Sai Sanjay Balaji awarded UMII-MnDRIVE Graduate Assistantship

Sanjay's research will overcome one of the limitations of existing work on automated seizure localization by developing a graph neural network for identifying the seizure onset zone.

Doctoral candidate Sai Sanjay Balaji was recently awarded the UMII-MnDRIVE Graduate Assistantship. The assistantship will support him in his research on causal brain network analysis using machine learning, particularly the development of a graph neural network for the identification of seizure onset zones. He is working under the guidance of Distinguished McKnight University and Erwin A. Kelen Chair and Professor of Electrical and Computer Engineering Keshab Parhi.

The research community’s relatively recent approach towards the brain as a graph network has opened up new pathways to understand and describe the interactions among its different regions. One area of research that has particularly benefited from this is seizure detection and prediction using signal processing and machine learning.

While there has been significant work done on automating seizure identification and prediction, seizure localization is a relatively new topic. Currently, physicians locate the seizure onset zone (SOZ), the region within the brain responsible for inducing seizures, by visual inspection of electroencephalogram (EEG) signals. In fact, almost all existing research on the automated identification of SOZ is done retrospectively. However the recent massive increase in available data, and the success of neural networks for biomedical datasets has unlocked new avenues for the identification of SOZ, factors that have motivated Sanjay to pursue this area of research.

A significant limitation of existing work on automated seizure localization is the extensive process of extracting viable biomarkers from EEG signals (the features that are effective indicators for SOZ identification) and then choosing the ones that would be helpful in accurate identification. The process depends on a successful model built on substantial knowledge of the domain (knowledge that is typically the province of a specialist medical professional). Sanjay intends to overcome this limitation by developing a graph neural network for identifying SOZ using intracranial EEG signals, which will learn brain connectivity unaided, without much preprocessing.

Prior to receiving the assistantship, Sanjay was a teaching assistant, a position that occupied a sizable chunk of his time each week. With the UMII-MnDRIVE assistantship he will now be able to devote his time solely to his research. The assistantship also includes a travel grant which will support his travel to a relevant conference to present his research.  

Sanjay earned his bachelor’s degree in electronics and instrumentation from Anna University in India. His interest in biomedical devices and systems goes back to his high school days. Keen on honing his skills in the area to prepare for a career in biomedical systems and devices he joined the University of Minnesota in 2019.

The University of Minnesota Informatics Institute MnDRIVE Graduate Assistantship program supports UMN PhD candidates pursuing research at the intersection of informatics and any of the five MnDRIVE areas: Robotics, Sensors and Advanced Manufacturing; Global Food Ventures; Advancing Industry, Conserving Our Environment; Discoveries and Treatments for Brain Conditions; and Cancer Clinical Trials.