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

Quantifying Entrainment Evidence: A Comparison of Frequentist and Bayesian Approaches for Information Processing Pathway Maps

Kaibo Zhang¹, Ji Wu¹, Chao Zhang¹, Andrew Thwaites²; ¹Tsinghua University, ²University College London

Presenter: Kaibo Zhang

Information Processing Pathway Maps (IPPMs) formalize sequences of mathematical transformations in sensory processing by statistically linking encoding model outputs with neural activity. Traditionally, IPPMs rely on frequentist hypothesis testing (P(D|H0)). However, evaluating competing computational models is fundamentally a problem of model adjudication, which is better suited to probabilistic inference (P(H|D)). Here, we present a direct comparison between the established frequentist approach and a novel Bayesian framework for mapping cortical entrainment. We evaluate both approaches using an empirical EMEG dataset to reconstruct a known auditory loudness-processing pathway. Our results show that while both frameworks converge at major processing hubs, the Bayesian framework alters the selection criterion through a competitive hypothesis space (the dilution effect). Crucially, the naive Bayesian implementation introduces anti-causal artifacts, highlighting the necessity of shifting from uniform priors to biologically informed regularizing priors for robust brain mapping.

Topic Area: Methods, Tools, Theory & Neural Coding