1991 Volume 34 Issue 3 Pages 149-157
A phase spectral analysis developed by Fridman (1982) was applied to detect the amplitude-modulation following response (AMFR), and its clinical usefulness was evaluated. A sinusoidally amplitude-modulated (SAM) tone was delivered to the right ear of 20 normal-hearing subjects. According to Fridman's technique, an ensemble of sweeps was divided into 10 groups to obtain group averages. Then phase variances were calculated using FFT for the group averages. In order to determine the most suitable coditions of the analysis for constructing an automatic detection program using phase spectral analysis, frequency spacing, number of sampling points and number of sweeps averaged for group averages were investigated. The change of phase spectrum as a function of the modulation frequency (20-70Hz) was also investigated. Furthermore, the sensitivity of phase spectral analysis for the detection of AMFR was evaluated.
From the results, it was suggested that the most suitable numbers of sampling points and frequency spacing were 512 points and 4.9Hz (observation window was 204.8msec), respectively. Concerning the modulation frequency, 40Hz was the most appropriate frequency for the stimulation from the standpoint of phase spectral analysis. It was also demonstrated that the threshold determination by phase spectral analysis was more sensitive than the detection of the response by waveform configuration.
