Signal-to-noise ratio and frequency analysis of continuous loop averaging deconvolution (CLAD) of overlapping evoked potentials

Quick Estimation of Minimum Hearing Levels Using a Binaural Multifrequency Stimulus Paradigm: Proof of Concept

OBJECTIVES

Objective estimation of minimum hearing levels using auditory brainstem responses (ABRs) elicited by single frequency tone-bursts presented monaurally is currently considered the gold standard. However, the data acquisition time to estimate thresholds (for both ears across four audiometric frequencies) using this method usually exceeds the sleep time (ranging between 35 and 49 minutes) in infants below 4 months, thus providing incomplete information of hearing status which in turn delays timely clinical intervention. Alternate approaches using faster rate, or tone-burst trains have not been readily accepted due to additional hardware and software requirements. We propose here a novel binaural multifrequency stimulation paradigm wherein several stimuli of different frequencies are presented binaurally in an interleaved manner. The rationale here is that the proposed paradigm will increase acquisition efficiency, significantly reduce test time, and improve accuracy by incorporating an automatic wave V detection algorithm. It is important to note that this paradigm can be easily implemented in most commercial ABR systems currently used by most clinicians.

DESIGN

Using this binaural multifrequency paradigm, ear specific ABRs were recorded in 30 normal-hearing young adults to both tone-bursts, and narrow-band (NB) iChirps at 500, 1000, 2000, and 4000 Hz. Comparison of ABRs elicited by tone-bursts and narrow-band chirps allowed us to determine if NB iChirps elicited a more robust wave V component compared with the tone-bursts. ABR data were characterized by measures of minimum hearing levels; wave V amplitude; and response detectability for two electrode configurations (high forehead-C7; and high forehead-linked mastoids).

RESULTS

Consistent with the research literature, wave V response amplitudes were relatively more robust for NB iChirp stimuli compared with tone-burst stimuli. The easier identification and better detectability of wave V for the NB iChirps at lower stimulus levels contributed to their better thresholds compared with tone-burst elicited responses. It is important to note that binaural multifrequency hearing levels close to minimum hearing levels were determined in approximately 22 minutes using this paradigm-appreciably quicker than the 45 to 60 minutes or longer time required for threshold determination using the conventional single frequency method.

CONCLUSIONS

Our novel and simple paradigm using either NB iChirps or tone-bursts provides a reliable method to rapidly estimate the minimum hearing levels across audiometric frequencies for both ears. Incorporation of an automatic wave V detection algorithm increases objectivity and further reduce test time and facilitate early hearing identification and intervention.

Multi-response deconvolution of auditory evoked potentials in a reduced representation space

The estimation of auditory evoked potentials requires deconvolution when the duration of the responses to be recovered exceeds the inter-stimulus interval. Based on least squares deconvolution, in this article we extend the procedure to the case of a multi-response convolutional model, that is, a model in which different categories of stimulus are expected to evoke different responses. The computational cost of the multi-response deconvolution significantly increases with the number of responses to be deconvolved, which restricts its applicability in practical situations. In order to alleviate this restriction, we propose to perform the multi-response deconvolution in a reduced representation space associated with a latency-dependent filtering of auditory responses, which provides a significant dimensionality reduction. We demonstrate the practical viability of the multi-response deconvolution with auditory responses evoked by clicks presented at different levels and categorized according to their stimulation level. The multi-response deconvolution applied in a reduced representation space provides the least squares estimation of the responses with a reasonable computational load. matlab/Octave code implementing the proposed procedure is included as supplementary material.

Subspace-constrained deconvolution of auditory evoked potentials

Auditory evoked potentials can be estimated by synchronous averaging when the responses to the individual stimuli are not overlapped. However, when the response duration exceeds the inter-stimulus interval, a deconvolution procedure is necessary to obtain the transient response. The iterative randomized stimulation and averaging and the equivalent randomized stimulation with least squares deconvolution have been proven to be flexible and efficient methods for deconvolving the evoked potentials, with minimum restrictions in the design of stimulation sequences. Recently, a latency-dependent filtering and down-sampling (LDFDS) methodology was proposed for optimal filtering and dimensionality reduction, which is particularly useful when the evoked potentials involve the complete auditory pathway response (i.e., from the cochlea to the auditory cortex). In this case, the number of samples required to accurately represent the evoked potentials can be reduced from several thousand (with conventional sampling) to around 120. In this article, we propose to perform the deconvolution in the reduced representation space defined by LDFDS and present the mathematical foundation of the subspace-constrained deconvolution. Under the assumption that the evoked response is appropriately represented in the reduced representation space, the proposed deconvolution provides an optimal least squares estimation of the evoked response. Additionally, the dimensionality reduction provides a substantial reduction of the computational cost associated with the deconvolution. matlab/Octave code implementing the proposed procedures is included as supplementary material.

Latency-dependent filtering and compact representation of the complete auditory pathway response

Auditory evoked potentials (AEPs) include the auditory brainstem response (ABR), middle latency response (MLR), and cortical auditory evoked potentials (CAEPs), each one covering a specific latency range and frequency band. For this reason, ABR, MLR, and CAEP are usually recorded separately using different protocols. This article proposes a procedure providing a latency-dependent filtering and down-sampling of the AEP responses. This way, each AEP component is appropriately filtered, according to its latency, and the complete auditory pathway response is conveniently represented (with the minimum number of samples, i.e., without unnecessary redundancies). The compact representation of the complete response facilitates a comprehensive analysis of the evoked potentials (keeping the natural continuity related to the neural activity transmission along the auditory pathway), which provides a new perspective in the design and analysis of AEP experiments. Additionally, the proposed compact representation reduces the storage or transmission requirements when large databases are manipulated for clinical or research purposes. The analysis of the AEP responses shows that a compact representation with 40 samples/decade (around 120 samples) is enough for accurately representing the response of the complete auditory pathway and provides appropriate latency-dependent filtering. MatLab/Octave code implementing the proposed procedure is included in the supplementary materials.

