Active Inference for Adaptive BCI: application to the P300 Speller

TL;DR

This paper introduces a Bayesian Active Inference framework for adaptive Brain-Computer Interfaces, specifically applied to P300 spellers, demonstrating significant performance improvements and unification of adaptive methods.

Contribution

It presents a novel, flexible Bayesian approach using Active Inference for adaptive BCIs, outperforming traditional algorithms in P300 speller simulations.

Findings

Bit rate increased by 18-59% with AI

Unified adaptive framework for BCIs

Enhanced performance in realistic simulations

Abstract

Adaptive Brain-Computer interfaces (BCIs) have shown to improve performance, however a general and flexible framework to implement adaptive features is still lacking. We appeal to a generic Bayesian approach, called Active Inference (AI), to infer user's intentions or states and act in a way that optimizes performance. In realistic P300-speller simulations, AI outperforms traditional algorithms with an increase in bit rate between 18% and 59%, while offering a possibility of unifying various adaptive implementations within one generic framework.