CNN-based framework using spatial dropping for enhanced interpretation of neural activity in motor imagery classification

D F Collazos-Huertas,C D Acosta-Medina,G Castellanos-Dominguez,A M Álvarez-Meza,G A Castaño-Duque

doi:10.1186/s40708-020-00110-4

Abstract

Interpretation of brain activity responses using motor imagery (MI) paradigms is vital for medical diagnosis and monitoring. Assessed by machine learning techniques, identification of imagined actions is hindered by substantial intra- and inter-subject variability. Here, we develop an architecture of Convolutional Neural Networks (CNN) with an enhanced interpretation of the spatial brain neural patterns that mainly contribute to the classification of MI tasks. Two methods of 2D-feature extraction from EEG data are contrasted: Power Spectral Density and Continuous Wavelet Transform. For preserving the spatial interpretation of extracting EEG patterns, we project the multi-channel data using a topographic interpolation. Besides, we include a spatial dropping algorithm to remove the learned weights that reflect the localities not engaged with the elicited brain response. We evaluate two labeled scenarios of MI tasks: bi-class and three-class. Obtained results in an MI database show that the thresholding strategy combined with Continuous Wavelet Transform improves the accuracy and enhances the interpretability of CNN architecture, showing that the highest contribution clusters over the sensorimotor cortex with a differentiated behavior of rhythms mu and beta .

Highlights

The motor imagery (MI) paradigm is a form of brain– computer interface (BCI) that performs the imagination of a motor action without real execution, relying on the similarities between imagined and executed actions at the neural level
Due to the Convolutional Neural Networks (CNN)-model training back-propagates the discriminating information, through the tied weights, from the hidden spaces in the input data, we propose to assess the relevance of input feature mappings, employing the matrix W (q)∈RD×Nh that holds the row vectors wwthqdqde∈mhRiedNadhseuwnreitsshptahDcee=chGornG(tq′rI)iN,btuf h .teiBonraesleoedfvaionnnpcuetthofefeadfta-uctrhtestfhetaaottuberueaiclihds assessed as the generalized mean of its corresponding reverse projection vector, the vectorq = {̺dq ∈R+; that is, ∀d∈D}
5 Discussion and concluding remarks We present an approach using CNN models to improve the interpretability of spatial contribution in discriminating between MI tasks, preserving an adequate classification accuracy

Summary

Introduction

The motor imagery (MI) paradigm is a form of brain– computer interface (BCI) that performs the imagination of a motor action without real execution, relying on the similarities between imagined and executed actions at the neural level. There is an increasing interest in deep learning models that are composed of multiple processing layers of inference using data representations with multiple levels of abstraction. Convolutional Neural Networks (CNN) become the leading deep learning architectures due to their regularization structure and degree of translation invariance [7], yielding an outstanding ability in transferring knowledge between apparently different tasks of classification [8, 9]. For applications in MI tasks, designing an available end-to-end CNN architecture remains a challenge due to several restrictions: their large number of hyperparameters to be learned increase the computational burden (being unsuitable for online processing [11]), and complicated multilayer integration to encode relevant features at every abstraction level of the input EEG data [12]

Methods

Results

Conclusion