Selective processing of auditory evoked responses with iterative-randomized stimulation and averaging: A strategy for evaluating the time-invariant assumption.

The recording of auditory evoked potentials (AEPs) at fast rates allows the study of neural adaptation, improves accuracy in estimating hearing threshold and may help diagnosing certain pathologies. Stimulation sequences used to record AEPs at fast rates require to be designed with a certain jitter, i.e., not periodical. Some authors believe that stimuli from wide-jittered sequences may evoke auditory responses of different morphology, and therefore, the time-invariant assumption would not be accomplished. This paper describes a methodology that can be used to analyze the time-invariant assumption in jittered stimulation sequences. The proposed method [Split-IRSA] is based on an extended version of the iterative randomized stimulation and averaging (IRSA) technique, including selective processing of sweeps according to a predefined criterion. The fundamentals, the mathematical basis and relevant implementation guidelines of this technique are presented in this paper. The results of this study show that Split-IRSA presents an adequate performance and that both fast and slow mechanisms of adaptation influence the evoked-response morphology, thus both mechanisms should be considered when time-invariance is assumed. The significance of these findings is discussed.

Auditory brainstem and middle latency responses recorded at fast rates with randomized stimulation.

Randomized stimulation and averaging (RSA) allows auditory evoked potentials (AEPs) to be recorded at high stimulation rates. This method does not perform deconvolution and must therefore deal with interference derived from overlapping transient evoked responses. This paper analyzes the effects of this interference on auditory brainstem responses (ABRs) and middle latency responses (MLRs) recorded at rates of up to 300 and 125 Hz, respectively, with randomized stimulation sequences of a jitter both greater and shorter than the dominant period of the ABR/MLR components. Additionally, this paper presents an advanced approach for RSA [iterative-randomized stimulation and averaging (I-RSA)], which includes the removal of the interference associated with overlapping responses through an iterative process in the time domain. Experimental results show that (a) RSA can be efficiently used in the recording of AEPs when the jitter of the stimulation sequence is greater than the dominant period of the AEP components, and (b) I-RSA maintains all the advantages of RSA and is not constrained by the restriction of a minimum jitter. The significance of the results of this study is discussed.

A study of adaptation mechanisms based on ABR recorded at high stimulation rate.

OBJECTIVE: This paper analyzes the fast and slow mechanisms of adaptation through a study of latencies and amplitudes on ABR recorded at high stimulation rates using the randomized stimulation and averaging (RSA) technique. METHODS: The RSA technique allows a separate processing of auditory responses, and is used, in this study, to categorize responses according to the interstimulus interval (ISI) of their preceding stimulus. The fast and slow mechanisms of adaptation are analyzed by the separated responses methodology, whose underlying principles and mathematical basis are described in detail. RESULTS: The morphology of the ABR is influenced by both fast and slow mechanisms of adaptation. These results are consistent with previous animal studies based on spike rate. CONCLUSIONS: Both fast and slow mechanisms of adaptation are present in all subjects. In addition, the distribution of the jitter and the sequencing of the stimuli may be critical parameters when obtaining reliable ABRs. SIGNIFICANCE: The separated responses methodology enables for the first time the analysis of the fast and slow mechanisms of adaptation in ABR obtained at stimulation rates greater than 100 Hz. The non-invasive nature of this methodology is appropriate for its use in humans.

The effect of stimulus rate and time-separation on Volterra slices of otoacoustic emissions.

Volterra slices (VSs) of otoacoustic emissions are temporal non-linear interaction components which can be measured using the maximum length sequence technique of stimulation. Previous studies have found, but not explained, non-monotonic variations in the amplitude of VSs when stimulus rate is increased. In this study, a simple phenomenological model is investigated which provided possible insights into the effect of rate on VS amplitudes. Resulting theoretical considerations suggest that the effect of rate on VS amplitude is best examined when the time-separation parameter of the VS is held constant. To test these suggestions, data on VSs of order 2 and 3 were measured in 24 normal hearing ears in which the rate is varied while holding constant the time-separation. Under these conditions, the results reveal a monotonic reduction in the amplitude of the VSs with increasing rate. The phenomenological model offers a possible explanation of some of these results in terms of the derivatives of the input-output function of the non-linearity. In addition, measured cross-correlations between waveforms of VS of different order and slice number were obtained, revealing a complex dependence on time-separation that has not been explained.

