeed an age change. The other source of : evidence indicating that the slowing of response speed is ; an age change was provided by Botwinick and Birren (1965) I who did a five-year longitudinal study of reaction time. n 11 j When reaction time on a card sorting task was retested In i seventeen men of a mean age of seventy-five, there was a : statistically significant Increase In reaction time over ' the five-year span. These empirical data, coupled with I i the layman's typical acknowledgement that one does Indeed slow down with age leave little doubt that the slowing of | psychomotor speed is a reliable age change. Several investigations have indicated that reaction time can be improved with practice regardless of age (Leonard, 1953; Rutherford, 1894). Mowbray and Rhoades (1959) demonstrated that with an extensive practice (45,000 reactions) one young subject improved at a two- choice reaction time task by 25% (from 392 msec, to t 218 msec.) and at a four-choice reaction time task by 27% ; (from 303 msec, to 222. msec.). Such studies clearly ; indicate that practice is an important factor in the speed i of response, and the data suggested to Murrell (1970) that reported age differences in reaction time may repre sent differences in capacity to perform an unfamiliar task i as well as differences in the ability being tested. Murrell (1970) reasoned that the old may take longer to | adapt to the new situation and therefore be handicapped in | typical reaction time study paradigms which involve | samplings of ten to fifty reaction time trials. To test 12 jthis notion, Murrell (1970) sampled 12,500 to 16,200 ! reaction time trials on three subjects aged 17, 18, and 57 iyears. Initial age differences in reaction time were i largely eliminated with practice, although the older subject did perform slightly more slowly on two- and eight-choice jreaction time tasks. Younger subjects improved at the I beginning of the task while the older subject took up to I I 1300 responses (on the complex tasks) before improvement ishowed. These data indicate that practice is an important variable in reaction time tasks, but they also demonstrate that, even with extensive practice, age differences in reaction time are not eliminated. i j Complex Reaction Time | On complex reaction time tasks, age-related slowing is slightly greater than it is for simple reaction time tasks (Goldfarb, 1941; Griew, 1959; Miles, 1931; Singleton, 1955; Rabbitt, 1965; Talland, 1961). Increasing the number of choices makes the task more difficult for the subject, iand his reaction time increases. The old seem to be relatively more hampered by the increase in task difficulty than are the young. For example, Goldfarb (1941) found ;that the difference between young (18 to 24 years) and old (55 to 64 years) subjects on a simple reaction time task | was 11 msec., on a two-choice task the difference was i i 57 msec., and on a five-choice task there was a 66 msec. 13 difference. The addition of information appears more j embarrassing to the old than to the young subject. This j finding suggests that the older organism is less efficient f ; in processing information. The older nervous system does i . not handle input as well as the younger nervous system, i and numerous studies have been undertaken to identify the | parts of the nervous system which change with age and i alter the information processing capabilities of the organism. Movement Time One possible difference between young and old 1 individuals which could account for age differences in | reaction time is an age difference in movement time. Precisely how much movement time slows with age, compared with decision or reaction time, is difficult to determine. The reason for this lies in the fact that movements are usually measured in a context where decisions are involved. That is, movement time measurements can be easily con- taminated with an undetermined contribution of time taken i j in the decision process governing the intent of the move- ; ment. Tapping speed is a measure of movement time which ! is probably least involved with decisions. Hence, tapping ! ; speed may be a relatively "pure" measure of movement time. I i < 14 Talland (1962) found small increases in age in tapping speed. Miles (1931) and Pierson and Montoye (1958) found i that the old move more slowly, but Szafran (1951) found no age difference in movement speed when subjects were blind- I ifolded. Hence, physical changes in effectors cannot account for the slowing of movement time as the old are i apparently capable of moving as fast as the young in icertain conditions. Szafran (1951) attributed the slower movement time of older subjects in normal visibility con ditions to their greater tendency to visually monitor their : movements. In spite of the observed increases in movement time ;with age, Birren (1955) argued that movement time alone cannot account for increases with age in reaction time. When movement time is held constant while task difficulty is increased, the difference in reaction time between young and old increases as the task becomes more complex. Hence, movement time accounts for only a small proportion of the age-related slowing. j I Sensory Acuity j Although a large number of Independent changes occur in receptors of all sensory modalities (for summaries of these data see Birren & Szafran (1969; Jakubczak, 1967; Magladery, 1959), these age-related ' changes can account for only a small proportion of the ; total slowing of reaction time. Koga and Morant (1923) clearly demonstrated that reaction time had little relation 1 to the acuity of sensory receptors, and the association ! between reaction time and visual and auditory stimuli was i much greater than the relation between sensory acuity and reaction time. These results suggested that a process : common to all sensory modalities was related to the age , change in psychomotor speed. ; Conduction Velocity Since all information from sensory receptors is i coded in nerve impulses and travels to the brain in | afferent pathways, the slowing in nerve conduction velocity j with age could lead to the results observed by Koga and Morant (1923). Several groups of investigators have measured conduction velocity in human peripheral nerves and have found decreases in mean conduction velocity from the ages of 30 to 80 to be 3 m/sec. (Wagman & Lesse, 1952) and 10 m/sec. (Norris, Shock & Wagman, 1953). Birren and Wall (1956) found no change in conduction velocity in the sciatic nerve of rats. Norris, et al^ , (1953) pointed out |that small age changes in human peripheral nerve conduction velocity can account for only 4 msec, (assuming a 1 meter 16 'pathway), and this accounts for only a fraction of the ! I observed age change in reaction time. I Testing the possible significance of age changes in conduction velocity in another manner, Birren and / jBotwinick (1955) measured simple auditory reaction time of |the foot, finger, and jaw in old and young subjects. The |investigators reasoned that if conduction velocity was a I factor in age-related slowing, the difference between old i and young in foot reaction time (transmitted over a long I peripheral pathway to and from the brain) would be jrelatively greater than the age difference in jaw reaction |time (transmitted over a short pathway). The age I jdifference in foot, finger, and jaw reaction time was equal with the old always slower than the young, and Birren and Botwinick (1955) concluded that age changes in peripheral conduction velocity could not account for age changes in i reaction time. Inference led to the conclusion that the slowness of older subjects was a function of the central inervous system rather than of peripheral structures. Synaptic Delay The synapse is another structure in the nervous system which might change with age, and age changes in synapses could lead to observed age change in reaction time. 17 I Wayner and Gamers (1958) measured synaptic delay in a mono- | synaptic reflex in rats and found a significant increase in ; synaptic delay from .97 msec, in young rats to 1.36 msec. i 1 in old rats. This represents an increase of 40% suggesting that a large proportion of the slowing of behavior may be | accounted for by the suomation of synaptic delays in the : central nervous system. Since a greater number of neurons ; and hence a greater number of synapses would probably be involved in a complex than in a simple reaction time task, age changes in synapses could also account for the greater slowing observed in complex reaction time tasks. Central Factors The evidence clearly indicates that peripheral factors (movement time, sensory acuity, conduction velocity) alone cannot account for the oiagnitude of the age j change in reaction time. Age changes in synapses and in the functioning of the brain and brainstem where sensory input and motor output are integrated seem to be the loci where the major changes leading to slowing occur. It would seem appropriate, therefore, to focus on the measure* ment of central factors which might be related to the age I I ! changes in psychomotor speed. One such measure of brain activity is the electro- i | encephalogram (EEG). The EEG is recorded with surface 18 electrodes and reflects cortical and subcortical neuronal activity (e.g., Andersen & Andersson, 1968; Lindsley, 1960) More specifically, the EEG reflects the post-synaptic and local dendritic potentials. Since the EEG is a reliable and valid measure of the activity of the brain, it is a ! ' useful tool for assessing age changes in brain activity. Age Changes in the Electroencephalogram (EEG; The description of age changes in the EEG has been the topic for numerous investigations, and a number of parameters of the EEG have been found to change with age. Since the present experiment involves only one EEG para meter, it is beyond the scope of this dissertation to present a detailed description of age differences in the large number of EEG parameters on which developmental data is available. Therefore, only a brief description of the major age differences in EEG parameters will be presented. Following this discussion, the age difference in EEG whichhas been most prominently reported in the developmental EEG literature, the slowing of the dominant EEG alpha frequency, will be discussed in greater detail. The extensive literature on age differences in parameters of the EEG has been the subject for several excellent review papers (Obrist & Busse, 1965; Thompson, 1972), and I 19 I i I the material in this section represents an encapsulation of those reviews. For thorough documentation of the results reported in this section, the reader is referred to the exhaustive review papers cited above. General EEG Changes The major changes which have been observed in rest ing EEG are: the slowing of the dominant brain wave (alpha) rhythm, decrease in the amplitude and abundance of the dominant rhythm, increase in focal slow activity (localized episodic bursts of high voltage activity in the delta and theta range), slight increases in diffuse slow activity (delta and theta activity distributed in several areas of the brain) in subjects over 75 years, increase in fast activity (activity in the frequency range of 18 to 30 hz.) in middle age, and becoming less apparent in old age. The averaged evoked potential (EP) is a method of measuring brain activity which involves averaging the on going EEG. The EP is a summation of brief segments of EEG activity immediately following the presentation of a stimulus. Thus, the EP provides an assessment of electro- cortical responsivity to a stimulus. A growing number of recent investigations report consistent trends in EP data collected in cross-sectional designs (Dustman & Beck, 1969; 20 ! Luders, 1970; Schenkenbergt 1970; Shagass & Schwartz, 1965; | Straumanis, Shagass & Schwartz, 1965). Changes in EP | which have been observed with age are: latencies of most EP components to visual, auditory, and somatosensory stimuli increase with age, and latencies during the first 1 200 msec, show the greatest increases; amplitude of the early components during the first 100 msec, following the stimulus tend to increase with age for all three sensory modalities while amplitude of late components (up to 400 msec, following the stimulus) tend to decrease with age in all sensory modalities. These age changes become apparent in the age groups forty to fifty and become progressively more apparent in the older groups. Resting EEG and EP's have been the most extensively studied parameters, but several other parameters have received some attention. Andermann and Stoller (1961) and Verdeaux et al. (1961) measured electrocortical reactivity (desynchronization or alpha blocking as a result of sensory input) as a function of age and found it to decrease in older subjects. Wilson and Obrist (1963) corroborated this result when they compared the EEG reactivity in healthy elderly subjects selected for superior intelligence to reactivity in normal young controls. These data suggest i that the older brain may be less excitable. 21 A relatively new EEG measure, Contingent Negative Variation (CNV) involves the measurement of DC activity in an interval between a warning signal and a signal to which the subject has to respond. Typically, in this interval there is a slow rise in negativity which peaks Immediately before presentation of the second stimulus and shifts to a positive direction after the second stimulus has been presented. The CNV has been related to a number of psychological constructs, including attention or arousal (Tecce, 1971), but the meager available data (Thompson & Nowlin, 1971) suggest that healthy aged men develop CNV's which are comparable jto the CNV*s observed in young subjects. Taken as a whole, these EEG data suggest that the aging brain may function differently, and probably less efficiently, than the young brain. Such age differences in brain function have important implications for age differences in behavior, and it will be useful for investigators to examine the behavioral significance of each of the age differences which has been described. Since the purpose of this dissertation is to examine the significance of the slowing of the dominant brain wave rhythm for reaction time behavior, the following discussion will elaborate on the description of this major age 22 difference in EEG. i Age Changes in EEG Alpha Rhythm i When an individual is awake, but in a relaxed state with closed eyes and/or lacking visual stimulation, his EEG from the occipital area contains easily detectable large |amplitude waves of 8 - 12 hz. These are called alpha waves, or alpha rhythm. As early as 1931 Berger noticed that patients with senile dementia had slower alpha rhythm than normal individuals, but he assumed that the slowing was related to the patients' illness. Berger (1931) did not associate the slowing with aging. Davis (1941), however, demonstrated that the alpha rhythm slowed with age, and her work has been upheld in later research (Busse & Obrist, 1963; Brazier & Finesinger, 1944; Fredlander, 1958; Harvald, 1958; Matousek, Volavka, Roubicek & Roth, 1967; iMundy-Castle, Hurst, Beerstecher & Prinsloo, 1954; Obrist, 1954, 1963; Otomo, 1966; Silverman, Busse & Barnes, 1955). Thus, the mean EEG frequency in young adults is 10.2 - 10.5 hz. (Brazier & Finesinger, 1944), whereas the average frequency for people in their sixties in 9.1 hz., and for people beyond eighty, it is further decreased to 8.6 hz. (Obrist, 1954, 1965). Alpha recordings from healthy :Community volunteers between 28 and 99 years of age showed 23 that the alpha period (reciprocal of alpha frequency) ! increased at a rate of four msec, per decade (Surwillo, j 1963). I 1 Since the studies reported above were cross- sectional, they did not rule out the possibility that the ! decreases in alpha frequency resulted from cohort dif ferences or from selective dropout of subjects. However, Obrist, Henry & Justiss (1961) found in a longitudinal study that two-thirds of the subjects showed a progressive decline in frequency over a ten year period. Hence, the ‘ results of the cross-sectional studies seemed to represent i an age change rather than a cohort difference. In one individual the alpha rhythm declined from 9.4 hz. at the age of 79 to 8.0 hz. at the age of 89. In addition, it was found that dropouts due to death occurred more often j with individuals with slow alpha than with individuals having fast rhythms. This finding suggests that data from cross-sectional studies may underestimate the magnitude of alpha slowing that may occur in a given individual. In a subsequent longitudinal study, Wang and Busse (1969) corroborated the results of Obrist, Henry i and Justiss (1961). Other reported changes in alpha pertain to its | general abundance and amplitude. Mundy-Castle et al. i i (1954) found a decrease in percent-time alpha, particularly ! with evidence of increased dementia. Obrist and Henry ; (1958a, b) also noted a pronounced decrease in the amount ' of alpha activity which was replaced by slower wave forms I in elderly deteriorated patients. The factors leading to the observed age changes in the EEG alpha rhythm are not clearly understood. Obrist (1963, 1964) presented impressive evidence which suggested that vascular disease is an important factor in alpha slowing. Comparing arteriosclerotic and normal groups matched in age, Obrist and Bissell (1955) found slower alpha frequencies in the diseased group. These results were replaced by Obrist (1963) when he compared a highly select group of healthy old men to a matched age group with mild asymptomatic arteriosclerosis. Dastur et. al. (1963) studied cerebral circulation and metabolism in these same healthy old men and found no significant differences between the elderly men and a group of normal young controls. The group with asymptomatic disease showed a 10 to 16% decline in cerebral blood flow, i and cerebral oxygen consumption was also decreased in this i group though the decrease was not statistically significant. Such results led Obrist and Busse (1965) to suggest 25 : that age-related slowing of the alpha rhythm is related to vascular pathology. Hence, they argued that alpha slowing is a result of disease rather than a part of the normal aging process. The issue of whether the declining alpha rate results from normal aging or from pathology has not ! | been settled. The mean alpha frequency of the healthy elderly men observed by Dastur et. al. (1963) and by Obrist (1963) had slowed to 9 hz., and this decrease in frequency was statistically significant (Obrist, 1963). Since alpha slowing has been observed in even the health iest of aged samples, it is questionable that all of the slowing is related to disease processes. Obrist (1963) argued, however, that even this healthy group may have ; had subclinical cardiovascular pathology which caused the slowing to occur. This is, however, an open-ended | hypothesis in the sense that it may imply that all age changes are due to undetermined pathology, circulatory or otherwise. In any case, the slowing of the dominant EEG rhythm with age appears to be a highly reliable i phenomenon occurring even in the healthiest of aged samples. Sunmarizing this issue, Thompson (1972) con- i eluded that the likelihood that there are at least two l contributing factors to alpha slowing (age and pathology) should not be overlooked. 26 The Relationship between EEG and Reaction Time A domain of behavior in which alpha slowing may j ! have great significance is in speed and timing. Investigators have frequently attempted to link EEG frequency to the timing of behavior, and in a number of studies they have presented evidence for a relationship between parameters of the alpha rhythm and response speed. Early EEG-reaction time studies involved the attempt to relate the latency of alpha-blocking (the desynchronization of the alpha rhythm which occurs at the onset of a novel stimulus) to the speed of response. In a number of studies (Bakes, 1939; Jasper & Cruickshank, 1937; Knott, 1939; Travis, Knott & Griffith, 1937) signi ficant positive correlations were found between reaction time and alpha blocking to both visual and auditory stimuli. Stamm (1952) pointed out, however, that alpha blocking was causally related to the motor response as the motor response typically occurred before the desyn chronization. A likely explanation for the correlation between the events is that they are both related to a third neurophysiological mechanism such as the reticular activating system. A number of EEG-reaction time studies were under- | i i ! 27 r [ i taken to test hypotheses in terms of reticular activation theory. Speculating that alpha desynchronization j represented a state of excitation in the cortex produced i I by reticulo-cortical impulses while synchronization of alpha represented a state of cortical inhibition, Lansing, Schwartz, and Linds ley (1959) undertook an experiment in which they compared speed of response to stimuli presented during alpha activity to response speed to stimuli presented after a warning signal which desynchronized alpha activity. No surprisingly, Lansing et. al. (1959) observed significantly faster responses during desyn chronized alpha, and Fedio, et. al. (1961) replicated these results. The problem with the design of these studies was that reaction times to stimuli preceded by a i I warning signal were compared to reaction time to stimuli I presented with no warning or presented with a warning at a different inter-stimulus interval. More recent studies (Thompson & Botwinick, 1966; Leavitt, 1968) demonstrated that there was no relationship between degree of alpha desynchronization and reaction time. Subjects perform equally well regardless of the state of alpha desynchron ization. A third factor, the time interval between the warning signal and the reaction time stimulus, accounts for the correlation between alpha desynchronization and 28 reaction time as both variables are dependent on the inter-stimulus interval (Leavitt, 1968). Busk (1971a) argued that the lack of differential effect on reaction time of the widely divergent EEG i patterns used by Thompson and Botwinick (1966) and i Leavitt (1968) did not rule out the validity or usefulness of the EEG as a measure of central integration. The point made by Busk (1971a) was that much of the information in the EEG was lost by merely dichotomizing it into states of alpha and non-alpha. It is possible that theta (4-7 hz.