Auditory event related potentials were recorded from neonatal, 3-month, and 3-year old rhesus monkeys. Auditory brainstem evoked responses (ABRs) were reliably recorded at all ages. ABR latencies decreased with age. Age effects were greater the more centrally generated the wave. Wave I amplitude decreased with age, Wave II increased, and Wave IV remained about the same. Stimulus rate effects were greater in neonates than older monkeys. Stimulus frequency also affected the ABR, but not differentially as a function of age. Recording montage had a significant effect on the recorded waveform. Wave I tended to be larger in amplitude in horizontal recordings and front-back recordings, while the later waves were relatively more prominent in more vertical montages. Middle latency evoked responses and late potentials were less reliably recorded than the ABR. Their reproducibility improved with age. Auditory event related potentials are promising measures of auditory function for research requiring nonhuman primate models of the developing human.

When special pseudo-random stimuli sequences (maximum length sequences: MLS) are combined with a deconvolution analysis technique, it is possible to derive new evoked potential components that are called kernels. The kernels give a measure of the temporal interactions that take place between the responses to successive stimuli. This may provide an objective neurophysiological test for the exploration of a dimension of hearing which has hitherto been limited to psychophysical methods. Until now, auditory short-latency kernels obtained by the MLS method have been related to the late portion of the brainstem auditory evoked potential (BAEP), suggesting that temporal interactions occur rather late in the auditory pathways. We report 4 children without any BAEP neural components, who all retained isolated cochlear microphonic potentials. Three of them produced click-evoked otoacoustic emissions and two of them demonstrated only moderately impaired audiometric thresholds. This combination of absent BAEP neural components with preserved otoacoustic emissions and cochlear microphonic potential corresponds to a peculiar pattern of auditory dysfunction recently coined "auditory neuropathy'. All 4 children exhibited well-defined kernels at latencies consistent with the microphonic potential. These data indicate that the cochlea itself can generate kernels at a presynaptic level. They open up the question of the identification of the physiological site(s) responsible for the generation of MLS-evoked kernels.

We report 3 children without any brainstem auditory evoked potential (BAEP) neural component who all retained isolated cochlear microphonic potentials as well as click-evoked otoacoustic emissions. Two of them demonstrated only moderately impaired audiometric thresholds. These features correspond to a peculiar pattern of auditory dysfunction recently coined `auditory neuropathy'. In contrast with the published previous cases of auditory neuropathy presenting with an acquired hearing deficit as children or young adults, all 3 children had a history of major neonatal illness and the auditory neuropathy was already demonstrated in the first months of their lives.

Early (ABRs) and middle (MLRs) surface-recorded auditory evoked potentials were compared in eight adult monkeys (Macaca mulatta) and eight adult humans. Responses whose probable generators were the cochlear nucleus and lateral lemniscus were of shorter latency and larger amplitude in monkeys. Relative to humans, ABR response latencies in monkeys were less affected by stimulus intensity, stimulus rate, and masker level. In contrast, monkey amplitudes were relatively more affected by those same stimulus parameters. The most prominent MLR wave was longer in latency and greater in amplitude in humans than the homologous wave in monkeys. The reduction in amplitude of that wave with increasing rate was greater for humans than monkeys. Temporal interactions (the effect of prior stimuli on the response to current stimulation) were investigated from a non-linear systems identification framework using maximum length sequences (MLSs). Both monkey and human auditory systems were second and probably third-order systems at the levels assessed. As the separations between the stimulus pulses decreased, evidence for temporal interactions became more prominent, reached a maximum, and then decreased with further decreases in stimulus pulse separation. At the highest stimulus rates presented, variations in temporal spacing among stimuli had less of an effect on monkey than human evoked responses.

