Opinion Article - Journal of Neuroinformatics and Neuroimaging (2022) Volume 7, Issue 2

Spatial pattern technique for motor imagery based brain computer interface design.

Luis Perez ^*

Department of Pharmacology and Neuroscience, University of North Texas Health Science Center, Texas, United States

*Corresponding Author:

Jenna Merenstein
Department of Pharmacology and Neuroscience,
University of North Texas Health Science Center,
Texas, United States
E-mail: franco13@utexas.edu

Received: 31-Mar-2022, Manuscript No. AANN-22-110; Editor assigned: 02-Apr-2022, PreQC No. AANN-22-110(PQ); Reviewed: 15-Apr-2022, QC No. AANN-22-110; Revised: 20-Apr-2022, Manuscript No. AANN-22-110(R); Published:27-Apr-2022, DOI:10.35841/ aann-7.2.110

Citation: Perez L. Spatial pattern technique for motor imagery based brain computer interface design. J NeuroInform Neuroimaging. 2022;7(2):110

Abstract

A Brain Computer Interface (BCI) system is capable of reading human brain electrical activities and transforming it to meaningful commands. Brain Computer Interface (BCI) is a useful artificial neural mechanism for patients suffering from partial or complete body paralysis. It helps in interacting and controlling assistive applications and devices such as computer cursor, wheelchair and robotic arm by simply interpreting human brain activities.

Keywords

Magneto-Encephalogram, Electroencephalography, Functional magnetic resonance imaging.

Introduction

These human brain activities can be measured by the tools such as Electroencephalogram (EEG), Magneto-Encephalogram (MEG), Functional Magnetic Resonance Imaging (fMRI) and Electrocorticography (ECoG) imaging techniques. However, EEG is the most preferred brain imaging technique in BCI system design, since it is highly portable, non-invasive and having high temporal resolution.

The most usually utilized electrophysiological peculiarity in BCI framework configuration incorporates Visual Evoked Potential (VEP), engine symbolism movement, Slow Cortical Potential and P300. Notwithstanding, engine symbolism action based EEG signals have been thought of and broke down for BCI framework configuration in this work [1]. Engine Imagery is a psychological interaction, where an individual re-enacts given activities/developments in the brain without really performing it. The engine symbolism action causes variety in spatial dissemination of EEG signals for various envisioned developments. It is habitually involved mental action causing high motions in alpha and beta rhythms of EEG signals. The rhythms of EEG signal display peculiarity of Event Related Desynchronization comparing to every symbolism state. Occasion Related Desynchronization is the variety in continuous EEG movement described by an abatement in the control over the cortex region of the mind. It includes cortical regions during development planning and execution, and helps BCI framework to perceive engine symbolism task.

Effective acknowledgment of any BCI framework relies upon exact acknowledgment of engine symbolism action, which thus requires productive component extraction calculations. It is required that the separated highlights ought to be exceptionally discriminative in nature to achieve a higher acknowledgment rate. Accordingly, notwithstanding highlight extraction, studies are additionally accessible to choose ideal elements to lessen include vector size and achieve exceptionally discriminative elements [2]. Different productive EEG highlight extraction approaches have been recommended previously. For example, Pfurtscheller originally recommended Adapted Autoregressive model for engine symbolism based BCI framework plan. In another review, Roamser recommended ideal spatial separating procedure for breaking down single preliminary EEG during envisioned hand development. Li and Wen fostered an altered CC-LR calculation for engine symbolism assignments characterization in EEG based BCI. It is ordered in wide sense that these methodologies supported utilization of time examination (fleeting component), recurrence investigation (otherworldly element) and time-recurrence examination as the rule include extraction approach. Notwithstanding, the time and recurrence space based procedures are tracked down inadmissible in highlight extraction of nonlinear and non-fixed EEG exercises. Likewise, the most regularly utilized time-recurrence strategy Wavelet Transform is a non-versatile methodology and requires ideal base wavelet choice before activity.

