An analysis of the effect of basilar membrane nonlinearities on noise suppression.

Computational models of the peripheral auditory system have largely modeled basilar membrane (BM) mechanics as a linear filter-bank-like entity. Recent mathematical work on the nature of auditory system noise suppression allows us to analyze and argue for the incorporation of BM nonlinearities into these models. This analysis shows that vowel perception improves with increasing presence of BM nonlinearities whereas consonant perception degrades with increasing influence of BM nonlinearities. Experimental results on tones and real speech corrupted by noise corroborate the analysis as well as suggest a novel approach to speech processing.

Power spectral analysis of EEG in a multiple-bedroom, multiple-polygraph sleep laboratory.

OBJECTIVES: We describe the methods for power spectral analysis (PSA) of sleep electroencephalogram (EEG) data at a large clinical and research sleep laboratory. The multiple-bedroom, multiple-polygraph design of the sleep laboratory poses unique challenges for the quantitative analysis of the data. This paper focuses on the steps taken to ensure that our PSA results are not biased by the particular bedroom or polygraph from which the data were acquired. METHODS: After describing the data acquisition system hardware, we present our signal amplitude calibration procedure and our methods for performing PSA. We validate the amplitude calibration procedure in several experiments using PSA to establish tolerances for data acquisition from multiple bedrooms and polygraphs. RESULTS: Since it is not possible to acquire identical digitized versions of an EEG signal using different sets of equipment, the best that can be achieved is data acquisition that is polygraph-independent within a known tolerance. We are able to demonstrate a tolerance in signal amplitude of +/- 0.25% when digitizing data from different bedrooms. When different data acquisition hardware is used, the power tolerance is approximately +/- 3% for frequencies from 1 to 35 Hz. The power tolerance is between +/- 3 and +/- 7% for frequencies below 1 Hz and frequencies between 35 and 50 Hz. Additional data demonstrate that variability due to the hardware system is small relative to the inherent variability of the sleep EEG. CONCLUSION: The PSA results obtained in one location can be replicated elsewhere (subject to known tolerances) only if the data acquisition system and PSA method are adequately specified.

Time-frequency analysis of postural sway.

Postural sway during quiet stance has been used to characterize the postural control system. Most studies have used center of pressure (COP) measurements and have assumed stationarity, however, recent research has indicated that COP is not stationary. The purpose of this study is to introduce and demonstrate a nonstationary spectral estimation technique to examine the time-varying nature of postural sway. Data from two experiments were used to verify the usefulness of the spectral estimator for the analysis of COP. The first data set contains COP recorded from normal subjects swaying about their ankles in response to a metronome as it was gradually changed from 2 to 1 Hz. The time-frequency distribution reveals time-varying spectral changes corresponding to frequency changes made by the subjects. The second set consists of COP from normal subjects and vestibularly impaired patients standing quietly on a force plate with eyes closed for 100 s. The time-frequency distributions for the COP were estimated for both sets of data. The COP's appear to be nonstationary with the energies at a given frequency modulating through time.