Upper Limb Movement Execution Classification using Electroencephalography for Brain Computer Interface.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Accurate classification of upper limb movements via electroencephalogram (EEG) signals is crucial for brain-computer interfaces (BCIs).
• Decoding EEG signals can significantly aid individuals with spinal cord injury (SCI) and neuro-muscular diseases, restoring independence in daily activities.
• Current research focuses on detecting and classifying executed or imagined upper limb movements using BCIs.

Purpose of the Study:

• To decode movement execution (ME) of the upper limb using EEG signals.
• To classify four distinct upper limb ME classes from EEG data.
• To evaluate the efficacy of deep learning models applied to EEG spectrograms for movement classification.

Main Methods:

• Utilized a publicly available 61-channel EEG dataset from fifteen subjects.
• Proposed a classification method using spectrograms derived from EEG data.
• Employed pre-trained deep learning (DL) models, fine-tuned for the specific task of classifying upper limb ME.

Main Results:

• Achieved a highest average classification accuracy of 87.36% for four ME classes.
• One subject demonstrated a peak classification accuracy of 97.03%.
• The proposed method using EEG spectrograms and DL models proved effective for classifying upper limb ME.

Conclusions:

• Movement execution of upper limbs can be classified with significant accuracy using EEG signal spectrograms and fine-tuned pre-trained deep learning models.
• This approach holds promise for developing advanced BCIs for individuals with motor impairments.
• The findings highlight the potential of EEG-based BCI for restoring functional capabilities and improving quality of life.