A central issue in speech recognition is how contrastive phonemic information is stored in the mental lexicon. The conventional view assumes that this information is closely related to acoustic properties of speech. Considering that no word is ever pronounced alike twice and that the brain has limited capacities to manage information, an opposing view proposes abstract underspecified representations where not all phonemic features are stored. We examined this proposal using event-related brain potentials, in particular mismatch negativity (MMN), an automatic change detection response in the brain that is sensitive to language-specific phoneme representations. In the current study, vowel pairs were presented to subjects, reversed as standard and deviant. Models not assuming underspecification predict equal MMNs for vowel pairs regardless of the reversal. In contrast, enhanced and earlier MMNs were observed for those conditions where the standard is not phonologically underspecified in the mental representation. This provides the first neurobiological evidence for a featurally underspecified mental lexicon.

This study further elucidates determinants of vowel perception in the human auditory cortex. The vowel inventory of a given language can be classified on the basis of phonological features which are closely linked to acoustic properties. A cortical representation of speech sounds based on these phonological features might explain the surprisingly inverse correlation between immense variance in the acoustic signal and high accuracy of speech recognition. We investigated timing and mapping of the N100m elicited by 42 tokens of seven natural German vowels varying along the phonological features tongue height (corresponding to the frequency of the first formant) and place of articulation (corresponding to the frequency of the second and third formants). Auditoryevoked fields were recorded using a 148-channel whole-head magnetometer while subjects performed target vowel detection tasks. Source location differences appeared to be driven by place of articulation: Vowels with mutually exclusive place of articulation features, namely, coronal and dorsal elicited separate centers of activation along the posterior-anterior axis. Additionally, the time course of activation as reflected in the N100m peak latency distinguished between vowel categories especially when the spatial distinctiveness of cortical activation was low. In sum, results suggest that both N100m latency and source location as well as their interaction reflect properties of speech stimuli that correspond to abstract phonological features.