Use of the matching pursuit algorithm with a dictionary of asymmetric waveforms in the analysis of transient evoked otoacoustic emissions

Characteristic of Mayer Waves in Electrophysiological, Hemodynamic and Vascular Signals

We evaluated the properties of oscillations in the Mayer waves (MW) frequency range (∼0.1Hz) detected in blood pressure, heart rate variability, cerebral blood oxygenation changes and evolution of electroencephalographic (EEG) rhythms to elucidate the mechanisms of MW generation. We examined the persistence of MW in different signals and stability of their oscillations on the level of individual MW waveforms, which was achieved by applying matching pursuit (MP). MP yields adaptive time-frequency approximation of signal's structures in terms of frequency, amplitude, time occurrence, and time-span. The number of waveforms contributing to 95% of the energy of the signals was vastly different for the time series, but the average number of waveforms conforming to the MW criteria was almost the same (3.5 ± 0.4 per 120s epoch). In all the investigated signals, MW had the same distributions of frequency and the number of cycles. We show that the MW energy ratios in different signals varied strongly, p < 0.001. The highest percentage of MW energy was observed in blood pressure signals, heart rate variability, and reduced hemoglobin, in contrast to brain signals and oxygenated hemoglobin. The percentage of MW energy was related to the strength of causal influence exerted by them on the other signals. Our results indicate existence of a common mechanism of MW generation and support the hypothesis of MW generation in the baroreflex loop; however, they do not exclude the action of a central pacemaker.

Analyzing transient-evoked otoacoustic emissions by concentration of frequency and time

The linear part of transient evoked otoacoustic emission (TEOAE) is thought to be generated via coherent reflection near the characteristic place of constituent wave components. Because of the tonotopic organization of the cochlea, high frequency emissions return earlier than low frequencies; however, due to the random nature of coherent reflection, the instantaneous frequency (IF) and amplitude envelope of TEOAEs both fluctuate. Multiple reflection components and synchronized spontaneous emissions can further make it difficult to extract the IF by linear transforms. This paper proposes to model TEOAEs as a sum of intrinsic mode-type functions and analyze it by a nonlinear-type time-frequency (T-F) analysis technique called concentration of frequency and time (ConceFT). When tested with synthetic otoacoustic emission signals with possibly multiple oscillatory components, the present method is able to produce clearly visualized traces of individual components on the T-F plane. Further, when the signal is noisy, the proposed method is compared with existing linear and bilinear methods in its accuracy for estimating the fluctuating IF. Results suggest that ConceFT outperforms the best of these methods in terms of optimal transport distance, reducing the error by 10% to 21% when the signal to noise ratio is 10 dB or below.

The 1.06 frequency ratio in the cochlea: evidence and outlook for a natural musical semitone

A frequency ratio of about 1.06 often appears in cochlear mechanics, and the question naturally arises, why? The ratio is close to that of the semitone (1.059) in music, giving reason to think that this aspect of musical perception might have a cochlear basis. Here, data on synchronised spontaneous otoacoustic emissions is presented, and a clustering of ratios between 1.05 and 1.07 is found with a peak at 1.063 ± 0.005. These findings reinforce what has been found from previous sources, which are reviewed and placed alongside the present work. The review establishes that a peak in the vicinity of 1.06 has often been found in human cochlear data. Several possible cochlear models for explaining the findings are described. Irrespective of which model is selected, the fact remains that the cochlea itself appears to be the origin of a ratio remarkably close to an equal-tempered musical semitone, and this close coincidence leads to the suggestion that the inner ear may play a role in constructing a natural theory of music. The outlook for such an enterprise is surveyed.

Spontaneous otoacoustic emissions in schoolchildren

OBJECTIVE

Spontaneous otoacoustic emissions (SOAEs) are one of the least studied types of otoacoustic emissions (OAEs). The purpose of this study was twofold: first, to determine the prevalence of SOAEs in schoolchildren, and second to test whether there was dependence between the presence or absence of SOAEs in a subject and the corresponding level of their transiently evoked OAEs (TEOAEs).

METHODS

Measurements were made on a group of normally hearing children of age 7-13 years. A technique which detects synchronized SOAEs (SSOAEs) was used in which the response to repetitive clicks (12.5/s) was analyzed in the 60-80 ms time window following each click. The matching pursuit method was used to detect SSOAEs components above the noise in this window. For comparison, TEOAEs evoked by clicks (40/s) were obtained using the standard nonlinear protocol (20 ms time window).

