University of Southern California Dissertations and Theses

Auditory brainstem responses (ABR): variable effects of click polarity on auditory brainstem response, analyses of narrow-band ABR's, explanations

(USC Thesis Other)

Auditory brainstem responses (ABR): variable effects of click polarity on auditory brainstem response, analyses of narrow-band ABR's, explanations

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content AUDITORY BRAINSTEM RESPONSES (ABR): VARIABLE EFFECTS OF CLICK POLARITY ON AUDITORY BRAINSTEM RESPONSE t ANALYSES OF NARROW-BAND ABR’s, EXPLANATIONS. by Konstantinos Alataris A Thesis Presented to the FACULTY OF THE SCHOOL OF ENGINEERING UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree MASTER OF SCIENCE IN BIOMEDICAL ENGINEERING August 1995 Copyright 1995 Konstantinos Alataris T h is thesis, w r itte n by ....................K0N5TANTINr S ALATAR IS........................ under the gu id a n ce of F a c u lty C o m m itte e a n d a p p ro v e d by a l l its m e m b e rs, has been presented to and accepted by the S chool of E n g in e e rin g in p a rtia l f u lf illm e n t of the re quirem ents f o r the degree o f MASTER OF SCIENCE IN BIOMEDICAL ENGINEERING D a te S/4/95 l: acuity C om m ittee C h a irm a n ACKNOWLEDGEMENTS The experiments were conducted at the Electrophysiology Lab of the House Ear Institute, under the supervision of Dr Manuel Don, to whom I am grateful for giving me the opportunity to work in this collaborative project and for his guidance and suggestions. I would also like to thank Dr. Curtis Ponton for his significant assistance, and Ann Masuda who very patiently taught me how to conduct the experiments I express my gratitude to Dr. Vasilis Marmarelis for establishing the collaboration between the University of Southern California’s Department of Biomedical Engineering and the House Ear Institute, as well as for his support throughout the project. I would also like to thank Dr Jean- Michael Maarek and Dr Michael Khoo for their suggestions. CONTENT ACKNOWLEDGEMENTS ii LIST OF FIGURES v LIST OF TABLES vi 1 INTRODUCTION 1 2 FUNCTIONAL ANATOMY OF THE AUDITORY SYSTEM 2 3. BRAINSTEM RESPONSE AUDIOMETRY 10 3 A BASIC ABR MEASUREMENTS 15 4 RECORDING PROCEDURE 16 4 1 ELECTRODES 17 4 1 A Electrode sites 17 4 1 B. Electrode application 19 4 2 ACOUSTIC STIMULATION 20 4 2 A. Acoustic stimuli 22 4.2.B.Acoustic representation and calibration 24 4.3.AMPLIFIER 27 4.4 FILTERS 30 4.5. A - D CONVERTER 32 4 6 AVERAGING TECHNIQUE 35 5 STUDY 39 5 A POLARITY EFFECTS 40 5 B.DERIVED-BAND ABRS AND HIGH-PASS MASKING 45 iii 6. METHODS 51 6.A.INITIAL STUDY 51 I Subjects 51 II Stimuli 52 III.ABR recordings 53 6 B.FINAL STUDY 54 6 C THE ART OF PEAK PICKING' 57 7 RESULTS 61 8 DISCUSSION 78 REFERENCES 83 iv LIST OF FIGURES Figure 2.1. Schematic of the auditory system 3 Figure 2.2.An unrolled cochlea 4 Figure 2.3.A schematic view of one turn of the cochlea. 5 Figure 2.4.Characteristics of human cochlea 6 Figure 2.5.Tonotopic organization 7 Figure 2.6.Ascending and descending paths 9 Figure 3.1.Auditory evoked potentials 10 Figure 3.2.Classification of ERA. 11 Figure 3.3.Auditory Brainstem Response. 12 Figure 3.4.ABR in comparison to the ACAP 13 Figure 4.1. Schematic diagram of instrumentation 16 Figure 4.2.The International 10-20 Electrode System 17 Figure 4.3.Shielded transducer (earphone) 22 Figure 4.4.Acoustic stimuli. 23 Figure 4.5.a),Condensation and Rarefaction stimuli. 25 b).Physical properties of stimuli. Figure 4.6.a),Differential amplifier 28 b). Suppression of common mode signals Figure 4.7.Principle of operation of a differential amplifier. 29 Figure 4.8.Bandpass filters. 30 Figure 4.9.Other type of filters 31 Figure 4.10.Effect of A-D amplitude resolution 33 Figure 4.11.Effect of A-D time resolution 34 Figure 4.12.Signal Averaging 36 Figure 4.13.Averaging of ABR’s 37 Figure 5.1.Polarity effects 40 Figure 5.2.High pass masking technique 49 Figure 6.1. Derived ABR’s. 56 Figure 6.2.Analysis of ABR 59 Figure 7.1.Male and female Wave V latencies. 62 Figure 7.2. C-R signed and absolute latency difference compared to theoritical half-period of CF for derived ABR’s. 63 Figure 7.3.Six superimposed runs of the unmasked and derived band responses for one individual 64 Figure 7.4.Distribution of wave V latencies between runs 66 Figure 7.5.Mean signed wave V latency differences for C-C, R-R and C-R are plotted in comparison to the expected theoretical half-period differences. 68 Figure 7.6.|C-C + theo| plotted against the signed (C-R) and the theoretical half-period differences 69 Figure 7.7.Mean absolute wave V latency differences 70 for C-C, R-R, and C-R are plotted in comparison to the expected theoretical half-period differences. Figure 7.8.The absolute differences for the simulated data (|C-C + theo|) and the |C-R| are compared to the theoretical half-period differences. Figure 7.9.Resulting distribution when absolute values are taken from a normal distribution with mean = 0. Figure 7.10. (a) |C-C| and (b) ]R-R| against the compound standard deviations of those distributions Figure 7.11. (a) |C-R against the compound standard deviations of those distributionsand the |C-C + theo| LIST OF TABLES Table 1.ANOVA results testing for polarity effects on wave V latency 1.INTRODUCTION The auditory electrophysiology is often referred to as electric response audiometry (ERA) and during the past 10 years electric response audiometry {ERA),particularly brainstem audiometry, has become an important clinical tool The aim of ERA is to record the potentials that arise in the auditory system as a result of sound stimulation Electric response audiometry (ERA) is divided into several classes This classification is useful since the various methods record evoked potentials (EP’s),produced or stimulated by sounds, originating from different locations along the auditory pathway The location where the individual evoked potential is generated ,is called the generator site. The ERA consists from the following three major recording methods: 1 Electrocochleography (ECoG). 2 Brain stem response audiometry (BRA) 3,Cortical response audiometry(CRA) Of all the above three we’ re going to focus to the BRA and the Auditory Brainstem Responses (ABR’s). 1 2.FUNCTIONAL ANATOMY OF THE AUDITORY SYSTEM The human auditory system can be divided into three main portions : the conductive mechanism,the sensorineural mechanism and the central mechanism ( figure 2.1). The primary function of the outermost portion ,the conductive mechanism, is to bring the vibrational sound energy from outside the head to the inner portions of the ear so they can be be used by the sensorineural mechanism The outer ear consists of the pinna and the external auditory canal and ends at the lateral border of the tympanic membrane The three ossicles (malleus,incus stapes) in the middle ear couple the tympanic membrane to the inner ear Two muscles and several ligaments help support these ossicles The resonance of the external auditory canal and the tympanic membrane, along with the lever action of the ossicular chain and the area difference between the tympanic membrane and the oval window membrane, help amplify the air pressure at the external auditory meatus so that the air pressure can drive the dense fluids of the inner ear For higher intensities motion of the ossicular chain varies as the frequency and intensity of the input stimulus change. 2 c o rM e u m ic k le ear ___ J J l____ J L Central Senaort-neural Conductive Figure 2.1.Schematic of the auditory system. The main part of the three inner ear structures is the cochlea .which is divided into three sections:scala vestibuii, scala media ,scaia tympany (by the basilar membrane and Reissner's membrane).The scala vestibuii and scala tympani contain the fluid perilymph and the scala media contains endolumph The basilar membrane is supported on the modiolar side by the osseous spiral lamina and on the stria vascularis side by the spiral ligament (figure 2 .2 ). The basilar membrane is widermore flaccid and under no tension at the apical end.The base end is narrower and stiffer than the apical end and may 3 be under a small amount of tension In the scala media is the organ of Corti On the inner (modiolar) side of the tunnel of Corti are the inner hair cells. On the outer side of the tunnel of Corti are the three rows of outer hair cells;their cilia are in contact with the tectorial membrane The hair cells and nerve fibers are held in place by supporting cells ( figure 2.3 ). The basilar membrane has a traveling-wave motion when exited by the stapes.The traveling wave yields maximal displacement only at the base for high-frequency stimuli and maximal displacement , after a time delay, at the apex for low frequency stimulation (figure 2 4) l»#licotrem a batal mm! Figure 2.2.An unrolled cochlea 4 vnr m w bony spiral l ami na Reissner's membrane stria vascularis spiral ligamenr tectond membrane outer Hair cells inner hair cells canai of Cam basiiar memarane Figure 2.3. A schematic view of one turn of the cochlea The differential motion of the basilar and tectorial membranes results in a shearing motion of the cilia of the hair cells , which in turn triggers the nerve fibers. 5 CHAIlACTEItlSTICS OF HUMAN COCHLEA Length (mm) 9 12 *4 16 *8 2U 22 24 20 28 .'» > 32 34 r m i 11111 h 11 i t i ii 1 1 i i 111 i i 11 ii i Ov:il Window llusu(;!um Second 0.U 0.25 Apex 4.0 2.0 1.0 Frequency (kHz) .5 ,:i . i 0.5 1.0 2.0 4.0 Travelling Wave Delay (msec) Figure 2.4. Characteristics of human cochlea To resume the cochlea is a spiral-shaped organ that is responsible for converting the sound energy carried inwards by the conductive mechanism into a neurological code that can be interpreted by the central auditory mechanism. This neurological code that now represents the auditory stimulus is carried by the VIII cranial nerve (the auditory nerve) to the central auditory mechanism.The cochlea and the VIII nerve are known jointly as the sensorineural portion of the auditory mechanism (figure 2 5) 6 Tonotapic Orfaaizatiofi Figure 2.5.Highly schematic diagram of the bilateral central auditory system. The main pathways and nuclei are shown for both cochleas Bilateral representation from binaural stimulation occurs at the superior olive and in all regions above. 7 The central auditory mechanism is responsible for the the recognition , interpretation and integration of auditory information.The auditory message, after being encoded by the cochlea .passes through the brainstem on his way to the cortex.The auditory pathways through the brainstem and the auditory areas in the cortex are refereed to as the central auditory mechanism We mentioned the very important concept of frequency being represented by a particular place along the basilar membrane, the place being determined by the location of maximum displacement of the basilar membrane in response to a particular frequency of stimulation Higher frequencies were represented in the base of the cochlea and lower frequencies in the apex.This spatial representation of frequency along the cochlea partition was maintained by the afferent VIII th n e rve fib e rs.F ib e rs with high characteristics frequencies innervated the base of the cochlea and those with low CF's the apex The maintenance of this spatial representation of frequency throughout the nuclei of the central auditory pathway is refereed to as Tonotopic organization(figure 2.6). 8 Corp iaitosum \\\inf*ricr csU ktjUjs T*cto ‘Lateral lefnnwcu* pontai M and T^~TK^n»uc(»i r*^ > S P F J ^ ll\ I I ) P o n s .Cocnear Nuclei Supen o r/v —f-y oli^ Oescericir.g Figure 2.6. Top. Scheme of the tonotopic organization of the dorsal and ventral cochlear nuclei. Cochlear nerve fibers establish an orderly correspondence of successively more apical regions of the cochlea with progressively more ventrolateral sectors in each part of the cochlear nucleus. Bottom. The main ascending and descending pathways of the central auditory system Key: corticolateral tract (*) ;crossed olivocochlear tract (arrowhead);medial lemniscus (ML) ;reticular formation (RF);motor trigeminal nucleus (5) ;motor facial nucleus (7); spiral ganglion (8). 