The Mechanism of Auditory Evoked EEG Responses

Abstract

ANALYSIS of the electroencephalogram (EEG) recorded over 0.5–1.0 s immediately following an auditory stimulus, is a well known objective method for evaluating auditory perceptual threshold, for example, in very young subjects. The low-level and visually indistinct individual responses occur in a spontaneous EEG background; hence repeated stimuli are used and the resulting ensemble of post-stimulus EEG signals averaged at each instant of post-stimulus time (PST). An exactly similar method is widely used in studying EEG visual evoked responses and other neuroelectric phenomena¹. This averaging procedure is effective in clarifying any consistent response in the presence of spontaneous activity if identical responses are evoked by all stimuli and if the response waveform is uncorrelated with the spontaneous activity; however, neither requirement seems fully to be met in EEG evoked responses generally or in evoked-response audiometry (ERA) specifically. Correlated waveforms imply common Fourier spectral components, and it is recognised that a linear filter method like averaging could only achieve limited success in separating two such waveforms since it requires the waveforms to have different spectral distributions², in which case, averaging reduces the additive background due to components with frequency outside the spectral band in which the main power of the response waveform occurs. When all consequent improvement has been made in this way, the response waveform may still seem to be degraded by the presence of background activity within the same spectral band, and linear filter methods cannot assist further; the following remarks refer to this situation³.