Keynotes | K&Ts | GACs | Talks | Posters | Search

Poster C in Poster Session C: Wednesday, August 5, 9:30 – 11:15 am, Kimmel Center, Shorin & Rosenthal Rooms

Reservoir Network Model of Asymmetric Acoustic Priming reveals Chaotic Dynamics in Auditory Brainstem

Praveena Satkunarajah¹, M Ganesh Kumar², Benjamin Rich Zendel¹; ¹Memorial University of Newfoundland, ²Max Planck Institute for Biological Cybernetics

Presenter: Praveena Satkunarajah

Auditory scene analysis requires the segregation of concurrent sound sources, yet how subcortical circuits use acoustic context to facilitate this process remains poorly understood. Frequency following responses (FFRs) from 24 non-musician adults were recorded while they listened to sequences consisting of a solo instrument tone (guitar or piano) immediately followed by a combined guitar-and-piano tone. Guitar priming selectively enhanced stimulus-to-response correlations, while piano priming produced comparatively weaker modulation. To investigate the biological mechanisms underlying this asymmetry, we modeled this perceptual process as a leaky-integrator recurrently connected reservoir network. Sweeping the network's time constant (τ) and spectral radius (g) revealed that a (τ) value of approximately 70 ms and (g) value of 1.2 was required to reproduce the empirically observed priming asymmetry. This provides computational support for the hypothesis that brainstem processing is dynamic rather than passive stimulus encoding alone.

Topic Area: Auditory, Speech & Language Processing