A Gap Prepulse with a Principal Stimulus Yields a Combined Auditory Late Response

BACKGROUND AND OBJECTIVES

The gap prepulse inhibition of the acoustic startle response has been used to screen tinnitus in an animal model. Here, we examined changes in the auditory late response under various conditions of gap prepulse inhibition.

SUBJECTS AND METHODS

We recruited 19 healthy adults (5 males, 14 females) and their auditory late responses were recorded after various stimuli with or without gap prepulsing. The N1 and P2 responses were selected for analysis. The gap prepulse inhibition was estimated to determine the optimal auditory late response in the gap prepulse paradigm.

RESULTS

We found that the gap per se generated a response that was very similar to the response elicited by sound stimuli. This critically affected the gap associated with the maximal inhibition of the stimulus response. Among the various gap-stimulus intervals (GSIs) between the gap and principal stimulus, the GSI of 150 ms maximally inhibited the response. However, after zero padding was used to minimize artifacts after a P2 response to a gap stimulus, the differences among the GSIs disappeared.

CONCLUSIONS

Overall, the data suggest that both the prepulse inhibition and the gap per se should be considered when using the gap prepulse paradigm to assess tinnitus in humans.

Matrix-based formulation of the iterative randomized stimulation and averaging method for recording evoked potentials

The iterative randomized stimulation and averaging (IRSA) method was proposed for recording evoked potentials when the individual responses are overlapped. The main inconvenience of IRSA is its computational cost, associated with a large number of iterations required for recovering the evoked potentials and the computation required for each iteration [involving the whole electroencephalogram (EEG)]. This article proposes a matrix-based formulation of IRSA, which is mathematically equivalent and saves computational load (because each iteration involves just a segment with the length of the response, instead of the whole EEG). Additionally, it presents an analysis of convergence that demonstrates that IRSA converges to the least-squares (LS) deconvolution. Based on the convergence analysis, some optimizations for the IRSA algorithm are proposed. Experimental results (configured for obtaining the full-range auditory evoked potentials) show the mathematical equivalence of the different IRSA implementations and the LS-deconvolution and compare the respective computational costs of these implementations under different conditions. The proposed optimizations allow the practical use of IRSA for many clinical and research applications and provide a reduction of the computational cost, very important with respect to the conventional IRSA, and moderate with respect to the LS-deconvolution. matlab/Octave implementations of the different methods are provided as supplementary material.

A BCI Gaze Sensing Method Using Low Jitter Code Modulated VEP

Visual evoked potentials (VEPs) are used in clinical applications in ophthalmology, neurology, and extensively in brain-computer interface (BCI) research. Many BCI implementations utilize steady-state VEP (SSVEP) and/or code modulated VEP (c-VEP) as inputs, in tandem with sophisticated methods to improve information transfer rates (ITR). There is a gap in knowledge regarding the adaptation dynamics and physiological generation mechanisms of the VEP response, and the relation of these factors with BCI performance. A simple, dual pattern display setup was used to evoke VEPs and to test signatures elicited by non-isochronic, non-singular, low jitter stimuli at the rates of 10, 32, 50, and 70 reversals per second (rps). Non-isochronic, low-jitter stimulation elicits quasi-steady-state VEPs (QSS-VEPs) that are utilized for the simultaneous generation of transient VEP and QSS-VEP. QSS-VEP is a special case of c-VEPs, and it is assumed that it shares similar generators of the SSVEPs. Eight subjects were recorded, and the performance of the overall system was analyzed using receiver operating characteristic (ROC) curves, accuracy plots, and ITRs. In summary, QSS-VEPs performed better than transient VEPs (TR-VEP). It was found that in general, 32 rps stimulation had the highest ROC area, accuracy, and ITRs. Moreover, QSS-VEPs were found to lead to higher accuracy by template matching compared to SSVEPs at 32 rps. To investigate the reasons behind this, adaptation dynamics of transient VEPs and QSS-VEPs at all four rates were analyzed and speculated.

Signal-to-noise ratio of auditory brainstem responses (ABRs) across click rate in the bottlenose dolphin (Tursiops truncatus)

Although the maximum length sequence (MLS) and iterative randomized stimulation and averaging (I-RSA) methods allow auditory brainstem response (ABR) measurements at high rates, it is not clear if high rates allow ABRs of a given quality to be measured in less time than conventional (CONV) averaging (i.e., fixed interstimulus intervals) at lower rates. In the present study, ABR signal-to-noise ratio (SNR) was examined in six bottlenose dolphins as a function of measurement time and click rate using CONV averaging at rates of 25 and 100 Hz and the MLS/I-RSA approaches at rates from 100 to 1250 Hz. Residual noise in the averaged ABR was estimated using (1) waveform amplitude following the ABR, (2) waveform amplitude after subtracting two subaverage ABRs (i.e., the "±average"), and (3) amplitude variance at a single time point. Results showed that high stimulus rates can be used to obtain dolphin ABRs with a desired SNR in less time than CONV averaging. Optimal SNRs occurred at rates of 500-750 Hz, but were only a few dB higher than that for CONV averaging at 100 Hz. Nonetheless, a 1-dB improvement in SNR could result in a 25% time savings in reaching criterion SNR.