Auditory functioning was assessed in two groups of adult rhesus monkeys (11 years of age). One (n = 11) received modest exposure to lead early in life and the other (n = 8) served as controls and did not receive any lead supplementation. Two lead-exposed monkeys had abnormal distortion product otoacoustic emissions (DPEs) and smaller amplitude or absent evoked potentials. These monkeys had the highest blood levels recorded in their respective groups. For the remaining lead-exposed monkeys there was little difference between their DPEs and the DPEs of the control monkeys with one exception. DPE amplitudes of the control monkeys increased more rapidly as a function of stimulus level than those of the lead-exposed monkeys at most frequencies. There was also a significant but modest effect of lead exposure on the auditory brain stem evoked responses (ABRs) of these lead-exposed monkeys. There was no apparent effect on the middle latency evoked responses (MLRs), although that result could be due to the relatively greater variability of the MLR.

Distortion product otoacoustic emissions (DPOAEs) were compared in eight rhesus monkeys (Macaca mulatta) and eight normal hearing humans. DPOAEs were recorded in three conditions. In the first condition, DPOAEgrams were generated for monkeys and humans from approximately f2 = 0.5-20 kHz. Monkeys had larger amplitude DPOAEs at all frequencies except around f2 = 1 kHz. In the second condition, DPOAE amplitudes increased and then decreased as the separation between the primaries increased. These functions were similar in the two species except at the lowest frequencies assessed. In the third condition, the levels of the primaries were varied independently. Monkeys had steeper input/output (I/O) functions than humans. The slopes of DPOAE I/O functions increased with frequency in both species. When the levels of both primaries were increased simultaneously, DPOAE I/O functions were well described by power functions throughout the intensity range assessed (from threshold to 65 dB SPL). Monkey I/O functions tended to be expansive power functions at all but the lowest frequencies, while human I/O functions tended to be compressive power functions except at the highest frequencies assessed. Other differences in I/O functions f2 = 8 kHz may indicate species specific differences at high (for human) frequencies.

This study compared monaural and binaural maximum length sequence auditory-evoked brain-stem responses (MLS ABRs) in normal hearing adults. The first experiment demonstrated that reliable binaural MLS ABRs could be recorded which were essentially the same as those recorded monaurally. The second experiment generalized this finding by assessing a range of intensities including threshold stimuli. ABR thresholds, wave V latency x intensity and amplitude x intensity functions, wave V latency and amplitude reproducibility, and waveform reproducibility were comparable for the monaural and binaural MLS ABRs with some minor qualifications. In the third experiment, comparability of monaural and binaural MLS ABRs was generalized to a range of rates from those used conventionally to rates far faster than possible with signal averaging. Again, there was little difference between the binaural and monaural MLS ABRs over the range of rates assessed.

A spatial feature extraction and regularization model is developed to represent free-field-to-eardrum transfer functions (FETFs). A Karhunen-Loeve expansion is used to derive a low dimensional eigen-transfer function (EF) subspace for the measured FETFs. The coordinates of each FETF in the subspace are determined by projecting all measured FETFs onto the EFs. These coordinates represent samples of the FETFs' spatial features. Functional representations of the spatial features, termed spatial characteristic functions (SCFs), are obtained by applying a thin-plate general spline smoothing model to regularize the samples. A functional representation for the FETF is thus obtained by linearly combining the EFs with the SCFs. Typical errors between the measured and modeled FETFs for a KEMAR are on the order of a hundredth of 1%.<>

A functional representation for head-related transfer functions (HRTFs) is desirable as it overcomes many of the limitations associated with use of measured HRTFs. It provides a continuous representation of auditory space, such that the synthesis of HRTF at any given spatial location can be performed by functional evaluation of the model. Such a model that establishes a mathematical representation of the external ear transformation characteristics based on spatial feature extraction and regularization is proposed.<>

Two simultaneously presented maximum length sequences (MLSs) were used to investigate temporal nonlinearities. Not only did the recorded auditory evoked brain stem responses to these stimuli predictably increase in latency and decrease in amplitude as a function of the temporal interactions between MLSs, but thresholds were elevated by more than 20 dB. Simultaneous MLS paradigms make it possible to investigate a number of nonlinearities in the auditory system efficiently. This study also demonstrated that binaural MLS techniques can be used to assess auditory function even in individuals with asymmetric hearing losses without fear of crossover effects.

