Medial Wall's Potential in Enhancing Finger Movement Decoding from Electrocorticography (ECoG): A Single-Subject...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Engineering

Background:

• Next-generation motor brain-computer interfaces (BCIs) require multi-region brain recordings.
• The medial wall's role in finger movement decoding remains understudied.
• Electrocorticography (ECoG) offers high-resolution brain activity monitoring.

Purpose of the Study:

• To assess the medial wall's potential for multiclass finger movement decoding.
• To investigate the impact of combining medial wall and lateral cortical ECoG recordings on decoding accuracy.
• To identify key neural features for motor decoding.

Main Methods:

• Utilized ECoG recordings from medial and lateral cortical areas during a 5-finger + rest classification task.
• Analyzed frequency features, including Local Motor Potentials (LMP) and alpha band power (8-12 Hz).
• Compared decoding performance using medial wall channels alone versus combined with lateral channels.

Main Results:

• Achieved a significant six-class accuracy of 0.46 for medial wall decoding, with rest and thumb movements classified best.
• Local Motor Potentials (LMP) showed pre-movement onset activity, and alpha band power aided rest trial decoding.
• Combining medial wall with lateral channels improved average accuracy from 0.36 to 0.42, with minimal improvement near the primary motor cortex.

Conclusions:

• The medial wall is a valuable, underutilized region for motor decoding in BCIs.
• Integrating medial wall recordings enhances BCI performance, particularly for individuals with compromised motor function.
• Future motor BCIs should consider targeting the medial wall for improved decoding capabilities.