Health Intelligence Lab
Neurological diseases disrupt the brain's healthy function and cause severe personal and societal impact on patients and caretakers. Recent advances in wearables, brain implants, and sensing technology (e.g., EEG) enable us to design systems that continuously monitor patients' brain health and determine patient-specific treatments (e.g., brain stimulation, adaptive pharmacotherapy, surgery). However, there is a lack of efficient methods that translate continuous physiological data streams into meaningful biological models of underlying diseases, relate them to existing clinical knowledge and biomarkers, and provide actionable treatment parameters. Our research goals are to leverage rich multichannel EEG recordings and other clinical data (e.g., imaging, omics) under a robust ML-based framework to a) discover new insights into spatio-temporal patterns of brain pathophysiology, early disease processes, and biomarkers of brain health, b) design domain-guided ML models for neurological disease diagnosis, prognosis, and treatment recommendations, and c) utilize those insights/models to build trustworthy ML systems that improve the brain health of patients living with neurological diseases.
From a machine learning perspective, we address questions such as a) how to develop domain/data-centric ML models in low labeled-data settings, b) how to extract meaningful patterns from data in a self/un-supervised manner, c) how dataset shifts in the target environment impact the performance of ML models during deployment, and d) what strategies can help anticipate and protect against such scenarios. The clinical areas we focus on are epilepsy and dementia, for which we work closely with domain experts at several clinical institutions, including Mayo Clinic, Cleveland Clinic, OSF Healthcare, University of Chicago, and Carle.
Lab Members
Lead Investigator
Assistant Professor, Computer Science & Engineering, University of Minnesota
Visiting Scientist, Neurology, Mayo Clinic
Graduate Students
Gayal Kuruppu - Ph.D. Student in CS&E, UMN
Neeraj Wagh - Ph.D. Student in BIOE, UIUC
James Evans - Ph.D. Student in BIOE (co-advised with Prof. Brad Sutton)
Huzaifa Suri - MS Student in ECE
Teja Gupta - MS Student in ECE (now at Lockheed Martin)
Medical Students
Samarth Rawal - MD Student in Carle Illinois College of Medicine
Undergraduate Students
John Zhang (2021 - Present)
Michael Zeng (2021 - Present)
John Wei (2020 - Present)
Collaborators
Prof. Ravi Iyer, University of Illinois
Prof. Brad Sutton, University of Illinois
Dr. Gregory Worrell, Mayo Clinic
Dr. Brent Berry, Neurology, Mayo Clinic
Dr. David Jones, Mayo Clinic
Dr. Lara Jehi, Cleveland Clinic
Dr. Matt Bramlet, OSF Healthcare
Dr. Naoum Issa, University of Chicago
Dr. Graham Huesmann, Carle
Publications
Machine Learning
Neeraj Wagh, Jionghao Wei, Samarth Rawal, Brent Berry, Yogatheesan Varatharajah, "Assessing Robustness of EEG Representations under Data-shifts via Latent Space and Uncertainty Analysis", Neurips (accepted), 2022.
Teja Gupta, Neeraj Wagh, Samarth Rawal, Brent Berry, Yogatheesan Varatharajah, "Tensor decomposition analysis of a large scalp EEG corpus reveals physiologically meaningful and clinically correlated/consistent factors", Manuscript in Preparation, 2022.
Neeraj Wagh, Jionghao Wei, Samarth Rawal, Leland Barnard, Benjamin Brinkmann, Brent Berry, Gregory Worrell, David Jones, Yogatheesan Varatharajah, "Domain-guided Self Supervision of EEG Data Improves Downstream Classification Performance and Generalizability", Machine Learning for Health Symposium, 2021.
Rawal, Samarth, Yogatheesan Varatharajah. "SCORE-IT: A Machine Learning Framework for Automatic Standardization of EEG Reports", IEEE Signal Processing in Medicine & Biology Symposium, 2021.
Wagh, Neeraj, Yogatheesan Varatharajah. "EEG-GCNN: Augmenting Electroencephalogram -based Neurological Disease Diagnosis using a Domain-guided Graph Convolutional Neural Network", Machine Learning for Health Symposium Neurips Workshop, 2020.
Clinical
Wentao Li*, Yogatheesan Varatharajah*, Ellen Dicks, Leland Barnard, Benjamin Brinkmann, Gregory Worrell, Bradley Boeve, David Knopman, and David Jones, “Data-driven retrieval of population-level EEG features and their role in neurodegenerative diseases: Correlations with cognitive exam and fluid biomarkers,” Under Review, 2022.
Yogatheesan Varatharajah, Boney Joseph, Benjamin Brinkmann, Marcia Morita‐Sherman, Zachary Fitzgerald, Deborah Vegh, Dileep Nair, Richard Burgess, Fernando Cendes, Lara Jehi, Gregory Worrell. "Quantitative Analysis of Visually Reviewed Normal Scalp EEG Predicts Seizure Freedom Following Anterior Temporal Lobectomy". Epilepsia, 2022.
Krishnakant V Saboo, Chang Hu, Yogatheesan Varatharajah, Scott A Przybelski, Robert I Reid, Christopher G Schwarz, Jonathan Graff-Radford, David S Knopman, Mary M Machulda, Michelle M Mielke, Ronald C Petersen, Paul M Arnold, Gregory A Worrell, David T Jones, Clifford R Jack Jr, Ravishankar K Iyer, Prashanthi Vemuri. "Deep learning identifies brain structures that predict cognition and explain heterogeneity in cognitive aging". NeuroImage, 2022.
Yogatheesan Varatharajah, Brent Berry, Boney Joseph, Irena Balzekas, Tal Pal Attia, Vaclav Kremen, Benjamin Brinkmann, Ravishankar Iyer, Gregory Worrell. "Characterizing the electrophysiological abnormalities in visually-reviewed normal EEGs of drug-resistant focal epilepsy patients". Brain Communications, 2021.
Lab Pictures
Lab Lunch at Lao Sze Chuan
Summer High School Research Poster Presentation by Annie and Trish
Mayo SURF Poster Presentation by John and Dr. Iezzi
Neurips Poster Presentation by Neeraj and Brent