Cortical, auditory, evoked potentials in response to changes of ...

Cortical, auditory, evoked potentials in response to changes of spectrum and amplitudea) Brett A. Martin Program in Speech and Hearing Sciences, Graduate Center, City University of New York, 33 West 42 Street, New York, New York 10036 and Department of Speech-Language-Hearing Sciences, Hofstra University, 106 Davison Hall, Hempstead, New York 11550

Arthur Boothroyd Program in Speech and Hearing Sciences, Graduate Center, City University of New York, 33 West 42 Street, New York, New York 10036

共Received 28 April 1999; revised 4 October 1999; accepted 28 December 1999兲 The acoustic change complex 共ACC兲 is a scalp-recorded negative–positive voltage swing elicited by a change during an otherwise steady-state sound. The ACC was obtained from eight adults in response to changes of amplitude and/or spectral envelope at the temporal center of a three-formant synthetic vowel lasting 800 ms. In the absence of spectral change, the group mean waveforms showed a clear ACC to amplitude increments of 2 dB or more and decrements of 3 dB or more. In the presence of a change of second formant frequency 共from perceived /u/ to perceived /i/兲, amplitude increments increased the magnitude of the ACC but amplitude decrements had little or no effect. The fact that the just detectable amplitude change is close to the psychoacoustic limits of the auditory system augurs well for the clinical application of the ACC. The failure to find a condition under which the spectrally elicited ACC is diminished by a small change of amplitude supports the conclusion that the observed ACC to a change of spectral envelope reflects some aspect of cortical frequency coding. Taken together, these findings support the potential value of the ACC as an objective index of auditory discrimination capacity. © 2000 Acoustical Society of America. 关S0001-4966共00兲01704-5兴 PACS numbers: 43.64.Qh, 43.64.Ri 关RDF兴

INTRODUCTION AND PURPOSE

The study reported here is one of a series in which a specific electrophysiological response is being explored in terms of its ability to demonstrate peripheral discrimination capacity. This response is a negative–positive complex that is elicited by a change that occurs during an ongoing acoustic stimulus. This complex has been named the acoustic change complex 共ACC兲共Martin and Boothroyd, 1999兲. In appearance and timing, the ACC is very similar to the well-known negative–positive (N1 – P2) complex that occurs in response to stimulus onset 共Hillyard and Picton, 1978; Naä¨ta¨nen, 1992; Naä¨ta¨nen and Picton, 1987; Onishi and Davis, 1968; Pantev et al., 1996兲. It has been demonstrated that both amplitude and frequency modulation during an ongoing sound can evoke an N1 – P2 complex 共Clynes, 1969; Jerger and Jerger, 1970; Naä¨ta¨nen and Picton, 1987; Kohn et al., 1978, 1980; McCandless and Rose, 1970; Spoor et al., 1969; Yingling and Nethercut, 1983兲, as can an acoustic change during a sustained speech sound 共Kaukoranta et al., 1987兲. In the first study of the present series, the ACC was demonstrated in response to the transition from fricative to vowel in a naturally produced syllable 共Ostroff et al., 1998兲. a兲

Portions of this paper were presented at the Twenty-Second Midwinter Research Meeting of the Association for Research in Otolaryngology, St. Petersburg Beach, FL, February 13, 1999, and at the Biennial Meeting of the International Evoked Response Audiometry Study Group, Tromso, Norway, May 31, 1999.

2155

J. Acoust. Soc. Am. 107 (4), April 2000

The transition, however, included at least three kinds of change: intensity, periodicity, and spectral envelope. In subsequent studies, it was shown that the ACC can be elicited by a change of spectral envelope alone or by a change of periodicity alone 共Martin and Boothroyd, 1999兲 when rms amplitude is held constant. These findings raise the possibility that this response can be used as an index of the capacity for perceiving the kinds of acoustic cues that are important in differentiating speech sounds. If the acoustic change complex is to be useful as an index of peripheral spectral resolution, and therefore of the potential for development of speech perception skills, it is important to establish that the observed response to a spectral change is, indeed, attributable to spectral change and not simply to a difference in the magnitude of excitation. In an earlier study rms amplitude remained constant for the duration of the stimuli. There is no guarantee, however, that the rms level of the external stimulus is the sole determinant of the amount of internal excitation. The amount of synchronous neural excitation may depend on some property of the acoustic waveform other than its rms amplitude 共for example, peak-to-peak amplitude兲. It is possible that the spectral change is accompanied by a change in the magnitude of excitation and that this, alone, is responsible for the observed response. The present study had three goals. The first was to confirm that the ACC is elicited by amplitude change alone—in the absence of changes of spectral envelope or periodicity. It will be recalled that the ACC was observed in response to a

