Assessing stroke rehabilitation degree based on quantitative EEG index and nonlinear parameters

Area of Science:

• Neuroscience
• Rehabilitation Medicine
• Biomedical Engineering

Background:

• Motor function assessment is crucial for stroke rehabilitation.
• Current behavioral scoring lacks direct neurological indicators of brain function.
• There is a need for objective, brain-based measures in stroke recovery evaluation.

Purpose of the Study:

• To determine if resting-state electroencephalography (EEG) indicators can enhance stroke rehabilitation assessment.
• To explore the utility of quantitative EEG (QEEG) and nonlinear parameters for evaluating motor function post-stroke.

Main Methods:

• Recruited 68 participants with varying stroke severity (severe, moderate, mild) based on Brunnstrom stages.
• Recorded resting-state EEG data and calculated ten QEEG and five nonlinear parameters.
• Employed statistical tests and a genetic algorithm-support vector machine for feature selection and classification.

Main Results:

• Significant differences in QEEG parameters (Delta, Alpha1, Alpha2, DAR, DTABR) were observed across stroke severity groups (P < 0.05).
• Significant differences in nonlinear parameters (ApEn, SampEn, Lz, C0) were also found (P < 0.05).
• The optimal feature combination achieved an 85.3% classification accuracy rate.

Conclusions:

• Resting-state EEG parameters, both quantitative and nonlinear, show significant differences related to motor function levels in stroke patients.
• These EEG indicators hold promise as objective biomarkers for evaluating stroke rehabilitation progress.
• The findings suggest that EEG-based metrics can complement or potentially replace traditional behavioral assessments in stroke recovery.