Estimation of auditory steady-state responses based on the averaging of independent EEG epochs

The amplitude of auditory steady-state responses (ASSRs) generated in the brainstem of rats exponentially decreases over the sequential averaging of EEG epochs. This behavior is partially due to the adaptation of the ASSR induced by the continuous and monotonous stimulation. In this study, we analyzed the potential clinical relevance of the ASSR adaptation. ASSR were elicited in eight anesthetized adult rats by 8-kHz tones, modulated in amplitude at 115 Hz. We called independent epochs to those EEG epochs acquired with sufficiently long inter-stimulus interval, so the ASSR contained in any given epoch is not affected by the previous stimulation. We tested whether the detection of ASSRs is improved when the response is computed by averaging independent EEG epochs, containing only unadapted auditory responses. The improvements in the ASSR detection obtained with standard, weighted and sorted averaging were compared. In the absence of artifacts, when the ASSR was elicited by continuous acoustic stimulation, the computation of the ASSR amplitude relied upon the averaging method. While the adaptive behavior of the ASSR was still evident after the weighting of epochs, the sorted averaging resulted in under-estimations of the ASSR amplitude. In the absence of artifacts, the ASSR amplitudes computed by averaging independent epochs did not depend on the averaging procedure. Averaging independent epochs resulted in higher ASSR amplitudes and halved the number of EEG epochs needed to be acquired to achieve the maximum detection rate of the ASSR. Acquisition protocols based on averaging independent EEG epochs, in combination with appropriate averaging methods for artifact reduction might contribute to develop more accurate hearing assessments based on ASSRs.

Fast acquisition of full-range auditory event-related potentials using an interleaved deconvolution approach

This work relates to recent advances in the field of auditory event-related potentials (ERP), specifically deconvolution-based ERP acquisition and single-trial processing. An efficient stimulus sequence optimization method for ERP deconvolution is proposed, achieving consistent noise attenuation within a broad designated frequency range. Furthermore, a stimulus presentation paradigm for the fast, interleaved acquisition of auditory brainstem, middle-latency and late responses featuring alternating periods of high-rate deconvolution sequences, and subsequent low-rate stimulation is investigated in 20 normal hearing subjects. Deconvolved sequence responses containing early and middle-latency ERP components are fused with subsequent late responses using a time-frequency resolved weighted averaging method based on cross-trial regularity, yielding a uniform signal-to-noise ratio of the full-range auditory ERP across investigated timescales. Obtained average ERP waveforms exhibit morphologies consistent with both literature values and reference recordings acquired in 15 normal hearing subjects using a prior art approach to full-range auditory ERP acquisition, with all prominent waves being visible in the grand average waveforms. Results suggest the proposed interleaved stimulus presentation and associated ERP processing methodology to be suitable for the fast, reliable extraction of full-range auditory processing correlates in future ERP studies.

The Parallel Auditory Brainstem Response

The frequency-specific tone-evoked auditory brainstem response (ABR) is an indispensable tool in both the audiology clinic and research laboratory. Most frequently, the toneburst ABR is used to estimate hearing thresholds in infants, toddlers, and other patients for whom behavioral testing is not feasible. Therefore, results of the ABR exam form the basis for decisions regarding interventions and hearing habilitation with implications extending far into the child's future. Currently, responses are elicited by periodic sequences of toneburst stimuli presented serially to one ear at a time, which take a long time to measure multiple frequencies and intensities, and provide incomplete information if the infant wakes up early. Here, we describe a new method, the parallel ABR (pABR), which uses randomly timed toneburst stimuli to simultaneously acquire ABR waveforms to five frequencies in both ears. Here, we describe the pABR and quantify its effectiveness in addressing the greatest drawback of current methods: test duration. We show that in adults with normal hearing the pABR yields high-quality waveforms over a range of intensities, with similar morphology to the standard ABR in a fraction of the recording time. Furthermore, longer latencies and smaller amplitudes for low frequencies at a high intensity evoked by the pABR versus serial ABR suggest that responses may have better place specificity due to the masking provided by the other simultaneous toneburst sequences. Thus, the pABR has substantial potential for facilitating faster accumulation of more diagnostic information that is important for timely identification and treatment of hearing loss.

Comparison of maximum length sequence and randomized stimulation and averaging methods on the bottlenose dolphin auditory brainstem response

The purpose of the present study was to compare auditory brainstem responses (ABRs) using two approaches that allow the use of high stimulation rates, but with different temporal variability in the interstimulus interval: maximum length sequences (MLS) and iterative randomized stimulation and averaging (I-RSA). ABRs were obtained to click stimuli in six bottlenose dolphins (Tursiops truncatus). In experiment 1, click level was held constant and click rate varied from 25 to 1250 Hz. For MLS, interstimulus intervals varied by a factor of 6 at each rate, while for I-RSA the interstimulus intervals varied by ± 0.5 ms regardless of rate. In experiment 2, stimulus rates ranged from 100 to 1000 Hz and click level varied from 105 to 135 dB re: 1 μPa. For experiment 1, MLS and I-RSA showed similar decreases in ABR peak amplitudes and increases in ABR peak latencies and interwave intervals with increasing rate. For experiment 2, there was an increase in peak latency and a decrease in peak amplitude with decreasing click level; however, the effects of click level were reduced at higher rates. The results indicate that the greater jitter for MLS compared to I-RSA does not substantially affect the dolphin ABR.

Auditory brainstem, middle and late latency responses to short gaps in noise at different presentation rates

OBJECTIVE

The effects of rate on auditory-evoked potentials (AEP) to short noise gaps (12 ms) recorded at high sampling rates using wide-band filters were investigated.

