Adaptive neural network classifier for decoding MEG signals

TL;DR

This paper presents an interpretable CNN model optimized for decoding MEG signals, outperforming traditional classifiers and generalizing well across individuals in brain state classification tasks.

Contribution

The study introduces a neurophysiologically interpretable CNN tailored for MEG data, demonstrating superior decoding accuracy and robustness to individual differences.

Findings

Decodes event-related brain responses effectively

Outperforms traditional classifiers and complex neural networks

Generalizes across subjects in offline and online settings

Abstract

Convolutional Neural Networks (CNN) outperform traditional classification methods in many domains. Recently these methods have gained attention in neuroscience and particularly in brain-computer interface (BCI) community. Here, we introduce a CNN optimized for classification of brain states from magnetoencephalographic (MEG) measurements. Our CNN design is based on a generative model of the electromagnetic (EEG and MEG) brain signals and is readily interpretable in neurophysiological terms. We show here that the proposed network is able to decode event-related responses as well as modulations of oscillatory brain activity and that it outperforms more complex neural networks and traditional classifiers used in the field. Importantly, the model is robust to inter-individual differences and can successfully generalize to new subjects in offline and online classification.