} activity might be associated with slower reaction time, beta (13 - 30 hz.) activity could be related to faster reaction time, and alpha (8 - 13 hz.) activity might be associated with reaction time in the middle range. By grouping beta and theta together as is done in the alpha, i non-alpha dichotomization, differences in reaction time occurring in alpha as opposed to non-alpha states will be eliminated. Hence, a more valid test of the EEG-reaction time relationship would be to relate specific EEG frequencies to reaction time. A number of studies have proved this approach to be useful. Specific frequency ! bands in auto- and cross-spectra have been related to variables such as level of activation (Walter, Rhodes & 29 1 Adey, 1967), levels of stress (Berkhout, Adey, & Walter, 1969), performance on a visual-motor coordination task (Busk, 1971b), and reaction time (Galbraith, in press). On the basis of physiological evidence, Andersen and Andersson (1968) hypothesized that stimulus infor- | mation may be registered in the cortex only after the i cortex has been "primed” by two or more rhythmic impulses from the thalamus. The investigators suggested that thalamo-cortical volleys are the physiological basis of the alpha rhythm. This hypothesis implies that alpha frequency and reaction time would be related. Two or more thalamo-cortical volleys at a slower alpha frequency would take longer and hence be registered in the cortex ; after a greater time interval than would thalamo-cortical ; volleys at a faster frequency. Output time could be the same for fast and slow alpha frequencies, but reaction time would be slower in the case of a slower alpha frequency as it would take longer for the reaction time stimulus to be registered in the cortex. According to this hypothesis, reaction time in old individuals would be slower because their dominant brain wave rhythm is slower. Another line of EEG-reaction time research involved 1 the attempt to relate specific phases in the alpha cycle i : 30 / i | to faster reaction times. Kibbler, Boreham, and Richter > (1949) noted that voluntary movement tended to occur at the same point in the phase of the alpha cycle recorded from scalp electrodes over the motor cortex, and Bates (1951) corroborated this finding. Venables (1960) observed j that responses occurred in periods spaced by one-tenth 4 of a second which is the average duration of an alpha wave. These studies supported the notion that the EEG cycle reflects a modulation of responsiveness in cortical cells, and additional research was undertaken in an attempt to relate phasic cortical excitability to the alpha rhythm. With alpha rhythm recorded from the motor cortex, Lansing (1957) demonstrated that there was a phase in the I occipital alpha cycle during which stimulus incidence favored shorter visual reaction time. Callaway and Yeager (1960) supported these results. In a subsequent study Callaway (1962) found that, for a given individual, there was an enduring tendency for particular phases of the alpha cycle to be associated with fastest or slowest reaction times, but this phase was different for different individuals. Dustman and Beck (1965) related the phase of the alpha rhythm to reaction time and to characteristic!! i of the visual evoked potential and found that reaction 31 time was fastest and evoked potential amplitude was r 1 greatest when alpha phase was at the negative maxima. < Although these results for visual reaction time indicate a 1 relationship between alpha phase and reaction time, no relationship between alpha phase and auditory reaction i time has been found (Callaway, 1962; Cooper & Mundy-Castle, ; 1960). The data relating reaction time to phases of the alpha cycle give some credence to the hypothesis that brief phases of the alpha cycle might be associated with the admission of inputs and emission of outputs by a central processor. The existence of such a mechanism would lead to the prediction that reaction time would covary with the frequency of the brain wave rhythm. The implications of | these data for the aging organism are clear. A slower j dominant brain wave rhythm in advancing age would provide i fewer phases of optimal excitability per unit of time and hence fewer occasions for optimal reaction time. Thus, the data imply that the observed slowing of response speed with age is directly related to the slowing of the alpha rhythm. | The studies which have been reviewed have provided two different hypotheses about the coding of information ! in the nervous system, and both hypotheses lead to the prediction that reaction time will vary as a function of brain wave frequency. The thalamo-cortical priming hypothesis of Andersen and Andersson (1968) implies that I 1 a fixed number of volleys from the thalamus to the cortex are required to prime the cortex before a stimulus is perceived. These thalamo-cortical volleys are the physiological substrate of scalp recordings of EEG alpha, and a fixed number of volleys at a slower alpha frequency will be of longer duration and hence lead to slower stimulus perception and thus slower reaction time. The phasic cortical excitability hypothesis predicts periods of maximal responsiveness coincident with certain phases of the alpha rhythm. According to this hypothesis with slower alpha waves there will be fewer optimal phases available for the fastest possible transmission of information. In this manner the phasic cortical excitability hypothesis also leads to the prediction of slower reaction time with slower EEG frequency. The descriptive data which suggest that EEG alpha frequency and reaction time are related were provided in a series of experiments by Surwillo. Because alpha waves seemed to be related to reaction time, and because older individuals had slower alpha waves and slower reaction | times, Surwillo (1960, 1961) attempted to determine if ! 33 i i reaction time and duration of alpha rhythm cycle were ' related. Measuring the duration of waves occurring be* tween the onset of an auditory signal and the initiation of the subject's response. Surwillo found a statistically significant rank order correlation when he related subjects' mean reaction time to their mean EEG frequency. | The inter-individual correlation between reaction time ; and alpha period was .81 in a group of thirteen subjects between the ages of eighteen to seventy-two years. On the basis of these results, Surwillo hypothesized that period of the alpha rhythm, or some multiple of the alpha cycle, serves as the master timing mechanism in behavior. In other investigations, Surwillo attempted to ; replicate his first results with a larger sample and more ! sophisticated reaction time tasks. Simultaneously 1 measuring reaction time and average period of EEG between stimulus and response for 100 subjects ranging in age from 28 to 99 years, Surwillo (1963a) confirmed his previous findings. In this study an inter-individual i correlation of .72 was obtained between average reaction time and average EEG period. Intra-individual correlations ; averaged to .41. Disjunctive reaction time as related to : age and duration of alpha cycle was also studied by Surwillo (1964a). Fifty-four subjects, ranging in age r 34 from 34 to 92 years practiced pressing a button when they | heard tones of different frequencies, and they were then i instructed to press only when they heard a specific I i | frequency. The inter-individual correlation coefficient for brain wave period and disjunctive reaction time was .76, and Surwillo concluded that subjects with slow brain : waves require more time to decide between two alternatives than subjects with fast brain waves. In both experiments there was a low but statistically significant correlation between reaction time and age which disappeared when brain wave period was partialled out. The presence of slow brain potentials appeared to be necessary for the occurrence of slow reaction time in old age, and this fact suggested to Surwillo that EEG frequency is the factor behind age- I associated drops in processing capacity of the brain. All j findings were in agreement with Surwillo*s original i 1 hypothesis that the alpha rhythm is the timing mechanism for behavior. Measuring EEG period and reaction time in 110 chil dren ranging in age from 46 to 207 months, Surwillo (1971) found some support for his hypothesis in children over ! 140 months, but the data were not as conclusive as in his i previous studies of adult subjects. Developmental changes | in EEG frequency could not account for the magnitude of 35 ! developmental change in reaction time. ! To adequately explain age changes in speed of res- 1 ponse, Surwillo's hypothesis had to account for increased ! variability with age as well as increased slowing. Obrist (1965) demonstrated that, even when practice effect, fluctuations in motivation, fatigue, and variations in performance due to sensorimotor factors were ! controlled, mean reaction time variability within in- i dividuals increased fifty per cent from a group of young | subjects (mean age 27.5 years) to a group of old subjects I (mean age 80.6 years). Surwillo felt that his hypothesis | could explain increased reaction time variability as i resulting from two sources; cycle-to-cycle variations in alpha rhythm frequency, and trial-to-trial variations in the total number of alpha waves required to run off the events between stimulus and response. Mundy-Castle i et al. (1954) found the standard deviation of alpha rhythm frequency of two groups of mean ages of 22 and 75 years to be 0.98 and 1.11 cps respectively, but the 1 j magnitude of increase in alpha rhythm variability was not ! great enough to account for the large variability in reaction time reported in senescence. Surwillo (1963b) claimed that the effect of trial-to-trial variation in 36 \total number of alpha waves could account for Increased reaction time variability in senescence, and he demon strated this point hypothetically by comparing old and young subjects with mean alpha frequencies of 8 and 12 hz., jrespectively. He argued that, if a subject's fastest {reaction time takes two alpha cycles while his slowest ] reaction time takes three alpha cycles, then the difference between the maximum and minimum reaction times of young subjects is 83 msec., and the difference between the maximum and minimum reaction times of old subjects is 125 msec. Hence, the difference for the old subjects is 42 msec, greater, or 50%, greater than the variability in the young (see Table 1). I TABLE 1 (SURWILLO, 1968) i I ! EFFECT OF VARIATION II NUMBER OF ALPHA ! CYCLES WITHIN A RT INTERVAL ON RT VARIABILITY (HYPOTHETICAL DATA) Type S of Alpha Freq. (cps) RT in 2 waves (msec) RT in 3 waves (msec) Max RT - Min. RT (msec. ) Young 12 167 250 83 Old 8 250 375 125 What Surwillo overlooked in his example was the fact that the average alpha rhythms for old and young 37 subjects are not 8 and 12 cps respectively. The magnitude i of alpha slowing in senescence has been established at slightly more than one and at most two cycles per second. Substituting the reported mean alpha frequencies for young \ adults and adults beyond 80 (Obrist, 1954) in Surwillo's 2 and 3 cycle example of minimum and maximum reaction time, trial-to-trial differences in number of waves can | account for only 25% of the variability (see Table 2). | Table 2 EFFECT OF VARIATION IN NUMBER OF ALPHA 1 CYCLES WITHIN A RT INTERVAL ON RT VARIABILITY Type of S Alpha Freq. (cps) RT in 2 waves (msec) RT in 3 waves (msec.) Max. RT - Min. RT (msec.) Young 10.5 191.0 286.5 95.5 Old • 00 238.8 358.2 119.4 i Hence Surwillo*s hypothesis cannot account for the : observed magnitude of variability in the reaction time of old subjects, though it might explain behavioral slowing with age. In an attempt to replicate Surwillo*s work, Birren I | (1965) related reaction time to the characteristics of the EEG at the time of stimulus presentation. No correlation was found between mean alpha frequency and mean reaction 38 time in young subjects. There were two main differences between Birren*s and Surwillo*s work: 1.) the period in which EEG was measured, and 2.) the age of the subjects. Birren measured alpha frequency in the period immediately preceding the reaction time signal while Surwillo measured alpha frequency in the interval between the onset of the signal and the initiation of the response. It is dif ficult to understand why the EEG preceding the stimulus would not be related to reaction time while EEG immediately after the st‘ mulus would be highly correlated with reaction time. A more likely explanation of the discrepant results is based on the age differences in the subjects. Psychomotor slowing and the shift to slower dominant EEG rhythm may be two phenomena caused by a common central nervous system (CNS) aging process or they may be | independent phenomena related to age. Surwillo argued i that independent age changes in reaction time and EEG rhythm could not account for the high correlation between EEG frequency and reaction time because EEG frequency and reaction time were highly related even when the variance due to age was statistically partialled out. However, if individuals "age" at different rates so that EEG and reaction time slow in some individuals at age fifty and 39 in others at age seventy, then the statistical partialling out of chronological age would not weaken the correlation between EEG frequency and reaction time. Hence, by testing young subjects before CNS changes led to slowing of reaction time and EEG rhythm, Birren concluded that EEG frequency and reaction time were not related. Surwillo tested subjects ranging in age from 28-99, and he may have observed the relationship between EEG frequency and reaction time because of age differences in his sample rather than because EEG alpha rhythm paces reaction time. An additional difficulty in Surwillo*s data is one Surwillo (1968) reports himself. Subjects for all studies * were selected only if they produced well-defined EEG, and only samples with well-defined alpha waves in the interval between stimulus and response were included in the analysis. Surwillo admitted that this arbitrary selection of data imposed restrictions on interpretation, as there was, for example, no way to determine if any relation existed between response latency and alpha period when the EEG was desynchronized. Because stimulus onset frequently blocked the alpha rhythm, Surwillo might have selected for analysis those exceptional trials in which subjects were less aroused by the stimulus. Although Surwillo*s hypothesis receives impressive support from his 40 work, the limitations of this work must not be overlooked. Other investigators have also attempted to assess the value of Surwillo's hypothesis. Boddy (1971) under took two experiments which represent partial replications of Surwillo (1963a). The first experiment reported by Boddy involved recording EEG and simple visual and auditory reaction time successively (rather than simultaneously as Surwillo, 1963a). Boddy*s purpose was to test if EEG data were useful in predicting behavioral performance remote in time. Correlation coefficients (presumably Pearson product-moment correlation coef ficients --although Boddy did not state which coefficient he used) were computed between mean reaction times and mean alpha periods. For several reasons this first experiment was not useful in the assessment of Surwillo*s hypothesis. First, a methodological weakness in Experiment 1 was the in clusion of female subjects in the sample. Boddy did not analyze the data separately by sex, and since Surwillo studied men exclusively, the data are not comparable. Women have slower mean reaction time and greater reaction time variability than do men (Beilis, 1933; Anonymous, 1962) while there seem to be no brain-wave frequency differences between sexes. Studies relating EEG frequency, 41 reaction time, and age in women would provide useful information, and such studies may provide a good test of Surwillo*s hypothesis. Since women have greater reaction j time variability, Surwillo might predict that they would ! have slower dominant brain wave rhythms. This inference | [ is not validated by empirical findings, hence Surwillo should modify his hypothesis to explain sex differences in . i reaction time variability. While there are sex differences! in reaction time, reaction time variability and EEG frequency may be important to Surwillo* s hypothesis, con- j f founding sex in the study does not clarify the issue. j Indeed, since sex differences do exist on these variables, i i separate comparisons for men and women must be made. The inclusion of women in Boddy*s sample clouds rather than clarifies the issue of brain wave frequency-reaction time relationships. Another reason that Boddy*s Experiment I was not useful was that Obrist (1963) reported an experiment almost exactly comparable to Boddy*s Experiment I. Obrist concluded in that study that there was little relationship between occipital alpha frequency and reaction time when the two are recorded independently, and Boddy did not acknowledge the existence of that study in his report. 42 ' Surwillo*s hypothesis involves the contiguous relationship I I between reaction time and brain wave frequency; inferences i | drawn from non-contiguous measurement of EEG and reaction | time cannot conclusively support or disprove the hypo- thesis. Hence, Boddy*s first experiment was of relatively I little use in testing the hypothesis. Boddy*s second experiment represented an attempt ! j to accurately replicate Surwillo (1963a). EEG was j monitored while simple visual and auditory reaction time tasks were performed by male and female subjects. Brain wave frequency was sampled in the one-second period immediately prior to each stimulus, and correlations were computed between brain wave period and reaction time for i i | each subject. In this experiment Boddy was unable to ! replicate Surwillo (1963a). Inter-individual correlations ; were low and statistically non-significant as were mean intra-individual correlations. Boddy concluded that Surwillo*s hypothesis was not particularly useful in explaining brain-behavior relationships. I Since Boddy failed to replicate Surwillo on several major features of the experiment, his data do not con clusively rule out the proposed relationship between reaction time and brain wave frequency. Several discre- i paneles between the two designs may account for the dis crepant results. For example, as in his first experiment, Boddy included women in the sample. For the reasons mentioned previously, the inclusion of women in the sample may have made the results less definitive. Another difference between Boddy (1971) and Surwillo (1963a) involved the age range of subjects. Sur willo was interested in the effect of chronological age on the magnitude of the relationship between brain wave frequency and reaction time, and he included in his experiment subjects ranging in age from 28 to 99 years. | The age range of Boddy's subjects was from 18 to 38 years. Since brain wave period decreases with age with the greatest increases after age 60 to 65 (Matousek, et, al., 1967; Obrist, 1954, 1963), Boddy had a much narrower range of brain wave periods in his sample than did Surwillo. The mean brain wave periods sampled by Boddy ranged from 91.2 to 117.0 msec.* while mean brain wave periods in There is some question as to the actual range of brainwave period reported by Boddy on Tables Ila ana lib in his article. The range, 91.2 to 117.0 msec., was taken from Table lib, as the data reported on Table Ila must represent some kind of typographical error. Boddy reports an EEG period of 0.0 for one subject and of 12.5 for another. A brain wave period of 0.0 is logically impossible, while a period of 12.5 represents an EEG frequency of 80 hz. Since the EEG frequency range in humans is usually considered to extend from .5 to 40 hz., an EEG period of 12.5 is not feasible. In addition to____ 44 ; Surwillo's study ranged from 85 to 120 msec. Hence, the ! range of brain waves in Surwillo*s study was over 30% greater than the range in Boddy's study. If Boddy*s data 1 were corrected for range, they probably would yield a higher inter-individual correlation between reaction time and GEG period. Still another difficulty was that Boddy sampled anywhere from 6 to 40 reaction times per subject on which he based the intra-individual correlations between EGG period and reaction time. The assumptions of the Pearson product-moment correlation test may have been violated as some intra-individual correlations were based on such a small number of observations. Surwillo based his intra individual correlations on 90 simultaneous measurements per subject. Because he was afraid that EEG sampled after the onset of the stimulus would be confounded by the sensory evoked potential to the stimulus, Boddy sampled EEG in the one-second interval preceding the stimulus. Surwillo these errors, Boddy reports a standard deviation of brain wave period of 6.4 msec. Computation of the standard I deviation from the data presented on the table indicates that the standard deviation is 42.75. For these reasons Table Ila is difficult to interpret. 45 clearly predicted that the speed of response at a particular Instant would be dependent on the frequency of the EEG In the Interval between stimulus and response. i Although he maintained that one would not expect dramatic frequency changes In underlying alpha activity as the stimulus Is Introduced, Boddy did not sample EEG in the I period most relevant to Surwillo1s hypothesis. In a subsequent analysis, Boddy (1970) found that the evoked potential was not a source of correlation between EEG period and reaction time. Hence, his original concern about sampling the EEG in the interval between the onset of the stimulus and the initiation of the response was unfounded. Since there are so many discrepancies between Boddy's replication and Surwillo's original study, Boddy did not fulfill his stated purpose which was to resolve i ambiguities in Surwillo*s work and to assess the value of the hypothesis that brain wave frequency is the master timing mechanism in behavior. The intent of Boddy*s studies was worthwhile as the implications of the ! hypothesis that Surwillo proposed are great. Hence, an i accurate replication of Surwillo (1963a) would be a t valuable contribution to the literature. In addition, there are other strategies which would also be useful in j testing Surwillo*s hypothesis. 46 ! Surwillo (1963a) concluded, "It is worth noting here j that our hypothesis also demands that experimental i alterations of brain wave frequency should be accom- ■ panted by corresponding changes in speed of response. This interesting proposition also deserves investigation." ! (P. 113) In an attempt to modify EEG frequency while simul taneously measuring reaction time, Surwillo (1964b) used photic stimulation to alter EEG frequency. Photic "driving" involves inducing a change in the EEG with flashing lights. Presenting the flickering stimulus over a range of 6 to 15 flashes per second, Surwillo attempted to measure reaction time when subjects* EEG was syn- j chronized with the different frequencies of flashing light. i | The experiment proved difficult to carry out, however, i I as only 5 of the 48 subjects tested showed evidence of I EEG synchronization of more than a narrow range of frequencies with the flashing light. Of the five subjects whose EEG could be synchronized within a mean range of ; four flashes per second, inter-individual correlations between reaction time and the driving signal were positive j and ranged from 0.10 to 0.55. The control group of | subjects showed EEG synchronization to the flashing light under resting conditions but not while taking part in the 47 reaction time experiment. Intra-individual correlations j between EEG period and reaction time in this group were ; negative as well as positive. The findings of this i j experiment were extremely limited, and while the evidence i does not rule out the possibility of a causal relationship I between EEG and reaction time, it provides only very : limited support for the hypothesis. Surwillo reported in a subsequent paper (Surwillo, 1968) that he received a personal communication from Waszak in 1965 reporting that i i the experiment was replicated at Duke University on two j subjects. Inter-individual correlations were 0.27 and 0.54. Since Surwillo*s hypothesis that the alpha frequency ! is the master timing mechanism in behavior implies a causal relationship between alpha frequency and reaction | time, correlational evidence alone is not sufficient to support the hypothesis. Causality can only be established by experimental manipulation, but Surwillo was unable to ; manipulate alpha frequency in more than a few select subjects. Until quite recently, the techniques necessary : to experimentally test Surwillo's hypothesis were not | available, but developments in the emerging field of 1 biofeedback training provide a new means with which the f | hypothesis can be tested. 