NEUROTOKEN: Joint Source and Directional AAD with Envelope Decoding via Conditional Flow Matching

2026-09-30

AI Summary

This study addresses the challenge of EEG-based auditory attention decoding (AAD) in noisy environments by proposing the NEUROTOKEN framework. This framework leverages conditional flow matching to model the conditional likelihood between EEG signals and speech envelopes, thereby unifying source and direction decoding tasks. It introduces a generative scoring head, ATTUNEFLOW, alongside QUADTRACK and ENV-FLOW integration strategies, coupled with a multi-task shared frontend, to effectively overcome the noise susceptibility and fragmentation inherent in conventional statistical features. Experimental results demonstrate that the proposed method improves source AAD accuracy by 9%–16% within a 5-second window, reduces cross-subject variance by approximately threefold, and achieves fusion accuracies exceeding 93% on select datasets.

Abstract

Identifying which speaker a listener is attending to in a noisy room -- the cocktail-party problem -- is the missing ingredient for next-generation hearing aids and brain-computer interfaces: it tells the device whose voice to amplify. Auditory attention decoding (AAD) reads this answer from EEG, but the literature splits into disconnected pieces: directional-AAD classifies side but does not map side to stream; regression-based source-AAD ranks candidate streams by a single Pearson correlation that is intrinsically noisy at the 1-5 s windows real devices need; and envelope reconstruction has no native AAD rule. We argue the right object is not any single statistic but the conditional likelihood of the attended envelope given EEG, and we make this practical with NEUROTOKEN: a single network whose three heads share one EEG front-end, with a conditional flow-matching head (ATTUNEFLOW) that scores candidates by an integrated velocity-residual likelihood ratio. Two inference-time ensembles -- QUADTRACK (four complementary statistics) and ENV-FLOW (z-normalised QUADTRACK+ATTUNEFLOW) -- absorb per-statistic failure modes for free. On KU Leuven, DTU, and NJU at 5 s, ATTUNEFLOW lifts per-segment source-AAD by 9%-16% over the strongest non-generative baseline and shrinks across-subject variance by ~3x; trial-level fusion exceeds 93% on two of three datasets. In parallel reproductions we show that canonical 95-97% direction-AAD numbers collapse by 17%-45% under a strict trial-disjoint protocol, clarifying both the true ceiling and why a likelihood-based formulation is needed.

Problem

• Auditory attention decoding
• Cocktail-party problem
• EEG
• Source separation
• Hearing aids

Innovation

• Auditory Attention Decoding
• Conditional Flow Matching
• EEG
• Neurotoken
• Envelope Reconstruction