Performance Modeling for Correlation-based Neural Decoding of Auditory Attention to Speech

2025-03-12

AI Summary

In auditory attention decoding (AAD), a fundamental trade-off exists between temporal resolution—determined by the decision window length—and decoding accuracy; conventional approaches require exhaustive evaluation across multiple window lengths to characterize performance, incurring high computational cost. This paper proposes a novel method that models the full window-length performance curve using only a single window length, by computing label-aligned correlation coefficients. We introduce the Fisher transformation to map attended and unattended neural–speech correlations to approximately normal distributions, then integrate a linear decoder with the VLAAI network to predict classification accuracy across window lengths. Our method achieves a modeling error of only ~2 percentage points, with 94% of ground-truth accuracies falling within the 95% confidence interval. This enables efficient real-time performance monitoring and parameter adaptation in neurocontrolled hearing aids.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Hardware-aware ML
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• User Modeling, Personalization and Recommendation: User modeling for targeted and personalized online advertising
• Search and Retrieval-Augmented AI: Ad search and search for Web retail
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Correlation-based auditory attention decoding (AAD) algorithms exploit neural tracking mechanisms to determine listener attention among competing speech sources via, e.g., electroencephalography signals. The correlation coefficients between the decoded neural responses and encoded speech stimuli of the different speakers then serve as AAD decision variables. A critical trade-off exists between the temporal resolution (the decision window length used to compute these correlations) and the AAD accuracy. This trade-off is typically characterized by evaluating AAD accuracy across multiple window lengths, leading to the performance curve. We propose a novel method to model this trade-off curve using labeled correlations from only a single decision window length. Our approach models the (un)attended correlations with a normal distribution after applying the Fisher transformation, enabling accurate AAD accuracy prediction across different window lengths. We validate the method on two distinct AAD implementations: a linear decoder and the non-linear VLAAI deep neural network, evaluated on separate datasets. Results show consistently low modeling errors of approximately 2 percent points, with 94% of true accuracies falling within estimated 95%-confidence intervals. The proposed method enables efficient performance curve modeling without extensive multi-window length evaluation, facilitating practical applications in, e.g., performance tracking in neuro-steered hearing devices to continuously adapt the system parameters over time.

Problem

• Model trade-off between temporal resolution and AAD accuracy.
• Predict AAD accuracy using single window length correlations.
• Enable efficient performance curve modeling for neuro-steered devices.

Innovation

• Models trade-off curve with single window data
• Uses Fisher transformation for normal distribution modeling
• Validates on linear and non-linear decoders