9 3.BRAINSTEM RESPONSE AUDIOMETRY Brainstem response audiometry records the electrical activity in the cochlear nerve and parts of the brain stem in 1 to 15 msec time window The auditory brainstem response (ABR) is a complex response ,which reflects activity from the cochlear nerve (N VIII ) ,the cochlear nucleus (CN) and possibly from the superior olivary complex (SOC) the lateral lemniscus (LL) and the inferior colliculus (IC) (figure 3.1, 3.2) MGS CN B O CN SG 400 ma £0 ms s ~ x_ S P 10 ms Figure 3.1. Auditory evoked potentials. 10 CLASSIFICATION Method: Bscmc fl« p o n ss Audiometry (ERA). OF ERA. Weeponee: Audftoiy Eisetnc Reaponee (AER). y y l / W V tO ms L Peripheral activity (0 -10 ms). Method: EisetrococWsogtaphy (ECoGi Reaponee: CM. SP and ACAP 11 Activity from the hair cells: Cochlear mtcrophonic (CM) Summenng potential (SP) 2) Activity from the cochlear nerve Auditory compound action potential i ACAP) I u ill iv V vi vii '5 ms 50 m j A C S I T O O 400 ms II. Brain stem activity (1 -15 ms). Method: Brain stem response audiometry (BRA) Reaponee: Auditory drain stem response (ABR) The A8R consists or six to seven oeaxs - indicated Dy i Vfl The activity s cossiciv generated m cccniear nerve cccm ear nucleus, superior olivary ccmotex .aterai lemnisc-is ■ntenor coihcuius ACS ID . Cortical resoonse activity (10 - 400 m il. Metnod: Conical response auciom etrv CRAi Response: Auotorv ccrticai esoonse i ACRi ‘ i Miccie -esccnse 'C ■ =0 n s 1 a mixture of myogenic- arc neurai ictiviiv _.ne atte' s oqssioiv generaiec n re raian-LS anc. 2' he ormarv aucucrv cortex 2' Care -esccrse ;50 - -CO m si is oossiciv ceneratec n :re orimarv- and associated cortical areas Figure 3.2. Classification of ERA In these figures we also see the differences between ABR’s and the responses recorded with the other techniques ie., ECoG and CRA.In the first 15 msec after a click stimulus , a series of small waves can be recorded between an electrode on the scalp or forehead and one in the ipsilataral 11 mastoid or earlobe.These waves are designated using Roman numerals l-Vlll for the vertex positive peaks (figure 3.3) I II III IV /V ( Nl N 2 N 3 N4 VII N 6 ) N 5 0.5 mV CLK, ( P 5 ) o 5 to m s e c I 1 1 1 I !____! I L i i I ■ t i 1 Figure 3.3. Auditory Brainstem Response Wave I Based on data from several studies , there is a general agreement that the first (I) vertex positive potential in the ABR sequence is produced by acoustic nerve activity The most common basis for relating the wave I to the acoustic nerve is its temporal correspondence with the acoustic nerve potential Nl. In the upper part of figure 3 4 is shown an auditory brainstem response (ABR) recorded in a normally hearing subject (The response has been elicited by a fairly strong 2 kHz half sinusoid acoustic stimulus) For comparison is also shown an auditory compound action potential (ACAP) from the same 12 subject. That’s recorded by the ECoG and is composed of the summed activity from all the individual fibers in the cochlear nerve in response to the applied stimulus. |N2 ACAP CM 0 2 4 6 8 10 12 14 16 ms T + 250 nV T + 1000 nV Figure 3.4. ABR in comparison to the ACAP It is obvious that the ABR is much smaller than the ACAP (in this case about 4 times ) , and that the brainstem potential can be characterized as a pattern of almost uniform peaks, the so called ‘Jewett waves' (after Don L Jewett who first labeled the human ABR) ,and designated with the Roman numerals l-VIII From the figure it can be seen that the first positive wave (I) is identical to the negative wave " ‘ ' he ACAP This means that it is possible also with surface electrodes to pick up the response generated in the auditory nerve 13 84 Wave I appears as a relatively small positivity when a vertex electrode is referred to the pinna, mastoid Also many studies have shown that wave I persists while all subsequent waves are abolished Wave II ABR wave II appears as a small-to-moderately-large positivity when a vertex electrode is referenced to the pinna or mastoid. It’s generated from a different far field locus than wave I.It is believed to arise from the central end of the eighth nerve Wave III ABR wave III recorded as as a small-to-relatively-large positivity when a vertex electrode is referenced to the pinna or mastoid It's one of the most prominent and consistent components (along with wave V) and it occurs approximately 2 msec later than the wave I.A variety of data links the MSO (medial superior olive) and its field potentials with wave III Wave III is variable in shape between individuals. Wave IV ABR wave IV may overlap with wave V to variable extend Wave IV is variable in shape between individuals Wave V ABR wave V is the most prominent and consistent component.lt occurs approximately 4 msec later than the wave I. It is much less affected by 14 increasing stimulus presentation than the earlier waves.Wave V can be considered a compound response deriving from connections from different cochlear areas High pass noise masking can be used to separate these different subcomponents of the compound wave V 3.A. BASIC ABR MEASUREMENTS The two parameters of the ABR waveform that usually are measured are amplitude and latency.Amplitude typically is measured between a positive peak and the following negative “peak” or trough Peak-to-peak measures are favored because they avoid the difficulty of determining the baseline of the potential There are several latency measures of interest The most basic is absolute latency, which is defined as the time difference between stimulus on-set and the peak of the wave. Interwave latencies (or intrepid intervals) are the differences between absolute latencies of two peaks.such as l-V,l-lll and l-V. The latency differences between l-lll and l-V have proven to be important measures in the diagnosis of retrocochlear disorders.The latency difference between l-lll is almost constant and approximately 2.2 msec.The difference between l-V depends on age,sex and hearing loss. 15 4.RECORDING PROCEDURE A schematic diagram of the instrumentation used in the ABR recordings is given in figure 4.1 We 'II go through the main components of that configuration and discuss the main factors influencing the ABR measurements. SICNAl. AVERAGING COMPUTER STIMULUS GENERATOR OSCILLOSCOPE Iaa/vajuJ | a A> CONVERTOR | DISK STORAGE AMPLIFIER PLOTTER Non Pathologic Subjsot Foetors Stimulus Factor* Acquisition Factors Pathologic Sub)*cl Factors . ago . gender . body tem perature . i l a t * of orousol . drug* . m u te le activity . frequency . duration . Intensity . ro t* . polarity . tronaducer . mashing . p r**«ntallon m od* . slsclrod** . amplification . mi wing . analysis tlms . signal averaging . conductive hearing lass . cochleor hearing loee . Blh nerve dysfunction . brainstem dysfunction . cerebral dysfunction Figure 4.1. Schematic diagram of instrumentation. 16 4.1.ELECTRODES 4.1.A. Electrode sites Appropriate electrode sites can make the difference between recording a well-formed response and not observing a response at all In ABR measurement in particular and in clinical neurophysiology in general electrode sites are usually defined according to the International 10-20 Electrode System The rational between this system is evident in figure 4 2 R C N T TOP (IF HEAD Figure 4.2. The International 10-20 Electrode System. 17 There are several general principles relating electrode site and AER components.First, the closer the electrode is to the anatomic generator , the larger the responce.For example, the ECochG AP component recorded from the promontory (lateral wall of the cochlea ) may be 20 times larger than the AP recorded from the earlobe or mastoid. Second , in recording far-field responses from sites equidistant from the generator, such as the ABR , the exact location of the non-inverting electrode is not crutial The response is essentially comparable when recorded anywhere along the midline, from vertex (Cz) to forehead (Fz) Inverting electrodes located on the earlobe are preferable to mastoid sites, however , because wave I tends to be larger , and the electrode picks up less electromagnetic artifact from bone vibrators, as typically used , when electrodes are located on the earlobe instead of on the mastoid bone.lt is possible that wave V amplitude may be slightly reduced with earlobe placement of the inverting electrode. Either of these sites (mastoid or earlobe ) are, however, active with reference to electrical activity arising from the auditory system (especially cochlea, eight nerve and lower brainstem) and are not properly termed reference or or indifferent electrodes. 18 If interactions among noninverting and earlobe electrodes is suspected, on the basis of waveform morphology and/or difficulty identifying waves beyond I, then a noncephalic electrode site (nape or the neck) is indicated 4.1.B. Electrode application Electrode application is a technical factor that is extremely important for successful evoked response measurment.The over objectives are 1. Consistent (among subjects) and anatomically accurate placement 2 Low interelectrode impedance (less than 5000 ohms). 3 Balanced interelectrode impedance (difference between electrodes or less than 3000 ohms) 4. Secure and consistent attachment throughout the test session 5 Minimal discomfort and no risk to the subject. The electrode site is first prepared by scrubbing vigorously with an abrasive liquid substance. Alcohol or acetone can also be used This removes the natural oils of the skin, as well as the superficial layers of skin,thus improving the interelectrode impedance. Then each electrode is dabbed with conducting paste or gel (in order to enhance the electrode - to - skin connection ), applied and and taped. 19 Placing a cotton ball over the affixed electrode and then securing the electrode with tape may help to form a better connection. Some kind of electrode paste ( as opposed to gel ) actually abhere the electrode to the skin although tape is usually needed Also , taping the lead wire to the skin several inches from the electrode, provides added security to prevent electrode slippage or unintentional movement After the end of the test electrodes can be removed easily and the sites cleaned (ie , gel or paste removed ) with alcohol swab,diaper wipes, face cloth soaked in warm water, or any other suitable way. 4.2.ACOUSTIC STIMULATION In the auditory electrophysiology the acoustic stimuli are produced by a transducer, which most often is specially designed sound receiver. Due to calibration problems and the limited ability to reproduce acoustic transients, loudspeakers are seldom used. If the evoked potential has a frequency range that coincides with the frequency of the stimulus , the averaging procedure is extremely sensitive to electromagnetically induced interference or artifacts from the sound generating system-predominantly the receiver. 20 Especially at high sound levels the transducer is surrounded by a strong magnetic field created by the stimulus current in the receiver In order to attenuate these magnetic artifacts , the sound transducer is enclosed in a p- metal case which has the ability to reduce the magnetic radiation The shielding has the disadvantages that the receiver gets more heavy and clumsy and that the acoustic specifications to some extend are changed.Details of the inside of a shielded transducer in figure 4.3. Important transducer characteristics : I. The frequency characteristic is a measure of the receiver’s ability to reproduce sounds of different frequencies.lt should be reasonable flat without any sharp resonance s. II The impulse response describes the receiver's ability to deliver brief sound pulses or transients.lt should show a reasonable damped oscillation. III The maximum sound pressure indicates the highest sound levels the receiver can reproduce without distortion This parameter is especially important when brief sound stimuli with peak-to-peak sound pressure levels of around 130-140 dB are delivered by the transducer 21 Circumaurai Cushion / MX ai -AR Cushion Noise Excluding Cover Foam Rubber Riling Mu-Metal Boxes Copper Shielding Transducer ctoicicxo cicictcrcicic; Figure 4.3. Shielded transducer (earphone). 