Feindel Seminar Series: The resonant nature of auditory neural activity

The resonant nature of auditory neural activity

Human speech contains information across multiple hierarchical timescales from phonemes to syllables, words and sentences. Neural oscillations at different frequencies have been suggested as a putative neural mechanism supporting processing and segmentation at these timescales. However, there has been a debate in the field, with critics arguing that the apparent oscillatory nature of auditory neural activity can be explained by sequences of evoked responses. In this talk, I want to show a few examples from my work with MEG to argue why the oscillatory viewpoint is important in human electrophysiology. These include results from 1) natural speech listening, where I show how predictions modulate neural activity overlapping in time but distinct in frequency, and 2) targeted auditory probes that reveal the intrinsic resonant nature of auditory neural activity, as well as the limits of a linear evoked response model. I will show novel ways to parameterize the auditory neural response that distinguishes oscillatory from aperiodic activity and allows linking classical evoked response phenomena with oscillatory interpretations.

Peter Donhauser

Peter Donhauser, PhD, received his PhD in Neuroscience from McGill University working with Sylvain Baillet on MEG imaging and predictive mechanisms in speech processing. He spent time as a postdoctoral fellow with Denise Klein at the Neuro working on neural network models of bilingualism. Right now he is a postdoctoral researcher in David Poeppel's group at the Ernst Strüngmann Institut in Frankfurt, Germany.

For more information, visit the web page To watch online, click here