Their paper was entitled “Development of an Electroencephalogram-based (EEG) Brain Computer Interface for Upper Limb Control in Spinal Cord Injury Patients”, a robotic assistive device that decodes and translates brain signals into commands that guide the movement of artificial limbs attached to SCI patients.

“In our research study, we investigate whether the use of data transformation and augmentation techniques, together with Hybrid-scale Convolutional Neural Network (CNN), could allow us to record a few data from SCI patients, and then create a synthetic training data to compensate for the training of CNN while maintaining the desirable degree of accuracy,” Sigua said in the presentation.

Moreover, the ten participants, aged between 20 to 69 years old, were patients with subacute and chronic cervical SCI, and were experiencing restricted hand movements.

In summary, the team proposed a hybrid-scale convolutional neural network (HS-CNN) architecture with integrated data augmentation and, thereby, showed significant improvement in the accuracy of EEG motor-imagery classification of hand movements for SCI patients. Moreover, their proposed method has shown a 23.25% improvement rate to patients with SCI.

Department of Electronics Engineering faculty member Engr. Seigfred V. Prado, MRes, MSc, SMIEEE served as the team’s adviser.

Watch their presentation here:

The post BS-ECE students recognized in NAST 2021 conference appeared first on University of Santo Tomas.