Exponentially Embedded Family to Detect High-Frequency Oscillations in Human Tripolar Laplacian EEG

Undergraduate #366
Discipline: Technology and Engineering
Subcategory: Biomedical Engineering

Frederick J. Lee - Diné College
Co-Author(s): Mark Musngi, Diné College; Oleksandr Makeyev, Diné College, AZ

Epilepsy affects approximately 67 million people worldwide with up to 75% from developing countries [1]. Diagnosing epilepsy using electroencephalogram (EEG) is complicated due to its poor signal-to-noise ratio, high sensitivity to various forms of artifacts, and low spatial resolution. Laplacian EEG signal via novel and noninvasive tripolar concentric ring electrodes (tEEG) is superior to EEG via conventional disc electrodes due to its unique capabilities which allow automatic attenuation of common movement and muscle artifacts in applications including detection of high-frequency oscillations (HFOs) and seizure onset zones [1] and seizure onset detection [2]. In [1], detection of HFOs, a promising bio-marker of epileptogenesis and ictogenesis, was performed manually exposing HFOs in tEEG preceding 100% of seizures. In [2], an exponentially embedded family (EEF) based detector was used for automatic detection of seizure onset with 100% accuracy on human tEEG data. In this work, we apply EEF to automatically detect HFOs in tEEG data from human patients with epilepsy. This research is important because it allows assessing the potential of automatically detecting HFOs in tEEG with the ultimate goal of using them as auxiliary features for seizure onset detection to improve diagnostic yield of tEEG for epilepsy. To create an automatic HFO detector EEF is applied to power in individual and/or combined HFO frequency bands (gamma activity, ripples, fast ripples) rather than to power across the spectrum as was done for seizure onset detection in [2]. The dataset from [2] is used to validate the proposed detector including tEEG data from 7 human patients with epilepsy (26.3 hours of data total including 5 seizures). Automatic HFO detection is performed on complete recordings (not just up to an hour prior to clinical seizures as in [1]) both for patients with clinical seizures, patients with epileptiform activity but no seizures, and patients with neither epileptiform activity nor seizures using non patient-specific model. Accuracy of EEF based automatic HFO detection on tEEG dataset from [2] will be presented and discussed. Preliminary results suggest the potential of the approach and feasibility of detecting HFOs in tEEG data using the EEF based detector. Further investigation on a larger dataset is needed for a conclusive proof. References: 1. Besio W.G., Martínez-Juárez I.E., Makeyev O., Gaitanis J.N., Blum A.S., Fisher R.S., Medvedev A.V., High-frequency oscillations recorded on the scalp of patients with epilepsy using tripolar concentric ring electrodes, IEEE J. Transl. Eng. Health Med., Vol. 2, Article 2000111, June 2014, p. 11. 2. Makeyev O., Ding Q., Martínez-Juárez I.E., Gaitanis J., Kay S., Besio W.G., Multiple sensor integration for seizure onset detection in human patients comparing conventional disc versus novel tripolar concentric ring electrodes, 35th Annual International Conference of the IEEE EMBS, Osaka, Japan, July 3-7, 2013, pp. 17-20.

Not Submitted

Funder Acknowledgement(s): This research was supported, in part, by the National Science Foundation (NSF) Division of Human Resource Development (HRD) Tribal Colleges and Universities Program (TCUP) award number 1622481 to O. Makeyev.

Faculty Advisor: Oleksandr Makeyev, omakeyev@dinecollege.edu

Role: As an undergraduate research assistant and a part of the group working on this NSF award, I have been involved in all aspects of this research project. In particular I have taken the lead on theory and application of EEF detector to this particular problem. I am in charge of running the HFO detection experiments.