4.2.A. Acoustic stimuli 1. Broadband clicks: A broadband click is provided by driving the sound transducer with a short square wave (50 -100 psec ).Depending upon the frequency characteristic of the transducer , it will give a short oscillatory acoustic signal (close to the 22 3416 impulse response) which, if measured in the frequency domain , will have a broad frequency spectrum capable of stimulating most of the basilar membrane in the inner ear. A broadband click is a frequency non-specific stimulus (figure 4.4). — ^ A / W W W W W V ^ ■ 1 0 0 n s CBck 2 kHz Hail Sinusoid fill. Click 1.5-0-1.5 Tone Bunt 1.5-5-1.5 Tone Eurat j 8 TIS 25 5 8 4 h Frequency ikHz) Figure 4.4. Acoustic stimuli. 2.Half sinusoid clicks. 3.Filtered clicks. 4.Tone bursts. 23 4.2.B. Acoustic representation and calibration Besides the choice of stimulus type and frequency , some other parameters also have to be concidered.These are : I Acoustic polarity II.Inter stimulus interval (ISI) II Sound level I Acoustic polarity The acoustic polarity or the phase is a very important parameter especially when brief stimuli are used.Three different possibilities e xist: (a) Rarefaction (b) Condensation (c) Alternating When a rarefaction stimulus is applied the first part of the stimulus will give an outward movement of the tympanic membrane and thereby an upward movement of the basilar membrane in the cochlea However, when an acoustic condensation stimulus is used, then the first part of the stimulus will force the eardrum to move inward and the basilar membrane will move downward. The alternating polarity is a mixture of rarefaction and condensation stimulation.This stimulation is commonly used to eliminate e.g. stimulus artifacts and cochlear microphonics (figure 4.5 ). 24 r- '« r« r.n c ? ic r cdnd*rtttlic (a) CONDENSATION CLICK E n r r m l *ud«o rv c jr u l d M O "rm *rn lounojid dhpla«nMM) 1 T ym o»i»e himhE m *"* r , ■ \ J im *d *4 *v d iip litr d » I " " \\ / RAREFACTION c lick HPidpfcon* dipo*»j*m (diwvd dhpU(«rv*r<J T rrtiO IK m # ri> O iA h e dtlftUctd’ (b) Figure 4.5. a).Condensation and Rarefaction stimuli.b).Physical properties of stimuli. II Inter stimulus interval (ISI) The inter stimulus interval is the time between successive stimuli. Often, however.the term repetition rate is used, which can be defined as the average 25 number of stimuli per second.An ISI o f , say, 40 msec corresponds for a brief stimulus to a repetition rate of approximately 0 25/sec. Ill Sound level The sound level of a stimulus can be expressed in two different types of units : a) Physical units b), Psychoacoustic units When long-lasting acoustic stimuli (>100 msec) are applied the physical unit dB SPL (sound pressure level) is used DB SPL is the sound pressure in dB relative to 20 iiPa . For psychoacoustic purposes the unit, dB HL (hearing level), is used to indicate the sound level in dB relative to the hearing thresholds in a population of normally hearing subjects. When short acoustic signals (<100 msec ) are used , no standards exist.Normally, however, dB p.e SPL -peak equivalent sound pressure level - is applied If we take a click with a level of 100 dB p.e.SPL , then the peak-to- peak sound pressure of the click will be equivalent to that of a pure tone with a sound pressure level of 100 dB SPL 26 4.3.AMPLIFIER In the recordings of auditory evoked potentials two kind of signals are measured: a desired (the target signal ) and the interference signals, which mostly consist of 50-60 Hz activity from external sources coupled electro statically to - or induced magnetically in - the amplifier. As stated above , the main purpose of the amplifier is first of all to amplify the weak electrode signals but also to suppers the interfering signals.lt is always designed as a differential amplifier with high input impedance, low internal noise and a high common mode rejection. A differential amplifier (figure 4.6a) has 3 input terminals which are connected to 3 electrodes.Two of these are connected to the non-inverting input terminal (+) and the inverting input terminal (-) respectively, while the third electrode is connected to the ground terminal. A differential amplifier measures the difference between the signals that appear at the (+) and (-) terminal. Signals that appear at the two inputs in the same phase and with identical amplitude - so-called common mode signals - will be suppressed by the differential amplifier. The ability of the amplifier to suppress common mode signals is called the common mode rejection ratio (CMRR) and is a very important property of the 27 differential amplifier (figure 4.6b and figure 4.7) The CMRR is normally given in dB and in modern amplifiers it amounts to around 8 0 - 120 dB. DMmruMJ-AmpMier ( + ) 0 Output 0 o- X A M o (a) Output Signal (b) Figure 4.6.a) Differential amplifier b) Suppression of common mode signals This means, that if our measurements are disturbed by e g 50 or 60 Hz signals from power lines, then theoretically this interference is reduced 10.000 - 1.000.000 times by the differential amplifier. It is important that no extra noise is added to the weak electrode signals and , therefore the (biological) preamplifier should have a low internal noise The input impedance of the amplifier is also of great importance.lt should be first of all so high that it does not present any load to the electrodes. 28 Secondly ,it should be high enough to take advantage of the amplifier's common mode rejection, which will be reduced by a combination of low impedance and differences in the electrode impedances The gain of the amplifier is defined as the ratio between the output signal and the input signal. 1 A ME ’ Q l A R i r Y / V / in « tuf 1 O U T g M Q S lT f P O lA d lT T J " 1 - = 0 , 4 M » i _ L J ' a ^ p o tirc p T P O L A R IT ? ! --------- - 1 : O U T n O U T n u Figure 4.7. Principle of operation of a differential amplifier The differential amplifier substracts voltages applied at its two inputs So the common parts of the two voltages are cancelled out , as the “differential” signals are passed through and amplified 29 4.4.FILTERS Most biological signals do not appear as clean signals but are more or less embedded in noise.Therefore it is necessary to include an electronic filter to eliminate the frequency components which do not correlate with the signal to be measured In more technical terms , one could also say that the purpose of a filter is to improve the SNR An electronic filter is specified by different parameters: The filter slope or roll off refers to the rate at which a filter attenuates the frequencies beyond/below the its cut-off frequency (fc)(figure 4.8) and is measured in db per octave or db per decade Band Pass Filter Attenuation (dB) A ’ / 3d8 3dB 20 F 100 50 500 1000 2000 Hz Log Frequency Figure 4.8: Bandpass filter. 30 The cut-off frequency (fc) is defined as the point where the filter has attenuated the signal by 3 dB. If one wants to exclude the low frequency content of a signal, then a high-pass filter (HP) is applied. However, if it is the high frequencies which disturb the recordings then a low-pass filter (LP) should be used. If a low -pass and a high-pass filter are combined , then the result will be a so called band-pass filter (BP) which is a commonly used filter in auditory response recordings. A typical band -pass filter for recording the auditory brainstem responses would have the low cut-off frequency set to 100 Hz and the upper cut-off frequency to 3.000 Hz (that is a band width of 2 900 Hz , as measured between the cut-off frequencies ) and with slopes in the range of 12-24 dB/octane(figure 4 9). H & i A U m A Q on <dS) te w * P a n Ffflwr *nwnuaoan (Ofil Low P u t P iitf AMnutfon (dB) F Figure 4.9. Other type of filters. 31 4.5. A - D CONVERTER The A-D converter samples the signal from the low level analog section at equal intervals Sampling begins when initiated by a timing pulse that is time locked to the click stimulus, and continuous for a preset period of time (the epoch 10-15 msec). At each sample point the A-D converter converts the signal into a number.The output of the A-D converter actually consists of binary numbers The amplitude resolution of the digitized waveform is defined as the number of discrete amplitude steps (bit resolution ) available to represent signal voltage In specifying amplitude resolution requirements for ABR’s one must consider the fact that the background noise is typically 10 times that of the ABR ; hence only about 10% of the A-D converter amplitude resolution is available to resolve the ABR waveform. Generally the larger the A-D converter amplitude resolution, the better or measurements are(figure 4 10). 32 TmTfTTTTTFrntTTTlTTTTT . iT u o e TTTTrnTTTTTT 1M K I M M M T t U M n l « IN T S (a) (b) Figure 4.10. Effect of the amplitude resolution on the fidelity of waveform reproduction.The records are outputs of hypothetical D - A converters, a) An A - D converter with only 2-bit word length, allowing it to resolve amplitude into only 4 steps, b). An A - D converter capable of resolving amplitude into 32 steps (The amplitude limits of the A - D converter are also shown) Also the sampling rate is very important, since it must be : fs> 2*fm a * where : fs (=1/Ts ) = sampling rate, f = maximum frequency component of the sampled signal Also the bigger the sampling rate the more accurate the reconstruction of the analog signal from the digital will be assuming of course that we have enough bit resolution (figure 4.11.). 33 r • ^ m r n f f m r r r m m T f T T # tTTTTTTTTTTTT I l ] I < I ( I • » « I I Q (a) (b) Figure 4.11. Effect of A - D time resolution (sample rate) on the fidelity of the waveform reproduction. 4.6. AVERAGING TECHNIQUE When auditory evoked potentials are recorded from the human scalp they will always be accompanied by interfering signals In more simple terms ,one could say that the activity picked up by the the electrodes is a mixture of a target' signal (the evoked potential) and signals from other generators -the background noise We wouldn't have any problem if the evoked potential was much larger than the background noise. But that's not true and to make things worst the background noise is greater than the evoked potential by a factor of 10 or more. 34 Making the assumption that the evoked potential and the background noise don't have any mutual interference the signal picked by the electrodes can be written as; S=EP + BN In ABR where the evoked potentials are measured with surface electrodes in afar field situation some typical values are: EP=0 3 uV (rms) BN= 3 pV (rms) So the SNR will be; SNR=(EP)2 /(BN)2 =0.01 Most often the SNR is given in dB : SNR(dB) = 10 * log( SNR) = -20 dB So in our example the signal is 20 dB below the noise level A very effective way to improve our SNR is to use signal averaging,Synchronously with the acoustic stimulus ,the computer begins to sample the amplified activity from the electrodes. The sampling takes place at some discrete points distributed over a certain time period (10-15 msec ).Such a measurement is called a sweep and the time frame is designated the sweep length.The sampled values are stored at adjacent locations in the memory of the computer and the whole procedure is repeated. 