0001-4966/2000/107(4)/2155/7/$17.00

© 2000 Acoustical Society of America

2155

fricative-vowel transition, which involved simultaneous changes of spectral envelope, periodicity, and amplitude. It has already been established that the changes of the first two parameters alone can elicit the complex. Although there was every reason to expect that a change of amplitude alone would also be effective as a stimulus, it was appropriate to confirm this prediction. The second goal of the present study was to determine the sensitivity of the ACC to changes of rms amplitude. There were two reasons for seeking this information. First, a finding that sensitivity is close to that reported in the psychoacoustic literature would support the possibility of electrophysiological measures replacing behavioral measures in the clinical evaluation of auditory capacity in young children. Second, data on amplitude sensitivity can help establish the likelihood that an ACC in response to spectral change is actually produced by a change in some measure of amplitude. The third goal was to measure the effect of simultaneous amplitude change on the magnitude of the ACC elicited by a change of spectral envelope. If the response to spectral change is, in fact, a response to some aspect of waveform amplitude then it should be possible to offset the effect by an opposing change of rms amplitude. In other words, there should be some value of amplitude change for which the spectrally elicited ACC is weakened or even eliminated. The absence of such a finding would lend support to the conclusion that the ACC in response to a change of spectral envelope, in the absence of a change of rms amplitude, is not confounded by a change in some other aspect of waveform amplitude. I. METHOD A. Subjects

Three men and five women, aged 26 to 35 years (mean⫽29 years), participated. All subjects had normal hearing sensitivity 共thresholds⭐25 dB HL from 250 through 8000 Hz bilaterally兲 and no history of neurological disorder. B. Stimuli

The acoustic change used in this study was a transition from /u/ to /i/ in synthesized vowels. The vowels were synthesized with a constant fundamental frequency of 150 Hz. Each vowel lasted for 400 ms and included three formants. The first and third formants were fixed at 300 and 3000 Hz, respectively, for the two vowels. The second formants for the /u/ and /i/ were 900 and 2400 Hz, respectively. Onsets and offsets of the two vowels were shaped with raised cosine functions lasting for one cycle. The stimulus containing the acoustic change was created by concatenating the /u/ and the /i/ with one cycle of overlap. The overlap was introduced to minimize spectral splatter at the transition. This stimulus will be referred to as /ui/. A reference stimulus without spectral change was also created by concatenating two samples of /u/. This stimulus will be referred to as /uu/. The amplitude of the first half of both stimuli 共/ui/, /uu/兲 was 70 dB SPL. Eleven versions of each stimulus were created in which the rms amplitude changed at the midpoint by an amount rang2156

J. Acoust. Soc. Am., Vol. 107, No. 4, April 2000

FIG. 1. The acoustic waveforms for the /uu/ and /ui/ stimuli are shown for all amplitude change conditions. In the 0-dB amplitude change condition, the /uu/ stimulus contains no acoustic change, while the /ui/ stimulus contains a change of spectrum alone. For the remaining amplitude change conditions, /uu/ contains a change of amplitude alone, while /ui/ contains changes of spectrum and amplitude.

ing from ⫺5 dB to ⫹5 dB in 1-dB steps. The waveforms of the resulting stimuli are shown in Fig. 1. All stimuli were digitized at 12 bits and 22 050 samples per second, and were presented to subjects via a Neuroscan STIM system. Note that when no amplitude change is introduced at stimulus midpoint 共the 0-dB change condition兲, the /uu/ stimulus contains no acoustic change, and the /ui/ stimulus contains only a change of spectrum. The /uu/ stimuli across the remaining amplitude change conditions contain only a change of amplitude, while the /ui/ stimuli in the remaining amplitude change conditions contain a change of both spectrum and amplitude.

