Degraded discrimination between speech-like sounds by patients with multiple sclerosis and Friedreich's ataxia

We previously found that some patients with multiple sclerosis are selectively 'deaf' to changes in the pitch of a tone, even when audiometric sensitivity to pure tones is unimpaired. This subtle form of deafness is not experienced by patients with noise-induced hearing loss of exclusively peripheral origin. It was suggested that this auditory defect may be one possible cause for difficulties in discriminating speech, on the grounds that frequency changes in the speech waveform are known to be important for intelligibility. This implication is not self-evident; our earlier studies tested hearing with a single pure tone that was either frequency-modulated or amplitude-modulated, while even a simple approximation to speech sounds involves not one, but three narrow bands of noise (formants) whose frequencies and intensities change from instant to instant. The present study has investigated the ability of subjects to discriminate between speech-like sounds. These consisted of three formant frequencies generated by computer. The only difference between the sounds was that the lowest-frequency formant rose or fell in pitch by different amounts. In order to ensure that subjects used frequency (pitch) cues rather than any associated loudness cues were mixed different loudness shifts with the frequency shifts. Nineteen control subjects, 25 patients with multiple sclerosis (MS) and 4 patients with Friedreich's ataxia (FA) were tested. Nine of the patients with MS and all 4 patients with FA gave results that fell outside the range of the control subjects. A possible pathophysiological basis for this observation is the finding that some neurons in the auditory pathway of animals respond preferentially to changes in tone frequency: homologues of these neurons might be functionally impaired in some patients with MS and FA.