The purpose of this investigation was to describe the properties of averaged auditory evoked potential distortion products (AEP-DPs) in guinea pigs. This study provided a step toward developing a clinical index of nonlinear processing of auditory signals and supplied a baseline for studies evaluating the effect of cochlear damage on AEP-DPs.The amplitude of the AEP-DPs was evaluated as a function of f2/f1 ratio (1.12-1.52) and primary frequency (500 Hz-2000 Hz). The amplitude of the AEP cubic difference tone (AEP-CDT) increased with increasing f2/f1 ratio for the 500-Hz f1 primary and remained constant for the 800-Hz and 1700-Hz f1 primaries. The AEP-CDT generated by the 1100-Hz and 1400-Hz f1 primaries was maximum for the middle f2/f1 ratios (1.22, 1.32, and 1.42). The AEP-CDT could not be distinguished from the noise floor for the 2000-Hz f1 primary. The AEP difference tone (AEP-DT) was larger and more frequently identified than the AEP-CDT. The amplitude of the AEP-DT decreased with an increase in f2/f1 ratio. The decrease was more pronounced for low-frequency f1 primaries than for high-frequency f1 primaries.

Spectral analysis of auditory-evoked potential recordings from ten normal-hearing subjects to two-tone signals revealed energy at difference tone (DT = f2-f1) and cubic difference (CDT = 2f1-f2) frequencies that was not present in the acoustic signal. Control experiments and calibrations provided substantial evidence supportive of the biological nature of these auditory nonlinearities, suggesting that they are not the result of electromagnetic, acoustic, or analytic artifact. Amplitudes of DT- and CDT-evoked responses were evaluated for rarefaction and condensation signals with f1 = 510 and 800 Hz across frequency ratios (f2/f1) of 1.16, 1.26, 1.36, and 1.46. Additionally, time-domain summation and subtraction of separately collected evoked responses to rarefaction and condensation signals were performed to demonstrate that these electrophysiological DT and CDT responses reflect their expected quadratic and cubic nature. Suggestions for development of clinical applications of assessing auditory nonlinearities using this methodology are provided.

Brain-stem auditory-evoked responses (BAERs) were obtained in six normal-hearing adults using single-tone and two-tone stimuli arithmetically centered around 4000 Hz. Two-tone stimuli varied in frequency separation from 200 to 3200 Hz, and started in-phase (homophasic) or 180 deg out-of-phase (antiphasic) with each other. Responses to each of the single-tone components of the two-tone stimuli were elicited and then summed for comparison with responses to the two-tone stimuli. Results indicated no significant difference in wave V latency between homophasic or antiphasic two-tone conditions, and summed single-tone conditions. Under the homophasic condition, the mean latency for the widest frequency separation of the tones was significantly longer than those for narrower separations. A significant difference in wave V amplitude between two-tone phase conditions was found for frequency separations of 200, 400, and 3200 Hz only. Summed single-tone BAERs demonstrated a significantly larger wave V amplitude than responses from either two-tone phase condition at all frequency separations.

Experiments were performed in which brain-stem auditory evoked responses (BAERs) were elicited by two types of pseudorandom pulse trains: maximum length sequences (MLS) and Legendre sequences (LGS). In experiment 1, each pulse sequence was presented at 50 dB nHL with minimum pulse intervals varying from 1 to 10 ms. Wave V latency increased and wave V amplitude decreased with decreasing minimum pulse intervals, with no significant effect of the type of sequence (MLS vs LGS), and no significant interaction between sequence and minimum pulse interval in terms of wave V amplitude or latency. In a second experiment, the minimum pulse interval was held constant at 4 ms, while MLS and LGS levels were varied from 20 to 60 dB nHL. With increasing click intensity, there is a decrease in wave V latency and an increase in wave V amplitude. There was no significant effect of type of sequence (LGS vs MLS) or interaction between type of sequence and stimulus intensity for wave V amplitude or latency. Despite the obvious violation of the assumptions (linearity and stationarity) underlying the application of maximum length sequence analysis and Legendre sequence analysis, both techniques produced reliable responses remarkably similar in morphology to evoked responses obtained by conventional averaging. The results of these experiments support the possibility that analysis methods based on pseudorandom pulse sequences may prove more efficient in data collection and provide a more thorough description of the electrophysiologic behavior of the auditory system compared to conventional averaging.

