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

2026-03-27

AI Summary

This work addresses the critical challenge of accurately decoding the speech stream attended by a listener in multi-speaker scenarios using electroencephalography (EEG). The authors propose the CA-TCN model, which introduces—for the first time—a dual-branch temporal convolutional network comprising causal and anticausal pathways to explicitly model the bidirectional temporal alignment between EEG neural responses and speech stimuli. This architecture maintains online processing capability while enhancing decoding accuracy. By integrating EEG spatial filtering with an end-to-end classification framework, the model consistently outperforms existing methods across multiple datasets, achieving accuracy gains of 0.5%–3.2% in subject-independent settings and 0.8%–2.9% in subject-dependent settings, along with excellent spatial robustness.

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.

Problem

• Auditory Attention Decoding
• EEG
• speech stream selection
• neural decoding
• multi-speaker environment

Innovation

• Causal-Anticausal Convolution
• Temporal Convolutional Network
• Auditory Attention Decoding
• EEG-based Classification
• Online Neural Decoding