35 After N repetitions (figure 4 12 ) the computer memory will hold the sum of N sweeps.By dividing the values in each memory cell by N ,the average is obtained There are two basic assumptions that allow us to use the averaging technique: 1 The evoked potential is invariable (constant shape and amplitude) and is elicited at exactly the same time during each sweep 2 The background noise is assumed to be zero mean random noise and that its characteristics don't change during the recording session That is normally fulfilled if the patient remains relaxed during the investigation. S - EP -f- BN E ? evoked potential BN avg. background noise 200 nV Figure 4.12. The avaraged signal, evoked potential and background noise 36 So after the averaging S a v g = EP/ BN a v B As a rule of thumb, after N sweeps, (figure 4.13 ) the reduction will amount the square root of N: BN a v fl = BN / VN No. of 250 500 1000 2000 4000 8000 I 400nV -i 6 a to 12 14 16 ms Figure 4.13. Averaging of ABR’s 37 So the improvement in the SNR will be SNRa v g = SNR*N SNRa v g ,d B , = 10 * log(SNRa v fl )= 10 * log{SNR * N)= =SNR(d B t +10*log(N) Except for the standard averaging most often we are using the Bauessian averaging which is a weighted averaging which assigns small weights to the most noise sweeps By this way we improve the SNR even more, than we do with the standard averaging There is also another facility called the artifact rejection that we use in order to have clearer and more accurate measurements By this method the amplitude of the incoming signal is measured and if it exceeds a certain preset level { e g patient moves and the myogenic activity produces noise sweeps) the sweep will be rejected because otherwise would corrupt the useful information already collected. 38 5.STUDY Over the past twenty years there have been numerous studies investigating the effect of stimulus polarity on the latency of wave V of the ABR. As a whole, the results have varied widely and will not be reviewed here. Reviews of these variable findings can be found in [23],[15],[14],[22], Part of the confusion in understanding and interpreting the results of polarity studies can be attributed to differences in the types of stimuli (i.e. clicks and tonebursts) used and the hearing status of the subjects (normal and hearing impaired) In order to sort out the results of all these studies, we need to view them with a theoretical framework that is based on our knowledge of the cochlear process that leads to neural activation and the representation of that neural activation in evoked potentials recorded from the surface of the head. 39 5.A. POLARITY EFFECTS. I Clicks and Transient Stimuli. The work of another investigator [6] .demonstrated that for a complex stimulus, firing of eighth nerve fibers occur when the basilar membrane is displaced upwards, or rarefaction phase, and not downward or condensation phase(figure 5.1). N e ^ a tlv a d e i t r teel p u lM Pooitlve •leetri p u la * Middle ear A ceuvtlc waveform Car ewei Cochlea Ractlcular lamina It baalar mom brara H er ceil footplate A ctlvotlor Condensation wavwe -7 * )))).; Activation Figure 5.1. Schematic representation of events produced by rarefaction versus condensation click stimulus polarities. 40 Thus, it is expected that the latency of response to a click stimulus produce by applying a voltage pulse whose initial phase is rarefaction will demonstrate shorter latencies than when the initial phase is condensation How short this latency will be depends on the where in the cochlea the nerve fiber is located which is revealed by its characteristic frequency (CF) Kiang [16] showed quite clearly with click stimuli that temporal separation of peaks in the post-stimulus time histograms (PSTHs) was equivalent to the reciprocal of the fiber's CF, particularly for low CF fibers Moreover, the peaks shifted by half the period when the polarity of the click was reversed. The relative heights of the peaks varied with stimulus level. Thus, at the single unit level, the latency shift with polarity is observed and is equivalent to the half period of the CF Given that fibers in the same region of the cochlea have similar PSTHs, then the latency of the largest peak in the PSTH might be closely related to the latency of Wave I. Because wave I has a relatively constant temporal relationship to wave V [9], it would seem reasonable that polarity effects might be seen in the latencies of wave V 41 However, frequently, in addition to the half-period latency shift of the peaks, the relative magnitudes of the peaks in the PSTH changed with polarity Thus, it is possible that despite shifts in the latency of the peaks of the PSTH, the peak of the envelope may not. Importantly, whether or not a half-period latency shift is detected depends, to a large degree, on which frequency region dominates the ABR If the response is dominated by contributions from the high frequency regions, as is the case with broad-band high level signals, then it may be difficult to detect half-period changes. Therefore, the results of the many studies using click stimuli in normal hearing adults demonstrating no polarity effects (e g [21], [2],[23]) are not surprising. We will also argue later that the use of unmasked clicks to study polarity may be inappropriate. II Tonebursts Stimuli. Recognizing that high-frequency stimulation or click stimulation is dominated by high-frequency fibers and half period shifts would be difficult to detect, some investigations used low frequency tonebursts [15],[14] or single cycle tones [18] 42 Although these studies demonstrated latency shifts with polarity changes, these results are to be expected because the stimulus itself is driving the delay. With click stimuli, the polarity effect is a result of the difference in time that a given frequency regions moves upward initiating neural activation. This time difference is related to response characteristics of that segment of the cochlea and not to the differences of the driving stimuli, which for clicks is often only 0.1 ms difference Even so, fibers with CFs of 500 Hz will show a 1.0 ms (half-period) shift to a reversal of the click polarity. For low frequency tones, the differences in the stimulating phase is delayed by a half-period. Thus, the onset of stimulation is delayed and one expects to see the polarity effect but it is related to the onset of the stimulating rarefaction phase rather than to its own natural CF characteristics. Units tuned higher than 500 Hz will show delays related to the stimulus as oppose to its half-period CF Since it is already known that onset of neural activation is related to the upward displacement of the basilar membrane, latency shifts owing to the delays of the rarefaction phase in the stimulus is to be expected and does not shed light on the issue as to whether latency differences related to onset of natural responses initiated from a given segment rather than temporal differences in stimulating phase "onset" can be demonstrated. 43 Thus, one should not mix the results of studies using low frequency tonebursts with those using transient click stimuli for the expectation and mechanisms are different III Hearing Loss and Absolute Latency Differences Some investigators [8],[4] have demonstrated that polarity effects are observed in patients with hearing loss. Although in [8] did not necessarily see a dependence on high-frequency loss, [4] suggested that the slope of the high-frequency loss was an important predictor. As suggested in [22] the selective loss of cochlear high frequency channels may be responsible for the increased dependence of the ABR on click polarity seen in patients with high-frequency hearing loss Two previous studies [8],[5] also found that wave V responses to rarefaction stimuli do not always precede those to condensation stimuli This inconsistency, which varies from subject to subject, led them to analyzing the absolute difference between condensation and rarefaction stimulation as a more revealing indicator of the latency shift produced by changing stimulus polarity. The use of absolute differences were applied in these studies demonstrating a polarity effect in patients with hearing loss. 44 Apparently, it may be necessary to record the absolute values because other studies not using absolute values (e.g. [23]) could not demonstrate a polarity effect on wave V in hearing impaired patients The contradictory studies may be the result of different analyses (absolute vs signed differences) or the use of unmasked clicks and the unpredictability of the expected latency shift with click polarity inversion 5.B..Derived-Band ABRs and High-pass Masking To answer the issue of problems with the use of click stimuli, another investigator [22] studied polarity effects on the derived-band latencies of normal hearing subjects. The derived-band approach seeks to eliminate contributions of the cochlea above and below a specified frequency band The derived ABR technique [9] yields narrow-band contributions to the compound ABR through successive subtraction of waveforms obtained with successive highpass cutoff masking noise conditions. This is a more appropriate way of investigating the polarity effect as the response will reflect a given frequency region and not the sum total having varying contributions in amplitude and time from different frequency regions The expectation is that with lower CF derived bands, the wave V latency differences for condensation-rarefaction clicks should become greater in 45 accordance with the half-period of the CF. Another study [22] concluded that even though he found "substantial latency differences" for individual subjects, he did not find polarity effects for group data. Thus, perhaps, it is necessary to use absolute latency differences to show the polarity effect Another study [24] evaluated polarity effects with stimulus and response analyses using highpass masking with normal hearing subjects and showed a good correlation between absolute polarity latency differences and the half-period of the frequency of stimulation. However, as we will demonstrate in this paper, the use of absolute latency differences to assess polarity effects may be inappropriate and that polarity differences in wave V from click stimulation may not be revealed by peak latency measures. The goal of this study was to define stimulus polarity effects on the derived band ABRs of normals in order to understand the variable effects observed in the past At this point it is necessary to give a brief introduction to Derived-Band ABRs and High-pass Masking technique that was introduced dy Don and Eggermont. 46 High -Pass masking Technique and Derived-Band ABR s It has been demonstrated that the whole of the basilar membrane contributes to the brain stem response to a broad-frequency click [9],[19].The technique of deriving the contribution initiated from each portion of the basilar membrane is illustrated in figure 5.2 In this figure the cochlea is rolled out flat and marked off in sections A through F Section A represents the area of the cochlea whose maximum sensitivity is 8 kHz and above; section B , from 4 to 8 kHz; section C from 2 to 4 kHz ; section D, from 1 to 2 kHz ; section Et from 0.5 to 1 kHz ;and section F, the region below 500 Hz A click stimulus presented at moderate hearing levels and above stimulates the entire cochlea because of its broad-band spectral nature The brainstem response R+1 (line 1 of figure 5.2) represents the sum of brainstem activity initiated by stimulation of the whole cochlea (i.e. from section A to F ) Next , as seen in line 2 , the level of continuos broad-band noise that is sufficient to desynchronize and thereby obliterate the response to the click is determined. This masked activity is denoted as MR After the appropriate noise level has been determined, the noise is steeply high-pass-filtered at 8 kHz (the high-frequency component of the 47 noise above 8 kHz is allowed to pass) , and the clicks are presented in this noise As seen in line 3 of fig ure 5 2 , the brainstem response (R+2) obtained under these conditions results from click-synchronous activity initiated from the unmasked region below 8 kHz The subtraction of this response (R+2) from the response obtained without any masking noise (R+1) in the computer results in the derived narrow band response ,DR+1, seen in line 4. This subtraction procedure eliminates the common contributions from regions below 8kHz (stippled area in line 4 ) and results in the contribution from the cochlea that was masked by the 8 kHz high-pass noise (section A ). Next the high-pass cut-off of the noise is lowered by an octane to 4 kHz , and the clicks are presented in this noise.The brainstem response recorded (R+3) shown in line 5 of figure 5.2 , results from click-synchronous activity from the unmasked portion of the cochlea, i.e.,the region below 4 kHz Subtraction of the response (R+3) from that obtained with the 8 kHz high- pass noise (R+2) eliminates the common contribution from the region below 4 kHz (stippled area,line 6 ). The response derived from this subtraction (DR+2) is initiated from the narrow band region of the cochlea that is not masked by 8 kHz high-pass noise but was masked by the 4 kHz high-pass noise (section B ). 