International Journal on Advanced Science, Engineering and Information Technology, Vol. 13 (2023) No. 2, pages: 666-673, DOI:10.18517/ijaseit.13.2.18299

EEG Time-Frequency Domain Analysis for Describing Healthy Subjects and Stroke Patients during Stroke Rehabilitation Motion Tasks

Adhi Dharma Wibawa, Yuri Pamungkas, Monica Pratiwi, Rosita Devi Kusumastuti, Wardah Rahmatul Islamiyah, Diah Risqiwati

Abstract

To overcome the long-term impact of stroke attacks on society, stroke rehabilitation is the only solution WHO and many healthcare organizations suggested. Until recently, stroke rehabilitation monitoring has been done using visual observation, which has several drawbacks. EEG is a new approach to understanding how the central nervous system controls motion. This study compares the motion pattern done by a group of 12 healthy subjects and nine stroke patients during the rehabilitation motion tasks using the OpenBCI system. Time-frequency domain features, namely PSD, MAV, and STD are used to explore how the patterns differ. Three rehabilitation motions are implemented: grasping, elbow flexion extension, and shoulder flexion-extension. The result shows that the healthy cross-brain correlation happens in healthy subjects. This means that when the left-side arm does the motion, the EEG feature values from the right hemisphere are higher, and vice versa. However, this healthy cross-brain correlation pattern did not happen within the stroke patient group. The overall value of PSD, MAV, and STD from both hemispheres during all motions is higher in the healthy group than in stroke patients. The type of motion also contributes to describing the time-frequency domain feature comparison. In conclusion, this gap value using time-frequency domain features can be used as a target for stroke rehabilitation programs by implementing the EEG technology to monitor it.

Keywords:

Stroke rehabilitation monitoring; EEG time-frequency domain; PSD; MAV; STD; home plasticity training; electroencephalogram

Viewed: 74 times (since abstract online)

cite this paper     download