DESIGN

Auditory brainstem (ABR), middle latency (MLR), late latency (LLR) and steady-state (ASSR) responses were simultaneously recorded in adult subjects at four gap rates (0.5, 1, 5 and 40 Hz). Major components (V, Na, Pa, Nb, Pb, N1 and P2) were identified at each rate and analysed for latency/amplitude characteristics. Gap responses at 40 Hz were recovered from Quasi-ASSRs (QASSR) using the CLAD deconvolution method.

STUDY SAMPLE

Fourteen right ears of young normal hearing subjects were tested.

RESULTS

All major components were present in all subjects at 1 Hz. P1 (P50) appeared as a low-pass filtered component of Pa and Pb waves. At higher rates, N1 and P2 disappeared completely while major ABR-MLR components were identified. Peak latencies were mostly determined by noise onsets slightly delayed by offset responses.

CONCLUSIONS

Major AEP components can be recorded to short gaps at 1 Hz using high sampling rates and wide-band filters. At higher rates, only ABR and MLRs can be recorded. Such simultaneous recordings may provide a complete assessment of temporal resolution and processing at different levels of auditory pathways.

Improved Transient Response Estimations in Predicting 40 Hz Auditory Steady-State Response Using Deconvolution Methods

The auditory steady-state response (ASSR) is one of the main approaches in clinic for health screening and frequency-specific hearing assessment. However, its generation mechanism is still of much controversy. In the present study, the linear superposition hypothesis for the generation of ASSRs was investigated by comparing the relationships between the classical 40 Hz ASSR and three synthetic ASSRs obtained from three different templates for transient auditory evoked potential (AEP). These three AEPs are the traditional AEP at 5 Hz and two 40 Hz AEPs derived from two deconvolution algorithms using stimulus sequences, i.e., continuous loop averaging deconvolution (CLAD) and multi-rate steady-state average deconvolution (MSAD). CLAD requires irregular inter-stimulus intervals (ISIs) in the sequence while MSAD uses the same ISIs but evenly-spaced stimulus sequences which mimics the classical 40 Hz ASSR. It has been reported that these reconstructed templates show similar patterns but significant difference in morphology and distinct frequency characteristics in synthetic ASSRs. The prediction accuracies of ASSR using these templates show significant differences (p < 0.05) in 45.95, 36.28, and 10.84% of total time points within four cycles of ASSR for the traditional, CLAD, and MSAD templates, respectively, as compared with the classical 40 Hz ASSR, and the ASSR synthesized from the MSAD transient AEP suggests the best similarity. And such a similarity is also demonstrated at individuals only in MSAD showing no statistically significant difference (Hotelling's T² test, T² = 6.96, F = 0.80, p = 0.592) as compared with the classical 40 Hz ASSR. The present results indicate that both stimulation rate and sequencing factor (ISI variation) affect transient AEP reconstructions from steady-state stimulation protocols. Furthermore, both auditory brainstem response (ABR) and middle latency response (MLR) are observed in contributing to the composition of ASSR but with variable weights in three templates. The significantly improved prediction accuracy of ASSR achieved by MSAD strongly supports the linear superposition mechanism of ASSR if an accurate template of transient AEPs can be reconstructed. The capacity in obtaining both ASSR and its underlying transient components accurately and simultaneously has the potential to contribute significantly to diagnosis of patients with neuropsychiatric disorders.

Bottlenose dolphin (Tursiops truncatus) auditory brainstem responses recorded using conventional and randomized stimulation and averaging

Auditory brainstem response (ABR) measurements using conventional averaging (i.e., constant interstimulus interval, ISI) are limited to stimulus rates low enough to prevent overlapping of the ABRs to successive stimuli. To overcome this limitation, stimuli may be presented at high rates using pseudorandom sequences (e.g., maximum length sequences) or quasi-periodic sequences; however, these methods restrict the available stimulus sequences and require deconvolution to recover the ABR from the overlapping responses. Randomized stimulation and averaging (RSA) is an alternate method where evoked responses at high rates are obtained by averaging responses to stimuli with ISIs drawn from a random distribution. The RSA method enables precise control over stimulus jitter, is flexible with respect to stimulus sequence parameters, and does not require deconvolution to extract the ABR waveform. In this study, ABRs were measured in three normal-hearing dolphins using conventional averaging and RSA. Results show the RSA method to be effective in dolphins if the ISI jitter ≥ ∼1.5 ms and that the influence of stimulus artifacts in the averaged ABR can be substantially reduced by alternating stimulus polarity on successive presentations rather than employing digital blanking or iterative processes.

The effects of click rate on the auditory brainstem response of bottlenose dolphins

Rate manipulations can be used to study adaptation processes in the auditory nerve and brainstem. For this reason, rate effects on the click-evoked auditory brainstem response (ABR) have been investigated in many mammals, including humans. In this study, click-evoked ABRs were obtained in eight bottlenose dolphins (Tursiops truncatus) while varying stimulus rate using both conventional averaging and maximum length sequences (MLSs), which allow disentangling ABRs that overlap in time and thus permit the study of adaptation at high rates. Dolphins varied in age and upper cutoff frequency of hearing. Conventional stimulation rates were 25, 50, and 100 Hz and average MLS rates were approximately 50, 100, 250, 500, 1000, 2500, and 5000 Hz. Click peak-equivalent sound pressure levels for all conditions were 135 dB re 1 μPa. ABRs were observed in all dolphins, at all stimulus rates. With increasing rate, peak latencies increased and peak amplitudes decreased. There was a trend for an increase in the interwave intervals with increasing rate, which appeared more robust for the dolphins with a full range of hearing. For those rates where ABRs were obtained for both conventional and MLS approaches, the latencies of the mean data were in good agreement.