48 (A+B+C+D+ E+F) Masked (B+C+D+E+ F) R1-R2=A (C+D+E+F) R2-R3=B Figure 5.2. Diagram illustrating high pass masking technique In similar fashion, by successive subtraction of the response, one obtains the derived narrow band contribution to the brainstem response for the other sections of the cochlea.This procedure is repeated for different click intensities, and in this manner the contribution from each portion of the basilar membrane at each intensity is derived. Click Click and Broadband noise Click and 8kHz HP noise Click and 4kHz HP noise B -c 7 ) r. f / £ © ■ % 7 \ B 3 MR R2 - A -C - C D R 3 B - E - - F ^ ) or2 49 In patients with normal hearing , contributions to the brainstem response to the click can be detected down to the 30 db sensation level for the 8 kHz and above region and 500 hz and below regions of the cochlea contributions to the brainstem response from 4,2, and 1 kHz octave-wide regions can be detected down to at least the 10 db sensation level The differences in the threshold levels between a patient and those from normal hearing individuals for each of the derived frequency regions are used as the estimate of the hearing loss for constructing an estimate of the audiogram. The major advantage of this technique is that it provides information about the place in the cochlea that the losses occur, which is not always the case for techniques using tonal stimuli of moderate sound levels. The major disadvantages are that it requires a little more time than most other procedures, special filters for the masking noise , and equipment capable of storing and subtracting waveforms However, most newer equipment has that capability either hardware or software wise. These are small prices to pay for a technique that can often accurately assess peripheral hearing function in the very young or otherwise difficult-to-test patient 50 6.METH0DS 6.A. INITIAL STUDY I SUBJECTS Six females and 6 males aged 18 to 37 years served as subjects Subjects were recruited from the staff at the House Ear Institute, House Ear Clinic and from the student bodies of local colleges and universities. All subjects were in good general health and reported normal neurological status. Otoscopic examinations were performed to identify existing conditions that would preclude audiometric and ABR testing Subjects had normal hearing as defined by pure-tone thresholds at or less than 10 dB (ANSI 1969) for frequencies between 500 to 4000 Hz and less than 15 dB for 6000 and 8000 Hz at the time the ABR data were collected Pure-tone audiometric testing was accomplished with a Grason-Stadler GSI 16 audiometer and Telephonies TDH 50P earphones in P/N 10C017-1 cushions. Hearing thresholds were evaluated in 2 dB steps with use of a modified Hughson-Westlake procedure [7], 51 II. STIMULI Rarefaction and condensation click stimuli were produced by applying 100 ps rectangular voltage pulses to a TDH-50P earphone with a P/N 10C017-1 ear cushion Clicks were presented at regular intervals of 22msec and at 93 dB peak-peak equivalent sound pressure level (p -p.e SPL) with a 1 kHz tone as the reference The transducer clicks were calibrated and measured with a Bruel and Kjaer 4152 artificial ear, 6 cc coupler, and 2209 sound level meter. Perceptual detection thresholds were determined for 1 second bursts of clicks presented at the same inter-stimulus interval (ISI) used in recording the ABRs This perceptual threshold was defined as the 79% point on the psychometric detection function obtained in a modified block up-down procedure [25] For the group, 93 dB p -p.e. SPL was just under 65 dB above the average perceptual detection threshold. Ipsilateral white noise masking was used for obtaining derived ABRs [9] The noise was produced by a General Radio Noise Generator (Type 1310) and presented at a level sufficient to mask the ABR to the clicks. For the 83 dB p.e. SPL clicks, the required broad-band white-noise RMS level was 92 dB SPL. The noise was attenuated with the click, thereby 52 maintaining a fixed click-to-noise masking ratio There were 6 stimulus conditions: clicks presented alone (unmasked condition), and clicks presented with ipsilateral noise high-pass filtered at 8, 4, 2, 1, and 0 5 kHz with a slope of 96 dB/octave The high-pass filtering of the masking noise was achieved by cascading both channels of a Krohnhite (Model 3343) dual filter, each with a 48 dB/octave slope All subjects reported that the clicks and the masking noise were loud but tolerable Previous work [11] has shown little effect of adaptation induced by the level of noise exposure for this protocol III ABR RECORDINGS Subjects were placed in a reclining chair in a sound-treated, double walled sound room (IAC) ABRs were obtained by recording differentially between electrodes applied to the vertex (Cz) and the ipsilateral mastoid (M1 or M2). The contralateral mastoid was used as ground. This scalp activity was amplified by 5 X 10^ and filtered with a bandpass of 0.1 to 3 kHz The activity was sampled at a rate of 20 kHz for 15 milliseconds after stimulus onset with use of an Ariel DSP-16 A/D-D/A board and an NEC 386 computer. 53 After each block of 256 sweeps, the RMS value of the averaged background noise was estimated according to procedures of [9], The noise estimate was used in a Bayesian estimation technique [13],[12] to form a weighted average This technique reduces the destructive effects of noise variation on the ABR average by weighting the average towards those blocks of sweeps with low background noise Data collection for each run was terminated when the estimated averaged background noise reached 20 nV or less. Background noise level was used as the stopping criterion rather than a fixed number of sweeps to reduce the effect of physiological background noise variation on the interpretation of the ABR recordings Thus, all recordings had approximately the same estimated low background noise levels. 6.B.FINAL STUDY In the final study, the above protocol was administered repeatedly at 83 dB p.-p.e. SPL to 3 additional subjects (two males and one female). For two of these subjects, the protocol was repeated 6 times and for the third, 7 times The repeated testing was necessary to assess the variability of the response measures for each polarity of stimulation. 54 Representative unmasked and derived band response waveforms with the latency of wave V marked are shown for one subject in figure 6 1 As demonstrated by early studies [9],[19] ,the latencies of the peak of wave V increased as the theoretical center frequency of the derived band decreased. The theoretical center frequency is estimated to be the square root of the product of the two high-pass cut-off frequencies used to form the derived band Both the signed and the absolute differences in wave V latency between condensation and rarefaction click responses, (C-R) and |C-R| respectively, were measured. At this point it necessary to talk about the “peak picking " procedure ,a procedure of great importance for correct interpretation of the measurements 55 Condensation Click Rarefaction Click 83 dB p. • p. SPL 6.4 Unmasked 5.95 11.3 kHz 6.25 6.8 7.55 0.7 kHz 0 2 4 6 8 10 m s 6.3 Unmasked 11.3 kHz 6.3 5.7 kHz 6.9 2.8 kHz 7.85 0.7 kHz 4 6 8 10 12 14 ms Figure 6.1. Derived Narrow - Band ABR s to condensation and rarefaction clicks.The unmasked,derived band and 0 5 kHz HPM responses are shown 56 6.C.THE ART OF “PEAK PICKING”. An important skill is the ability to “pick the peak “ That is, the ability to select , consistently and accurately, the single representative data point on a waveform that will be used in labeling the wave and in calculating latency and amplitude values There are two fundamental approaches to this type of wave analysis ,as illustrated in figure 6.2. A One is to select as the peak the point on the wave component that produces the greatest amplitude In waveforms with sharply peaked components , this selection is simple and unequivocal (waveforms in graph A in figure 6.2). Although intuitively appealing , this approach can present analysis problems. One problem occurs when the point of greatest amplitude clearly does not best represent the wave. Perhaps the most frequent example of this limitation , even in normal subjects , is found with patterns of the wave IVA/ complex that do not have two actual peaks,With a prominent wave IV and a relatively minor wave V pattern , selecting the maximum amplitude as the peak essentially substitutes wave IV latency for wave V latency 57 The clinical consequences of this type of waveform misinterpretation would include calculation of an unusually short latency difference for wave V and the wave l-V latency interval , and possibly the presumption that the nonsuspect ear is abnormal. Another problem with defining peaks on the basis of maximum amplitude arises when the top portion of the wave is rounded or even a plateau, rather than sharply peaked This morphology may occur spontaneously , or it may be the result of a restricted low-pass filter setting.An apparent solution to this problem is to take as the peak the point at which lines extended from the two slopes of the wave intersect, as shown in figure 6.2 . Several disadvantages of the technique are evident First, the point of intersection of the two lines does not correspond to an actual peak Also , slight variations in either the leading or the following slope may produce important variations in the arbitrary defined peak B The second fundamental “peak picking “approach is to select the final data point on the waveform before the negative slope that follows the wave.This point may be the final peak , or a plateau or shoulder in the downward slope. This technique virtually eliminates the incorrect selection of wave IV versus V, but it introduces its own problems Some waves have multiple 58 shoulders on the downward slope , caused by background activity Other waves have shoulders that are extremely subtle and ill defined.The initial solution to these intricacies in wave morphology .again, is to adhere to consistent analyses criteria ABR LATENCY P aok S hieuldar 1.60 ------ 2.70 ------ 3.80 ------ 3 8 0 5 75 1.10 ------- 2.20 ------ 1.80 1 .9 ! +.00 *.15 1 — Y r-a i-m m-'f l-V i_n AMPLITUCc a. SN10 - 0 .5 a 1.0 2.0 3.0 +.0 3.0 > 0 7.0 8 .0 B.Q Latency in milliseconds Figure 6.2.Analysis of ABR.For waveforms (top), latency values are determined for the peaks of waves I, II and the shoulder of wave V. For the waveform (lower), peaks are determined for all waves, Wheather the peak or the shoulder is selected in analysis may have an important effect on the latency calculation. 