Fingerprints of data compression in EEG sequences

TL;DR

This paper introduces a novel statistical model selection method using context tree models and a projective clustering technique to investigate how the brain compresses data, based on EEG responses to auditory stimuli.

Contribution

It presents a new approach combining context tree models and a projective clustering method to analyze EEG data for understanding neural data compression.

Findings

EEG segments can be clustered based on stimulus models

The method distinguishes different stimulus sequence models

Results support the brain's use of probabilistic models for data compression

Abstract

It has been classically conjectured that the brain compresses data by assigning probabilistic models to sequences of stimuli. An important issue associated to this conjecture is what class of models is used by the brain to perform its compression task. We address this issue by introducing a new statistical model selection procedure aiming to study the manner by which the brain performs data compression. Our procedure uses context tree models to represent sequences of stimuli and a new projective method for clustering EEG segments. The starting point is an experimental protocol in which EEG data is recorded while a participant is exposed to auditory stimuli generated by a stochastic chain. A simulation study using sequences of stimuli generated by two different context tree models with EEG segments generated by two distinct algorithms concludes this article.