The neural coding of relational invariance in speech: Human language analogs to the barn owl
Harvey Martin Sussman · The Journal of the Acoustical Society of America · 1988
This paper presents a brain-based perspective on the noninvariance problem in speech research, namely, that physically different auditory stimuli come to elicit an invariant perception of a phoneme category. Extrapolating from the sound localization system in the barn owl, two speculative human models are offered that account for the emergence of relational invariance for (1) bilabial and alveolar stop consonants across vowel contexts, and (2) vowel identity across speakers with different sized vocal tracts. The barn owl's extraction of interaural time differences (ITDs) to signal azimuth is based on disambiguating interaural phase information coded by frequency specific neurons. Only by spanning across a broad frequency spectrum can a neuronal functional array signal an unambiguous ITD to higher centers. A similar principle is invoked to model both stop consonant place invariance and vowel normalization in human speech processing. Formant manipulation metrics, capable of distinguishing contrastive phonemic categories, are described to illustrate the operational features of the modeling scheme.