48 Biofeedback Manipulation of the EEG | In a discussion of scientific creativity, Lord j Adrian, the famous biologist, stated, "New ideas in | science are induced by new discoveries, and at the present i ! time it seems to me that the most potent factor in t i promoting new discoveries has been the introduction of ! i some new technique, some new tool that can be used for I exploring natural phenomena." (Adrian, 1961) Biofeedback easily qualifies as the kind of promising new technique of : which Lord Adrian was speaking. The basic mechanism employed in feedback condition- i ing is an electronic system which serves to amplify and I inform an individual about ongoing activity of selected ! physiological processes. The actual feedback is a signal I , or stimulus presented to the subject contingent upon the presence of a certain physiological state. For example, if a subject is being trained to produce brain waves in the alpha frequency, the feedback might be a tone sounded whenever the subject produced an alpha wave. ' The biofeedback technique is based on the funda- 1 mental learning principle that a response is learned when : an individual received reinforcement (feedback) when he i makes a correct response. While physiological feedback i | loops exist for some systems in the body (e.g., pro- 49 1 prioceptors provide information about position, movement, j and tension), we normally do not receive rapid feedback | information concerning the functioning of our internal i j status. For example, we usually are unaware if our brain waves are in the alpha range. In this sense biofeedback \ functions like a sensor which provides information about the state of our internal organs. Biofeedback is a relatively new research tool which has been developed in the last five or six years. In the late 1960*5, researchers asked the question: If a subject is given information (biofeedback) which tells him that his brain waves are within the alpha range, that his heart is beating slowly, or that his blood pressure is low, can he learn to prolong the occurrence of alpha rhythm, or i to keep his heart rate slow, or to maintain a low level | of blood pressure? A number of investigations have pro vided an affirmative answer to this question, and the research literature in this area has grown so quickly that two extensive volumes of biofeedback research have i i already been published (Barber, et. al., 1971a, b). ' Biofeedback research with the EEG which has i i received its primary impetus from Kamiya (1962, 1967, | 1968. 1969) has led to the demonstration that subjects can I learn to voluntarily control their EEG alpha rhythm. In 50 ; 1962, Kamiya presented a paper suggesting that different ! stages of EEG frequency could be discriminated by subjects, i i and he later demonstrated that subjects could learn to ! reliably produce or suppress EEG activity in the band- i width of the alpha rhythm (Kamiya, 1968). He also demon- strated that subjects could learn to control both the j amplitude and frequency of alpha, although he did not | specify the range of frequencies which could be controlled (Kamiya, 1967, 1969). Numerous investigators have replicated Kamiya's work (e.g., Dewan, 1964, 1966; : Bundzen, 1966; Brown, 1968, 1970; Hart, 1967; Mulholland ; & Evans, 1966; Runnals & Mulholland, 1965). These i investigators all focused on the alpha rhythm of the EEG, j and all reported some success in training subjects to i manipulate the percentage of time in which they produced ; the alpha frequency. The typical procedure was to place subjects in an electronic feedback loop giving them information about the presence of their alpha rhythm, i Feedback was visual or aural, and subjects learned either I I to control turning on and off short alpha bursts (Bundzen, Dewan) or to shift to greater periods of alpha (Brown, | Hart, Kamiya). I ! The EEG feedback training apparatus has typically consisted of a minimum of two channels of EEG, a bandpass 51 filter set at the alpha frequency, and a trigger con- j nected to a signal which can be perceived by the subject. I ! The experimenters also have used various means of simul- i 1 taneously recording the signal and the EEG, and this has ! usually been achieved with additional polygraph channels. I | The alpha rhythm is most prominent in recordings | from occipital leads, and subjects seem to produce more alpha when they are relaxed with their eyes closed in a quiet, darkened room. Under these standard conditions there are great individual differences in amount and frequency of alpha, so exact baselines of each subject's EEG must be determined before feedback training begins. Usually subjects are given a session of non-feedback trials before the actual experiment can commence, and it i I is after determining the baseline EEG frequency that the | investigator attempts to alter the brain wave pattern. All of the investigators working with autocontrol of brainwaves have tested subjects under standard con ditions, but they have used different methods of rein forcement and different instructions, thus giving subjects different expectations and various levels of understanding I of the experiment. Some investigators report successful | alteration of alpha production by placing "dummy" I electrodes on all areas of the subjects' bodies and 52 ! instructing them to turn on a stimulus light ^Reese (1969), j personal communicationj. The light is seen through the i | closed eyelids as an increase in illumination in a | darkened room, and it does not usually produce alpha blocking. Subjects are given no additional information I | about what is expected of them, yet they are able to increase the time that the light is on by producing more I alpha. As investigators attempt variations on the biofeed back technique they become more proficient at training subjects. Dewan (1966) was able to learn to control the presence or absence of alpha in his own EEG record so well that he used his EEG to send messages to a computer in Morse code. Hart (1967) showed that subjects given i I several types of feedback performed better than subjects I given only one source of information. Subjects presented with both two-minute totals and immediate feedback for presence of alpha performed better than subjects receiving immediate feedback alone. Nowlis and Kamiya (1970) i demonstrated that subjects left free to learn control of j their brain waves in any manner they chose could indeed | master the task of controlling their alpha waves. By the end of the experiment all twenty-six subjects succeeded i in producing more alpha in "on" trials than during ”off" 53 trials. Green, Green, and Walters (1969, 1970) have obtained positive results with a triple feedback training program in which subjects simultaneously learn to reduce \ forearm muscle tension, increase hand temperature, and increase percentage of alpha in the EEG record. ! The physiological mechanisms which make instrumental j | conditioning of the EEG alpha rhythm possible are largely I unknown. Mulholland (1968) hypothesized that, in at I least some subjects, alpha rhythm is related to eye position. He speculated that alpha increases when the eyes are moved to an extreme side or up position. Fenwick (1966) demonstrated in a study of sixteen subjects that alpha was not significantly related to the eye , positions suggested by Mulholland, although a few of the subjects did show the hypothesized effect. Furthermore, i Kamiya (1967) stated that his subjects could learn to control alpha with their eyes in an up or down position. Thus, although the Mulholland effect has not been ruled out for all subjects, it does not seem to be the typical ! response of most subjects. Mulholland himself (1968) recognized that the hypothesized relationship is not characteristic of all subjects. With animals, Olds and Olds (1961) and Olds (1963) ! were able to demonstrate that single cortical cells in 54 lightly anesthetized rats could be brought to increase i their rate of firing through instrumental conditioning. I Electrical stimulation of the medial forebrain bundle in the hypothalamus ("pleasure center") was used for positive reinforcement. Single unit responses could be j j k | reinforced when measured with micro-electrodes placed in various sensory and motor areas of the paleocortex* Fetz (1969) trained unanesthetized monkeys to increase the activity of single neurons in the precentral cortex by : reinforcing high rates of discharge with food pellets. The monkeys learned to fire newly isolated cells up to j 500% faster than the rate before reinforcement. Sterman and Wrywicka (1967) have shown that cats i can learn to control two EEG patterns through instrumental , conditioning with food as a reinforcer. Carmona (1967) ; reinforced cats with electrical stimulation of the medial forebrain bundle to produce high or low voltage EEG activity. Cats rewarded for high voltage activity : showed more high voltage slow waves, while cats rewarded for low voltage activity showed much more low voltage fast activity. This type of training modified both the i electrical brain activity and the behavior of the cats. ^ For this reason Carmona (1967) performed another experiment I with curarized rats to rule out the possibility of skeletal L j mediation. Curarized rats were able to change the pattern 55 of their brain waves in the rewarded direction, and Carmona j concluded that sensory consequences of changes in skeletal i activity were not the means by which animals change their ! ' brain wave patterns. Another EEG parameter which has been modified with i biofeedback is the averaged evoked potential (EP). Fox j ■ and Rudell (1968) trained hungry cats to change the amplitude of a late component of their visual EP by re inforcing them with milk whenever the response reached a specified amplitude. A similar experiment was carried out ; with the human auditory EP (Rosenfeid, et al,, 1969). i Instrumental conditioning was used to train subjects to ; change the amplitude of a late component of the auditory I | EP. This same result, an increase in the amplitude of a | late component of the auditory EP was achieved by Spilker, j et al. (1969) when they measured EP*s after training sub jects to produce a greater abundance of alpha waves. Since the abundance of alpha decreases with age, and since age changes in the EP include decreases in the ; amplitude of the late components, the results of Spilker, ! et al., (1969) lead to speculation about a causal re- j lationship between these variables in elderly individuals. i I Perhaps decreases in alpha abundance lead to decreases in , amplitude of late components of the EP. Descriptive 56 studies relating alpha abundance and amplitude of EP late components as well as biofeedback manipulation of alpha abundance with concurrent measurement of EP's in old sub- jects might provide answers to this question. In this new area of biofeedback, there is much to be | discovered, and the implications and applications for brain ! I wave control are almost totally unexplored. One important application of the phenomena relates to research design. Fields such as electroencephalography which, heretofore had, of necessity, relied on correlational approaches, I can supplement their methods with experimental procedures. Since subjects can learn to produce specific brain wave frequencies, causal relationships between EEG frequency and behavior can be observed. Because investigators have considered the slowing i of the EEG alpha rhythm to be of great significance In the process of aging, the opportunity to manipulate alpha frequency has important implications. If slowing of the dominant EEG frequency is the cause of slower reaction : time in the aged, then training subjects to produce more slow brain waves should cause those subjects to have | slower reaction times. Conversely, it might be possible I to speed the reaction time of old subjects by training them to produce more fast brain waves. r 57 Biofeedback manipulation of the alpha frequency is ; also a means of eliminating some of the problems inherent ! in the methods used by Surwillo (I960, 1961, 1963a). i i 1 Since the subjects can learn to produce alpha, perhaps by teaching subjects to always produce alpha between stimulus I | onset and response initiation, less data would have to j ! be eliminated due to alpha blocking. i 1 i ; I Reaction Time-Aging-EEG Relationships and BloteedbackManipulation I The material presented in the first section of this j i dissertation leaves little doubt that psychomotor slowing is an invariant biological transformation--a primary aging factor. Clear evidence has also been presented i which indicates that the slowing of the dominant brain i : wave rhythm occurs in the healthiest of elderly subjects i : and results from both normal aging and disease. That the slowing of response speed with age is related to the slowing of the dominant brain wave rhythm was strongly suggested by the large body of literature presented in ! | the third section. Reaction time and brain wave frequency i ; have been related in a large number of studies, and | Surwillo (1968) summarized a number of studies indicating | an association between the age-related slowing of alpha • rhythm and reaction time. Although Surwillo inferred a 58 : causal relationship between alpha frequency and reaction | time and hypothesized that the alpha rhythm is the master I timing mechanism in behavior, he was not able to offer I ' strong experimental support for his hypothesis. Reliable techniques for the manipulation of alpha frequency were i not available to Surwillo, but recently such techniques ' have been developed. Biofeedback is an instrumental conditioning procedure, and with it subjects can be trained to produce or suppress brain waves in selected frequency bands. Hence, the biofeedback technique pro- ; vides the opportunity to experimentally manipulate brain ; wave frequency and to experimentally test the hypothesis of Surwillo. ! Hypotheses The purpose of the present investigation was to | test the hypothesis that period of the EEG alpha rhythm is the master timing mechanism in behavior. The first two hypotheses followed directly from the predictions of Surwillo (1963a, 1963b). These hypotheses are: I I 1. When subjects are trained to produce slow i alpha frequencies, they will have slower reaction time i 1 than they will have when they are trained to produce fast brain wave frequencies. 59 2. Reaction time variability will be greatest in | subjects when they are producing slow brain wave I frequencies. i Surwillo (1963b) considered the increase in reaction time variability in the aged to be an important j phenomenon to explain. Arguments were presented in a i previous section of this dissertation which indicated : that Surwillo's explanation of increased reaction time variability in terms of alpha waves could not account for the magnitude of increase in reaction time variability which occurs with age. An alternative or perhaps a i supplementary explanation to Surwillo*s hypothesis in- ! volves EEG variability. Old subjects produce less alpha ; activity (Matousek, et. al.t 1967), and this fact might I I be used to explain the increase in reaction time | variability in the old. If reaction time and EEG alpha rhythm are related, then subjects consistently producing alpha activity should have reaction times which are less variable than subjects producing less alpha activity. < Hence, it was hypothesized that: 3. Variability in the reaction times of subjects I who have a greater abundance of alpha will be smaller i than reaction time variability of subjects producing a j | lesser abundance of alpha. 60 Biofeedback training provides a method for experi mental manipulation of brain wave frequency, but it may impose on the subjects certain limitations which affect their reaction time. Hence, the subject may be concentrat ing so intently on producing the feedback signal that he i ! ignores the reaction time signal. To control for this possibility, a group of subjects whose brain waves do not ! control the feedback signal were compared to the subjects ! ' who did receive feedback training. ' In the present study age was simulated inasmuch as I young subjects trained to produce slow brain wave fre- 1 quencies were purported to simulate the brain waves of old i ; individuals while these young subjects were thought of i | as simulating the brain wave frequencies of young individuals when they produced brain waves of their mean frequency and faster. In this sense, chronological age was not a relevant variable. The present experiment would represent a developmental study in the age-simulation i sense if only young subjects were included in the sample. i There are additional factors, however, which made it I S desirable to include an old as well as a young group of l : subjects. I Because subclinical pathology may account for the J observed slowing of EGG rhythms with age (Obrist, 1963), the relationship between EEG and reaction time may be different in old than in young subjects. For example, if 1 age changes involve a feature of the central nervous system which leads to slowing of reaction time and to EEG slowing, a relationship between EEG frequency and reaction time would occur in the old but not in the young subjects. The data of Blrren (1965) suggest this possibility as no EEG-reaction time relationship was | found in a group of young subjects. The complete absence of data on biofeedback con ditioning of old subjects made it difficult to predict | j age differences in this experiment. On the one hand, it j might be predicted that, since age changes occur in the j nervous system which affect EEG and reaction time, these two variables will be more highly related in older than in ; younger subjects. Thus, one would predict greater i reaction time differences for old subjects between the | Speed and Slow brain wave conditions. On the other hand, i it might be expected that the aging nervous system is less plastic than the young nervous system, and old subjects would not be able to shift EEG frequency or reaction time ! to as great a degree as young subjects. This assumption would lead to the prediction of greater reaction time i differences for young subjects between the Speed and Slow | conditions. While the author was aware of the possibility i ' of a differential relationship between EEG frequency and 62 'reaction time in the two age groups, no hypotheses j regarding differential age relationships were made. | Age differences on biofeedback tasks had not been 'investigated, and data from this investigation provided , comparisons of rates of biofeedback conditioning between | young and old subjects. Numerous investigations have | demonstrated on a variety of tasks that old subjects learn i and condition more slowly than do young subjects. For this reason it was hypothesized that; 4. Old subjects will take more trials than young subjects to learn the biofeedback tasks to a criterion. CHAPTER II METHOD 1 Subjects I Fifteen male subjects between the ages of 18 to 29 i (mean age was 23.7) years and 15 male subjects in the age ; range of 60 to 81 (mean age was 72.5) years participated in the study. Young subjects were students at Long Beach State College or Vietnam war veterans who were patients at the Long Beach Veterans Administration Hospital. Old subjects were recruited from employees and volunteers working at the hospital, from members of a retirement organization, or in two cases from patients at the hospital All but two of the old subjects resided in the community ; at the time of testing. Subjects were asked to volunteer ; for an experiment dealing with reaction time and brain 1 waves. Assignment to experimental and control groups was randomized. Apparatus The basic pieces of equipment used for biofeedback training were a Beckman type R dynograph, an Intelex 1 filter-feedback system (with a bandpass filter capable of filtering frequencies from 1 to 20 hz. at a bandwidth of + 0.5 to + 2.0 hz.), a Tektronix oscilloscope, an 64 ' Ampex FM tape recorder, a Grass electronic stimulator, a I Hunter Klockounter and a Wollensak tape recorder. This I equipment was arranged in a configuration depicted in i i ' Figure 1. Silver-silver chloride cup electrodes were attached ; to the subject's scalp, and the EEG signal was amplified in the dynograph and recorded on graphic paper output and on magnetic tape. Output of raw EEG was also channeled through the bandpass filter, and filtered EEG was routed through an oscilloscope. When the filtered EEG signal reached a predetermined amplitude, it triggered a signal on the oscilloscope which in turn triggered the timer. In this manner, each filtered EEG wave represented one unit 1 on the timer. The filtered EEG signal also triggered a ; Mallory sonalert which produced a tone whenever the j filtered EEG output exceeded the critical amplitude. The subjects heard this tone in the right earphone. The tone was amplified in the tape recorder. Filtered EEG was also recorded graphically on paper output, j The reaction time apparatus consisted of a micro- ' switch held in the hand of the subject and depressed with i the thumb. The microswitch was connected to a signal ^ marker on the graphic paper output and to a Hunter Cloc- | Kounter accurate to the nearest millisecond. The auditory r : signal was a click generated by the Grass stimulator SUBJECT MAGNETIC TAPE DYNOGRAPH FILTER SIGNAL • MARKER EARPHONES I MICRO PHONE OSCILLO- qrnPF TAPE GRAPHIC SIGNAL RECORDER OUTPUT • MARKER SONALERT (FEEDBACK SIGNAL) KLOC- Fig. 1.--Block diagram of Biofeedback Conditioning Apparatus 66 1 and presented to the subject through the left earphone, j The experimenter depressed a switch to sound the click, and the onset of the click was marked on the graphic I paper output with the same stimulus marker which marked the microswitch. Reaction time was also recorded on mag- < netic tape simultaneously with EGG. Hence, two channels (raw EGG and reaction time) were recorded on magnetic tape, and three channels (raw EGG, filtered EEG, and reaction time) were recorded on the graphic paper output, j The bandpass filter and a General Radio audio frequency microvolter were used to perform a spectral analysis on the raw EEG recorded on magnetic tape. The taped EEG signal was sent to the microvolter where the : amplitude of the signal was attenuated. From the micro- i | volter the signal went to the bandpass filter and | filtered output and reaction time were reproduced on graphic output. Raw EEG was filtered at four frequencies at a bandwidth of + 1 hz. EGG was filtered at 6, 8, 10, and 12 hz. for subjects with a mean frequency of 8 or 10 hz., and the filter was set at 7, 9, 11, and 13 hz. for subjects with a mean frequency of 9 or 11 hz. Procedure | Baseline Condition ! In the first session, the experimental and control 67 1 subjects received the same treatment. This preliminary I testing involved measuring baseline EEG and mean reaction j time. For the first fifteen minutes of the session, sub- i | jects sat in a darkened room with their eyes closed while one channel of EEG was recorded from surface electrodes placed at - 0^ according to the "10 - 20" International System (Jasper, 1958). For two-minute intervals the i | experimenter set the bandpass filter at frequencies of ; 6, 8, 10, and 12 hz. with a bandwidth of + 1 hz. If a | subject's mean frequency was judged to be 9 or 11 hz., EEG j was sampled at filter settings of 7, 9, 11, and 13 hz. I There was an ascending and descending sequence, hence four I frequencies. Time in each frequency was recorded on the Hunter timer. Only those subjects with a baseline of 25% alpha were included in the sample. Four young and three old subjects were rejected as they spent less than 25% time ' in the alpha brain wave state during the baseline con- , dition. i After baseline EEG was sampled, subjects were l ] allowed to practice the simple reaction time task. They were instruated to press the microswitch at the onset of i a tone controlled by the experimenter. After twenty ! practice trials, simple reaction time was recorded simul taneously with EEG. The auditory signal was presented 68 when the subject was producing EGG in the alpha frequency I range (8 - 13 hz.)