RESULTS

The prevalence of SOAEs was 37%, and higher in females and right ears. There was an average of 2.3 SOAEs per emitting ear. TEOAE levels were higher for ears that had SOAEs and were lower for ears that did not have any SOAEs.

CONCLUSION

Although not all normal human have SOAEs, they appear to reflect an important aspect of cochlear function. Their presence is strongly related to elevated levels of TEOAEs which are routinely used in audiological tests.

Otoacoustic emissions in newborns evoked by 0.5kHz tone bursts

BACKGROUND

Otoacoustic emissions (OAEs) are currently widely used in newborn hearing screening programs. OAEs evoked by transients (TEOAEs) in newborns are usually characterized by large response levels at higher frequencies but lower frequencies are affected by physiological noise. The purpose of the present study was to acquire responses at lower frequencies by measuring OAEs evoked by 0.5kHz tone bursts (TBOAEs).

METHODS

Otoacoustic emissions (OAEs) were recorded from 49 newborns. Measurements were made using the ILO 292 equipment from Otodynamics. In each ear, three measurements were made: first with a standard click stimulus at 80dB pSPL (CEOAEs), a second using a 0.5kHz tone burst at 80dB pSPL (TBOAEs), and a third recording of spontaneous OAEs (SOAEs). Global and half-octave-band values of OAE signal-to-noise ratio (SNR) and response level were used to assess statistical differences between CEOAEs and 0.5kHz TBOAEs. Additionally, time-frequency (TF) analysis of signals was performed using the matching pursuit method.

RESULTS

Global levels were highest for CEOAEs. However, at low frequencies (0.7-1kHz), 0.5kHz TBOAEs had significantly higher levels and SNRs than CEOAEs. At these frequencies, SNRs of CEOAEs were usually below 0dB. At 0.5kHz there were no statistically significant differences between CEOAEs and TBOAEs. In ears with recordable SOAEs, CEOAEs and TBOAEs had higher levels and SNRs than in ears without SOAEs.

CONCLUSIONS

Use of 0.5kHz TBOAEs may be a useful addition to standard CEOAE tests in newborns. They provide information about lower frequencies, a region where CEOAEs are usually prone to noise. The presence of SOAEs affects the magnitudes of both CEOAEs and TBOAEs.

Tone burst evoked otoacoustic emissions in different age-groups of schoolchildren

INTRODUCTION

Otoacoustic emissions (OAEs) are believed to be good predictors of hearing status, particularly in the 1-4kHz range. However both click evoked OAEs (CEOAEs) and distortion product OAEs (DPOAEs) perform poorly at 0.5kHz. The present study investigates OAEs in the lower frequency range of 0.5-1kHz evoked by 0.5kHz tone bursts (TBOAEs) in schoolchildren and compares them with emissions evoked by clicks.

METHODS

Measurements were performed for two groups of normally hearing schoolchildren. Children from 1st grade (age 6-7 years) and children from 6th grade (age 11-12 years). Tympanometry, pure tone audiometry, and OAE measurements of CEAOEs, 0.5kHz TBOAEs, and spontaneous OAEs (SOAEs) were performed. Additionally, analysis by the matching pursuit method was conducted on CEOAEs and TBOAEs to assess their time-frequency (TF) properties.

RESULTS

For all subjects OAEs response levels and signal to noise ratios (SNRs) were calculated. As expected, CEOAE magnitudes were greatest over the range 1-4kHz, with a substantial decrease below 1kHz. Responses from the 0.5kHz TBOAEs were complementary in that the main components occurred between 0.5 and 1.4kHz. In younger children, TBOAEs had SNRs 4-8dB smaller in the 0.5-1.4kHz range. In addition, CEOAEs had lower SNRs in the 0.7-1.4kHz range, by 3-5dB. TBOAEs in younger children had maximum SNRs shifted toward 1-1.4kHz, whereas in older children it was more clearly around 1kHz. The differences in response levels were less evident. The presence of SOAEs appreciably influenced both CEOAEs and TBOAEs, and TF properties of both OAEs did not differ significantly between grades.

CONCLUSION

TBOAEs evoked at 0.5kHz can provide additional information about frequencies below 1kHz, a range over which CEOAEs usually have very low amplitudes. The main difference between the two age groups was that in older children CEOAEs and 0.5kHz TBOAEs had higher SNRs at 0.5-1.4kHz. Additionally, for ears with SOAEs, 0.5kHz TBOAEs had higher response levels and SNRs similar to CEOAEs.