59 There is one further complication in ABR peak picking A set of criteria may be legitimately used in analysis of some wave components within the waveform , but not for others. Thus ,as seen in figure 6.2, the maximum amplitude (mid) point may be selected as the wave peak for certain components ,such as wave I and III , while the shoulder is selected for other components Within a laboratory or clinical facility , this apparent complexity , confusion, and uncertainty in ABR interpretation can be minimized by specifying which of these two fundamental waveform analysis approaches- wave peaks or shoulders- must be applied with each major wave component by all persons interpreting waveforms. In our study in order to minimize reading bias, the latency of wave V was simply the time sampled point whose digitized value was greatest for the peak waveform component identified as wave V 60 7.RESULTS Figure 7.1 shows plots separately for males and females of the wave V latencies for the unmasked and derived band ABR responses to condensation and rarefaction clicks. Two points are immediately obvious from this graph: 1). The latency differential between the highest and lowest CF bands is greater for males than females as previously described in detail by our earlier work [11], 2) There appears to be no effect of polarity on peak latencies for any of the derived bands An analysis of variance (ANOVA) was performed and, as shown in Table I, no effect of click polarity on wave V latency was found for the unmasked or any of the derived bands. Derived Band CF F - Value P - Value Unmasked 113 kHz 5.7 kHz 2 8 kHz 14 kHz 0.7 kHz .511 365 .097 .030 .056 .020 482 .552 .759 .864 815 888 Table I. ANOVA results testing for polarity effects on wave V latency. 6 1 11 — O— Cond-Wave V: Male — Cond-Wave V: Female — Rare-Wave V: Male — Rare-Wave V: Female 1 0 9 8 7 6 5 Unmasked 11.3 5.7 2.8 1.4 0.7 Figure 7.1. Males and females wave V latencies for the unmasked and derived band ABR responses to condensation and rarefaction clicks Note the gender effect. This lack of polarity effect is shown graphically in figure 7 2 where the wave V latency difference, (C-R), is plotted (filled squares) for the unmasked and derived band responses. The mean values for the (C-R) latency difference for all bands fall within ±0 1 ms and do not follow the theoretical latency difference, based on the half-period of the CF of the derived band, curve (thick solid line) 62 Also plotted in figure 7.2 are the means of the absolute differences, |C -R|, (open diamonds) for each of the derived bands. The means of the absolute difference appear to follow the theoretical difference V ) S a * u a S i .8 Half period C-R mean -m - C-R s.d. - 0 “ Absolute C-R mean .7 6 .5 A .3 2 4 .1 0 1 Unmasked 11.3 5.7 2.8 1.4 Derived Band CF 0.7 Figure 7.2. C - R signed and absolute latency difference compared to theoretical half-period of CF for derived band ABR s Also shown is the s.d. for the signed values. As seen by the plot of the standard deviations (x’s), the lower the frequency, the greater the standard deviation. This would be expected since the half-period increases as CF frequency decreases. However, it is also possible that peak latency of the lower CF derived responses are more variable and accounts for the increase in the absolute values for lower Cfs. 63 To ascertain if the increase in the absolute latency differences for the lower CFs is due to polarity effects or variability of the measurement, 3 additional subjects were tested repeatedly as described in the final study Figure 7.3 shows for one individual, six superimposed runs of the unmasked and derived band responses Condensation Rarefaction 83 dB p.-p. SPL Unmasked 11.3 kHz 5.7 kHz 2.8 kHz 1.4 kHz 0.7 kHz i ' i 1 i '"' i 1 i 1 i "*' i i 1 0 2 4 6 8 10 12 14 ms l 1 I ' l ' l ’ l 1 I l l » 0 2 4 6 8 10 12 14 m s Figure 7.3. Six superimposed runs of the unmasked and derived band responses for one individual.Note the increase variability of wave V latency peaks for lower CF derived bands. 64 Typically, the bands showing the most variability in peak latency are the 113, 14 and 0.7 CF bands. This variability is reflected in the distribution of wave V latency differences between runs of the same polarity, i.e. C-C or R-R shown in Figure 7.4.a for each of the derived bands. The mean difference is, as expected nearly zero. Since the theoretical half-period for the 113 kHz band is so small (less than the 50 psec sampling period of the 20 kHz digitizing rate) and the variability was relatively high for repeated runs of the same polarity, this band will be eliminated in the subsequent analyses of polarity effects. The standard deviations increase as the CF of the derived band decreases Thus, there appears to be an increasing variability in reading the latency of lower derived bands that is not related to polarity differences. For comparison, figure 7 4b shows the distribution of differences between condensation and rarefaction runs wave V latencies for all the possible combination of runs for the three subjects. Again, the same pattern of an increasing variability for decreasing derived band CF is observed. However, the variability as indicated by the standard deviation and the nearly zero mean are very similar to that for differences between runs of the same polarity (figure 7.4a). Another graphical representation of the data in figure 7.4 is shown in figure 7 5 where the mean signed latency differences are plotted in comparison to the expected theoretical half-period differences 65 Distributions of wave latency differences for multiple runs. sor 40 J £ 30 £ 20 “■ 1 0 1 2 5 i e 20 i 4 1 a is £ 10 0 - 25 s 2 0 a is £ 10 5 C £201 a is £ 101 51 0 25 S 20 £ 15 1 £ io 5 0 C-C , R -R Unmasked 1 X - .017 s.d. - .172 C F-11.3 kHz X - .013 s.d. - .403 CF-S.7 kHz CF-2.8 kHz A A X - -.034 s.d. - .204 X - .027 s.d. - .223 CF-1.4 kHz j k X - .075 s.d. - .303 CF-0.7 kHz X - -.003 f s.d. - .422 t -2.5 -2 -1.5-1 -.5 0 .5 A m s C-R 40 £ 3 0 S 20 1 0 0- 25 - 20 S is 1 0 Unmasked X - .019 S.d. - .192 CF-11.3 kHz X - -.094 s.d. - .393 o 25 c 20 a is £ io 5- (T CF-5.7 kHz 2 5 CF-2.8 kHz g 201 1 0 1 5 » 25- c 20- £ * * £ io- 5 25 £20 a is £ ioi X - -.048 s.d. - .210 X - -.109 s.cL - .212 CF-1.4 kHz J L X - -.083 s.d. - .296 CF-0.7 kHz X - -.035 s.d. - .539 -2.5 -2 -1.5 -1 -.5 0 .5 1 1.5 2 2.5 A m s (a) (b) Figure 7.4. Distribution of wave V latency differences between runs of (a) the same polarity , i.e. C-C or R-R , in comparison to (b) different polarity, i e.,C- R Note the increase variability of wave V latency peaks for lower CF derived bands. 66 If there was a consistent polarity effect (e g. condensation always lagging rarefaction) then the mean value would not be nearly zero for C-R and would increase for lower CFs This is illustrated in figure 7.6 where a consistent polarity difference based on the theoretical half-period of the derived band CF is added to C-C differences. By using C-C differences, normal run-to run sampling variability is included along with the expected polarity effect The mean values of this simulated polarity effect (C-C + theo) are plotted as open circles in figure 7.6 The mean values follow the simple theoretical differences very closely as expected since the mean C-C differences are nearly zero for all derived bands The possibility that polarity latency differences are washed out in group data because the responses to condensation clicks may be longer in comparison to rarefaction clicks in some individuals but shorter in others, has led to the suggestion that mean absolute d iffe re n ce s may b e m ore revealin g as observed in figure 7.2 for the initial study. Figure 7.7 plots the mean absolute differences for the C-C, R-R, and C-R data of study 2 Although there is a trend toward higher mean absolute values, again, the differences are less than the theoretically expected half period values. 67 Difference i n ms .8 Theo Half Period Mean (C-R) Mean (C-C) Mean (R-R) .7 .6 .5 4 .3 .2 .1 0 .1 -.2 -.3 Unmasked 5.7 0.7 2.8 1.4 Derived Band CF in kHz Figure 7.5. Mean signed wave V latency differences for C-C , R-R and C-R are plotted in comparison to the expected theoritical half-period differences. 68 < L > a C ,< u s Q .8 Theo Half Period Mean (C-R) Mean (C-C+theo) .7 .6 .5 .4 .3 .2 1 ■ 0 -.1 -.2 -.3 Unmasked 5.7 0.7 2.8 1.4 Derived Band CF in kHz Figure 7.6. A consistent polarity difference based on the theoritical half period of the derived band CF is added to C-C differences, ( I C-C + theo I), and plotted against the signed (C-R) and the theoritical half-period differences 69 Difference i n ms Theo Half Period Mean IC -R J Mean IC-Cl Mean IR -R I -.2 -.3 Unmasked 5.7 0.7 2.8 1.4 Derived Band CF in kHz Figure 7.7. Mean absolute wave V latency differences for C-C, R-R and C-R are plotted in comparison to the expected theoritical half-period differences 70 Difference i n ms .8 Theo Half Period Mean IC -R I Mean IC-C + theo I .7 .6 .5 .4 .3 .2 .1 0 .1 -.2 -.3 0.7 Unmasked 2.8 1.4 5.7 Derived Band CF in kHz Figure 7.8. The absolute differences for the simulated data ( I C-C + theo I), and the IC-R I are compared to the theoritical half-period differences 71 Furthermore, if the mean absolute differences for the simulated data (jC-C + theof) are plotted (filled circles in figure 7.8), they also follow closely the theoretical values as well The fact that the means of the absolute C-R differences (filled squares) increase with decreasing derived band CF but are still less than the theoretical values, suggests that the effect is not one of polarity but instead is mostly related to the greater variability. To examine this possibility, the relationship of absolute values to variability will now be derived. For this derivation, we will hypothesize that there is no polarity effect such that the distribution formed by differences in peak latencies between condensation and rarefaction click responses has a mean of zero and is normally distributed. As a starting point and by way of review we state the general case. If x represents the difference in latency between condensation and rarefaction, the mean value of x, px is given by [3], + t c M x, - f xp{ x)d Eq (1) - X where p(x) is the probability density function for x. 12 In the same way, the mean value of the absolute latency difference, |x|, may be expressed as + o c M|x|. = J \Ap(x)d Eq (2) If the probability density function is a normal distribution with a mean of 0, then p(x) 1 a Eq (3) [3] and substituting in Eq. (2), 1 M |x| = V2/r V2a 2 } Eq (4) Because both |x| and p(x) are symmetrical about x=0, the integral over all x values may be expressed as two times the integral over just the positive x values, as expressed in Eq. (5) and seen schematically in Figure 7 9. Area « 1 - o + Signed values Area - 1 o + Absolute values Figure 7.9. Schematic representation of the resulting distribution when absolute values are taken from a normal distribution with mean=0 73 Eq. (5) The latter integral may be readily evaluated: Eq. (6) Thus, the mean absolute latency difference between values randomly selected from a normal distribution with mean = 0 and standard deviation = s, is proportional to that s. If absolute differences are calculated between values randomly taken from two normal distributions with ss equal to si and S2, the proportionality will be to the compound standard deviation, sc, of the difference distribution, i.e. Thus, plotting the |C-R| (absolute) values against Sc will result theoretically (under the assumption that the distribution is normal with zero mean) in a straight line with a slope of nearly 0.8. For latency differences from runs of the same polarity (i.e. |C-C| and |R-R|) where the mean difference should be zero, such plots as shown in Figures 7 10a and 7,10b have slopes around 0.8 (0.810 and 0.835 respectively). Eq (7) 74 In Figure 7.11a, the plot is for |C-R| and again the slope is nearly 0 8 For all three plots, the r^ value is very high, demonstrating that the relationship is truly linear as suggested by Eq (6) For the |C-R| plot (Figure 7 11a), this close approximation to the theoretical 0 8 and highly linear relationship suggest that these differences come from a distribution of nearly zero mean, i.e., that there is no latency difference between condensation and rarefaction click stimulation. The increasing absolute mean latency difference for the lower derived CF bands is due to the greater variability of the measure, not to a polarity effect Data with the simulated polarity effect are shown in Figure 7.11b This plot demonstrates that if there was a polarity effect that could only be observed by the absolute values, the plot will yield a slope much greater than 0 8 and the amount of variance accounted for by a linear relationship would be poor since the mean of the absolute differences would be greatly dependent upon the half-period shift with polarity and not simply the standard deviation as shown in Eq. (6), 75 1 .9 .8 _ .7 V .6 ^ . 