This article describes the use of auditory-evoked potentials (AEPs) as a tool to assess nonlinear processes in the auditory system. Two-tone signals were used as stimuli to obtain AEPs in both animal and human subjects. Frequency analysis of the physiologic waveforms revealed frequencies in the evoked potential that were not present in the acoustic signal. The largest distortion product in the evoked potential corresponded to the difference between the two primary frequencies (f2-f1). This distortion product was present in all subjects tested. Other distortion products at frequencies defined by n(f2-f1), where n less than 5, were also present in some individuals. These frequencies represent distortion components generated from an even-order nonlinear system. Extensive acoustic and electric calibration procedures provided substantial evidence that the distortion products recorded in the AEP were biologic in origin and not the result of acoustic or recording artifact.

The effects of masking noise on wave V of the brain-stem auditory-evoked response (BAER) obtained to pseudorandom pulse sequences are evaluated in two experiments. In the first experiment, the level of broadband noise was covaried with minimum pulse interval (rate) using maximum length sequence analysis (MLSA). Both increasing noise level and decreasing minimum pulse interval decrease wave V amplitude and increase wave V latency. A nonadditivity of rate and noise level was observed such that, at the shortest interpulse intervals, simultaneous background noise produced virtually no latency change and minimal amplitude change, for the noise levels tested. In a second experiment, high-pass masking was performed to assess the feasibility of derived-band techniques using maximum length sequence analysis (MLSA) and to compare the frequency regions responsible for the BAER using MLSA versus conventional averaging. Results of experiment 2 showed that reliable responses across high-pass masker cutoff frequency could be obtained in normal-hearing listeners. The frequency specificity of the MLSA-based responses was nearly identical to that obtained by conventional averaging, although both amplitude and latency of wave V were affected by the high-pass masker cutoff and minimum pulse interval values. These studies suggest that the neuronal populations and frequency regions responsible for the BAER are virtually the same for MLSA and conventional averaging.

The auditory system is represented by a third-order Volterra series, and the cross-section function resulting from applying the cross-correlation method with an m pulse sequence as the input is derived in terms of Volterra kernels. The function can be used to predict system responses to pulse trains of different pulse rates. It is therefore concluded that the m pulse sequence may be an effective and efficient stimulus for studying nonlinear rate effects in neurological systems

The purpose of this paper is to describe the effect of broadband continuous noise on brain stem auditory evoked responses elicited from normal-hearing and hearing-impaired individuals. The motivation for this study derives from the increasing use of noise masking paradigms in diagnostic electrophysiology, the universal presence of background noise in everyday listening environments, and the frequent observation that background noise is more detrimental to the performance of the hearing-impaired individual than to the normal-hearing individual. Four studies were designed to evaluate: (1) the sensitivity and specificity of the latency-intensity series, (2) the sensitivity and specificity of the latency-noise series, (3) the dependence of the latency-noise series on signal-to-noise ratio near electrophysiologic threshold, and (4) the dependence of the latency-noise index on the signal level at which the test is performed. The results of the studies reported herein show that the electrophysiological response to increasing masker levels is more sensitive for identifying inner ear pathology than previously used latency-intensity series measures, without compromising specificity. It is suggested that simultaneous broadband masking should be considered as a test for localization of pathology in those subjects for whom routine behavioral measurements are not possible or when the results of such measurements are equivocal.