* A total of forty reaction time trials I i was recorded in this condition. i i ! Feedback Training Experimental groups of 10 young and 10 old subjects ! were trained to increase the per cent of time they spent in different frequencies of the alpha rhythm. In the first training session, the bandpass filter was set at the subject's mean frequency (as measured in the baseline session) with a bandwi3th setting of + 1 hz. During a I feedback trial, the subject heard a tone every time he produced an alpha wave in the selected frequency, and he was instructed to maintain the feedback tone as long as ' possible. Each feedback trial lasted two minutes (120 seconds). A session consisted of 20 two-minute trials with one five-minute break after the first 10 trials. When a subject met the criterion of producing his mean brain wave frequency two-thirds of the time for ; three consecutive trials (this meant that the subject i { spent 80 seconds in his mean frequency for three con- ! secutive 120-second Intervals), reaction time was tested. The procedure for testing reaction time during feedback ! | will be described in a following section. After reaction time was tested at the subjects' 69 i I mean brain wave frequency during feedback, experimental j groups of young and old subjects received one of two treatment conditions. Subjects were randomly assigned to the two treatments. One-half of the young and half of i the old subjects received additional feedback training j at a brain wave frequency 2 hz. above their mean fre- | quency (Speed condition), and one-half of the young and half of the old received feedback training at a frequency two hz. slower than their mean frequency (Slow condition). When a subject met a criterion of increasing the amount of activity in the selected frequency by 33% over the amount produced in the baseline condition for three consecutive two-minute trials in one session, reaction time was i ; measured during feedback. To control for the effect of practice on reaction time, subjects were tested at their j mean frequency after the Speed and Slow conditions. Finally, subjects in the original Speed condition were trained to increase per cent time in a brain wave fre quency two hz. slower than their mean, and subjects in the 1 original Slow condition were trained to increase per cent ' time at a frequency two hz. faster than their mean. Hence j the following two sequences of training and testing were I used in the young and old groups: Group 1 - Mean, Speed, ! Mean, Slow, Mean; Group II - Mean, Slow, Mean, Speed, 70 1 Mean. j j Dummy Feedback Training i j Control groups of five young and five old subjects were treated exactly as the experimental subjects except ! that their brain wave frequency did not control the feed back. Recordings of feedback produced in an experimental session were made and played to control subjects. Control subjects were matched to experimental subjects, and they ! proceeded with "training” at the same pace as their matched ! experimental subject. Reaction time was tested in control subjects after the same number of trials as it took their ! matched experimental subject to achieve criterion at the jmean frequency. The control group was "trained” only in the mean frequency condition. Figure 2 presents the experimental design for the investigation. i ■ Testing During Feedback Once the experimental subjects learned to produce ' a given frequency of brain waves, they were tested during I i feedback on the reaction time task which they practiced during the baseline session. Before testing began sub- | jects had an additional opportunity to practice responding | to the reaction time signal for ten practice trials. Reaction time signals were controlled by the PRETEST FB TRAIN FB TRAIN FB TRAIN FB TRAIN (EEG and RT) (at M ean freq.) 5ISE) (at Mean freq.J (Slow or Speed) (a t M ean freq.) TEST RT TEST RT TEST RT "TTST R T ^ TEST RT — □ B - B [ n "isj 8W | k «lo| Toung B Toung B Young .0 — B Q» □ E — B ' B — E Q ^ £ • ? ] Q ^ n W m r FB TRAIN TEST RT Old Young B B Fig. 2.--Experimental Design 72 experimenter, and an attempt was made to present the 1 signals when subjects were producing brain waves in the I selected frequency. The experimenter monitored the I graphic output to present reaction time signals during j ^ these periods. The inter-stimulus-interval (ISI) was random. To control for the possibility that subjects would associate production of the selected frequency with the onset of the reaction time stimulus, the reaction time I i j signal was also presented when subjects were not producing j brain waves in the selected frequency. This procedure I I also ensured that the ISI would not be a long time ; interval, as the experimenter did not always wait for the | production of brain waves in the selected frequency to i | present the reaction time signal. Analyses EEG Activity A spectral analysis of EEG frequency during reaction : time was achieved by measuring the amplitude of the fil tered EEG activity at the onset of the reaction time tone. i ; The EEG amplitude at 6, 8, 10, and 12 hz. (for subjects i i with a mean brain wave frequency of 8 or 10 hz.) or at 7, 9, 11, and 13 hz. (for subjects with a mean brain wave fre quency of 9 or 11 hz.) was measured for each reaction * time. Hence, four measurements of EEG amplitude were made 73 for each reaction time. These measurements were averaged for each reaction time, and in this manner brain wave | frequency for each reaction time was computed. Total ■ amplitude of EEG for each rer :tion time was computed by i summing the EEG activity of each of the four frequencies. Statistical Analyses Planned comparisons, analysis of variance, Pearson product moment correlations, and rank order correlations were the major statistical analyses carried out with the data. Most computations were made on an IBM 360 computer. i i CHAPTER III RESULTS EEG Frequencv-Reactlon-Time Relationships j ) The analysis of variance model used to test the major hypothesis was a 2 x 3 x 10 design comparing the | effects of age, brain wave frequency, and subjects on reaction time and on reaction time variability. Since data were originally collected Ina2x2x6x5 design comparing the effects of age, order, and frequency, and subjects (see Figure 2 for the original design), the final analysis of variance represented the collapse of the order variable and the averaging over the four mean brain wave conditions. The effect of order was not significant j (F - 1.55, and 0.32; df - 1/16; p<0.23 and 0.58 for ! reaction time and reaction time variability, respectively), i ; and there was no significant difference between the four mean brain wave frequency conditions in mean reaction time (F - 0.93; df - 3/54; p < 0.43), and in reaction time ; variability (F - 0.65; df - 3/54; p < 0.58). Hence, these ! modifications in the design were justified, and they were carried out to make the main analysis of variance more powerful. Mean reaction time and reaction time variability i j for young and old subjects in the three brain wave fre- j quency conditions are presented in Figure 3. I „ ______________ _ _ 7 4 ........... ....... .................................................... 75 ■•acttan E E C Fr*q. TIm O (H*. > 9 .0 0 • 310 300 290 280 ■EG 270 *T 230 9.30 - - 5 0 9.40 - -45 EEC 220 210 9.50 - -4 0 200 190 9.60 - -3 5 Slow Young Old Biofeedback Condition Fie. 3.— The Effect of Brain Wave Frequency on | Mean Reaction Time, Reaction Time Variability, and EEG Frequency for Young and Old Subjects in Three Biofeedback Conditions Hays (1965) asserts that when an experimental ques- 1 tion represented by a comparison is an important one for i j the interpretation of the experiment, and it is essential 1 that a Type II error not be made in the accompanying test, then a planned comparison should be carried out. For this I reason, the most important experimental question, the | effect of fast and slow brain waves on reaction time and i reaction time variability was tested with a planned com- ; parison between the reaction time data collected during : the Speed and Slow conditions. A planned comparison : between data collected in the Speed and Slow conditions, ; rather than an omnibus F test, was used to ensure maximum power. The summary table for the analysis of | variance and planned comparison is presented on Table 3. Reaction time was significantly slower in the Slow I brain wave condition than in the Speed brain wave con- : dition (F - 3.97; df « 1/36; p < .05). Thus, the first hypothesis was supported. When subjects were trained to produce slow alpha frequencies they had slower reaction j | times than when they were trained to produce fast brain wave frequencies. Reaction time variability was not ; significantly greater in tne Slow brain wave condition, j although the results were in the predicted direction (F - 3.37; df - 1/36; p < .08). 77 TABLE 3 SUMMARY TABLE FOR ANALYSIS OF VARIANCE OF THE EFFECT OF AGE, BRAIN WAVE FREQUENCY, AND SUBJECT ON SIMPLE REACTION TIME Source df SS MS F 1 Between Subjects Age 1 60641.74 60641.74 2.53 S/A 18 431427.62 23968.20 Within Subjects Frequency 1 2795.74 2795.74 3.97* F x A 2 1012.75 506.37 0.72 SF/A 36 25367.29 704.65 *p < .05 aplanned comparison Although age differences in reaction time were I ! large (mean reaction time was 274.8 msec, for old subjects, | and 208.9 msec, for young subjects), the between subjects j variability was great enough to render this difference non significant at the .05 level of confidence. The age differences in reaction time did approach significance, | however (F - 3.72; df - 1/18; p < .072), as did age dif- i 1 ferences in reaction time variability (F - 4.24; df » 1/18; j p< .056). Several of the young experimental subjects i I were Vietnam war veterans, and although they had no obvious I motor disturbances, their reaction time was somewhat slower 78 ! than the reaction of the other non-hospitalized young I | subjects. These experimental subjects were not among the | subjects matched to controls in the dummy feedback con- | dition, and when age differences in reaction time were i i compared in this condition, the effect of age was signi ficant (F « 14.87 and 14.33; df ■ 1/16; p < .001 and .002 i j for reaction time and reaction time variability, ! respectively). Hence, the absence of significant age i differences in reaction time probably occurred as a ; result of the slower reaction time in these young subjects, i The interaction effect between age and brain wave frequency was not statistically significant, although the results are in the direction of a differential effect ; with age. The difference in mean reaction time in the j Speed and Slow conditions for the old subjects was 26.0 j msec., while the difference was 7.4 msec, in the young. i : Reaction time was shifted almost three and a half times more in old subjects. The actual shifts in brain wave frequency in the Speed and Slow conditions were small, and they were not statistically significant. These shifts are also | presented on Figure 3. It was possible with the present ! data to test the hypothesis that greater shifts in brain ! waves would be accompanied by greater shifts in reaction 79 1 time. Ten reaction trials were selected from each subject I ! in the Speed and Slow conditions in which the subjects ! | had changed their brain wave frequency in the direction of ' training to the greatest degree. Hence, for the Speed condition, the reaction times during the trials in which i i brain waves were fastest were selected, and for the Slow condition, reaction times during the ten trials in which brain waves were slowest were selected. Mean reaction time and mean brain wave frequency for the two sets of ten trials were computed for each subject. The values for mean reaction time and brain wave frequency computed previously from averaging the four conditions in which reaction time was measured during mean brain wave fre quency were used for the Mean condition in this analysis. A 2 x 3 x 10 analysis of variance comparing the ! effects of age, brain wave frequency, and subjects was carried out with the selected data. A planned comparison between the Speed and Slow brain wave frequency condition was made instead of an omnibus F test of brain wave I frequency. As previously indicated, this procedure was ; taken to ensure maximum power for the most important ; effect. The results of the analysis of variance with the i selected data indicated that the effect of frequency was 80 ' significant (F - 7.49; df - 1/36; p < .01). Reaction time i was significantly faster in the Speed than in the Slow \ I condition. The effect of age and the age x frequency I ! interaction effect were not statistically significant. Using the same 2 x 3 x 10 design with brain wave frequency I as the dependent variable, the effect of frequency was also significant (F * 116.68; df - 1/36; p< .001). The effect | of age and the age x frequency interaction effect were not statistically significant. Figure 4 presents a comparison of the mean reaction : time in the Speed, Mean, and Slow conditions for the con dition in which all data were included and for the con dition in which fastest and slowest brain wave data were selected. Figure 5 presents a comparison of the selected i ; and non-selected EEG data. These figures show that j reaction time shifts were greater when maximum and minimum i ! EEG frequency were compared than when all data were compared. The existence of a causal relationship between 1 EEG frequency and reaction time is suggested by these data. Another analysis which suggested a causal relation ship between brain wave frequency and reaction time was a j chi square comparison of the number of subjects in which i reaction time was shifted in the predicted direction. The reaction time of nine of the ten old subjects and seven of 81 310 290 270 r 230 220 210 200 Slow Moan Blofttdbick Condition t • young • - old — • Total Data - --• Salactad Data Fig. 4 The Effect of Brain Wave Frequency on Reaction Time--A Comparison of Selected and Total , Reaction Time Data for Young and Old Subjects 82 »;00 8.50 N EC 8 u. 9.00 9.50 lo.otf 1 1 Speed Mean Slow Biofeedback Condition Young Old Total Data Selected Data Fig. 5.— The effect of Biofeedback on EEG Frequency— A Comparison of Selected and Total EEG Data for Young and Old Subjects 82 ! the ten young subjects was faster in the Speed than in j the Slow conditions. These results did not appear to be | random as significantly more subjects than would be ' expected by chance shifted reaction time in the predicted i direction. Chi squares were computed for the old and | young groups and for the total group of subjects. All chi i i o square tests were statistically significant {X - 4, 6.4, and 7.2; p < .05, .02, and .01; df - 9, 9, and 19 for young, old, and total subjects, respectively). These data suggest that shifts in EEG frequency were accompanied by shifts in reaction time. To test the effect of the biofeedback conditioning i on reaction time, a2x2x2x5 analysis of variance ■ comparing the effects of age, practice, and presence or absence of feedback and subjects was carried out. Control | subjects who heard dummy feedback were compared to matched experimental subjects. The effect of feedback was not significant (F - 0.15; df - 1/16; p < .705). Reaction time was not affected by the biofeedback procedure. The ; only significant effect was the effect of age (F - 14.87; df - 1/16; p < .001). None of the interaction effects { were significant. It might be speculated that slowing brain waves is ! a more difficult task than speeding brain waves, and this 84 increase in task difficulty led to slower reaction time during the Slow condition. Learning data to be presented I I in a following section Indicated that the task of increas- ' ing the abundance of slow brain waves was easier than I , producing fast brain waves. Subjects needed significantly more trials to reach the learning criterion In the Speed ! than in the Slow condition (F - 9.77; df - 2/32; p< .001). | Hence, slow brain waves rather than increased task difficulty appear to have caused reaction time to be slower in the Slow condition. It was hypothesized that alpha abundance may be related to reaction variability inasmuch as subjects spending less time in the alpha brain wave state would have greater reaction time variability. A rank order 3 correlation between abundance of mean alpha frequency and reaction time variability in the baseline condition was computed. This correlation was -.23 which was not significant at the .05 level of confidence. Although there was some tendency for greater reaction time I variability to be related to a lesser abundance of alpha i i activity, the relationship did not reach statistical I j significance. This explanation for the cause of increased ■ reaction time variability in old subjects did not prove I 1 useful. I I i 85 In a previous section of this dissertation argu- ! ments were presented to persuade the reader that | Surwillo*s hypothesis could not account for the magnitude | of increase in reaction time variability with age. The j I present data also refute an alternative explanation for i | Increased reaction time variability in old subjects. Hence, i age changes in the EEG alpha rhythm do not appear to | account for the magnitude of increase in reaction time variability with age, although the age changes in EEG alpha frequency and abundance may account for some of the i increase in variability. Correlations were computed between EEG period ; (inverse of EEG frequency) and reaction time for the data I j collected during the baseline condition to determine if the results obtained by Surwillo (1961) were replicated. Baseline data were used, as these data were collected in conditions most closely resembling the conditions in Surwillo (1961). Inter-individual comparisons were made by correlating mean reaction and mean EEG period for all j subjects. Since the N was relatively small for this i 1 computation (N - 30), a rank order correlation coefficient | was computed. The rank order correlation between EEG I period and reaction time was .40. This correlation was 86 j significant at the .05 level of confidence, but it was ! smaller than the rank order correlation of .81 obtained , by Surwillo (1961). i ! Intra-individual Pearson product-moment correlations i I between EEG period and reaction time were also computed, j These data were taken from the forty reaction times * measured in the baseline condition. The Pearson product- moment correlation for each subject in the baseline con dition is presented on Table 4. Intra-individual j correlations were averaged by means of an r to z i i transformation, and the mean inter-individual correlation j : between EEG period and reaction time was 0.018 which was ! i not significantly different from zero. Hence, the intra- i ! individual relationship between EEG period and reaction i | time in the present study was considerably smaller than i the inter-individual correlation of .41 reported by Surwillo (1963a). Intra-individual correlations between EEG period and reaction time were computed for each experimental I subject in the other five EEG frequency conditions in i 1 which reaction time was tested (Mean, Speed, Mean, Slow, Mean), and the intra-Individual EEG period-reactIon time correlation for all 240 reaction trials was also computed. 87 TABLE 4 CORRELATIONS BETWEEN EEG PERIOD AND REACTION TIME IN THE BASELINE CONDITION FOR YOUNG AND OLD EXPERIMENTAL AND CONTROL SUBJECTS Subj ect r Experimental Old 1 2 3 4 5 6 7 8 9 10 + .35 - .03 + .01 + .15 - .31 + .12 + .03 .12 Young 11 12 13 14 15 16 17 18 19 20 .00 + .19 - .11 + .17 - .09 .05 - .24 + .02 - .14 — * ’ Control Old I + .26 - .10 + .32 25 .00 + .24 88 TABLE 4 (CONT'D) Subject r Young 26 + .03 27 - .28 28 + .08 29 - .13 30 • .05 *The EEG data for these subjects was not recorded due to technical difficulties These correlations are presented on Tables 5 and 6. Host intra-individual correlations between these two variables were low, and they were not statistically different in the different EEG frequency conditions (F - 1.29; df - 4/64; ! p < 0.285). Hence, although analysis of variance indicated i a relationship between EEG frequency and reaction time, ; correlational analyses suggested that the relationship between EEG period and reaction time was small. These data indicate that brain wave period accounts for only a i I small portion of the intra-individual variance in reaction i i time, thus casting doubt on the hypothesis that EEG alpha ! rhythm is the master timing unit for behavior. Reaction I ; time seems to be a function of a number of variables of which EEG frequency is only one. TABLE 5 CORRELATIONS BETWEEN EEG PERIOD AND REACTION TIME IN THE MEAN, FAST, MEAN, SLOW, AND MEAN CONDITIONS FOR YOUNG AND OLD EXPERIMENTAL SUBJECTS Subject r Mean Fast Mean Slow Mean Old 1 .09 .28 + .07 _ .08 - .12 2 - .05 + .14 - .06 .00 + .01 3 + .02 + .21 - .05 + .10 .. * 4 .32 + .11 - .07 + .30 + .07 5 _ .07 + .20 + .07 + .14 + .24 6 — .23 .00 - .05 + .10 - .15 7 _ .03 + .14 — .04 - .13 + .05 8 - .12 .23 - .02 — * 9 + .28 + .25 - .11 - .40 10 + .25 + .04 + .13 + .32 + .08 Young 11 .03 + .06 - .09 - .08 - .12 12 _ .04 + .31 - .21 - .32 + .04 13 + .21 - .04 - .18 - .18 - .14 14 + .22 + .03 — * + .18 15 _ .05 - .18 - .10 - .07 16 + .15 - .05 - .13 + .14 - .33 17 + .25 + .10 .00 - .01 - .05 18 + .16 - .05 + .09 - .10 + .17 19 .30 .00 - .02 + .29 + .17 20 + .12 + .12 0 .17 + .44 + .20 * The EEG data for these subjects were not recorded due to technical difficulties j i i 90 TABLE 6 CORRELATIONS BETWEEN EEG PERIOD AND REACTION TIME FOR ALL CONDITIONS FOR YOU0B AND OLD EXPERIMENTAL AND CONTROL SUBJECTS Subject r Experimental Old 1 _ .10 2 - .01 3 + .09 4 + .05 5 + .08 6 .00 7 - .07 8 - .09; 9 + .03 10 + .01 Young 11 — .09 12 - .01 13 - • 15 14 + *21 15 - .14' 16 - .03 17 - .02i 18 + .01 19 + .03 20 + .11 Control Old 21 + .26 22 - .20 23 + .20 24 + .02 25 + .13 91 TABLE 6 (CONT'D) Subj ect r Control Young 26 + .16 27 - .12 28 + .02 29 - .03 30 - .02 Correlations for the experimental subjects were computed on the basis of 240 observations of EEG and ! reaction time unless otherwise noted. Correlations for control subjects were computed on the basis of 80 observations. ^Correlations were computed on the basis of 200 EEG and reaction time observations. i Correlations were computed on the basis of 160 EEG and reaction time observations. I ^Correlations were computed on the basis of i 120 EEG and reaction time observations. Taken as a whole, the data in this experiment support the hypothesis that slower reaction time is associated with slower brain wave frequencies. The data do not support the hypothesis that EEG frequency is the | master timing mechanism for behavior. 