C. EEG recordings

Using a Neuroscan SCAN system and Grass amplifiers, seven EEG channels were recorded from surface electrodes placed at Fz, Cz, Pz, T3, T4, A1, and A2. The EEG channels were referenced to an electrode at the tip of the nose 共Vaughan and Ritter, 1970兲. An eighth channel to monitor vertical eye movements and eye blinks 共EOG兲 was recorded from electrodes placed above and below the right eye. An electrode at Fpz served as ground. Electrode impedances were maintained below 5000 Ohms. During acquisition, the EEG channels were amplified using a gain of 20 000 共except for the EOG channel where B. A. Martin and A. Boothroyd: CAEPs to complex stimuli

2156

FIG. 2. The scalp distribution of the response is shown for the /ui/ stimulus in the ⫹5-dB amplitude change condition. The acoustic change complex 共indicated by the arrow兲 is clearly present. It is largest in amplitude at electrode site Cz, is smaller at the other electrode sites, and inverts in polarity at the earlobe sites 共A1, A2兲.

the gain was 5000兲, and band-pass filtered with a roll-off of 6 dB/octave below 0.1 Hz and above 100 Hz. D. Procedure

Stimuli were presented via a loudspeaker placed 70 cm from the subject at a 0 degree azimuth. They were presented in homogeneous blocks of 125 stimuli, using an onset-toonset interval of 3 s. Two replications were obtained for each condition. Thus each stimulus was heard 250 times. Stimulus presentation was randomized across conditions and subjects. Subjects watched a silent, captioned video during testing and were instructed to ignore the stimuli. The amplified EEG signals were digitized using the Neuroscan SCAN system at 341 Hz over a 1501-ms 共512 point兲 window, beginning 100 ms before stimulus onset. After acquisition, single trials were rejected from averaging when activity in any channel 共except the EOG channel兲 exceeded ⫾100 ␮V. Single trials were baseline corrected 共across the entire sweep duration兲 and an ocular artifact reduction algorithm was applied 共Semlich et al., 1986兲. Additional band-pass filtering was applied with a roll-off of 12 dB/octave below 0.1 Hz and above 30 Hz. Responses to each of the 22 stimuli were averaged separately for each subject. Each averaged waveform was then baseline corrected for the prestimulus period 共0–100 ms兲. Finally, individual averaged waveforms were combined to generate 22 group waveforms. II. RESULTS A. Scalp distribution of the ACC

Figure 2 shows the scalp distribution of the group mean waveforms in response to the /ui/ stimulus containing a ⫹5-dB amplitude change. Results are shown for electrode sites Fz, Cz, Pz, T3, T4, A1, and A2. Clear P1 – N1 – P2 potentials are seen in response to the onset of stimulation, followed by a sustained potential that continues for the duration of the stimulus, and a return to baseline in response to the offset of stimulation. There is an additional P1 – N1 – P2 complex seen in response to the change of both spectrum and amplitude at stimulus midpoint. This is the acoustic change complex 共ACC兲, and it is largest in amplitude at the vertex.1 2157


FIG. 3. The group mean waveforms are displayed for the 2 stimuli 共/uu/, /ui/兲 and the 11 amplitude change conditions. There is a clear N1 – P2 obtained to the onset of stimulation in all conditions. For amplitude change alone, the acoustic change complex 共shown in the boxes兲 is clearly present for changes of 2–3 dB or more. For spectrum plus amplitude change, the acoustic change complex is clearly present for all conditions.

The rms amplitude of the ACC is reduced by approximately 21% at Fz, 57% at Pz, and 73% at temporal electrode sites 共T3, T4兲. In addition, the response inverts in polarity at the earlobes 共A1, A2兲. B. Group mean waveforms

The group mean waveforms obtained for each acoustic change at Cz 共the electrode site giving the largest amplitude兲 are displayed in Fig. 3. The waveforms on the right show results for spectral and amplitude change 共i.e., the /ui/ stimulus兲. There is a clearly observable ACC for each level of rms amplitude change when combined with a spectral change. The waveforms on the left show results for amplitude change alone 共i.e., the /uu/ stimulus兲. In this case, the ACC is only clearly observable only for amplitude changes of 2–3 dB or more. It can also be seen that, for both stimuli, the amplitude of the ACC is higher for amplitude increments than for amplitude decrements. C. Waveforms for individual subjects

The waveforms obtained to the /ui/ stimulus in the ⫹5-dB amplitude change condition are shown for each individual subject in Fig. 4. Responses are shown at electrode site Cz 共the electrode site giving the largest amplitude in the B. A. Martin and A. Boothroyd: CAEPs to complex stimuli

2157

FIG. 5. Group mean rms amplitude of the acoustic change complex 共⫾1 s.e.兲 is shown as a function of amplitude change, with and without a simultaneous spectral change. Lines show least squares fits to hypothetical functions.

