Signal processing of the auditory brainstem response: clinical effects of variations in fast Fourier transform analysis

Abstract

Fast Fourier transformation (FFT) of the auditory brainstem response (ABR) has proven useful in improving this tests sensitivity and specificity as a diagnostic assessment tool, but has suffered from poor control of pre-FFT windowing procedures. This study compared the clinical significance, as versus the obvious mathematical significance, of using no Blackman, and modified Blackman window techniques for FFT analysis of ABRs in normal subjects. Windowing was shown to statistically reduce ABR FFT magnitudes with the differences between window results all being clinically significant. These findings warn against the clinical comparison of ABR FFT results obtained using different windows and reinforces the need to state specifically the FFT procedures used when conducting clinical ABR frequency analysis.