92 Biofeedback Conditioning Figure 6 presents the acquisition curves during ■ the Mean brain wave frequency training condition for | young and old experimental subjects. All of the subjects | i In both the young and old groups were able to increase the | abundance of alpha waves in their EEG. All but one old subject were able to reach the criterion of spending two- thirds of their time in their mean alpha frequency for three consecutive two-mirute trials. Data on Figure 7 represent the acquisition curves for the experimental and matched control groups in the ■ Mean brain wave condition. A2x2x5x5 analysis of variance comparing the effects of age, feedback practice, and subjects on per cent of learning was computed. i i I Table 7 presents the results of this analysis. The effect | of feedback was significant (F * 5.00; df - 1/16; ‘ p < .05). Hence, experimental subjects produced more alpha activity than control subjects who had dummy feed back. Increases in abundance of alpha activity in the I experimental subjects appear to have resulted from the I effect of feedback. The effect of age was also signifi- i cant (f ■ 4.47; df ■ 1, 16; p < .05). Old subjects pro- i | duced less alpha activity than did young subjects. Young and old experimental subjects also success- 120 •H110 Old 100 % 5 10 15 20 25 30 35 40 45 50 55 60 65 Trials i Fig. 6.--Acquisition Curves for Biofeedback Conditioning at Mean Frequency for ^I Young and Old Experimental Subjects < ■ * ■ > j 140 130 120 JllO ^00 u o 90 6 80 « *70 60 50 - Experimental St » Control Ss ■ Young * Old Young, experimental A Young, control > ^ ' •- • V « ►«V rv Old, control L - i A.1 ^ 1 I t I I __I 30 1 5 10 15 20 A * BessIon 1 Trials ’ Session 2 Fig. 7.--Acquisition Curves for Biofeedback Conditioning at Mean Frequency for Matched Experimental and Control Subjects 35 r i 95 TABLE 7 SUMMARY TABLE FOR ANALYSIS OF VARIANCE OF THE EFFECT OF AGE, FEEDBACK, AND PRACTICE ON LEARNING TO INCREASE PER CENT. TIME IN THE MEAN BRAIN WAVE FREQUENCY Source df SS MS F Between Subjects Age 1 13270.84 13270.84 4.47* Feedback 1 14835.18 14835.18 5.00* A x F 1 51.84 51.84 0.02 S/AF 16 47459.11 2966.19 Within Subjects Practice 4 1491.06 372.76 2.32 A x P 4 259.47 64.87 0.40 F x P 4 1238.67 309.68 1.93 A x F x P 4 963.86 240.96 1.50 SP/AF 64 10251.59 160.18 i ! *p < .05 | fully Increased the per cent, time they spent in fre quencies 2 hz. faster and slower than their mean frequeny. Figures 8 and 9 present the acquisition curves for the two age groups in the Speed and Slow conditions, | respectively. A 2 x 2 x 3 analysis of variance comparing the ! effects of age, order, and frequency on the number of ! learning trails to criterion was carried out, and Table 8 summarizes the results of this analysis. The hypothesis 1 2 0 110 \ * A* Young • - Old 115 135 Fig. 8.--Acquisition Curves for Biofeedback Conditioning in the Speed Condition for Young and Old Subjects 120 110 100 ! * & 80 “ a <m 1115 Trials Fig. 9.--Acquisition Curves for Biofeedback Conditioning in the Slow Condition for Young and Old Subjects 98 TABLE 8 SUMMARY TABLE FOR ANALYSIS OF VARIANCE OF THE EFFECT OF AGE, ORDER, AND FREQUENCY ON TRIALS TO CRITERION IN A BIOFEEDBACK TASK Source df SS MS F Between Ss Age 1 4067.16 4067.16 Order 1 707.24 707.24 A x 0 1 2281.62 2281.62 Error 16 22551.72 1409.48 thin Ss Frequency 2 37599.25 18799.62 A x F 2 641.74 320.87 0 x F 2 758.54 379.27 A x 0 x F 2 5603.72 2801.86 Error 32 61573.34 1924.17 2.87 0.50 1.62 9.77*** 0.17 0.20 1.46 ** * p < .001 that old subjects would take a greater number of trials to reach criterion on the biofeedback task was not supported. There were no significant age differences in the number of trials necessary to learn the three biofeed back tasks to criterion, and the order in which the frequencies were learned did not affect the trials to criterion. The absence of age differences in the num ber of trials to criterion is especially significant as the abundance of alpha activity during the baseline con dition was considerably less In old subjects than the alpha abundance in young subjects. Young subjects spent an average of 192,45 seconds in alpha while old subjects 99 spent an average of 138.54 seconds in alpha during the twelve minutes in which alpha activity was sampled in the baseline condition. This result confirms results of previous studies indicating a decrease in abundance of i alpha activity with age. I The effect of frequency on trials to criterion was significant. Both age groups took the greatest ! | number of trials to reach criterion in the Speed condition | and the least number trials to reach criterion in the | Mean frequency condition. Hence, increasing the abundance | of alpha activity at the mean frequency appears to be an | easier task than increasing abundance at frequencies i I faster and slower than the mean. Additional research is needed to determine the cause of the difference in task difficulty. i 3 i i CHAPTER IV i DISCUSSION Interpretations The results of this experiment indicate that slow brain wave frequencies may be causally related to slower , reaction time, and this relationship exists in young as well as in old individuals. While there was evidence for : a causal relationship between brain wave frequency and | reaction time, the effect of brain wave frequency on i reaction time was small. The data do not support the ; hypothesis that the EEG alpha rhythm is the master timing j : mechanism in behavior. Hence, such a hypothesis demands ! that most of the variance in reaction time must be I accounted for by the variance in brain wave frequency, both i i within and between subjects. This demand of Surwillo's i i hypothesis was clearly not met by data in the present j experiment. While the inter-individual rank-order correlation between EEG period and reaction time was fairly large (rg - .40), intra-individual correlations between these variables were small and ranged from +.35 to -.31 in ! the baseline condition. Only a small amount of the intra individual variability in reaction time was accounted for by variability in alpha period. 100 101 The range of alpha period in one individual is | limited, hence potentially leading to lower intra- | individual correlations between brain wave period and I reaction time. In the present investigation, the range I i of brain wave periods in one individual was increased as I i subjects were trained to produce brain waves faster and i slower than their mean brain wave frequency. In spite of this procedure intra-individual correlations between brain wave period and reaction time were small. For this reason it is difficult to conclude that the alpha rhythm is the master timing mechanism for behavior. Surwillo (1963a) reported intra-individual correlations between brain wave period and reaction time I I I which were much greater than the intra-individual correlations reported by Boddy (1971) or those reported i in the data in the present investigation. Two differences in method may account for these discrepancies. These differences in methodology both involve the manner in which the EEG data were analyzed as there were differences: 1.) in the segment of EEG data analyzed in relation to the reaction time stimulus, and 2.) in the I manner in which EEG frequency was determined. Boddy (1971) related the EEG activity occurring in the five ! seconds prior to the reaction time stimulus to the reaction time, and in the present data the EEG activity 102 occurring at the onset of the stimulus was related to 'reaction time. In contrast to these intervals, Surwillo (1963a) selected the EEG activity which occurred between the onset of the stimulus and the initiation of the response for analysis. Boddy stated that it is not likely j that EEG frequency would change dramatically in the time |intervals immediately preceding and following the I |stimulus onset. While there is not direct empirical i <evidence to support this statement, it is nevertheless difficult to imagine that the two or three alpha waves occurring between the onset of a stimulus and the I !initiation of a response would differ significantly in |frequency from the wave occurring at the onset of the I |stimulus. Indeed, the first of the two or three alpha |waves measured in Surwillo's (1963a) data is probably the i i |same alpha wave which was measured in the present I experiment. The second methodological difference between Surwillo and the other studies involves the manner in which ;alpha frequency was determined. Surwillo measured the raw complex wave form appearing on the EEG output paper and calculated the frequency by averaging the lineal | width between the waves occurring during the reaction time. Since the EEG paper output was running at a speed of 50 am./sec., Surwillo* s measurements were accurate to the nearest *5 msec. Spectral analysis | was used to estimate EEG frequency in the present EEG i | data and the EEG data of Boddy. Spectral analysis ! involves the measurement of specific frequencies within i the complex wave form and averaging these individual ! j frequency contributions to derive the frequency of the wave. Spectral analysis may have detected more subtle differences in the frequencies of alpha produced by the subjects, and perhaps these subtle differences in EEG alpha frequency are not reflected in small differences in reaction time. This difference in methodology might | have led to the discrepancy between the intra-individual ; correlations of Surwillo and the intra-individual i correlations in the studies in which spectral analysis I of EEG was undertaken. I The data in the present study suggest that age changes in the EEG alpha rhythm cannot account for all of the slowing of reaction time with age. The age difference in mean reaction time in the present study I was 66 msecs., and the age difference in mean brain wave i frequency was about .4 hz. The effect of a brain wave ! shift of 1.09 hz. on reaction time was 37 msecs. Hence, i a brain wave shift of almost three times the magnitude ! of the observed age difference can account for only half 3 of the age difference in reaction time. In addition to 104 this fact, even when young subjects produced brain waves | slower than old subjects, their reaction time was still | considerably faster than the reaction time of old | i subjects. The slowing of the dominant brain wave I frequency may account for some of the slowing of reaction i j time, but data from the present investigation suggest I that changes in alpha frequency cannot account for the magnitude of age change in reaction time. Implications The results of this project raise a number of issues which can only be resolved with additional re search. Several issues regarding research methodology i and design merit additional discussion in this j dissertation, and these issues could be clarified with i additional research data in subsequent projects. Substantive issues are raised in this research which also demand discussion and experimental testing. In this section both the methodological and substantive issues will be discussed. Methodology and Design i The results of this project indicate that the EEG alpha frequency can account for some, but by all means not all, of the variance in reaction time. This result implies that a number of other variables also slow reaction time in older subjects. One variable which is 105 known to affect reaction time is inter-stimulus*interval (ISI), and this variable has not been adequately j controlled in studies relating EEG frequency to reaction | | time. The typical procedure in these studies has involved waiting for alpha waves to occur and then ; presenting the reaction time stimulus. Hence, the i subject rather than the investigator is pacing the i stimulus presentation. The magnitude of the effect of ISI ! has not been measured in any of the investigations deal ing with EEG and reaction time, and it was not estimated in the present investigation. An attempt was made to control for this effect in the present experiment by presenting reaction time signals when alpha waves were not present as well as when they were present. This , procedure eliminated to some extent the very long ISI that would have occurred if the experimenter had always I waited for alpha to be present. The problem of long ISI intervals occurring as a result of the absence of alpha activity was also alleviated in another manner as alpha abundance was increased in the present investigation, and | subjects spent more time in the alpha brain wave state. I In spite of these precautions, ISI was not measured in the present experiment, and some of the variance in | reaction time may be accounted for by ISI effects, j Automation of stimulus presentation in a pre-selected 106 i I random ISI sequence would equate this variable for all | subjects. This procedure would also facilitate the analysis of the effect of ISI as the ISI intervals would i I be precisely measured. i Another difficulty in the present experiment i involved the presentation of the reaction time stimulus during the periods in which subjects were producing the fast or slow brain wave frequencies. Although subjects were trained to a criterion of a 33% increase in the output of brain waves two hz. faster and slower than their mean frequency, they apparently were unable to maintain this level of output in the selected frequency ; when they performed the reaction time task. Analysis i i of the EEG data during feedback indicated that during j | feedback old subjects shifted from the mean of 9.07 hz. | to 9.16 hz. in the Speed condition and from 9.07 hz. to | 9.04 hz. in the Slow condition. Young subjects shifted .14 hz. in the Slow condition, and in the Speed condition their brain waves were .03 slower than in the Mean condition. The procedure of presenting reaction time stimuli after a certain time interval irregardless of I brain wave frequency probably contributed to these small i differences as in some cases the reaction time stimulus I was presented when the subject was not producing the selected frequency. Hence, the data were selected to 107 represent the ten trials In the Speed and Slow conditions | in which subjects came closest to the selected frequency. | The mean EEG shift in the selected data condition was l I 1.09 hz. for old and 1.14 hz. for young subjects. The i j corresponding mean reaction time shifts were 37 msec, for old and 27 msec, for young subjects. These reaction time shifts were greater than the mean reaction time shifts ! of 26 msecs and 7 msec, for old and young subjects, respectively, in the condition in Which all the data were included. It appeared that in conditions in which brain I wave frequency was shifted to a greater degree, reaction time was also shifted to a greater degree. The consistent production of brain wave frequencies i faster and slower than the mean frequency appears to be a difficult task to maintain while responding to a reaction time signal. Since the task was not difficult at the mean brain wave frequency which had been learned to a higher criterion, it may be necessary to train subjects \ to a higher criterion at the fast and slow frequencies \ in order to test the effects of greater EEG shifts in : frequency on reaction time. One additional feature of the apparatus would also ensure that reaction time sampling occurred when the 108 1 subject was in the appropriate brain wave state. This j feature is a state-controlled trigger which would operate to suppress the reaction time signal unless it was i ! presented when the subject was in the selected brain wave I state. This piece of apparatus would preclude the use of preset ISI intervals, but it would also preclude the need to delete reaction time data collected when subjects | were not in the selected brain wave state. Ideally, if I the subject would be trained to spend most of his time in i | a given frequency, both the preset ISI interval I schedule and the state-controlled trigger could be used | as the subject would be producing the selected frequency | during most of the intervals in which the reaction time signal was set to trigger. A feature of the experiment which might have contributed to the fact that shifts in brain wave 1 frequency were relatively small during reaction time i testing was the frequency bandwidth involved in the criterion. The bandwidth setting on the Intelex filter- feedback device was set at a + 1 hz. at all times during the experiment. At this setting the feedback signal sounds whenever a subject produces a brain wave at the selected center frequency or at a frequency + 1 hz. from the center frequency. The apparatus which counted the number of alpha waves occurring per two-minute trial was set to ■ trigger whenever the feedback signal was heard by the | subject. Hence, subjects could reach a criterion in the ' Speed and Slow conditions by producing brain waves one hz. less than the criterion setting. In future studies, ! a more stringent criterion could be set by narrowing the bandwidth around the fast and slow frequencies. This I procedure will undoubtedly lengthen the training period which for some of the present subjects lasted for 22 one-hour sessions. In spite of this lengthy training period, it would be worthwhile to attempt to train subjects to increase the production of brain wave frequencies in a greater range than the range tested in j the present experiment. Such information would suggest I the limits of plasticity in the aged nervous system, j and it would also provide additional information on the degree of the relationship between EEG and reaction time. Several issues regarding the design and the validity of inferences from the design of this study merit discussion. An attempt was made to simulate age ! in this experiment in the sense that EEG frequency was manipulated. Inasmuch as chronological age is accom- 1 panied by slowing of the EEG alpha rhythm, slowing alpha frequency in young subjects was seen as simulating old 110 I age with age measured by brain wave frequency rather than by years. Simulation of youth in old subjects was attempted by conditioning old subjects to produce faster brain waves. It must be acknowledged that while the end i result of producing brain wave frequencies of different r | aged individuals was achieved to some degree by subjects, the means by which slowing was achieved may not simulate the aging process. Conversely, old subjects producing fast brain waves may produce those fast waves in a manner different from the way in which they produced fast brain waves in their youth. The cause for the slowing of brain wave frequency i with age is not well understood. Some of the slowing I I | seems to be a result of arteriosclerotic pathology | leading to insufficient cerebral oxygen supply. Alpha slowing occurs even in individuals without observable pathology, however, so other processes also appear to be involved in the slowing. Even less is understood about the means by which subjects learn to produce ; specific brain wave frequencies. Hence, the degree to which "normal" slowing and slowing produced by i biofeedback conditioning are similar is completely | unknown. Until more is understood about the physiological mechanism in "normal" slowing and in biofeedback slowing of the EEG alpha rhythm, inferences Ill ' must be limited to the fact that the present experiment j is an age simulation study only inasmuch as the brain ! wave frequencies of old age and youth were simulated. 1 Since groups of young and old subjects were com pared, the design of the study was cross-sectional. Cross-sectional studies confound generational with 1 ontogenetic effects, hence inferences about age changes i on the basis of cross-sectional evidence may not be valid. Indeed, recent investigations in the domains of intelligence (Schaie & Strother, 1968) and personality (Woodruff & Birren, 1972) suggest that cohort differences account for most of the variance observed in cross-sectional studies of these variables. Cross- sectional comparisons in the present investigation were i limited to comparisons of EEG frequency, reaction time, | j and rate of learning of the biofeedback task. That EEG frequency slows and reaction time slows with age has been demonstrated by longitudinal studies cited previously. ! For these variables, the inferences from the present investigation do not appear to be invalid. Differences between the two cohort groups on the biofeed- i back conditioning task might well be the result of cohort differences as well as the result of age changes, but the age differences in trials to criterion in the biofeedback 112 task were not statistically significant. Hence, neither * cohort differences or age changes seem to affect biofeed back conditioning to any great degree. Substantive Issues It is interesting to note that old subjects actually learned the biofeedback tasks in fewer trials than did young subjects. The average number of trials to criterion in the Speed, Mean, and Slow conditions, respectively, i I were 93, 24, and 63 for young subjects and 70, 16, and 45 i for old subjects. Although age differences in the number j | of trials to criterion were not statistically significant, the difference approached significance (F » 2.87; df - 1/16 p < .11) in favor of the old subjects conditioning more i quickly. These results did not support the hypothesis that the old would learn more slowly, and the results were contrary to the results of many learning and conditioning studies in which the young performed significantly better than the old. The present data supply few clues for explaining ! the success of the old subjects in learning the biofeed- i back tasks. Some age differences on verbal i learning tasks have been attributed to greater interference from past experience in old subjects 113 ! (e.g. Gladis & Braun, 1957; Ruch, 1934). It may be that j biofeedback conditioning represents a task on which ! the young and old subjects were equated in terms of ' interference. The biofeedback task was totally novel j in the sense that neither young nor old subjects had ever had any experience in controlling their brain waves. Because the old had not accumulated more experience over time than the young on this task, the old may not have I been at a disadvantage in terms of interference. | Investigators have provided an alternative j j explanation for age differences in performance on a | learning task which relates to age differences in I arousal in the autonomic nervous system. Eisdorfer (1968) provided impressive evidence that in the experimental setting old subjects may be autonomically over-aroused, and he has suggested that anxiety which appears to be a psychological concomitant of autonomic over-arousal may interfere with performance. The alpha brain wave state has typically been found to represent I j a psychological state of relaxed awareness. Hence, the ! learning task in the present experiment involved relaxation rather than arousal. Subjects who were not relaxed seemed to be unable to produce alpha activity. In a learning task in which subjects were conditioned 114 to relax, old subjects performed somewhat better than | young subjects. If a subject Is tense and anxious, he Is unable I to produce alpha activity. On the basis of Eisdorfer*s work, one might predict that old subjects would perform more poorly on alpha conditioning as they would not be relaxed in the experimental setting. Several differ ences in the subjects and procedure between the Duke and USC studies may account for the observed differ- I ences. Eisdorfer measured arousal in terms of the level of free fatty acid in the plasma component of blood, and i ; to perform this measure he took blood samples at various intervals in the experiment. Subjects had a indwelling needle in their arm, and this procedure alone may have caused subjects to be anxious. The stress coupled with having to perform a verbal learning task may have made 1 old subjects more anxious than young subjects. Given equal amounts of stress, the older nervous system ! appears to over-react to the stress and take longer to j return to a baseline level than does the younger nervous | system. In this sense, if Eisdorfer's experiments were i I perceived as stressful, the old subjects might be more | greatly aroused by the stress than the young. There was j no stress in the present experiment, and every attempt I ] was made by the experimenter to relax the subjects. 