SENSORY AND PERCEPTUAL CODES IN CORTICAL AUDITORY PROCESSING
Cervantes Constantino, Francisco Israel
Simon, Jonathan Z
MetadataShow full item record
A key aspect of human auditory cognition is establishing efficient and reliable representations about the acoustic environment, especially at the level of auditory cortex. Since the inception of encoding models that relate sound to neural response, three longstanding questions remain open. First, on the apparently insurmountable problem of fundamental changes to cortical responses depending on certain categories of sound (e.g. simple tones versus environmental sound). Second, on how to integrate inner or subjective perceptual experiences into sound encoding models, given that they presuppose existing, direct physical stimulation which is sometimes missed. And third, on how does context and learning fine-tune these encoding rules, as adaptive changes to improve impoverished conditions particularly important for communication sounds. In this series, each question is addressed by analysis of mappings from sound stimuli delivered-to and/or perceived-by a listener, to large-scale cortically-sourced response time series from magnetoencephalography. It is first shown that the divergent, categorical modes of sensory coding may unify by exploring alternative acoustic representations other than the traditional spectrogram, such as temporal transient maps. Encoding models of either of artificial random tones, music, or speech stimulus classes, were substantially matched in their structure when represented from acoustic energy increases –consistent with the existence of a domain-general common baseline processing stage. Separately, the matter of the perceptual experience of sound via cortical responses is addressed via stereotyped rhythmic patterns normally entraining cortical responses with equal periodicity. Here, it is shown that under conditions of perceptual restoration, namely cases where a listener reports hearing a specific sound pattern in the midst of noise nonetheless, one may access such endogenous representations in the form of evoked cortical oscillations at the same rhythmic rate. Finally, with regards to natural speech, it is shown that extensive prior experience over repeated listening of the same sentence materials may facilitate the ability to reconstruct the original stimulus even where noise replaces it, and to also expedite normal cortical processing times in listeners. Overall, the findings demonstrate cases by which sensory and perceptual coding approaches jointly continue to expand the enquiry about listeners’ personal experience of the communication-rich soundscape.