The main objective of this Scientist Development Award for Clinicians (SDAC) is to establish the Principal Investigator, Dr. Krystal, as an independent clinical researcher in neuropsychiatry with skill in quantitative EEG analysis. Richard D. Weiner, M.D., Ph.D., Associate Professor of Psychiatry at DUMC, Director of the DUMC ECT Program and Quantitative EEG Laboratory, and Chairman of the American Psychiatric Association Task Force on ECT will serve as Sponsor. The main objective will be met through: (1) course work in multivariate statistics, signal analysis, non-linear systems dynamics, and system modelling; (2) tutorials in research methodology, instruments for diagnosing major depression, measuring outcome of electroconvulsive therapy, clinical decision analysis, signal processing, non-linear dynamics, modeling, and quantitative EEG; (3) expert consultation; and (4) a supervised research experience. The proposed two-stage research program is designed to develop ictal electroencephalographic (EEG) markers of electroconvulsive therapy (ECT) treatment adequacy, which will allow ECT to be delivered in a safer and more effective manner. Stage I consists of 2 retrospective pilot analyses of data already collected, in which (1) a multivariate ictal EEG model of treatment adequacy based on manual ratings is developed for 2 channel data, and (2) existing 21-channel computer ictal EEG data are analyzed to determine an optimal recording electrode set for prediction of treatment adequacy. In the latter case, the electrode choice will be achieved by studying the inter-electrode relationships of ictal EEG data to choose the set of electrodes that records as much of the information available in the 21 channel data as possible, while minimizing redundancy. Stage II, building on the results of Stage I and three promising pilot studies, uses prospectively recorded ictal EEG data in depressed subjects receiving ECT to develop, refine, and test both computer derived and manually rated multivariate models of ECT seizure adequacy for their ability to predict therapeutic response. The final models will be compared for overall clinical utility taking into account simplicity, capacity for automation, and ability to predict treatment outcome.
Showing the most recent 10 out of 14 publications
