Kernel convolution model for decoding sounds from time-varying neural responses

TL;DR

This paper introduces a kernel convolution model that decodes natural sounds from high-dimensional neural responses like MEG, achieving high accuracy and revealing temporal sensitivity to spectral content in auditory processing.

Contribution

The study presents a novel kernel-based convolution model for decoding natural sounds from neural responses, with improved temporal and spectral analysis capabilities.

Findings

Decodes sounds with 70% accuracy from MEG data

Identifies neural response sensitivity to spectral content at 250-500 ms

Demonstrates the model's potential for understanding neural sound representations

Abstract

In this study we present a kernel based convolution model to characterize neural responses to natural sounds by decoding their time-varying acoustic features. The model allows to decode natural sounds from high-dimensional neural recordings, such as magnetoencephalography (MEG), that track timing and location of human cortical signalling noninvasively across multiple channels. We used the MEG responses recorded from subjects listening to acoustically different environmental sounds. By decoding the stimulus frequencies from the responses, our model was able to accurately distinguish between two different sounds that it had never encountered before with 70% accuracy. Convolution models typically decode frequencies that appear at a certain time point in the sound signal by using neural responses from that time point until a certain fixed duration of the response. Using our model, we evaluated several fixed durations (time-lags) of the neural responses and observed auditory MEG responses to be most sensitive to spectral content of the sounds at time-lags of 250 ms to 500 ms. The proposed model should be useful for determining what aspects of natural sounds are represented by high-dimensional neural responses and may reveal novel properties of neural signals.