Likewise, WT gives unfortunate time goal at lower frequencies. To beat the deficiencies of WT, specialists have proposed utilization of Stockwell-Transform/S-Transform for time-recurrence portrayal of EEG signals. S-Transform gives great restriction in the recurrence area for low frequencies and great confinement in time space for higher frequencies. Likewise, it jam abundancy and stage data of a sign. It is obvious from different investigations that ST is an effective time-recurrence portrayal strategy in BCI applications.

Aside from time-recurrence portrayal based highlight extraction draws near, Common Spatial Pattern is one more profoundly involved include extraction method in biomedical applications. Normal Spatial Pattern utilizes straight change to project multichannel EEG information to low layered space, which helps in accomplishing high separation capacity in grouping various classes of EEG information. Wu proposed CSP based Linear Dynamical System philosophy and built spatial channels for identifying ERD/ERS action utilizing Support Vector Machine classifier [3]. In another review, Wang and Zheng proposed Local worldly CSP based include extraction method for single-preliminary EEG signal arrangement consolidated WT with CSP calculations and announced high characterization rate in offbeat disconnected BCI application. In this work, Mausovi disintegrated EEG signal into wavelet parcels. Further, CSP is determined from decayed wavelet bundles and discriminative parcels were chosen utilizing fluffy rationale approach. Nonetheless, regardless of the way that countless examinations have been completed for EEG include extraction, no such review recommended ST based CSP highlight extraction system for BCI application. With an end goal to fill this exploration hole, the current work joins ST based time-recurrence portrayal procedure with CSP calculation for EEG include extraction. The composite methodology of EEG highlight extraction is created to accomplish high acknowledgment rate in engine symbolism signal arrangement. A similar table summing up the writing overview of the current work is given.

In present work, at first ST is utilized to decay EEG engine symbolism movement into five unmistakable recurrence sub-groups. From that point, CSP is determined on every recurrence sub-band to accomplish band-wise elements of decayed EEG action. These band-wise highlights are utilized to plan five CSP include vectors separate to every recurrence sub-band. Furthermore, band-wise highlights are joined to develop all groups CSP include vector. The preparation and approval of the delicate figuring procedure is performed independently, utilizing band-wise and all band/joined CSP include vectors. The proposed spatial separating strategy is utilized to recognize two classes of EEG information by amplifying the difference for one class and limiting fluctuation for another class. Three delicate registering methods viz. Least Square-Support Vector Machine (LS-SVM), Random Forest (RF) and Artificial Neural Network (ANN) have been utilized in present work to look at the adequacy of proposed highlight extraction approach in BCI application [4]. The proposed method of EEG highlight extraction is done on three human subjects for left and right hand engine symbolism movement. The construction of this work is as per the following: Section 2 gives point by point portrayal of EEG information utilized in present review. S-Transform (ST) calculation of time-recurrence portrayal is made sense of in Section 3. Segment 4 arrangements with the proposed ST based CSP highlight extraction procedure for engine symbolism grouping. Segment 5 presents results and conversation of the current review.

References

1. Dutta S, Singh M, Kumar A. Classification of non-motor cognitive task in EEG based brain-computer interface using phase space features in multivariate empirical mode decomposition domain. Biomed Signal Processing Control. 2018;39:378-89.

Indexed at, Google Scholar, Cross Ref

2. Mousavi EA, Maller JJ, Fitzgerald PB, et al. Wavelet common spatial pattern in asynchronous offline brain computer interfaces. Biomedical Signal Processing Control. 2011;6(2):121-28.

Indexed at, Google Scholar, Cross Ref

3. Wang H, Zheng W.Local temporal common spatial patterns for robust single-trial EEG classification. IEEE Transactions on Neural Systems and Rehabilitation Eng. 2008;16(2):131-39.

Indexed at, Google Scholar, Cross Ref

4. Upadhyay R, Manglick A, Reddy DK, et al. Channel optimization and nonlinear feature extraction for Electroencephalogram signals classification. Computers Electrical Eng. 2015;45:222-34.

Indexed at, Google Scholar, Cross Ref