Matching Pursuit with Asymmetric Functions for Signal Decomposition and Parameterization

The method of adaptive approximations by Matching Pursuit makes it possible to decompose signals into basic components (called atoms). The approach relies on fitting, in an iterative way, functions from a large predefined set (called dictionary) to an analyzed signal. Usually, symmetric functions coming from the Gabor family (sine modulated Gaussian) are used. However Gabor functions may not be optimal in describing waveforms present in physiological and medical signals. Many biomedical signals contain asymmetric components, usually with a steep rise and slower decay. For the decomposition of this kind of signal we introduce a dictionary of functions of various degrees of asymmetry – from symmetric Gabor atoms to highly asymmetric waveforms. The application of this enriched dictionary to Otoacoustic Emissions and Steady-State Visually Evoked Potentials demonstrated the advantages of the proposed method. The approach provides more sparse representation, allows for correct determination of the latencies of the components and removes the "energy leakage" effect generated by symmetric waveforms that do not sufficiently match the structures of the analyzed signal. Additionally, we introduced a time-frequency-amplitude distribution that is more adequate for representation of asymmetric atoms than the conventional time-frequency-energy distribution.

A bispectral approach to analyze nonlinear cochlear active mechanisms in transient evoked otoacoustic emissions

A new approach to study nonlinearity in cochlear active mechanisms, as evaluated in transient evoked otoacoustic emissions (TEOAEs), is presented. TEOAEs are signals generated in the cochlea by a mix of linear and nonlinear mechanisms. This new approach was designed to complement the traditional TEOAE analysis performed by currently available systems used in objective hearing screening and assessment. Nonlinearity of TEOAEs was studied by means of the bispectrum, which is able to find out quadratic frequency couplings (QFCs) that occur when a frequency is not only generated by an independent cochlear source, but it is the result of the interaction among a number of cochlear sources. To fit with the technical constraints of currently available TEOAE systems, the bispectrum was estimated by the third-order scaled polyperiodogram. The proposed method was characterized with synthesized TEOAEs as a function of the main TEOAE parameters and then used to analyze TEOAEs recorded in normal hearing adults and full-term neonates. Results revealed the presence of QFCs in both adult and neonatal TEOAEs, with peculiar patterns and significantly different frequency content in the two groups: adults had QFCs mainly around 2 kHz and neonates had QFCs mainly in the range 3.5-4 kHz.

Musical ratios in sounds from the human cochlea

The physiological roots of music perception are a matter of long-lasting debate. Recently light on this problem has been shed by the study of otoacoustic emissions (OAEs), which are weak sounds generated by the inner ear following acoustic stimulation and, sometimes, even spontaneously. In the present study, a high-resolution time–frequency method called matching pursuit was applied to the OAEs recorded from the ears of 45 normal volunteers so that the component frequencies, amplitudes, latencies, and time-spans could be accurately determined. The method allowed us to find that, for each ear, the OAEs consisted of characteristic frequency patterns that we call resonant modes. Here we demonstrate that, on average, the frequency ratios of the resonant modes from all the cochleas studied possessed small integer ratios. The ratios are the same as those found by Pythagoras as being most musically pleasant and which form the basis of the Just tuning system. The statistical significance of the results was verified against a random distribution of ratios. As an explanatory model, there are attractive features in a recent theory that represents the cochlea as a surface acoustic wave resonator; in this situation the spacing between the rows of hearing receptors can create resonant cavities of defined lengths. By adjusting the geometry and the lengths of the resonant cavities, it is possible to generate the preferred frequency ratios we have found here. We conclude that musical perception might be related to specific geometrical and physiological properties of the cochlea.

Time-frequency analysis of linear and nonlinear otoacoustic emissions and removal of a short-latency stimulus artifact

Click-evoked otoacoustic emissions (CEOAEs) are commonly recorded as average responses to a repetitive click stimulus. If the click train has constant polarity, a linear average results; if it contains a sequence of clicks of differing polarity and amplitude, a nonlinear average can be calculated. The purpose of this study was to record both protocols from the same set of ears and characterize the differences between them. The major features of CEOAEs were similar under both protocols with the exception of a region spanning 0-5 ms in time and 0-2.2 kHz in frequency. It was assumed that the signal derived from the linear protocol was contaminated by stimulus artifact, and so a simple procedure was used--involving high-pass filtering and time-windowing--to remove components of this artifact. This procedure preserved the short-latency, high-frequency responses; it also produced a marked similarity in the time-frequency plots of recordings made under the two protocols. This result means it is possible to take advantage of the better signal-to-noise ratio of the linear data compared to its nonlinear counterpart. Additionally, it was shown that CEOAEs recorded under the linear protocol appear to be less dependent on the presence of spontaneous otoacoustic emissions (SOAEs).