Resting-State-Based Spatial Filtering for an fNIRS-Based Motor Imagery Brain-Computer Interface

Yanchun Zheng,Daifa Wang,Ling Wang,Dan Zhang,Hao Deng,Shen Zhang,Yijun Wang,Deyu Li

doi:10.1109/access.2019.2936434

Abstract

Functional near-infrared spectroscopy (fNIRS) has attracted much attention in brain-computer interface (BCI) area due to its advantages of portability, robustness to electrical artifacts, etc. However, in practical applications, fNIRS-based BCI usually needs a labor-intensive and time-consuming training session (calibration procedure) to optimize the user-specific neural spatial and temporal patterns for further classification. Recently, studies revealed that neural spatial and temporal patterns extracted from user-specific resting-state brain signals were closely related to those of his/her task data. In this study, we proposed a resting-state independent component analysis (RSICA) based spatial filtering algorithm aiming at extracting individual task-related spatial and temporal brain patterns from the resting-state data. Specifically, independent component analysis (ICA) was applied to extract different independent components (ICs) from resting-state fNIRS data. The ICs with their spatial filter weights maximally lateralized over the sensorimotor regions were regarded as most relevant to motor imagery. These spatial filters were used to spatially filter the multi-channel motor imagery task data for feature extraction. Based on 8-minute resting-state data and a small training dataset (20 trials) from 10 participants, the proposed RSICA algorithm achieved an approximately 7% increase in left vs. right hand motor imagery classification accuracy, as compared to the conventional common spatial pattern (CSP)-based and shrinkage algorithms (69.8±12.1%, 63.3±10.3% and 63.4±11.8%, respectively). For acquiring a similar level of classification accuracy (i.e. 70%), the number of training data required could be reduced from 36 trials (CSP) to 22 trials (RSICA). As a relatively small training set is required to obtain a satisfactory performance, training burden is significantly reduced by RSICA, which might be useful for developing practical fNIRS-based motor imagery BCIs.

Highlights

Brain-computer interface (BCI) aims to establish a direct connection between its users and external devices through interpreting the users’ brain activities, towards improvingThe associate editor coordinating the review of this article and approving it for publication was Jafar A
We proposed and implemented a resting-state independent component analysis (RSICA) based spatial filtering method for an functional near-infrared spectroscopy (fNIRS)-based motor imagery BCI
The independent components (ICs) with highest lateralization index shows extraordinarily high values in the left and right sensorimotor areas, indicating its strong correlation to the motor imagery tasks. These results were consistent with our assumption that the lateralization index of IC is strongly related to motor imagery tasks

Summary

Introduction

Brain-computer interface (BCI) aims to establish a direct connection between its users and external devices through interpreting the users’ brain activities, towards improvingThe associate editor coordinating the review of this article and approving it for publication was Jafar A. Whereas electroencephalograph (EEG) is still the most widely used neuroimaging technique in BCI community, functional near-infrared spectroscopy (fNIRS) has attracted substantial interest in recent years. Different mental activities are associated with distinct spatial or temporal changes in users’ hemodynamic responses, as reflected. Y. Zheng et al.: Resting-State-Based Spatial Filtering for an fNIRS-Based Motor Imagery BCI by the concentration of oxygenated hemoglobin (HbO) and deoxygenated hemoglobin (HbR) measured by near-infrared light The most common brain areas used in fNIRS-based BCIs are the sensorimotor cortex and the prefrontal cortex and the most popular mental tasks include motor imagery, mental arithmetic, music imagery, verbal fluency, etc. FNIRS-based BCIs have mainly relied on spontaneous neural activities, rather than evoked neural responses (e.g. visual or auditory evoked responses), possibly due to its coverage effectiveness and temporal resolution

Objectives

Methods

Results

Discussion

Conclusion