5 g .4 1 3 .2 Slope - .010 .1 .973 0 .8 1.2 .6 1 2 .4 0 C-Compound SD (a) -6 .999 1.2 .6 .8 0 1 .4 .2 R-Compound SD (b) Figure 7.10. Plots of the mean absolute differences for the same polarity (a) I C-C I and {b) I R-R I against the compound standard deviations of those distributions The slopes are near the value .8 suggesting these distributions have a mean of 0 76 □ e ■ o c n w 2 . 0 6 9 0 + 0 .2 . 4 .6 .8 1 1.2 C & R Compound SO (a) S l o p e - 1 . 6 7 2 _ .8 V .6 X .2 . 3 1 7 C-Compound SD (b) Figure 7.11. Plots of the mean absolute differences for the same polarity {a) I C-R I and (b) I C-C+theo I against the compound standard deviations of those distributions.The simulated polarity effect, I C-C+theo I is plotted in (b) for comparison. 77 8.Discussion In reviewing the studies of the effect of click polarity on ABRs, the main issue to settle was whether it was necessary to use absolute differences to demonstrate the polarity effect. It has been amply demonstrated by nearly all investigators that there was no consistent effect of polarity in group data from normal or hearing impaired subjects if the simple differences are recorded. This is explained by most investigators by the fact that responses to broad spectrum clicks are dominated by the high frequency regions of the cochlea and, therefore, 180_ phase shifts would be too small to detect However, in [22] demonstrated with derived bands in normals with simulated high-frequency hearing loss, that no consistent effect of polarity could be seen even though individuals often showed large differences Our results are consistent with his in that we could not demonstrate polarity effects in derived bands If there was a consistent effect, it should be greatest and observable for the low CF derived bands. 78 This lack of phase sensitivity even for the low frequency derived bands is not contrary to studies [15],[18],[14] using low frequency tonal stimuli As we argued in the introduction, the use of low frequency tones to demonstrate polarity sensitivity is somewhat artificial because it could be viewed that the delay at the output of the cochlea can be viewed as being due to the delay of the stimulus input. That is, for low frequency tones, the initial rarefaction phase at the input is delayed for condensation signals and therefore, a latency shift equivalent to the half-period should be expected. With clicks, the time delay at the output of most of the cochlea is related to the response of the system only and not to the input delay of the stimulus which differs by about 0,1 ms for the two click polarities. We also observed larger polarity effects the lower the CF of the derived band when absolute differences are used. However, we demonstrate that this effect is not due to polarity but rather to the larger variability of peak measures for lower CF derived bands. Such increase variability is understandable as the responses are broader [9] and synchronization is poorer in the lower frequency regions owing in part to the slower traveling wave velocity. Why is it that we cannot see the effect of polarity for low-frequency derived bands whose half-period are relatively large? 79 In part, it may be due to the underlying neural variability that has been observed in the PSTH of eighth nerve fibers [20],[16] This is compounded by the fact that wave V of the ABR is generated well beyond the eighth nerve level. Although we agree with earlier observations of previous studies [8],[4],[5],[22] that some individuals demonstrate consistent significant effects, these observations are insufficient to dominate the results and the underlying variability that affects peak latency of wave V compromises any polarity effect even when using absolute measures. Other results [17] , intra-operative CAP recordings in man suggest variability. If there had been a true effect of polarity but was simply washed out in group data because individuals varied as to which phase actually causes stimulation, then absolute values should reveal that fact. However, this means that in large group data, the distributions of the C-R latencies should also be bimodal and observable for low frequency derived bands This was not observed. The distribution of signed values were just broader but with nearly zero mean. However, our simulated polarity effect (Figure. 7 11b) did show the bimodal nature since we simulated a true polarity effect Moreover, since we averaged consistently to low residual noise values (20 nV or less), variability due to noise, which can contribute appreciably, was minimized. 80 Although we can look for explanations of variability of effective stimulus, transfer functions etc., these are currently unnecessary as we must settle the issue of peak reading variability to the same polarity stimulus first Waveform morphology does often change with unmasked click polarity reversal but this is to be expected Unmasked clicks are not appropriate because absolute time is distorted when different bands are summed. For example, let us perform the analysis on two discreet frequency regions, 4 kHz and 2 kHz, whose contributions are summed The periods for these two frequencies are .25 and .5 ms. For rarefaction as the initial phase, peak excitation would theoretically occur at the first quarter period or .0625 and 125 ms respectively . Thus, the temporal difference in peak activity is .0625 ms. However, for condensation as the initial phase, the peak activity would be delayed a half period and would occur at .1875 and .375 ms. The temporal difference in peak activity is now .1875 ms instead of .0625 Even though both regions are contributing a half period later, their time difference is different for the two phases It is no wonder that the amplitude morphology can vary by polarity changes. Such temporal differences in contributions has been shown [10], to significantly affect amplitude and morphology. Thus, the variable effects observed with unmasked clicks in both normal and hearing impaired is bound 81 to be complex because many frequency regions are contributing to varying degree and their relative temporal contributions will be altered by polarity change. This is the likely reason for waveshape changes and possibly observed consistent differences for individuals, i.e. the temporal shifts cause shape and latency differences but varies for individuals The authors would recommend that unmasked clicks not be used as polarity reversals cause relative temporal differences which have unpredictable results In summary, we believe that click polarity effects on the derived brainstem responses are not revealed by peak wave V latency measures Furthermore, the use of absolute values to demonstrate the effect can be accounted for the relationship of the absolute values from distributions with different variance. 82 REFERENCES 1 ANSI (1969) Specification for audiometers (ANSI S3.6-1969), ANSI, New York 2.Beattie, R. C , and Boyd, R. L. (1984) "Early/middle evoked potentials to tone bursts in quiet, white noise and notched noise," Audiology. 24, 406-419, 3 Bendat, J S and Piersol, A.G. (1971) Random Data: Analysis and Measurement Procedures John Wiley & Sons, Inc 4 Borg, E , & Lofqvist, L (1981) "Brainstem response (ABR) to rarefaction and condensation clicks in normal hearing and steep high-frequency hearing loss," Scandinavian Audiology, 13, 99-101 5 Borg, E , & Lofqvist, L (1982) "Auditory brainstem response (ABR) to rarefaction and condensation clicks in normal and abnormal ears," Scandinavian Audiology, 11, 227-235. 6 Brugge, J F , Anderson, D J , Hind, J E. and Rose, J.E. (1969) Time structure of discharges in single auditory nerve fibers of the squirrel monkey in response to complex periodic sound" J. Neurophysiol. 32, 386- 7 Carhart, R , and Jerger, J F (1959) "Preferred method for clinical determination of pure-tone thresholds," J. Speech Hear. Disord 24, 330-345. 83 8 Coats, A C , & Martin, J L (1977) "Human auditory nerve action potentials and brainstem evoked responses," Archives of Otolaryngology, 103, 605-622 9 Don, M , & Eggermont, J. J. (1978). "Analysis of the click-evoked brainstem potentials in man using high-pass noise masking,” Journal of the Acoustical Society of America, 63, 1084-1092 10 Don, M , and Elberling, C. (1994) Evaluating residual background noise in human auditory brainstem responses. JASA, 96: 2746-2757. 11 Don, M., Ponton, C. W., Eggermont, J J , and Masuda, A. (1993) "Gender differences in cochlear response time: An explanation for gender amplitude differences in the unmasked auditory brain-stern response," J Acoust. Soc. Am. 94, 2135-2148 12.Elberling, and Don, M. (1984). "Quality estimation of averaged auditory brainstem responses," Scand. Audiol. 13, 187-197 13 Elberling, C , and Wahlgreen, D, (1985), "Estimation of auditory brainstem responses, ABR, by means of Bayesian reference," Scand. Audiol. 14, 89-96. 14.Fowler, C. G. (1992) "Effects of stimulus phase on the normal auditory brainstem response," Journal of Speech and Hearing Research, 35, 167-174. 15.Gorga, M. P., Kaminski, J. K , & Beauchaine, K. L. (1991). "Effects of stimulus phase on the latency of the auditory brainstem response," Journal of the American Academy of Audiology, 2, 1-6 84 16 Kiang, N. Y. S., Watanabe, T , Thomas, E C , and Clark, L. F (1965) "Dishcarge patterns of single fibers in the cat"s auditory nerve (Research Monograph No. 35). Cambridge & Massachusetts Institute of Technology. 17 Moller, A. R , and Jho, H D. (1991). "Compound Action Potentials Recorded from the intracranial portion of the auditory nerve in man Effects of stimulus intensity and polarity," Audiol 30, 142-163 18 Orlando, M S (1991) "The effects of reversing the polarity of frequency- limited single-cycle stimuli on the auditory brainstem evoked response of adults and infants," Unpublished doctoral dissertation, University of Washington, Seattle, WA 19.Parker, D.J., and Thornton, A R D (1978b): Frequency specific components of the cochlear nerve and brainstem evoked responses of the human auditory system Scand Audiol , 7:53-60 20 Peake, W. T , & Kiang, N Y S. (1962). "Cochlear responses to condensation and rarefaction clicks," Biophysical Journal, 2, 23-34. 21 Rosenhamer, H.J., Lindstrom, B , and Lundborg, T. (1978). "On the use of click-evoked electric brainstem responses in audiological diagnosis I The variability of the normal response." Scand. Audiol. 7, 193-205. 22 Schoonhoven, R (1992). "Dependence of auditory brainstem response on click polarity and high-frequency sensorineural hearing loss," Audiol 31, 72-86 23.Schwartz, D. M., Morris, M. D., Spydell, J. D., Ten Brink, C., Grim, M. A., & Schwartz, J. A, (1990). "Influence of click polarity on the brain-stern auditory evoked response (BAER) revisited," Electroencephalography and Clinical Neurophysiology, 77, 445-457. 85 24.Vermiglio, A J (1992) Master's thesis: "Effects of condensation versus rarefaction clicks in highpass masking on the auditory brainstem response," California State University, Long Beach 25 Weatherill, G. B , and Levitt, H (1965) "Sequential estimation of points on a psychometric function,"" Brit J Math Stat Psychol 18, 1-10 86 U MI MICROFILMED 1996 INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. U M I films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type o f computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction In the unlikely event that the author did not send U M I a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge Contact U M I directly to order. UMI A Bell & Howell Information Company 300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600 UMI N u m b e r: 1 3 7 8 3 9 4 UMI Microform 1378394 Copyright 1996, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Auditory brainstem responses (ABR): quality estimation of auditory brainstem responsses by means of various techniques