共4兲 Group means and standard errors of the rms amplitudes were then calculated for each of the stimuli. FIG. 4. The waveforms from each individual subject are displayed for the /ui/ stimulus in the ⫹5-dB amplitude change condition. In addition, the group mean waveform is also shown. The acoustic change complex 共shown in the box兲 is clearly present in all subjects in response to the change of spectrum and amplitude.

Figure 5 shows group mean response amplitude, ⫾1 ‘‘between-subject’’ standard error, as a function of stimulus amplitude change, for the /uu/ and /ui/ stimuli. The lines in Fig. 5 show the result of curve fitting procedures described below.

group mean waveforms兲. There is a clearly observable ACC for each subject in response to the change of spectrum and amplitude.

E. ACC to amplitude change only

D. Quantification of ACC amplitude

The data for amplitude change only were fit with exponential growth functions of the form:

For purposes of analysis, the root mean square 共rms兲 amplitude of the ACC was calculated for each subject and each of the 22 stimuli. The steps involved in this process were as follows: 共1兲 The latencies of the N1 and P2 maxima in the group mean waveforms were obtained for conditions with a clearly present response 共⫾3 to 5 dB兲. 共2兲 A ‘‘response window’’ was defined beginning 50 ms before the average N1 minimum and ending 50 ms after the average P2 maximum. The resulting window extended from 454 to 645 ms re: stimulus onset 共i.e., 61 to 252 ms after the onset of acoustic change兲. 共3兲 The standard deviations of the waveforms for each stimulus and each subject within this window were calculated. The result, which is the rms amplitude of the waveform within the response window, excluding any dc offset, was used as a measure of the amplitude of the complex.2 2158


y⫽a 共 1⫺e ⫺x/b 兲 ,

共1兲

where y is the rms amplitude of ACC in ␮V; a is the asymptote in ␮V; e is the base of natural logarithms; x is the amplitude change in dB; and b is a constant, in dB, providing an inverse measure of rate of growth of ACC amplitude with increasing amplitude change.3 A least-squares fitting procedure produced the following parameter values: a⫽2.2 ␮ V and b⫽2.5 dB for amplitude increments and a⫽1.1 ␮ V and b⫽⫺1.9 dB for amplitude decrements. The fitting procedure was limited to those conditions in which the ACC amplitude was clearly above the noise floor, as defined by the rms amplitude of the ACC for the /uu/ stimulus with 0-dB amplitude change 共i.e., no change of either spectrum or amplitude兲. Several observations can be made from the amplitude change data illustrated in Fig. 5. First, the ACC amplitude is greater for amplitude increments than for amplitude decreB. A. Martin and A. Boothroyd: CAEPs to complex stimuli

2158

ments. Second, for the stimuli and protocols used in this study, an amplitude change of at least ⫹2 or ⫺3 dB is required in order to produce an ACC that is clearly above the noise floor. Third, the standard errors 共a measure of intersubject variability兲 are small within the noise floor but increase with increasing ACC amplitude.

F. ACC to spectrum change plus amplitude change

The data for spectrum plus amplitude change were fit with a sigmoid transition function of the form: y⫽a⫹ 共 b⫺a 兲 / 共 1⫹e ⫺ 共 x⫺c 兲 /d 兲 ,

共2兲

where y⫽rms amplitude of ACC in ␮V; a⫽lower asymptote in ␮V; b⫽upper asymptote in ␮V; e⫽base of natural logarithms; x⫽amplitude change in dB; c⫽mid-point of transition in dB; and d⫽a constant in dB providing an inverse measure of rate of change of ACC amplitude in the transition region. A least-squares fitting procedure produced the following parameter values: a⫽1.3 ␮ V, b⫽2.9 ␮ V, c⫽2.9 dB, and d⫽1.6 dB. Several observations can be made from the spectrum-plus-amplitude change data illustrated in Fig. 5. First, the amplitude of the ACC is well above the noise floor when there is no change of amplitude 共group mean rms amplitude⫽1.5 ␮ V兲. Second, the addition of an amplitude increment has the effect of increasing amplitude of the ACC. Third, for amplitude increments, the ACC amplitude continues to increase as the amplitude change increases. Fourth, for amplitude decrements, there is no evidence to show that the ACC amplitude continues to increase with increasing amplitude change. Fifth, and most important, there is no evidence from these data that a small amplitude change can cancel the ACC produced by the change of spectrum.