New metric for optimizing Continuous Loop Averaging Deconvolution (CLAD) sequences under the 1/f noise model

Continuous loop averaging deconvolution (CLAD) is one of the proven methods for recovering transient auditory evoked potentials (AEPs) in rapid stimulation paradigms, which requires an elaborated stimulus sequence design to attenuate impacts from noise in data. The present study aimed to develop a new metric in gauging a CLAD sequence in terms of noise gain factor (NGF), which has been proposed previously but with less effectiveness in the presence of pink (1/f) noise. We derived the new metric by explicitly introducing the 1/f model into the proposed time-continuous sequence. We selected several representative CLAD sequences to test their noise property on typical EEG recordings, as well as on five real CLAD electroencephalogram (EEG) recordings to retrieve the middle latency responses. We also demonstrated the merit of the new metric in generating and quantifying optimized sequences using a classic genetic algorithm. The new metric shows evident improvements in measuring actual noise gains at different frequencies, and better performance than the original NGF in various aspects. The new metric is a generalized NGF measurement that can better quantify the performance of a CLAD sequence, and provide a more efficient mean of generating CLAD sequences via the incorporation with optimization algorithms. The present study can facilitate the specific application of CLAD paradigm with desired sequences in the clinic.

Estimation of a transient response from steady-state responses by deconvolution with built-in constraints

Evidence suggests that the steady-state response (SSR) elicited by a periodic train of auditory stimuli can largely be understood as a superposition of transient responses. This study is devoted to the problem of how to estimate that transient response from measured SSRs. The proposed method differs from previous approaches in that the solution can be constrained to be consistent with physiology-based prior knowledge or educated guesses. To achieve this goal, the transient response is not represented by a time series, but by a linear combination of auxiliary functions, called components. Constraints are introduced by assigning certain properties to the components. Only few parameters are required for that purpose, because the individual components are derived from a suitably designed mother component. After adjusting the components to the problem at hand, the component amplitudes are determined by optimizing the match between predicted and measured SSRs. This requires solving a linear inverse problem. A model simulation as well as an analysis of exemplary experimental data (auditory SSRs elicited by periodically presented clicks) prove the workability of the method. Since part of the theory is quite general, it would be relatively easy to refine and extend the method. Not only could responses other than SSRs be dealt with, it could also be realized that certain key parameters of the transient response, such as amplitude and delay, depend on stimulus repetition rate.

Comparison of Auditory Middle-Latency Responses From Two Deconvolution Methods at 40 Hz

GOAL

Auditory middle-latency responses (MLRs) are reported to be particularly susceptible to stimulation rate. Deconvolution methods are necessary to unwrap the overlapping responses at a high rate under the linear superposition assumption. This study aims to investigate and compare the MLR characteristics at high and conventional stimulation rates.

METHODS

The characteristics were examined in healthy adults by using two closely related deconvolution paradigms, namely continuous-loop averaging deconvolution and multirate steady-state averaging deconvolution at a mean rate of 40 Hz, and a conventional low rate of 5 Hz.

RESULTS

The morphology and stability of the MLRs can benefit from a high-rate stimulation. It appears that stimulation sequencing strategies of deconvolution methods exerted divergent rate effects on MLR characteristics, which might be associated with different adaptation mechanisms.

CONCLUSION

MLRs obtained by two deconvolution methods and the conventional reference feature differently from one another.

SIGNIFICANCE

These findings have critical implications in our current understanding of the rate effects on MLR characteristics which may inspire further studies to explore the characteristics of evoked responses at high rates and deconvolution paradigms.

Simultaneous acquisition of high-rate early, middle, and late auditory evoked potentials

Auditory evoked potentials (AEPs) are typically acquired at rates that facilitate their study as segregated by epochs relative to stimulus onset: early (ABR, 1.5-15 ms), middle (MLR, 15-60 ms), and late (LAEP, ≥60 ms) potentials. In particular, late AEPs are often acquired with stimulus repetition rates between 0.1 Hz and 1 Hz, and are bandpass filtered to contain information only within 1-30 Hz. These low repetition rates, filtering and low SNRs eliminate much of the potential contributions of the early and middle-latency responses in AEP recordings. This study aims to demonstrate a method for acquiring whole-AEP responses at higher stimulus repetition rates of 0.5 Hz to 10 Hz, by utilizing the Continuous Loop Averaging Deconvolution (CLAD) method, increasing the bandwidth of the recordings to 1-300 Hz to include early components, and using short-duration chirps to increase synchronous firing of the cochlear and auditory pathway neurons. Such a method may facilitate diagnostic or functional assessment of single AEP recordings for detection, identification, or evaluation of early, middle and late components of auditory responses.

Relationship between transient and steady-state pattern electroretinograms: theoretical and experimental assessment

PURPOSE

We determined if the overlap of transient (tr) pattern electroretinograms (PERG(tr)) can explain the generation of the steady-state (SS) pattern electroretinogram (PERG(SS)), and investigated the relationship between the two types of responses.

METHODS

Slightly jittered pattern reversals were used to generate quasi SS (QSS) PERG(SS) responses from eight normal subjects, recorded using lower eyelid skin electrodes, at rates between 6.9 and 26.5 reversals per second (rps). Jittered quasi PERG(SS) were deconvolved using the frequency domain continuous loop averaging deconvolution method. Additionally, conventional PERG(tr) at 2.2 rps and PERG(SS) at each of the QSS stimulation rates were obtained from all subjects. Two synthetic PERG(SS) responses were constructed at each stimulation rate, one using the PERG(tr) obtained at that rate, and the other using the conventional 2.2 rps PERG(tr). Synthetic responses then were compared to the recorded PERG(SS) using amplitude, latency, and spectral measurements.