PDF

Estimation of upper airway dynamics using neck inductive plethysmography

PDF

Cross-correlation methods for quantification of nonlinear input-output transformations of enural systems using a Poisson random test input

PDF

Respiratory system impedance at the resting breathing frequency range

PDF

A novel network for nonlinear modeling of neural systems with arbitrary point-process inputs

PDF

Design of a portable infrared spectrometer: application to the noninvasive measurement of glucose

PDF

Three-dimensional functional mapping of the human visual cortex using magnetic resonance imaging

PDF

A physiologic model of granulopoiesis

PDF

Comparison of evacuation and compression for cough assist

PDF

Human Skeletal Muscle Oxygenation And Perfusion: Non-Invasive Measurement By Near-Infrared Spectroscopy

PDF

Propofol Effects On Eeg And Levels Of Sedation

PDF

Fault tolerant characteristics of artificial neural network electronic hardware

PDF

Flourine-19 NMR probe design for noninvasive tumoral pharmacokinetics

PDF

A computational model of NMDA receptor dependent and independent long-term potentiation in hippocampal pyramidal neurons

PDF

Decomposition Of Neuronal Function Using Nonlinear Systems Analysis

PDF

A kinetic model of AMPA and NMDA receptors

PDF

Pulse oximetry failure rates

PDF

A proposed role for the superior colliculus within the feedback loop of the brainstem saccade generator

PDF

Processing And Visualization In Functional Magnetic Resonance Imaging (Fmri) Of The Human Brain

PDF

A multimodal screen reader for the visually impaired

Asset Metadata

Creator Alataris, Konstantinos (author)

Core Title Auditory brainstem responses (ABR): variable effects of click polarity on auditory brainstem response, analyses of narrow-band ABR's, explanations

School School of Engineering

Degree Master of Science

Degree Program Biomedical Engineering

Degree Conferral Date 1995-08

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

Tag biology, neuroscience,engineering, biomedical,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Marmarelis, Vasilis A. (committee chair), Khoo, Michael Chee-Kuan. (committee member), Maarek, Jean-Michel (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-6869

Unique identifier UC11357667

Identifier 1378394.pdf (filename),usctheses-c18-6869 (legacy record id)

Legacy Identifier 1378394-0.pdf

Dmrecord 6869

Document Type Thesis

Rights Alataris, Konstantinos

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