CA-TCN: A Causal-Anticausal Temporal Convolutional Network for Direct Auditory Attention Decoding

TL;DR

The paper introduces CA-TCN, a novel neural network architecture that improves auditory attention decoding accuracy by explicitly modeling neural and stimulus responses with causal and anticausal convolutions, advancing real-world auditory attention applications.

Contribution

It presents the CA-TCN model, integrating causal and anticausal convolutions for improved neural decoding, outperforming existing models in accuracy and robustness.

Findings

CA-TCN outperforms baseline models with 0.5%-3.2% accuracy gains.

Statistically significant improvements in 4 of 6 settings.

Model demonstrates spatial robustness across datasets.

Abstract

A promising approach for steering auditory attention in complex listening environments relies on Auditory Attention Decoding (AAD), which aim to identify the attended speech stream in a multiple speaker scenario from neural recordings. Entrainment-based AAD approaches, typically assume access to clean speech sources and electroencephalography (EEG) signals to exploit low-frequency correlations between the neural response and the attended stimulus. In this study, we propose CA-TCN, a Causal-Anticausal Temporal Convolutional Network that directly classifies the attended speaker. The proposed architecture integrates several best practices from convolutional neural networks in sequence processing tasks. Importantly, it explicitly aligns auditory stimuli and neural responses by employing separate causal and anticausal convolutions respectively, with distinct receptive fields operating in opposite temporal directions. Experimental results, obtained through comparisons with three baseline AAD models, demonstrated that CA-TCN consistently improved decoding accuracy across datasets and decision windows, with gains ranging from 0.5% to 3.2% for subject-independent models and from 0.8% to 2.9% for subject-specific models compared with the next best-performing model, AADNet. Moreover, these improvements were statistically significant in four of the six evaluated settings when comparing Minimum Expected Switch Duration distributions. Beyond accuracy, the model demonstrated spatial robustness across different conditions, as the EEG spatial filters exhibited stable patterns across datasets. Overall, this work introduces an accurate and unified AAD model that outperforms existing methods while considering practical benefits for online processing scenarios. These findings contribute to advancing the state of AAD and its applicability in real-world systems.