115 Since all but two of the old subjects in the present | experiment were thoroughly familiar with the j experimental setting (they worked as volunteers at the ! VA hospital or were patients there) while only half of the young subjects were familiar with the setting, | the environment and procedure of the experiment probably did not stress old subjects at all. One additional factor may account for the seemingly greater relaxation of the old subjects in the present experiment. While most of the young subjects volunteered for the experiment because they had heard j of biofeedback conditioning of the alpha rhythm and I wanted to try it, none of the old subjects had heard I | of biofeedback. Kamiya (1969) suggested that subjects | who consciously try to produce alpha brain waves j typically succeed in suppressing rather than enhancing i 1 alpha activity. It is only after the subject relaxes I and stops working at alpha production that he is success ful. Since most young subjects were highly motivated j to undertake the biofeedback task, they may have j | initially tried harder and hence taken more trials to reach criterion in the biofeedback task than the old I subjects who were totally naive. More important than the fact that old subjects 116 i performed slightly better on the biofeedback task Is the j fact that old subjects can learn to modify I physiological rates at all. Geropsychologists have i | often been concerned with the degree of plasticity In , the aging nervous system and with the extent to which | age changes In behavior can be reversed. While j questions of nervous system plasticity are common, empirical research data on the modiflability of the aging organism are scarce. Indeed, although the changes i | in the abundance and frequency of the dominant EEG i frequency are among the best established a^a changes in the gerontological literature, only one relatively i I | unsuccessful attempt has been made to modify the EEG I j activity in old subjects. Surwillo (1964b) attempted to i | modify the EEG of 48 subjects with photic stimulation, i Photic "driving" involves inducing a change in the EEG ' with flashing lights. Since he was only successful i in altering brain wave frequency in five subjects, i Surwillo concluded that this method of altering brain wave frequency was relatively ineffective. i The results of this investigation clearly establish the fact that old as well as young subjects can learn to increase the abundance of EEG alpha activity in several specific frequency ranges, and 117 these results have Important implicatIons for the study ! of aging. One implication of these results is that the i slowing of the EEG alpha rhythm does not always represent irreversible deterioration in the nervous system of elderly individuals. These results suggest that there is some plasticity in the aging nervous system as old subjects can learn to increase the abundance of alpha and to produce brain waves faster than their mean 1 frequency, thus reversing to some extent the age changes ! which had occurred. Additional research is needed to determine the limits of this plasticity and to establish I the extent to which the mean brain wave frequency can j be speeded in old adults. A second implication of this research involves behavioral changes which occur with age. Investigators have considered the slowing of the dominant brain wave frequency to be of great significance in the process of i aging. If age changes in brain wave frequency is an i antecedent for some of the consequent age-related changes i | in behavior, then the experimental manipulation of brain i ' wave frequency may lead to alterations in behavior ! which are similar to those alterations which normally i 1 occur developmentally. In this manner the behavioral t | significance of the slowing of the EEG alpha rhythm 118 i might be tested experimentally. It was stated previously that there is a wide range of behavioral variables which have been identified as changing with age, and a number of these variables i have been related to parameters of the EEG. The demon stration that old subjects can learn to increase the production of brain waves faster than their mean frequency I I suggests that it may be possible to reverse some of the : deleterious behavioral changes relating to aging. At i the very least, this biofeedback technique provides a | means by which EEG-behavior-age relationships can be i , experimentally tested. One EEG-behavior-age relationship i which might be experimentally tested is the relationship between EEG, depression, and aging. A number of behavioral indices indicate that depression increases with age. Elderly subjects score higher on depression on psychometric inventories (Botwinick & Thompson, 1966; Hathaway & McKinley, 1956; ; Kometsky, 1963), suicide rate increases dramatically ; with age steaming largely from depressive illness ' (Hedri, 1967; Walsh & McCarthy, 1965), and depressive affect is a common complaint among elderly admissions to oiental hospitals. The complexity and variety of i depressions which have been observed indicate a 119 I multivariate etiology comprised of biological, psychological and sociological variables, and there is suggestive evidence that EEG changes may be related to depression. Davis (1941) was the first to note the association between changes in the alpha rhythm and depression. In a group of young manic-depressive patients, Davis observed a tendency toward slower alpha frequency than the alpha observed in normals. Hurst, Mundy-Castle & Beerstecher (1954) confirmed Davis (1941) and concluded that manic-depressive patients with predominantly depressed phases have a slower alpha frequency than those with predominantly manic phases. No change in 3 ' alpha frequency was associated with the shift from i | manic to depressive phase in a given patient. Amplitude | and abundance of alpha were also significantly lower in the depressive patients than in a group of normal young adults, but since the group of depressives was | older (mean age for depressives was 55 years while the normal adults' mean age was 22 years), it was impossible i to determine whether the differences in alpha I I abundance and amplitude resulted from the age differences or from the mood differences between the two groups. 120 ! The euphoric, tranquil affect state achieved in a yoga trance is accompanied by high amplitude, continuous EEG aplha activity which cannot be blocked by novel visual or auditory stimulation (Bagchi & Wenger, 19S7, Kasamatsu, et al., 1957). A similar pleasant affect state has been reported to investigators by participants in alpha conditioning experiments (Kamiya, i 1968, 1969; Brown, 1970). Mundy-Castle, et al., (1954) i reported a decrease in percent time alpha activity with ! age, particularly with evidence of increased dementia, and Obrist & Henry (1958a,b) confirmed this result. Hence, depression seems to increase with age while percent time alpha activity decreases, and alpha activity has been associated with pleasant affective states. Attempts to relate the EEG measures of modal frequency and percent time slow activity to depression ' in a sample of healthy old men was not successful (Birren, et al., 1963) but the investigators did not I attempt to relate the EEG measure which may be of most I j significance to affect states, abundance of alpha, to i depression. Studies undertaken to precisely identify the relationship between alpha abundance and frequency | and depression would clarify this issue. It would also be useful to attempt to modify depressive mood by 121 altering frequency and abundance of alpha. The biofeedback technique may be a means with which to examine the behavioral significance of the slowing of the EEG alpha rhythm with age, but the biofeed back technique has even broader implications for ‘ geropsychology (Woodruff, 1971). While age changes in the EEG alpha rhythm are among the best documented age changes in EEG activity, a number of other EEG changes have been observed (Obrist & Busse, 1965; Thompson, 1972). Most of the investigations of changes over age have been concerned with alpha which is probably because the 8 - 13 hz. large amplitude regular waves are readily recognized and easily analyzed by visual i I inspection of oscillographic recordings. However, alpha is only dominant under special conditions, as when a subject is relaxed, has closed eyes, or is lacking visual stimulation. As soon as a subject is engaged in mental activity such as problem solving, the EEG becomes desynchronized into a composite of superimposed fre- | quencies of lower amplitudes (Lindsley, 1952). Hence, ! biofeedback control of other parameters and frequencies of the EEG may have even greater behavioral significance than biofeedback control of the alpha rhythm. Brain wave activity is not the only physiological 122 measure which has been successfully altered with biofeedback. Heart rate and blood pressure are other physiological functions which have been demonstrated to come under voluntary control with the biofeedback i technique. Shapiro and associates (1969, 1970a) monitored blood pressure with a conventional blood pressure cuff and reinforced one group of subjects for i increasing and another group for decreasing systolic I pressure. Pressure levels became differentiated between the two groups in a relatively short period of / j training (thirty minutes). The Shapiro group (1970b) i also used biofeedback techniques to train subjects to increase and decrease heart rate, and significant heart I I rate conditioning occurred in a single training session. Engel and Melmon (1968) had some success in treating cardiac arrythmias with biofeedback, Budzynski, Stoyva and Adler (1970) used biofeedback to relieve tension headaches, and Randt and his colleagues have had encouraging results in training epileptic patients to suppress paroxysmal spikes in their brain waves (Barber, 1971). Other researchers have been applying i i biofeedback techniques to various psychosomatic ! disorders. Working with animals, Miller (1969) and his | associates demonstrated learned control of heart rate, r 123 1 blood pressure, localized blood flow, urine secretion, I and stomach and intestinal contractions. The implications and potential applications for 1 biofeedback in the fields of gerontology and geriatrics | are almost totally unexplored. The descriptive literature of gerontology contains numerous examples of deleterious physiological changes which occur with age, but little attempt has been made to explore the extent to which these changes are reversible. Phenomena such as the increase in blood pressure are commonly accepted as concomitants of normal aging, but investigators have directed their research toward explanation of age ! functional relationships rather than attempting to modify the changes. With biofeedback it may be possible to train aged subjects to reverse some of the physiological age changes which may be deleterious. In ’ this manner, biofeedback offers a behavioral approach to the maintenance of optimal phsiological functioning in the aged. ! In addition to the implications of biofeedback as i a therapeutic technique are the implications for feedback as an experimental technique. As stated I previously, increasing emphasis has been placed on the need for experimental as well as descriptive reaserch in the psychology of aging (Balts & Goulet, 1971; 124 i Blrren, 1970), and biofeedback Is a tool which can be used for experimental manipulation. It is possible to train subjects with biofeedback to maintain physiological rates of young and old Individuals. In ’ this manner simulation of the observed physiological i age functions can be achieved. Using behavioral measures as the dependent variable in such age simulation studies j j might clarify the nature of the relationship between ! ' aging, physiology, and behavior. i Thus, the biofeedback technique is potentially | useful to geropsychologists in two respects: 1.) as a ; means of reversing deleterious physiological changes J : which occur with age, and 2.) as a technique for the experimental manipulation of physiological functions. To test the hypothesis that deleterious behavioral age changes are related to phyiological age changes, biofeedback might be used to train subjects to maintain i ; an optimal physiological rate hence reversing the ; change in behavior. In this manner the biofeedback I technique is potentially useful as a therapy and as a j technique for the scientific study of aging. I CHAPTER V i i SUMMARY AND CONCLUSIONS 1. Two age changes which figure prominently in the ! gerontology literature are the slowing of reaction time j with age and the slowing of the dominant EEG alpha ' rhythm. Surwillo (1961) speculated that these age changes might be related, and he obtained a rank-order correlation of .81 between EEG period and reaction time. Subsequent work on a larger sample (Surwillo, 1963a) replicated the original finding. From these studies, j Surwillo hypothesized that EEG alpha rhythm was the | master timing mechanism in behavior, j 2. Surwillo Inferred a causal relationship between | EEG frequency and reaction time on the basis of correlational data. This inference could only be proved if it could be demonstrated that the manipulation of | EEG frequency caused differences in reaction time. | Surwillo (1964) attempted to manipulate EEG frequency with a photic driving technique, but he was not successful. Hence, he was unable to prove the existence of a causal relationship between EEG frequency and | reaction time. 126 3. The advent of the biofeedback technique which Involves an operant conditioning procedure provided a means to experimentallly manipulate the EEG alpha frequency. The biofeedback technique Involved presenting the subject | with Information about his brain waves by means of an j electronic system which amplified and filtered the 1 brain wave activity. Feedback consisted of a tone pre sented to the subject contingent upon the presence of brain waves in a selected frequency range. Made aware of their brain wave states in this manner, subjects were able to increase brain wave activity in selected frequencies. 4. The present experiment represented an attempt to i experimentally test the hypothesis that alpha rhythm i j is the master timing mechanism for behavior. Specific [ hypotheses arising from this major hypothesis were tested. These hypotheses were: 1.) When subjects are trained to produce slow alpha frequencies, they will have slower i reaction times than they will have when they are i trained to produce fast brain wave frequencies. 2.) Reaction time variability will be greatest in subjects when they are producing slow brain wave frequencies. 3.) Variability in the reaction times of i subjects who have a greater abundance of alpha will be 127 smaller them reaction time variability of subjects pro ducing a lesser abundance of alpha. Result of studies of age differences in learning and conditioning led to i I the hypothesis that: 4.) Old subjects will take more ! trials than young subjects to learn the biofeedback tasks ] to a criterion. ! 5. Groups of young and old subjects were conditioned to increase the percent of brain wave activity in their mean brain wave frequency and in frequencies two hz. faster and slower than their mean frequency. Reaction time was measured during feedback. A yolked control group heard pre-recorded feedback, and their brain waves did not control the feedback signal. 6. Results indicated that reaction time differences ' in the Speed and Slow brain wave conditions were I statistically significant. Reaction time was faster in the fast brain wave condition than in the slow brain wave condition. These results suggested a causal relationship between brain wave frequency and reaction i time. ! 7. Although reaction time variability was greater in i the Slow than in the Speed brain wave condition, the j difference was not statistically significant. Thus the hypothesis that reaction time variability will be 128 ! greatest in subjects when they were producing slow brain wave frequencies was not supported. 8. Surwillo's hypothesis could not account for the magnitude of increase in reaction time variability j with age, and results from the present investigation | indicated that an alternative hypothesis relating de- ! creases in alpha abundance tc an increase in reaction ! time variabiltiy was also inadequate. The correlation between alpha abundance and reaction time variability was not statistically significant. Although age changes in EGG alpha frequency and abundance may account for some of the increase in reaction time variability, these age changes in EEG cannot account for the magnitude of * increase in reaction time variability with age. ! 9. The biofeedback conditioning procedure did not j affect reaction time. Reaction time in control subjects 1 receiving dummy feedback was not significantly different from the reaction time of experimental subjects. | 10. Intra-individual correlations between brain wave frequency and reaction time were small, and they did not reach statistical significance. Since Surwillo*s j hypothesis demanded that intra- as well as inter- ‘ individual differences in reaction time should be accounted for by differences in EEG frequency, the data 129 did not support the hypothesis the EEG frequency Is the master timing mechanism in behavior. 11. Old subjects did not learn the biofeedback tasks to criterion in fewer trials than young subjects. The old subjects actually learned the biofeedback tasks in fewer trials that the young subjects, though the age difference was not statistically significant. The absence of an age effect may have occured as a result of I the sample of young and old subjects used in this ! investigation, or it may relate to the fact that age groups were equated in terms of interference and/or autonomic arousal on the biofeedback task. 12. The fact that old subjects learned to increase the production of fast brain waves implies that the slowing of the dominant brain wave rhythm is to some degree a reversible phenomenon. These results imply that the aging nervous system retains some degree of plasticity, and future biofeedback studies could be undertaken to determine the limits of plasticity in the aging nervous I system. i 13, Since biofeedback training was successful with old subjects, a number of implications arise for experimentation and therapy. Biofeedback training can be applied to numerous physiological functions. 130 Investigators have altered blood pressure, heart rate, kidney filtration rate, peripheral blood flow rate, and stomach and intestinal contractions with biofeedback. Biofeedback can also be used to train subjects to simulate various physiological states in young and old i | individuals to examine the behavioral significance of 1 I these states. Hence, the technique may be used to i uncover relationships between behavior, physiology and aging as well as to modify deleterious physiological , age changes and the consequent age changes in behavior. i i REFERENCES i Adrian, Lord. Creativity in science, discussion on scientific creativity. Third World Congress of Psychiatry, 1961, _1, 41-44. I ' Andermann, K. & Stoller, A. EEG patterns in hospitalized i and non-hospitalized aged. Electroencephalography ! and Clinical Neurophysiology. 1961, ^3, 319. Andersen, P. & Andersson, A. Physiological Basis of the Alpha Rhythm. New York: Appleton-Century-Crofts, 1968. Anonymous. Handbook of Human Engineering Data. Tufts College, Medford, Mass., 1952. Bagchi, B. K. & Wenger, M. A. Electro-physiological correlates of some Yogi exercises. Electroen cephalography and Clinical Neurophysio1ogy. 1957, Supplement No. 7, 132-149. j Bakes, F. P. Effect of response to auditory stimulation on the latent time of blocking of the Berger rhythm. Journal of Experimental Psychology, 1939, 24, 406-418. i i Baltes, P. B. Longitudinal and cross-sectional sequences in the study of age and generation effects. Human | Development, 1968, 11, 145-171. i j i _______ _ ____ L31_________________________ 132 ' Baltes, P. B. & Goulet, L. R. Exploration of develop- j mental parameters by manipulation and simulation of I age differences in behavior. Human Development, I 1971, 14, 149-170. Baltes, P. B. & Nesselroade, J. R. Multivariate j longitudinal and cross-sectional sequences for analyzing ontogenetic and generational change. Developmental Psychology, 1970, 2, 163-168. Bankler, R. G. A correlative study of psychological and EEG findings in normal, physically ill and mentally ill seniles. Electroencephalography and Clinical Neurophysio1ogy, 1967, 22, 189-190. , Barber, T. X. Preface. In T. Barber et. al. (Eds.), 1 Biofeedback and Self-Control; 1970. New York; ! Aldine-Atherton, 1971. Pp. vii-xvi. I | Barber, T., DiCara, L. V., Kamiya, J., Miller, N. E., ' Shapiro, D. & Stoyva, J. (Eds.), Biofeedback and Self-Control. Chicago: Aldine-Atherton, 1971. (a) Barber, T., DiCara, L. V., Kamiya, J., Miller, N. E., i j Shapiro, D. & Stoyva, J. (Eds.), Biofeedback and j Self-Control. Chicago: Aldine-Atherton, 1971. (b) i Bates, J. A. V. Electrical activity of the cortex accom panying movement. Journal of Physiology (London), | 1951, 113, 240-257. 133 | Beilis, C. J. Reaction time and chronological age. i Proceedings of the Society for Experimental Biology and Medicine, 1933, 30, 801-803. Berger, H. Uber das electren-kephalogramm des menschen. i III. Archives of Psychiat. Nervenkr., 1931, 94, i | 16-60. ! Berkhout, J., Alter, D. 0., & Adey, W. R. Alternations of the human electroencephalogram induced by stress ful verbal activity. Electroencephalography and Clinical Neurophysiology, 1969, 27_, 547-556. Birren, J. E. Age differences in startle reaction time of the rat to noise and electric shock. Journal of Gerontology, 1955, 10, 437-440. Birren, J. E. Principles of research on aging. In i J. E. Birren (Ed.), Handbook of Aging and the j Individual: Psychological and Biological Aspects. ■ Chicago: University of Chicago Press, 1959. Pp. 3-42. Birren, J. E. Neural basis of personal adjustment in i aging. In P. F. Hansen (Ed.), Age with a Future. | Proceedings of the International Congress of i Gerontology, Philadelphia: F. A. Davis, 1964. 134 i ] Birren, J. E. Age changes in speed of behavior: Its central nature and physiological correlates. In ! A. T. Weiford & J. E. Birren (Eds.), Behavior, Aging an(* the Nervous System. Springfield, Illinois: Charles C. Thomas, 1965. Pp. 191-216. Birren, J. E. Towards an experimental psychology of | aging. American Psychologist. 1970, 25, 124-135. Birren, J. E. & Botwinick, J. Age differences in finger, jaw and foot reaction time to auditory stimuli. Journal of Gerontology, 1955, 10, 429-432. Birren, J. E., Butler, R. N., Greenhouse, S. W., Sokoloff, L. & Yarrow, M. R. (Eds.), Human Aging: A Biological and Behavioral Study. Washington, D.C.: U. S. Government Printing Office, 1963. i Birren, J. E. & Kay, H. Age and sex differences in i swimming speed of the albino rat. Journal of Gerontology, 1958, 13, 374-377. Birren, J. E., Riegel, K. F., & Morrison, D. F. Age dif ferences in response speed as a function of con trolled variations of stimulus conditions: Evidence of a general speed factor. Gerontologia. 1962, 6, 135 Birren, J. E. & Szafran, J. Behavior, aging and the ner- 1 vous system. In J. E. Birren (Ed.), Contemporary i Gerontology: Issues and Concepts. University of Southern California: Gerontology Center, 1969. Pp. 258-309. i Birren, J. E. & Wall, P. D. Age changes in conduction velocity, refractory period, number of fibers, con nective tissue space and blood vessels in sciatic nerve of rats. Journal of Comparative Neurology, 1956, 104, 1-16. Birren, J. E., Woodruff, D. S., & Bergman, S. The demand in social gerontology over the next decade for research, methodology, demonstration and training. ! Gerontologist, 1972, in press. Boddy, J. The behavioral significance of some EEG i j phenomena. Unpublished doctoral dissertation. ! University of Manchester, England, 1970. Boddy, J. The relationship of reaction time to brain wave period: A re-evaluation. Electroencephalography j and Clinical Neurophysiology. 