G. Statistical tests

Statistical tests were used to explore the generalizability of these findings to the means of the population represented by this sample of subjects. A two-factor repeated-measures analysis of the variance in the ACC amplitudes was performed. The two factors were spectrum change at 2 levels 共present or absent兲 and amplitude change at 11 levels. The main effect of spectrum change was highly significant 关 F(1,7)⫽44.90; p⬍0.001兴 as was the main effect of amplitude change 关F(10,70)⫽11.26; p⬍0.001兴. There was also a significant interaction between the two 关F(10,70)⫽1.97; p ⫽0.049兴. Post hoc testing, using the least significant difference test, confirmed the presence of a significant effect (p ⭐0.05) of spectrum change for most amplitude changes, with the exception of amplitude changes of ⫺3 dB (p ⫽0.192) and ⫺4 dB ( p⬍0.240). When the data for spectrum plus amplitude increment were examined separately, there was a significant main effect of amplitude change 关 F(1,5)⫽5.05; p⫽0.001兴. In the data for spectrum change plus amplitude decrement, however, the main effect of amplitude change did not reach the 5% level of significance 关 F(1,5)⫽1.75; p⫽0.15兴. 2159


III. DISCUSSION

The ACC was detectable in the group mean waveforms to amplitude changes of ⫹2 and ⫺3 dB. These thresholds compare favorably with those obtained from psychoacoustic studies, which range from 0.2 to 0.5 dB, depending on experimental paradigm, frequency, and sensation level 共Gelfand, 1990兲. Moreover, it may be possible to increase the sensitivity of the ACC by a change of protocol to increase signal-to-noise ratio. The finding of stronger responses to amplitude increments than to amplitude decrements is in keeping with those of studies using amplitude modulated stimuli 共Arlinger and Jervall, 1979; Clynes, 1969; McCandless and Rose, 1970兲. This finding can be explained if it is assumed that stimulus amplitude determines the total amount of synchronous cortical excitation. Amplitude increments would then be expected to produce an onset response from previously unexcited neural elements whereas decrements would produce offset responses from previously excited elements. The different amplitudes of the ACC waveform are in keeping with the known differences between the onset and offset N1 – P2 response 共Hillyard and Picton, 1978兲. There is some evidence, however, that amplitude is coded, at least partially, by locus of cortical excitation. In a classic study, Pantev et al. 共1989兲 measured neuromagnetic responses to 1-kHz tones presented at six amplitudes. The amplitude change produced both vertical and horizontal shifts in the hypothetical dipole generators of the M 100 wave 共equivalent to the N100, or N1 response in electrical recordings兲. Amplitopic organization would not explain why stimulus amplitude increments produce a larger ACC response than equivalent amplitude decrements, but it is possible that amplitopic effects are accompanied by changes in total amount of synchronous excitation. The demonstration of an ACC in response to spectral change, in the absence of a change of rms amplitude, is in keeping with the established tonotopic organization of the auditory cortex. The spectral change used in the present study involved an amplitude decrement at one frequency and a simultaneous amplitude increment at another frequency. The resulting ACC may, therefore, be assumed to reflect a combination of an onset response from one cortical region and an offset response from another. The alternative explanation—that the ACC in response to spectral change simply reflects a change of overall synchronous excitation—is not supported by the findings of the present study. There was no condition under which a small amplitude increment or decrement eliminated, or even reduced, the amplitude of the ACC response to spectral change. The present findings provide strong evidence for dissociation of cortical responses to changes of spectrum and overall amplitudes. In other words, the ACC elicited by spectral change in this and earlier studies can be assumed to reflect the effects of spectral change rather than an incidental change in the total amount of synchronous excitation. Amplitude increments enhanced the effects of spectral change, whereas amplitude decrements had little or no effect. This asymmetry has been observed previously. Ma¨kela¨ et al. 共1987兲 found evidence for different processing of amplitude B. A. Martin and A. Boothroyd: CAEPs to complex stimuli