Evaluation of a technique to measure latency jitter in event-related potentials.

Auditory selective attention results in larger event-related potentials (ERPs) than those recorded to unattended stimuli. Larger ERPs arise from either a greater number of neurons being stimulated or the same number of neurons with better synchrony. The synchrony aspect was studied in an attention experiment recording the N1 response and showed that the latency jitter was significantly less (p=0.002) for the responses to attended stimuli compared with those to unattended stimuli. Here the improved technique to measure latency jitter is the focus of study. Simulated responses were used to evaluate the technique's performance at different signal-to-noise ratios (SNRs) and compare it to the standard technique. Checks that the technique had performed satisfactorily in each case were achieved and a measure of reliability for each result was developed. The parameters defining the technique were varied and the optimum values chosen. This new technique opens the way for researchers to investigate the latency properties of lower SNR ERPs and gives a new insight into auditory selective attention.

Click-evoked otoacoustic emissions (CEOAEs) recorded from neonates under 13 hours old using conventional and maximum length sequence (MLS) stimulation.

Maximum length sequence (MLS) stimulation allows click evoked otoacoustic emissions (CEOAEs) to be averaged at very high stimulation rates. This enables a faster reduction of noise contamination of the response, and has been shown to improve the signal-to-noise ratio (SNR) of CEOAEs recorded from adult subjects. This study set out to investigate whether MLS averaging can enhance the SNR of CEOAEs recorded in newborns within the first day after birth, and so improve the pass rates for OAE screening in this period, when false alarm rates are very high. CEOAEs were recorded in a neonatal ward from 57 ears in 37 newborns ranging from 6 to 13h old, using both conventional (50/s) and high rate (5000/s) MLS averaging. SNR values and pass rates were compared for responses obtained within equal recording times at both rates. MLS averaging produced an SNR improvement of up to 3.8dB, with the greatest improvement found in higher frequency bands. This SNR advantage resulted in pass rate improvement between 5% and 10%, depending on pass criterion. A significant effect of age was found on both SNR and pass rate, with newborns between 6 and 10h old showing significantly lower values than those tested between 10 and 13h after birth, as well as a much greater improvement due to MLS averaging. The findings show that MLS averaging can reduce false alarm rates by up to 15% in very young neonates in a neonatal ward setting.

Selective attention increases the temporal precision of the auditory N100 event-related potential.

Selective attention increases the amplitude of the averaged N100 event-related potential (ERP). This increase may result from more neurons responding to the stimulus or from the same number of neurons better synchronised with the stimulus, or both. We investigated the synchronization mechanism using a new response latency jitter measurement algorithm that performed well for all the signal-to-noise ratios obtained in the experiment. We found that the significantly increased N100 amplitude is accounted for by a significantly decreased latency jitter variance for the attended stimuli.

Effects of maturation on parameters used for pass/fail criteria in neonatal hearing screening programmes using evoked otoacoustic emissions.

We aimed to investigate the incidence of false alarms that occurred with the pass/fail criteria used in a published series of neonatal hearing screening programmes, as a function of age. We analysed the database of 19137 normally hearing babies (38274 ears) tested in the Wessex Universal Neonatal Hearing Screening Project. Otoacoustic emissions were recorded prior to discharge from maternity units, using IL088 equipment. We assessed the pass/fail rate using the Wessex criteria and 10 other pass/fail criteria published in the literature. Using Pearson's correlation coefficient, a statistically significant correlation between signal-to-noise ratio at each of the frequency bands 1, 2, 3, 4 and 5 kHz and babies' age in hours at the 0.01 level was identified. The correlation was also significant (0.01 level) between age and frequency reproducibility in each of the bands at 1, 2, 3, 4 and 5 kHz as well as the whole reproducibility. The number of false alarms reduced significantly after the first 24 h of life with all the criteria examined. We conclude that in the first hours after birth due to insufficient maturation of the otoacoustic emission, there is a high rate of false alarms. This increase in the false alarm rate, whilst dependent on the criteria used, occurs with all criteria. This leads to the consideration of whether the establishment of age-dependent pass/fail criteria could reduce the false alarm rate and the subsequent strain on diagnostic centres.

