Decoding four hand gestures with a single bipolar pair of electrocorticography electrodes
Maxime Verwoert1, Mariska J Vansteensel1, Zachary V Freudenburg1
1Department of Neurology and Neurosurgery, University Medical Center Utrecht Brain Center, Utrecht University, P.O. Box 85500, 3508 GA Utrecht, The Netherlands.
Journal of Neural Engineering
|October 4, 2021
Summary
A single bipolar pair of electrocorticography (ECoG) electrodes can decode four hand gestures for brain-computer interfaces (BCIs). This finding supports minimally invasive ECoG implants for restoring communication in individuals with paralysis.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Rehabilitation Engineering
Background:
- Electrocorticography (ECoG) based brain-computer interfaces (BCIs) offer communication restoration for locked-in syndrome.
- Motor-based BCIs require discriminating neural activity for speed; minimizing implant size is crucial for minimally invasive approaches.
Purpose of the Study:
- To investigate decoding four hand gestures using a single ECoG strip with four electrodes.
- To compare decoding performance between unipolar and bipolar montages for minimally invasive BCI applications.
Main Methods:
- Collected ECoG data from seven individuals with epilepsy undergoing ECoG grid implantation.
- Generated virtual ECoG strips (unipolar/bipolar) from implanted grids to compare decoding accuracy for four hand gestures.
Main Results:
- Four hand gestures were decoded with similar accuracy using 'Unipolar Strips' (67.4%), 'Bipolar Strips' (66.6%), and 'Bipolar Pairs' (67.6%).
- 'Unipolar Electrodes' (61.6%) showed significantly lower decoding accuracy compared to other configurations.
Conclusions:
- A single bipolar pair is a viable option for minimally invasive motor-based BCIs.
- ECoG is a robust platform for multi-class movement decoding, supporting the development of advanced neuroprosthetics.