RESULTS

Findings indicate that the PERG(SS) obtained at SS rates can be predicted using the superposition of deconvolved tr PERGs at each particular rate. Although conventional PERG(tr) can explain PERG(SS) obtained at rates below 15.4 rps (≥ 97% correlation), for higher reversal rates only deconvolved responses obtained at that rate can produce the recorded SS responses (96% vs. 65% correlation at 26.5 rps).

CONCLUSIONS

The study shows that PERG(SS) results from the overlapping of tr PERG(tr) waveforms generated at that reversal rate. The first two peaks (N(SS) and P(SS)) of the PERG(SS) reflect N35 and P50 waves of the tr PERG(tr). The N95 amplitude is reduced at conventional (16 rps) SS rates, but contributes to the overall PERG(SS) amplitude.

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.

Automatic quality assessment and peak identification of auditory brainstem responses with fitted parametric peaks

The recording of the auditory brainstem response (ABR) is used worldwide for hearing screening purposes. In this process, a precise estimation of the most relevant components is essential for an accurate interpretation of these signals. This evaluation is usually carried out subjectively by an audiologist. However, the use of automatic methods for this purpose is being encouraged nowadays in order to reduce human evaluation biases and ensure uniformity among test conditions, patients, and screening personnel. This article describes a new method that performs automatic quality assessment and identification of the peaks, the fitted parametric peaks (FPP). This method is based on the use of synthesized peaks that are adjusted to the ABR response. The FPP is validated, on one hand, by an analysis of amplitudes and latencies measured manually by an audiologist and automatically by the FPP method in ABR signals recorded at different stimulation rates; and on the other hand, contrasting the performance of the FPP method with the automatic evaluation techniques based on the correlation coefficient, FSP, and cross correlation with a predefined template waveform by comparing the automatic evaluations of the quality of these methods with subjective evaluations provided by five experienced evaluators on a set of ABR signals of different quality. The results of this study suggest (a) that the FPP method can be used to provide an accurate parameterization of the peaks in terms of amplitude, latency, and width, and (b) that the FPP remains as the method that best approaches the averaged subjective quality evaluation, as well as provides the best results in terms of sensitivity and specificity in ABR signals validation. The significance of these findings and the clinical value of the FPP method are highlighted on this paper.

Unwrapping of transient responses from high rate overlapping pattern electroretinograms by deconvolution

OBJECTIVE

Due to overlapping, temporal information is mostly lost in high rate steady-state pattern electroretinograms (PERGSS). This study develops a deconvolution method and a display/recording system to "unwrap" PERGSS and obtain a transient, "per stimulus" response (PERGtr) regardless of reversal rate.

METHODS

Processing and instrumentation, including high temporal resolution display and acquisition were developed for deconvolving PERGs acquired at high rates by slight jittering of reversal onsets at a given mean rate.

RESULTS

The system was successfully tested at eight rates from 2.2 to 78.1rps. At medium rates (17.4-41.2rps) recordings with conventional morphology (N35-P50-N95) but earlier peaks and higher amplitudes were extracted up to 40rps. At higher rates, smaller triphasic responses were obtained, exhibiting similar peak latencies, but reversed polarity. Oscillating potentials (OPs) were also recorded at all rates after deconvolution.

CONCLUSIONS

Transient PERGs and OPs can be extracted from quasi steady-state PERG recordings obtained at high rates with a deconvolution algorithm using high temporal resolution display and acquisition systems.

SIGNIFICANCE

The methodology to extract transient and oscillatory responses from steady-state PERGs could be useful in understanding high rate responses and diagnosis of various retinal diseases by revealing temporal information on waveform components which cannot be normally observed.

Deconvolution of high rate flicker electroretinograms

Flicker electroretinograms are steady-state electroretinograms (ERGs) generated by high rate flash stimuli that produce overlapping periodic responses. When a flash stimulus is delivered at low rates, a transient response named flash ERG (FERG) representing the activation of neural structures within the outer retina is obtained. Although FERGs and flicker ERGs are used in the diagnosis of many retinal diseases, their waveform relationships have not been investigated in detail. This study examines this relationship by extracting transient FERGs from specially generated quasi steady-state flicker and ERGs at stimulation rates above 10 Hz and similarly generated conventional flicker ERGs. The ability to extract the transient FERG responses by deconvolving flicker responses to temporally jittered stimuli at high rates is investigated at varying rates. FERGs were obtained from seven normal subjects stimulated with LED-based displays, delivering steady-state and low jittered quasi steady-state responses at five rates (10, 15, 32, 50, 68 Hz). The deconvolution method enabled a successful extraction of "per stimulus" unit transient ERG responses for all high stimulation rates. The deconvolved FERGs were used successfully to synthesize flicker ERGs obtained at the same high stimulation rates.

Continuous- and discrete-time stimulus sequences for high stimulus rate paradigm in evoked potential studies

To obtain reliable transient auditory evoked potentials (AEPs) from EEGs recorded using high stimulus rate (HSR) paradigm, it is critical to design the stimulus sequences of appropriate frequency properties. Traditionally, the individual stimulus events in a stimulus sequence occur only at discrete time points dependent on the sampling frequency of the recording system and the duration of stimulus sequence. This dependency likely causes the implementation of suboptimal stimulus sequences, sacrificing the reliability of resulting AEPs. In this paper, we explicate the use of continuous-time stimulus sequence for HSR paradigm, which is independent of the discrete electroencephalogram (EEG) recording system. We employ simulation studies to examine the applicability of the continuous-time stimulus sequences and the impacts of sampling frequency on AEPs in traditional studies using discrete-time design. Results from these studies show that the continuous-time sequences can offer better frequency properties and improve the reliability of recovered AEPs. Furthermore, we find that the errors in the recovered AEPs depend critically on the sampling frequencies of experimental systems, and their relationship can be fitted using a reciprocal function. As such, our study contributes to the literature by demonstrating the applicability and advantages of continuous-time stimulus sequences for HSR paradigm and by revealing the relationship between the reliability of AEPs and sampling frequencies of the experimental systems when discrete-time stimulus sequences are used in traditional manner for the HSR paradigm.