Volterra Slice otoacoustic emissions recorded using maximum length sequences from patients with sensorineural hearing loss.

When normally hearing ears are stimulated with maximum length sequences (MLS) of clicks, a family of non-linear temporal interaction components of otoacoustic emissions (OAEs) can be derived, which have been named Volterra Slice OAEs (VS OAEs). This study investigates the sensitivity of VS OAEs to sensorineural hearing impairment in adults, compared to that of the widely used derived non-linear click evoked OAE (DNL CEOAE). VS OAEs and DNL CEOAEs were obtained from 24 normally hearing and 24 hearing impaired ears using a custom-built MLS system and a Otodynamics 'ILO88' OAE Analyzer, respectively. The results show that, based on waveform reproducibility, VS OAEs are as successful as DNL CEOAEs at separating normal from impaired ears at the audiometric frequencies of 1 and 2 kHz, where a strong correlation is found between the amplitudes of the two OAE types. At 4 kHz however, VS OAEs are a significantly better indicator of hearing loss than DNL CEOAEs. This difference at 4 kHz appears to be due to the lack of stimulus artefact contamination of VS OAEs in the early, high frequency portion of the response. The findings suggest that VS OAEs may provide a better diagnostic and monitoring tool for hearing loss at high frequencies than the conventional DNL CEOAE.

Transient evoked otoacoustic emissions recorded using maximum length sequences from patients with sensorineural hearing loss.

Much research has shown that transient evoked otoacoustic emissions (TEOAEs) can successfully separate normally hearing and hearing impaired populations. However, this finding comes from TEOAEs recorded using conventional averaging at low stimulation rates. Presenting clicks according to maximum length sequences (MLSs) enables TEOAEs to be recorded at very high stimulation rates. This study compares conventional and MLS TEOAEs in normally hearing and hearing impaired adults. Stimulus presentation rates of 40 clicks/s (conventional) and 5000 clicks/s (MLS) were used. The 'linear' TEOAEs (i.e., the directly recorded waveforms), the 'level nonlinear' (LNL) TEOAEs (i.e., those derived from two linear waveforms separated by a known difference in stimulus level) and the 'rate nonlinear' (RNL) TEOAEs (i.e., obtained by subtracting the emission recorded at 5000 clicks/s from that at 40 clicks/s at a fixed stimulus level) were examined to compare how they separated the normally hearing and hearing impaired subjects. When compared to the results for both conventional and MLS linear or LNL TEOAEs, the present study found that the RNL results best reflected the patients' hearing loss, although the conventional linear and LNL responses performed nearly as well. Only two impaired ears (2%), both with a best threshold of 30 dB HL at 1000 Hz, produced RNL responses with amplitude within the range produced by 95% of the normal group.

The order of testing effect in otoacoustic emissions and its consequences for sex and ear differences in neonates.

The amplitude values of transient-evoked otoacoustic emissions, recorded from a large sample of neonates, were used to examine the asymmetry between ears tested and the differences due to the sex of the subject. Whilst the sex difference, with females having larger responses than males, has been a consistent finding in previous reports, the right/left ear difference, with the right ear giving a larger response than the left, has produced variable results that differed between laboratories. In this study, the sex difference was confirmed with females giving a 1.2 dB greater response than males. It was not affected by the age of the neonate. A significant effect of test order was found. The measured right/left difference was enhanced when the right ear was tested first but was diminished when the left ear was tested first. If the left ear is tested first then the measured right/left difference would be about 0.5 dB whereas, if the right ear is tested first, the measured right/left difference would be about 1.5 dB. When male/female comparisons were made for right and left ears separately and for the same ear tested first, the sex differences were the same for all four conditions. The sex and right/left differences have been confirmed as statistically significant effects and the order effect could explain the discrepancies and variability of the right/left differences reported in the literature.