Speech comprehension and speaker recognition are generally accurate, fast and effortless. This suggests that we should be able to find robust neural correlates of the acoustic and phonetic information that listeners use to make these decisions. In this talk I will review a recent series of magneto-encephalographic (MEG) studies examining the timing and localizations of a major, early auditory cortical response, the M100 (also termed the N1m for its relation to the electro-encephalographic N1 response). This response occurs approximately 100ms after the onset of a well-defined acoustic event and various properties of this response are known to be correlated with phonetic properties of interest. For example, the latency of the M100 response varies with the frequency of both pitch (included inferred pitch, Monahan, de Souza & Idsardi 2008) and formant structure (Monahan & Idsardi 2010). In addition, the inferred location of the generator of the M100 within the auditory cortex (Scharinger, Merickel, RIley & Idsardi 2011; Scharinger, Poe & Idsardi 2011) forms a kind of tonotopic and articulo-topic map, with the anterior-posterior location reflecting the place of articulation (F2 and labial, coronal or dorsal), the superior-inferior axis reflects front vowel height and F1, and the medial-lateral axis reflects overall spectral gravity (e.g. rounding and F3) in vowels and in sibilants (Lago, Krorod, Scharinger & Idsardi 2010). In addition, the M100 also contains information relevant for other speech categories, including speaker identity and dialect affiliation (Scharinger, Monahan & Idsardi 2011).
Return to Colloquia Page || LSO Home || MSU Home