Auditory and visual speech perception: confirmation of a modality-independent source of individual differences in speech recognition.

Two experiments were run to determine whether individual differences in auditory speech-recognition abilities are significantly correlated with those for speech reading (lipreading), employing a total sample of 90 normal-hearing college students. Tests include single words and sentences, recorded on a videodisc by a male speaker [Bernstein and Eberhardt, Johns Hopkins Lipreading Corpus, The Johns Hopkins University, Baltimore, MD, 1986]. The auditory speech was presented with a white noise masker, at -7 dB Sp/N. The correlations between overall auditory and visual performance were 0.52 and 0.43 in the two experiments, consistent with the existence of a modality-independent ability to perceive linguistic "wholes" on the basis of linguistic fragments. Subjects in the second experiment also identified printed sentences, with 40%-60% portions of the printed characters deleted. Performance on this graphical "fragmented-sentences test" also correlated significantly with auditory speech recognition, providing a possible clue to the cognitive basis for the look-versus-listen correlation. The existence of a modality-independent source of variance in speech-recognition abilities may be a partial explanation of the difficulty in demonstrating strong associations between psychoacoustic measures of spectral or temporal acuity, and speech discrimination or identification. Female subjects in both experiments were significantly better lipreaders than their male counterparts.