2159

and frequency modulations up to the level of the supratemporal auditory cortex. Moreover, the order of presentation of the two types of modulation had an effect on response magnitude. The asymmetry observed in the present study can also be explained in terms of the amplitopic effects discussed earlier. Differences in locus and orientation of generators in response to amplitude increments and decrements could easily affect the results of vector summation. The foregoing attempts to explain the amplitude findings are clearly an oversimplification. It is well known that cortical encoding of amplitude modulated stimuli is complex 共Pickles, 1988兲, and it is likely that multiple mechanisms at different cortical generator areas are at work. Further, surface-recorded data, particularly using a small number of electrode sites, cannot equal the specificity of intracranial recordings for determining what is happening at the singlecell level of the auditory cortex. The findings of this study are encouraging in terms of the potential clinical use of the ACC as an index of spectral resolution in the impaired auditory system. The high amplitude of the ACC, and the fact that it is elicited by spectral changes of the type encountered in speech, are particularly important. Further, it has been shown in subsequent studies that the ACC can be demonstrated in response to small changes of second formant frequency—of the magnitude involved in vowel distinctions 共Ostroff, 1999兲. Further research will be required, however, to establish adequate reliability and sensitivity in individual waveforms as opposed to group waveforms. Approaches to improving sensitivity must address such issues as rate of spectral change 共Onishi and Davis, 1968; Ruhm, 1970兲, frequency region of spectral change 共Arlinger et al., 1976; Kohn et al., 1978兲, and the duration of stimulus on-time before the spectral change is introduced 共Budd and Michie, 1994; Hillyard and Picton, 1978兲. Application to persons with sensorineural hearing loss must also take account of the frequency-dependence of threshold and loudness, together with the effects of cochlear damage on the relationship between amplitude and loudness. In the present study, it was demonstrated that maintaining constant rms amplitude served to dissociate frequency and amplitude effects. In hearing-impaired subjects, it may be necessary to measure the effects of spectral change when accompanied by amplitude increments and decrements—as in the present study. IV. CONCLUSIONS

共1兲 The acoustic change complex was evoked by amplitude change alone, in the absence of changes of spectral envelope or periodicity. 共2兲 The amplitude change required to elicit an observable ACC in the group waveforms was 2–3 dB. This threshold is not far above those reported in the behavioral literature on amplitude discrimination. 共3兲 The acoustic change complex was evoked by a change of second formant frequency during a sustained synthetic vowel, regardless of the simultaneous presence of rms amplitude increments or decrements. The findings support the conclusion that the ACC produced in response 2160


to a change of spectrum, in a signal with unchanging rms amplitude, reflects the auditory systems’s response to spectrum and not to accompanying changes of perceived magnitude. ACKNOWLEDGMENT