Simultaneously extracted transient and steady-state evoked responses during general anesthesia: variability of different rates

Unintended intraoperative awareness occurs in one to two individuals out of every one thousand treated with general anesthesia. Patients that experience intraoperative awareness have significant post-operative psychological sequelae. The ability to detect intraoperative awareness is currently suboptimal because the mechanism employed by anesthetic drugs to impair consciousness remains poorly understood. Studies have suggested that evoked potentials (EP) may be used to monitor the depth of anesthesia. Both transient and steady state responses can be simultaneously extracted using the Continuous Loop Averaging Deconvolution (CLAD) method with specially designed CLAD sequences. 20 Hz and 30 Hz jittered CLAD sequences in addition to 5 Hz isochronic and 40 Hz jittered CLAD sequences were applied in baseline awake and general anesthesia conditions. A qualitative method to assess the extracted EPs was developed in this study, termed Randomized Split Set Average (RSSA). The results showed that EPs extracted during general anesthesia require a greater number of sweeps to obtain a signal-to-noise ratio comparable to that observed in EPs extracted during the awake state. Therefore, the development of a real time or quasi real time EP monitoring system for anesthesia provides an increased challenge. The RSSA employed in this study is a useful method for assessing the signal quality of EP responses.

Recording of auditory brainstem response at high stimulation rates using randomized stimulation and averaging

The recording of auditory brainstem response (ABR) at high stimulation rates is of great interest in audiology. It allows a more accurate diagnosis of certain pathologies at an early stage and the study of different mechanisms of adaptation. This paper proposes a methodology, which we will refer to as randomized stimulation and averaging (RSA) that allows the recording of ABR at high stimulation rates using jittered stimuli. The proposed method has been compared with quasi-periodic sequence deconvolution (QSD) and conventional (CONV) stimulation methodologies. Experimental results show that RSA provides a quality in ABR recordings similar to that of QSD and CONV. Compared with CONV, RSA presents the advantage of being able to record ABR at rates higher than 100 Hz. Compared with QSD, the formulation of RSA is simpler and allows more flexibility on the design of the pseudorandom sequence. The feasibility of the RSA methodology is validated by an analysis of the morphology, amplitudes, and latencies of the most important waves in ABR recorded at high stimulation rates from eight normal hearing subjects.

Innovative pattern reversal displays for visual electrophysiological studies

Pattern Reversal (PR) stimulation is a frequently used tool in the evaluation of the visual pathway. The PR stimulus consists of a field of black and white segments (usually checks or bars) of constant luminance, which change phase (black to white and white to black) at a given reversal rate. The Pattern Electroretinogram (PERG) is a biological potential that is evoked from the retina upon viewing PR display. Likewise, the Pattern Visual Evoked Potential (PVEP) is a biological potential recorded from the occipital cortex when viewing a PR display. Typically, PR stimuli are presented on a Cathode Ray Tube (CRT) or Liquid Crystal Display (LCD) monitor. This paper presents three modalities to generate pattern reversal stimuli. The three methods are as follows: a display consisting of array of Light Emitting Diodes (LEDs), a display comprised of two miniature projectors, and a display utilizing a modified LCD display in conjunction with a variable polarizer. The proposed stimulators allow for the recording of PERG and PVEP waveforms at much higher rates than are capable with conventional stimulators. Additionally, all three of the alternative PR displays will be able to take advantage of advanced analysis techniques, such as the recently developed Continuous Loop Averaging Deconvolution (CLAD) algorithm.

Auditory steady-state responses to 40-Hz click trains: relationship to middle latency, gamma band and beta band responses studied with deconvolution

OBJECTIVE

The nature of the auditory steady-state responses (ASSR) evoked with 40-Hz click trains and their relationship to auditory brainstem and middle latency responses (ABR/MLR), gamma band responses (GBR) and beta band responses (BBR) were investigated using superposition theory. Transient responses obtained by continuous loop averaging deconvolution (CLAD) and last click responses (LCR) were used to synthesize ASSRs and GBRs.

METHODS

ASSRs were obtained with trains of low jittered 40 Hz clicks presented monaurally and deconvolved using a modified CLAD. Resulting transient responses and modified LCRs were used to predict the ASSRs and the GBR.

RESULTS

The ABR/MLR obtained with deconvolution predicted accurately the steady portion of the ASSR but failed to predict its onset portion. The modified LCR failed to fully predict both portions. The GBRs were predicted by narrow band filtering of the ASSRs. Significant BBR activity was found both in the ASSRs and deconvolved ABR/MLRs.

CONCLUSIONS

Simulations using deconvolved ABR/MLRs obtained at 40 Hz predict fully the steady state but not the onset portion of the ASSRs, thus confirming the superposition theory.

SIGNIFICANCE

Click rate adaptation plays a significant role in ASSR generation with click trains and should be considered in evaluating convolved response generation theories.

