Cortical Oscillations in Auditory Perception and Speech: Evidence for Two Temporal Windows in Human Auditory Cortex

Huan Luo, David Poeppel

Natural sounds, including vocal communication sounds, contain critical information at multiple time scales. Two essential temporal modulation rates in speech have been argued to be in the low gamma band (∼20-80 ms duration information) and the theta band (∼150-300 ms), corresponding to segmental and diphonic versus syllabic modulation rates, respectively. It has been hypothesized that auditory cortex implements temporal integration using time constants closely related to these values. The neural correlates of a proposed dual temporal window mechanism in human auditory cortex remain poorly understood. We recorded MEG responses from participants listening to non-speech auditory stimuli with different temporal structures, created by concatenating frequency-modulated segments of varied segment durations. We show that such non-speech stimuli with temporal structure matching speech-relevant scales (∼25 and ∼200 ms) elicit reliable phase tracking in the corresponding associated oscillatory frequencies (low gamma and theta bands). In contrast, stimuli with non-matching temporal structure do not. Furthermore, the topography of theta band phase tracking shows rightward lateralization while gamma band phase tracking occurs bilaterally. The results support the hypothesis that there exists multi-time resolution processing in cortex on discontinuous scales and provide evidence for an asymmetric organization of temporal analysis (asymmetrical sampling in time, AST). The data argue for a mesoscopic-level neural mechanism underlying multi-time resolution processing: the sliding and resetting of intrinsic temporal windows on privileged time scales.