A TIME-SCALE FEATURE EXTRACTION SCHEME FOR THE AUTOMATED DETECTION OF BINAURAL INTERACTION IN AUDITORY BRAINSTEM RESPONSES

摘要

The detection of binaural interaction is of diagnostic interest in patients with central auditory processing disorders as binaural hearing tasks are frequently affected in these patients. Due to the comorbitity with disorders such as attention deficit hyperac-tivity disorder, pathological subjective test often represent extra-auditory factors such as reduced attention. Therefore, objective measures for auditory processing disorders are essential. The binaural interaction components (BICs), that is, the arithmetical difference of the summed monaural auditory evoked potentials of each side and the binaurally evoked brain-stem potentials, have been used as an objective measure of bin-aural interaction. As there is no clearly defined signal feature characterizing the difference between the summed monaural and the binaural responses the results obtained largely depend on the detection criteria used. Hence a reliable automated analysis of the BICs capable of clinical use is difficult. Using optimized wavelet packet decompositions, we were able to define a signal feature that accounts for most of the dissimilarity between summed monaural and binaurally evoked brainstem responses and found evidence that this feature represents binaural interaction. We were also able to show that it is possible to detect binaural interaction by exclusively analyzing the binaurally evoked brain-stem responses when utilizing this signal feature. This leads to a significant reduction in measurement time