This work was supported by NIH-NIDCD Grant No. 5P50DC00178. 1

The amplitude of P1 was quite small in some individual subjects. Data for P1 are not, therefore, analyzed in this paper. 2 The rms amplitude was used in preference to the more traditional peak-topeak measure in order to take full advantage of all data in the response window. 3 Note that b is the equivalent of ‘‘time constant’’ when the independent variable is time. It takes approximately three time constants for the dependent variable to come within 95% of the asymptote. Arlinger, S. D., and Jerlvall, L. B. 共1979兲. ‘‘Results of psychoacoustic and cortical evoked potential experiments using frequency and amplitude modulated stimuli,’’ in Models of the Auditory System and Related Signal Processing Techniques, Scandinavian Audiology Supplement 9, edited by M. Hoke and E. de Boer 共Nordic Audiology Society兲, pp. 229–239. Arlinger, S. D., Jerlvall, L. B., Ahren, T., and Holmgren, E. C. 共1976兲. ‘‘Slow evoked cortical responses to linear frequency ramps of a continuous pure tone,’’ Acta Physiol. Scand. 98, 412–424. Budd, T. W., and Michie, P. T. 共1994兲. ‘‘Facilitation of the N1 peak of the auditory ERP at short stimulus intervals,’’ Neuroreport 5, 2513–2516. Clynes, M. 共1969兲. ‘‘Dynamics of vertex evoked potentials: The R-M brain function,’’ in Average Evoked Potentials: Methods, Results, and Evaluations 共NASA SP-191兲, edited by E. Donchin and D. B. Lindsley 共Washington: U.S. Government Printing Office兲, pp. 363–374. Gelfand, S. 共1990兲. Hearing: An Introduction to Psychological and Physiological Acoustics 共Marcel Decker, New York兲, pp. 335–341. Hillyard, S. A., and Picton, T. W. 共1978兲. ‘‘ON and OFF components in the auditory evoked potential,’’ Percept. Psychophys. 24, 391–398. Jerger, J., and Jerger, S. 共1970兲. ‘‘Evoked responses to intensity and frequency change,’’ Arch. Otolaryngol. 91, 433–436. Kaukoranta, E., Hari, R., and Lounasmaa, O. V. 共1987兲. ‘‘Responses of the human auditory cortex to vowel onset after fricative consonants,’’ Exp. Brain Res. 69, 19–23. Kohn, M., Lifshitz, K., and Litchfield, D. 共1980兲. ‘‘Average evoked potentials and amplitude modulation,’’ Electroencephalogr. Clin. Neurophysiol. 50, 134–140. Kohn, M., Lifshitz, K., and Litchfield, D. 共1978兲. ‘‘Average evoked potentials and frequency modulation,’’ Electroencephalogr. Clin. Neurophysiol. 45, 236–243. Ma¨kela¨, J. P., Hari, R., and Linnankivi, A. 共1987兲. ‘‘Different analysis of frequency and amplitude modulations of a continuous tone in the human auditory cortex: A neuromagnetic study,’’ Hear. Res. 27, 257–264. Martin, B. A., and Boothroyd, A. 共1999兲. ‘‘Cortical, auditory, event-related potentials in response to periodic and aperiodic stimuli with the same spectral envelope,’’ Ear Hear. 20, 33–44. McCandless, G. A., and Rose, D. E. 共1970兲. ‘‘Evoked cortical responses to stimulus change,’’ J. Speech Hear. Res. 13, 624–634. Naä¨ta¨nen, R. 共1992兲. Attention and Brain Function 共Lawrence Erlbaum, New Jersey兲. Naä¨ta¨nen, R., and Picton, T. W. 共1987兲. ‘‘The N1 wave of the human electric and magnetic response to sound: A review and an analysis of the component structure,’’ Psychophysiology 24, 375–425. Onishi, S., and Davis, H. 共1968兲. ‘‘Effects of duration and rise time of tone bursts on evoked V potentials,’’ J. Acoust. Soc. Am. 44, 582–591. Ostroff, J. M. 共1999兲. ‘‘Parametric study of the acoustic change complex to synthetic vowel stimuli as a measure of peripheral auditory discrimination capacity,’’ Unpublished doctoral dissertation, Graduate Center of the City University of New York. Ostroff, J. M., Martin, B. A., and Boothroyd, A. 共1998兲. ‘‘Cortical evoked response to acoustic change within a syllable,’’ Ear Hear. 19, 290–297. Pantev, C., Eulitz, C., Hampton, S., Ross, B., and Roberts, L. E. 共1996兲. ‘‘The auditory evoked ‘‘off’’ response: Sources and comparison with the ‘‘on’’ and the ‘‘sustained’’ responses,’’ Ear Hear. 17, 255–265. B. A. Martin and A. Boothroyd: CAEPs to complex stimuli

2160

Pantev, C., Hoke, M., Lehnertz, K., and Lu¨tkenho¨ner, B. 共1989兲. ‘‘Neuromagnetic evidence of an amplitopic organization of the human auditory cortex,’’ Electroencephalogr. Clin. Neurophysiol. 71, 225–231. Pickles, J. O. 共1988兲. An Introduction to the Physiology of Hearing 共Academic, London兲, pp. 205–234. Ruhm, H. B. 共1970兲. ‘‘Rate of frequency change and the acoustically evoked response,’’ Journal of Auditory Research 10, 29–34. Semlitch, H. V., Anderer, P., Schuster, P., and Presslich, O. 共1986兲. ‘‘A solution for reliable and valid reduction of ocular artifacts applied to the P300 ERP,’’ Psychophysiology 23, 695–703.

2161


Spoor, A., Timmer, F., and Odenthal, D. W. 共1969兲. ‘‘The evoked auditory response 共EAR兲 to intensity modulated and frequency modulated tones and tone bursts,’’ International Audiology 8, 410–415. Vaughan, H. G., and Ritter, W. 共1970兲. ‘‘The sources of auditory evoked responses recorded from the human scalp,’’ Electroencephalogr. Clin. Neurophysiol. 28, 360–367. Yingling, C. D., and Nethercut, G. E. 共1983兲. ‘‘Evoked responses to frequency shifted tones: Tonotopic and contextual determinants,’’ Int. J. Neurosci. 22, 107–118.

B. A. Martin and A. Boothroyd: CAEPs to complex stimuli

2161