1971, 30, 229-235. i Botwinick, J. & Birren, J. E. A follow-up study of card- sorting performance in elderly men. Journal of Gerontology, 1965, 20, 208-210. i i 136 Botwinick, J. & Thompson, L. W. Pre-motor and motor com ponents of reaction time. Journal of Experimental ; Psychology, 1966, 71, 9-15. ! Brazier, M. A. B., & Finesinger, J. E. Characteristics of i I the normal electroencephalogram. I. A study of the ! occipital cortical potentials ih 500 normal adults. Journal of Clinical Investigations, 1944, 23, 303-311. Brody, E. B. The influence of age, hypophysectomy, j thyroidectomy, and thyroxin injection in simple reaction time in the rat. Journal of General Physiology, 194, 24, 433-436, Brown, B. B, Awareness of EEG-subjective activity re lationships detected within a closed feedback system. Paper presented at the Society for Psychophysiological Research, October, 1968. | Brown, B. B. Recognition of aspects of consciousness ! i through association with EEG alpha activity represent ed by a light signal. Psychophysiology, 1970, 6, 442-452. , Budzynski, T., Stoyva, J., & Adler, C. Feedback-induced | muscle relaxation: Application to tension headache. i i Journal of Behavior Therapy and Experimental Psychiatry, 1970, 1_, 205-211. Bundzen, P. U. Autoregulation of functional state of the brain: An investigation using photostimulation with feedback. Federal Proceedings Translation i 137 / I j Supplement, 1966, 25, 551-554. Busk, J. Electrophysiological properties of intellectual functioning and aging. Unpublished manuscript, Department of Psychology, University of Southern California, 1971. (a) Busk, J. EGG correlates of visual motor coordination practice. Unpublished master's thesis, University i of Southern California, 1971. (b) Busse, E. W. & Obrist, W. D. Significance of focal electroencephalographic changes in the elderly. Postgraduate Medicine, 1963, 34, 179-182. Callaway, E. Factors influencing the relationship between alpha activity and visual reaction time. Electro encephalography and Clinical Neurophysio1ogy, 1962, 14, 674-682. Callaway, E. & Yeager, C. L, Relationship between reaction time and electroencephalographic phase. Science, 1960, 132, 1765-1766. Carmona, A. B. Trial and error learning of the voltage of the cortical EEG activity. Unpublished doctoral dissertation, Yale University, 1967. Chown, S. M. Age and the rigidities. Journal of Geron tology, 1961, lji, 353-362. r 138 Cooper, R. & Mundy-Castle, A. C. Spatial and temporal i | characteristics of the alpha rhythm: a toposcopic | analysis. Electroencephalography and Clinical Neurophys iology» 1960, 12, 153-165. Dastur, D. K., Lane, M. H., Hansen, D. B., Kety, S. S., Butler, R. N., Berlin, S., & Sokoloff, L. Effects of aging on cerebral circulation and metabolism in man. In J. E. Birren, R. N. Butler, S. W. Green house, L. Sokoloff, & M. R. Yarrow (Eds.), Human Aging: A Biological and Behavioral Study. Washington, D. C.: U. S. Government Printing Office, 1963. Pp. 59-78. Davis, P. A. The electroenceophalogram in old age. i Diseases of the Nervous System. 1941, 2, 77. Dewan, E. M. Occipital alpha rhythm, eye position and lens accommodation. Nature, 1964, 214, 975-177. Dewan, E. M. Communication by voluntary control of the electroencephalogram. Proceedings of the Symposium on Biomedical Engineering. 1966, _1, 349-351. Dustman, R. E., & Beck, E. C. Phase of alpha brain waves, reaction time and visually evoked potentials. Electroencephalography and Clinical Neurophys iology, 1965, 18, 433-440. i 139 Dustman, R. E., & Beck, E. C. The effects of maturation and aging on the wave form of visually evoked potentials. Electroencephalography and Clinical I | Neurophvsiology. 1969, 26, 2-11. ! Eisdorfer, C. Arousal and performance: Experiments in verbal learning and a tentative theory. In G. A. Talland (Ed.), Human Aging and Behavior. New York: Academic Press, 1968. Pp. 189-216. Engel, B. T,, & Melmon, K. L. Operant conditioning of heart rate in patients with cardiac arrhythmias. Conditional Reflexes, 1968, 3, 130. Fedio, P. M., Mirsky, A. G., Smith, W. F., & Parry, D. Reaction time and EEG activation in normal and schizophrenic subjects. Electroencephalography anc i Clinical Neurophvs iology, 1961, 13, 923-926. j Fenwick, P. B. C. The effects of eye movement on alpha rhythm. Electroencephalography and Clinical Neurophyslology, 1966, 21, 618. (abstract) Fetz, E. E. Operant conditioning of cortical unit activity. Science, 1969, 163, 955-958. Flavell, J. H. The Deve1opmenta1 Psychology of Jean I Piaget. Princeton, New Jersey: Van Nostrand, 1963. 140 ! Fox, S. S. & Rudell, A. P. Operant controlled neural | event: Formal and systematic approach to electrical coding of behavior and the brain. Science, 1968, 162, 1299-1302. ; Friedlander, W. J. Electroencephalographic alpha rate In i adults as a function of age. Geriatrics, 1958, 13, 29-31. i Galbraith, G. C. Computer analysis of the electro encephalogram and sensory evoked response. Pro ceedings of NICHD symposium on Effects of Mal- I nutrition on Impaired Brain Function, in press. Gladis, M. & Braun, H. W. Age differences in transfer and ! | retroaction as a function of intertask response | similarity. Journal of Experimental Psychology, 1958, 55, 25-30. Goldfarb, W. An investigation of reaction time in older adults and its relationship to certain observed mental test patterns. (Teachers College Contri- I bution to Education, No. 831). New York: | Columbia University, 1941. ■ Green, A. M., Green, E. E., & Walters, E. D, Psycho- i | physiological training for creativity. Paper I ! presented at the American Psychological Association, Washington, D.C.: September, 1971. ; Green, E. E., Green, A. M. & Walters, E. D. Self- I regulation of internal states. In J. Rose | (Ed.) Progress of Cyberneticst Proceedings | of the International Congress of Cybernetics, i i London, 1969. London: Gordon & Breach, 1970. i ----------------- ----------- Green, E. E., Green, A. M. & Walters, E. D. Voluntary | control of internal states: Psychological and physiological. Transpersonal Psychology, 1970, 2, 1-26. i Griew, S. Complexity of response and time of initiating i i responses in relation to age. American Journal i of Psychology, 1959, 12. * 83-88. Hart, J. T. Autocontrol of EEG alpha. Paper presented to the Society for Psychophysiological Research, San Diego, October, 1967. Harvald, B. EEG in old age. Acta Psychiatrlca Scandanavica, 1958, 33, 193-196. Hathaway, S. R. & McKinley, J. C. Scale 2 (Depression). In G. S. Welsh & W. G. Dahlstrom (Eds.), I Basic Readings on the MMPI in Psychology and Medicine. Minnesota: University of Minnesota Press, 1956. Pp. 73-80. Hays, W. L. Statistics for Psychologists. New York: 142 Holt, Rinehart, and Winston, 1963. Hedrl, A. (Suicide in advanced ages.) Schweiz. Arch. i Neurol., Psychiat. 1967, 100, 179-202. Hicks, L. H. & Birren, J. E. Aging, brain damage and j psychomotor slowing. Psychological Bulletin, j 1970, 74, 377-396. I Hoagland, H. Studies of brain metabolism and electrical ! activity in relation to adrenocortical physiology. In G. Pincus (Ed.), Recent Progress in Hormone Research. Vol. 10. New York: Academic Press, 1954. Pp. 29-63. , Hugin, R., Norris, A. H. & Shock, N. W. Skin reflex and voluntary reaction times in young and old males. Journal of Gerontology, 1960, L5, 388-391. , Hurst, L. A., Mundy-Castle, A. C. & Berstecher, D. M. I The EEG in manic depressive psychosis. i Journal of Mental Science, 1954, 100, 220-240. Jakubczak, L. F. Psychophysiological aging. Gerontologist, 1967, 7_, 67-72. | Jasper, H. H. The ten-twenty electrode system of the International Federation. Electroencephalography and Clinical Neurophysiology. 1958, 10, 371-375. | Jasper, H. H. & Cruickshank, R. M. Electroencephalography: II Visual stimulation and the after-image as ] affecting the occipital alpha rhythm. Journal of General Psychology. 1937, L7, 29-48. Kamiya, J. Conditioned discrimination of the EEG alpha rhythm in humans. Paper presented at the ! Western Psychological Association, San Francisco, 1962. Kamiya, J. EEG operant conditioning and the study of states of consciousness. In D. X. Freedman (Chm,) Laboratory studies of altered psychological states. Symposium presented at the American Psychological Association, Washington, D. C., September, 1967. i j Kamiya, J. Conscious control of brain waves. i ! Psychology Today, 1968, .1, 56-60. Kamiya, J. Operant control of EEG alpha rhythm and | some of its reported effects on consciousness. In C. T. Tart (Eds.), Altered States of Consciousness. New York: Wiley, 1969. Pp. 507- 517. Kasamatsu, A., Okuma, T., Takenaka, S. Koga, E., Ikeda, I ! K. & Sugiyama, H. The EEG of "ZenM and "Yoga" 1 practitioners. E1ectroencephalography and Clinical Neurophysio1ogy. 1957, 9, 497-504. 144 Kerlinger, F. N. Foundations of Behavioral Research: Educational and Psychological Inquiry. New York: Holt, Rinehart & Winston, 1964. Kibbler, G. 0., Boreham, J. L. & Richter, D. Relation of the alpha rhythm of the brain to psychomotor phenomena. Nature (London), 1949, 164, 371. Koga, Y. & Morant, G. M. On the degree of association between reaction times in the case of different senses. Biometrika, 1923, L5, 346-371. Kometsky, C. Minnesota Multiphasic Personality Inventory results on an aged population. In J. E. Birren, R. N. Butler, S. W. Greenhouse, L. Sokoloff, and M. R. Yarrow (Eds.), Human Aging. Washington, D. C.: U. S. Government Printing Office, 1963, Pp. 253-258. Knott, J. R. Some effects of mental set upon the electrophyslological processes of the human cerebral cortex. Journal of Experimental Psychology, 1939, 24, 384-405. Lansing, R. W. Relation of brain and tremor rhythms to visual reaction time. Electroencephalography and Clinical Neurophysiology, 1957, 9, 497-504. Lansing, R. W., Schwartz, E. & Lindsley, D. B. Reaction time and EEG activation under alerted and i nonalerted conditions. Journal of Experimental i Psychology. 1959, 58, 1-7. i | Leavitt, F. EEG activation and reaction time. Journal of Experimental Psychology. 1968, 77, 194-199. ! Leonard, J. A. A reconsideration of the early history of reaction time studies. Applied Psychology Research Unit Report 460/53, 1953. Lindsley, D. B. Psychological phenomena and the electroencephalogram. Electroencephalography and Clinical Neurophysio1ogy, 1952, 4, 443-456. Lindsley, D. B. Attention, consciousness, sleep and wakefulness. In J. Field, H. W. Magoun, & ! V. E. Hall (Eds.) Handbook of Psysiology. Vol III. Baltimore: Williams and Wilkins, 1960. I Luders, H. The effects of aging on the wave form of the somatosensory cortical evoked potential. Electroencephalography and Clinical Neuro physiology, 1970, 29, 450-460. I Magladery, J. W. Neurophysiology of aging. In J. E. Birren (Ed.), Handbook of Aging and the Individual: Psychological and Biological i Aspects. Chicago: University of Chicago Press, 146 1959. Pp. 173-186. i I Magladery, J. W., Teasdale, R. D., & Norris, A. H. I Effects of aging on plantar flexor and super- i I ficlal abdominal reflexes In man a clinical and electromyographic study. Journal of ! Gerontologyi 1958, L3, 282-288. Matousek, M., Volavka, J., Roubicek, J., & Roth, Z. EEG frequency analysis related to age In normal adults. Electroencephalography and Clinical Neurophyslo1ogy, 1967, 23, 162-167. Miles, W. R. Correlation of reaction and coordination speed with age in adults. American Journal of Psychology, 1931, 43, 377-391. Miller, N. E. Learning of visceral and glandular I responses. Science, 1969, 163, 434-445. ; Mowbray, G. H., & Rhoades, M. V. On the reduction of choice reaction time with practice. Quarterly Journal of Experimental Psychology. 1959, 2, 16-23. i I | Mulholland, T* Feedback electroencephalography. Act!vitas Nervosa Superior, 1968, 10, 410-438. Mulholland, T. & Evans, C. R. Oculomotor function and 1 the alpha activation cycle. Nature, 1966, 211, 1278-1279. 147 Mundy-Castle, A. C., Hurst, L. A., Beerstecher, D. M., & Prinsloo, T. The electroencephalogram In the senile psychoses. Electroencephalography and Clinical Neurophys1o1ogy, 1954, 6, 245-252. ; Murrell, F. H. The effect of extensive practice on age I differences in reaction time. Journal of i Gerontology, 1970, 25, 268-274. Norris, A. H., Shock, N. & Wagman, I. Age changes in maximum conduction velocity of motor fibers of human ulnar nerves. Journal of Applied Physiology, 1953, 5, 589-593. | Nowlis, D. P. & Kamiya, J. The control of electro- I encephalographic alpha rhythms through i j auditory feedback and the associated mental j activity. Psychophysiology, 1970, 6, 476-484. j Obrist, W. D. Simple auditory reaction time in aged adults. Journal of Psychology. 1953, 35, 259-266. Obrist, W. D. The electroencephalogram of normal aged adults. Electroencephalography and Clinical j Neurophysio1ogy, 1954, 6, 235-244. ' Obrist, W. D. The electroencephalogram of healthy aged j males. In J. E. Birren, R. N. Butler, S. W. | Greenhouse, L. Sokoloff & M. R. Yarrow (Eds.), Human Agingr A Biological and Behavioral Study. 148 » ’ Obrist, | j Obrist, i I i ; Obrist, i l I Obrist, i ’ I 1 Obrist, I U. S. Government Printing Office, Washington, D.C.: 1963, Pp. 76-93. W. D. Cerebral ischemia and the senescent electroencephalogram. In E. Simonson & T. H. McGavack (Eds.), Cerebral Ischemia. Springfield, Illinois: Charles C. Thomas, 1964, Pp. 71-98. W. D. Electroencephalographic approach to age changes in response speed. In A. T. Welford & J. E. Birren (Eds.), Behavior, Aging and the Nervous System. Springfield, Illinois: Charles C. Thomas, 1965. Pp. 259-271. W. D. & Bissell, L. F. The electroencephalogram of aged patients with cardiac and cerebral vascular disease. Journal of Gerontology, 1955, 10, 315-330. W. D. & Busse, E. W. The electroencephalogram In old age. In W. W. Wilson (Ed.), Applications of Electroencephalography in Psychiatry: A Symposium. Durham: Duke University Press, 1965. Pp. 185-205. W. D., Busse, E. W., Eisdorfer, C. & Kleemeier, R. W. Relation of the electroencephalogram to intellectual function in senescence. Journal of 149 Obrist, | | Obrist, Obrist, Olds, J Olds, J Otomo, \ Gerontology, 1962, L7, 197-206. W. D. & Henry, C. E. Electroencephalographic findings in aged psychiatric patients. Journal of Nervous and Mental Disease, 1958, 126, 254-267. (a) W. D. & Henry, C. E. Electroencephalographic frequency analysis of aged psychiatric patients. Electroencephalography and Clinical Neuro physiology, 1958, K), 621-632. (b) W. D., Henry, C. E. & Justiss, W. A. Longitudinal study of EEG in old age. Excerpta Medical Inter national Congress, 1961, Serial No. 37, 180-181. Mechanisms of instrumental conditioning. In R. Hernandez-Peon (Ed.), The physiological basis of mental activity. E1ectroencephalography and Clinical Neurophysio1ogy, 1963, Monograph Suppl. 24, 219-234. . & Olds, M. E, Interference and learning systems. In J. F. Delafresnaye (Ed.), Brain Mechanisms and Learning. Springfield, Illinois: Charles C. Thomas, 1961. Pp. 153-187. E. Electroencephalography in old age: Dominant alpha patterns. Electroencephalography and Clinical Neurophysiology, 1966, 21, 489-491. 150 Pierson, W. R. & Montoye, H. J. Movement time, reaction J time and age. Journal of Gerontology. 1958, ! 13, 418-421. Rabbltt, P. M. A. Age and discrimination between complex I stimuli. In A. T. Welford & J. E. Birren (Eds.), i | Behavior. Aging and the Nervous System. Springfield, Illinois: Charles C. Thomas, 1965. Pp. 35-53. Rosenfeld, J. P., Ruddell, A. P. & Fox, S. S. Operant control of neural events in humans. Science, j 1969, 165. 821-823. | Ruch, F. L. The differentiative effect of age upon learning. Journal of Genetic Psychology. 1934, U, 261-286. Runnals, S. & Mulholland, T. Selected demonstrations of voluntary regulation of cortical activation. Bedford Research. 1965, 11, 26. i Rutherford, Lord. On the measurement of simple reaction ! time for sight, hearing and touch. Proceedings | of the Royal Society, Edinburgh, 1894, 20, 328-329. Schaie, K. W. A general model for the study of develop mental problems. Psychological Bulletin, 1965, 64, 92-107. 151 Schaie, J). W. & Strother, C. W. The cross-sequential I study of age changes in cognitive behavior. i ! Psychological Bulletin, 1968, 70, 671-680. | Schenkenberg, T. Visual, auditory and somatosensory evoked responses of normal subjects from child- ! hood to senescence. Unpublished doctoral i dissertation, University of Utah, 1970. Shagass, C. & Schwartz, M. Age, personality and somatosensory cerebral evoked responses. Science, 1965, 148, 1359-1361. ; Shapiro, D., Tursky, B., Gershon, E. et. al. Effects of feedback and reinforcement on control of human systolic blood pressure. Science, 1969, 163. 588-590. j i Shapiro, D., Tursky, B. & Schwartz, G. E. Control of | blood pressure in man by operant conditioning. i Supplement I, Circulation Research, 1970, 26. 27, 1-27 to 1-32. (a) Shapiro, D., Tursky, B. & Schwartz, G. E. i Differentiation of heart rate and systolic blood pressure in man by operant conditioning. | Psychosomatic Medicine. 1970, 32, 417-423. (b) Silverman, A. J., Busse, E. W. & Barnes, R. H. Studies 152 in the processes of aging: Electroencephalo graphic findings in 400 elderly subjects. J E1ectroencephalography and Clinical Neuro- physiology, 1955, 7[> 67-74. Singleton, W. T. Age and performance timing on simple ! skills. In anonymous. Old Age and the Modern World. London: E. & S. Livingstone, Ltd., | 1955. Pp. 221-231. I ' Spilker, B., Kamiya, J., Callaway, E., & Yeager, C. L. Visual evoked responses in subjects trained to control alpha rhythms. Psychophysiology, 1969, 5, 683-695. ! Stamm, J. S. On the relationship between reaction time i to light and latency of blocking of the alpha ! rhythm. E1ectroencepha1ography and Clinical Neurophysiology, 1952, 4, 61-68. i I Sterman, M. & Wrywicka, W. EEG correlates of sleep; I ! Evidences for separate forebrain substrates. Brain Research, 1967, 6, 143-163. I ! Stoller, A. Slowing of the alpha rhythm of the i I electroencephalogram and its association with mental deterioration and epilepsy. Journal I of Mental Science, 1949, 95, 972-984. Straumanis, J. J., Shagass, C. & Schwartz, M. Visually evoked cerebral response changes associated with 153 chronic brain syndromes and aging. Journal of Gerontology, 1965, 20, 498-506. Surwillo, W. W. Central nervous system factors in simple reaction time. American Psychologist, 1960, 15, 419. Surwillo, W. W. Frequency of the "alpha” rhythm, reaction time and age. Nature, 1961, 191, 823-824. Surwillo, W. W. The relation of simple response time to brain wave frequency and the effects of age. Electroencephalography and Clinical Neuro- physiology, 1963, L5, 105-114. (a) Surwillo, W. W. The relation of response time variability to age and the influence of brain wave frequency. Electroencephalography and Clinical Neurophvsio1ogy, 1963, 1^, 1029-1032. (b) Surwillo, W. W. The relation of decision time to brain wave frequency and to age. E1ectroencepha1ography and Clinical Neurophysiology, 1964, 16, 510-514. (a) Surwillo, W. W. Some observations on the relation of response speed to frequency of photic stimulation under conditons of EEG synchronization. E1ectroencephalography and Clinical Neuro phy Biology, 1964, T7, 194-198. (b) 154 Surwillo, W. W. On the relation of latency of alpha I attenuation to alpha frequency and to the I Influence of age. Electroencephalography and | Clinical Neurophvslo1ogy, 1966, 20, 129-132. | | Surwillo, W. W. Timing of behavior in senescence and the role of the central nervous system. In G. A. Talland (Ed.), Human Aging and Behavior. New York: Academic Press, 1968. Pp. 1-35. Surwillo, W. W. Human reaction time and period of the EEG in relation to development. Psychophysiology, 1971, 8, 468-482. ; Szafran, J. Changes with age and with exclusion of vision in performance at an aiming task. Quarterly Journal of Experimental Psychology, I 1951, 3, 111-118. Szafran, J. Limitations and reliability of the human operator of control systems to process information. Aerospace Medicine, 1966, ^7, 239. ’ Szafran, J. Psychophsiological studies of aging in i pilots. In G. A. Talland (Ed.), Human Aging ! and Behavior. New York: Academic Press, 1968. Pp. 37-74. Talland, G. A. The effect of age on speed of simple manual skill. Journal of Genetic Psychology, 1962, 100, 69-76. Tecce, J. J. Contingent negative variation and individual differences. Archives of General Psychiatry, 1971, 24, 1-16. Thompson, L. W. Psychophysiological studies of aging. In C. Eisdorfer & M. P. Lawton (Eds.), APA Task Force on Aging, Washington D. C.: American Psychological Association, 1972, in press. Thompson, L. W. & Botwinick, J. The role of the preparatory interval in the relationship between EEG alpha-blocking and reaction time. Psychophysiology, 1966, 3, 131-142. Thompson, L. W. & Nowlin, J. B. Cortical slow potential and cardiovascular correlates of attention during reaction time performance. In L. Jarvik, C. Eisdeofer & J. Blum (Eds.), Aging; Psychological and Somatic Changes. New York: Springer & Co., in press. Travis, L. E., Knott, J. R. & Griffith, P. E. Effect of response on the latency and frequency of the Berger rhythm. Journal of General Psychology. 1937, 16, 391-401. Venables, P. H. Periodicity in reaction time. British Journal of Psychology. 1960, 51, 37-73. 156 Verdeaux, G., Verdeaux, J. & Tunnel, J. Etude i statistique de la frequence et de la reactivite des electroencephalogramnes chez les suyets ages. Journal of the Canadian Psychiatric Association, 11 “ ™ ' — — —- 1961, 6, 28-36. Wagman, I. li. & Lesse, H. Maximum conduction velocities of motor fibers of ulnar nerve in human subjects i i of various ages and sizes. Journal of Neurophysiology, 1952, 15, 235-244. ; Walsh, D. & McCarthy, P. D. Suicide in Dublin's elderly. Acta Psychlatrica Scandinavica, 1965, 41, 227-235, ■ Walter, D. 0., Rhodes, J. M. & Adey, W. R. i Discrimination among states of consciousness by EEG measures: A study of four subjects. E1ectroencepha1ography and Clinical Neuro- physiology, 1967, 22, 22-29. j Wang, H. S. & Busse, E. W. EEG of healthy old persons-A longitudinal study. I. Dominant background 1 activity and occipital rhythm. Journal of Gerontology, 1969, 24, 419-426. Wang, H. S., Obrist, W. D. & Busse, E. W. Neuro- | physiological correlates of the intellectual \ j function of elderly persons living in the community. American Journal of Psychiatry, 157 1970, 126, 1205-1212. Wayner, M. J. & Eromers, R. Spinal synaptic delay In young and aged rats. American Journal of Physiology, 1958, 194, 403-405. Wilson, S. & Obrist, W. D. Age differences In EEG response to visual stimulation. Paper presented | at the American EEG Society, San Francisco, 1963. I I Wohwill, J. F. Methodology and research strategy in the i ! study of developmental change. In L. R. Goulet & P. B. Baltes (Eds.), Life-span Developmental i I Psychology. New York: Academic Press, 1970. Pp. 145-193. ! Woodruff, D. S. Biofeedback--Imp11cations for p f I gerontology. In D. S. Woodruff (Ohm.), Design i i strategies and hypotheses of psychoblological research in aging. Mini-symposlum presented at the 24th annual meeting of the Gerontological Society, Houston, October, 1971. Woodruff, D. S. & Birren, J. E. Age changes and cohort differences in personality. Developmental Psychology, 1972, 6, 252-259.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Simple Auditory Reaction Time Of Young Adult And Elderly Subjects In Relation To Perceptual Deprivation And Signal-On Versus Signal-Off Conditions