Signal-to-noise ratio of intraoperative tibial nerve somatosensory-evoked potentials

To reveal the intrinsic signal-to-noise ratio (SNR) of single-trial somatosensory-evoked potentials (SEP). SEP was recorded from 13 scoliosis patients during surgery. The power of SEP was estimated with least-square fitting to obtain the most accurate value and then to estimate the SNR of every trial of SEP. The SNR of cortical SEP from 13 cases presented individual difference among each other. According to the mean and standard deviation, the coefficients of variation of cortical and subcortical SEP were 4.2% and 23%, respectively. The SNR of SEP was estimated to be -24 +/- 1 dB in cortical SEP and -22 +/- 5 dB in subcortical SEP. The lowest SNR of individual case was found to be -30 dB in cortical SEP and -53 dB in subcortical SEP. The results showed that SNR of intraoperative SEP recordings varies from person to person and presents a higher variability in subcortical than that in cortical, with a broad range from -53 to -5 dB. The results from this study can be used to understand the nature of SEP signals, which could guide researchers and designers on SEP denoising method selection, extraction, and measurement, as well as equipment development.

Generation of the 40-Hz auditory steady-state response (ASSR) explained using convolution

OBJECTIVE

In this study, the superposition theory of the 40-Hz auditory steady-state response (ASSR) generation is investigated using auditory brainstem response (ABR) and middle latency responses (MLRs) obtained with 40 Hz jittered sequences with the continuous loop averaging deconvolution (CLAD) algorithm.

METHODS

Click sequences at around 40 Hz with high (maximum length sequence), medium and low jitters were presented to normal hearing awake adult subjects monaurally. Overlapping MLR responses were deconvolved using the frequency domain CLAD method. In addition, conventional auditory MLRs at 4.88 Hz and ASSRs at 39.1 Hz were obtained in all subjects. Synthetic ASSRs were constructed using different rate and jitter MLRs as base recordings. Contributions of the primary components were investigated by wave elimination using phasors.

RESULTS

Findings indicate that the generation of the 40-Hz ASSRs can be explained successfully by the superposition of the ABR and MLR waves generated at that stimulation rate. N(a)-P(a) and N(b)-P(b) components of the MLR contribute about equally (45% each), while the wave V of the ABR contributes a lesser amount (10%).

CONCLUSIONS

Forty-Hertz ASSRs are composite responses generated by the superposition of the major waves of the ABR and the MLR. Dramatic amplitude increase of the ASSR at 40Hz is primarily due to the superposition of the resonating P(b) component to the P(a) wave.

SIGNIFICANCE

Several unexplained properties of the 40-Hz ASSR can be explained by the stimulus and brain state dependent characteristics of the slow ABR, the P(a) and the P(b) components of the MLR.

Methodology to estimate the transient evoked responses for the generation of steady state responses

A method to acquire transient evoked responses at high rates, corresponding to traditional steady state responses (SSR) is developed. Continuous Loop Averaging Deconvolution (CLAD) method is used in conjunction with tailored, low-jitter stimulation sequences. A physiological brain convolution model for SSR generation is adapted and mathematically analyzed. A SSR synthesis method from acquired transient evoked potentials is proposed and implemented. The mathematical models are used to guide the stimulation sequence design method. Visual evoked potentials (VEP) at 10 Hz and auditory evoked responses (AER) at 40 Hz and 80 Hz are acquired using the specially designed and the traditional SSR sequences. Acquired and synthetically generated SSRs are then compared in time and frequency domains to asses the method consistency. The experimental results show an excellent agreement between the acquired and synthetic SSR in all three modalities.

Acquisition and analysis of high rate deconvolved auditory evoked potentials during sleep

Auditory Evoked Potentials (AEPs) have been recorded at high stimulus rates during sleep using Continuous Loop Averaging Deconvolution (CLAD) sequences. AEP transient signals are obtained via frequency domain deconvolution of overlapped responses. Simultaneous acquisition of Auditory Brainstem Response (ABR), Middle Latency Response (MLR), and Long Latency Response (LLR) is obtained at an average stimulation rate of 39.1 Hz, using 10, 20 and 100 second electroencephalography (EEG) recordings. Deconvolved responses confirm previous observations on the reduction and disappearance of the P1 MLR component during stage III and IV, obtained with standard averaging and stimulation methods. Results indicate that auditory stimulation at high rates during sleep, using short time sweeps, may help correlating the sleep EEG indicative of different arousal levels, with corresponding AEPs.

Pb(P1) resonance at 40Hz: Effects of high stimulus rate on auditory middle latency responses (MLRs) explored using deconvolution

OBJECTIVE

In this study, the effects of high stimulus rate on middle latency response (MLR) components P(a) and P(b) (P(1) or P50) were studied using high rate clicks in normal hearing awake subjects were investigated.

METHODS

Five jittered click sequences at different mean rates (24.4, 39.1, 58.6, 78.1, 97.7Hz) were presented to 10 subjects. Overlapping MLRs were deconvolved using the frequency domain continuous loop averaging deconvolution (CLAD) [Ozdamar O., Bohórquez, J., Signal to noise ratio and frequency analysis of continuous loop averaging deconvolution (CLAD) of overlapping evoked potentials. J. Acoust. Soc. Am., 119:429-438, 2006]. In addition conventional auditory transient MLRs at 4.88Hz were obtained using conventional averaging.

RESULTS

P(a) amplitude, latency and waveshape remained fairly constant up to 78.1Hz. P(b) component, however, showed a variable waveshape with latencies covering a wide range (50-70ms) and N(b)-P(b) amplitudes increasing at 39.1 and 58.6Hz and decreasing at other rates.

CONCLUSIONS

Recordings show that both P(a) and P(b) MLR components can be consistently recorded at all rates up to 100Hz. P(b) amplitude shows an increase at around 40Hz showing a resonance at that frequency.

SIGNIFICANCE

The dramatic increase of the P(b) component at around 40Hz may account for the high amplitude of the 40Hz ASSR.