Computer Science and Engineering
Date of Award
Master of Arts (MA)
Chair and Committee
Today, learning from the brain is the most challenging issue in many areas. Neural scientists, computer scientists, and engineers are collaborating in this broad research area. With better techniques, we can extract the brain signals by either non-invasive approach such as EEG: electroencephalography), fMRI, or invasive method such as ECoG: electrocorticography), FP: field potential) and signals from single unit. The challenge is, given the brain signals, how can we possibly decipher them? Brain Computer Interfaces, or BCIs, aim at utilizing the brain signals to control prothetic arms or operate devices. Previously almost all the research on BCIs focuses on decoding signals from the contralateral hemisphere to implement BCI systems. However, the loss of functionality in the contralateral cortex often occurs due to strokes, resulting in total failure to motor function of fingers, hands, and limbs contralateral to the damaged hemisphere. Recent studies indicate that the signals from ipsilateral cortex is relevant to the planning phase of motor movements. Therefore, it is critical to find out if human motor movements can be decoded using signals from the ipsilateral cortex. In the thesis, we propose using ECoG signals from the ipsilateral cortex to decode finger movements. To our knowledge, this is the first work that successfully detects finger movements using signals from the ipsilateral cortex. We also investigate the experiment design and decoding directional movements. Our results show high decoding performance. We also show the anatomical feature analysis for ipsilateral cortex in performing motor-associated tasks, and the features are consistent with previous findings. The result reveals promising implications for a stroke relevant BCI.
Liu, Yuzong, "Decoding Brain Activation from Ipsilateral Cortex using ECoG Signals in Humans" (2011). All Theses and Dissertations (ETDs). 513.
Permanent URL: http://dx.doi.org/10.7936/K7707ZGX