PDF

Psychomotor Performance And Change In Cardiac Rate In Subjects Behaviorally Predisposed To Coronary Heart Disease

PDF

Monocular Acquisition And Interocular Transfer Of Two Types Of Discriminations In Normal And Corpus Callosally-Sectioned Guinea Pigs

PDF

Electrophysiological Properties Of Reaction Time And Aging

PDF

Performance Evaluation Based On Multidimensional Job Behaviors

PDF

Age Differences In Serial Reaction Time As A Function Of Stimulus Complexity Under Conditions Of Noise And Muscular Tension

PDF

Signal Processing Time And Aging: Age Differences In Backward Dichoptic Masking

PDF

Protein synthesis in the hippocampus of rats during learning assessed by radioautography

PDF

Habituation Of The Multiple Unit Discharge Response To White Noise Stimulation In The Unanesthetized Rabbit

PDF

Influences Of Interoperative Experience And Age On Recovery Of Visual Function Following Two-Stage Lesions Of The Striate Cortex

PDF

Age Differences In Primary And Secondary Memory Processes

PDF

Changes In Memory As A Function Of Age

PDF

The Enhancement Of Eeg - Alpha Production And Its Effects On Hypnotic Susceptibility

PDF

The Use Of The Orienting Reflex To Test The Zeaman And House Theory Of Attention

PDF

Differential effects of epinephrine and propranolol on shuttle box avoidance learning in rats of different ages

PDF

The Effect Of Irrelevant Environmental Stimulation On Vigilance Performance

PDF

The Effects Of Prior Part-Experiences On Visual Form Perception In The Albino Rat

PDF

Experimenter Expectancy Effect Examined As A Function Of Task Ambiguity And Internal Versus External Control Of Reinforcement

PDF

Recognition Accuracy In The Method Of Single Stimuli: A Test Of An Operational Definition Of The Distinctiveness Of Stimuli

PDF

Human Performance As A Function Of The Joint Effects Of Drive And Incentive Motivation

Asset Metadata

Creator Woodruff, Diana Stenen (author)

Core Title Biofeedback Control Of The Eeg Alpha Rhythm And Its Effect On Reaction Time In The Young And Old

Degree Doctor of Philosophy

Degree Program Psychology

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

Tag OAI-PMH Harvest,psychology, experimental

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Birren, James E. (committee chair), Galbraith, Gary C. (committee member), Lindsley, David F. (committee member), Szafran, Jacek (committee member), Walker, James Paul (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-752086

Unique identifier UC11363534

Identifier 7226067.pdf (filename),usctheses-c18-752086 (legacy record id)

Legacy Identifier 7226067

Dmrecord 752086

Document Type Dissertation

Rights Woodruff, Diana Stenen

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