Date of Completion
5-12-2019
Embargo Period
3-28-2019
Advisors
Chi-Ming Chen
Field of Study
Psychological Sciences
Degree
Master of Science
Open Access
Open Access
Abstract
Ketamine provides a useful model for studying the physiopathology of schizophrenia as it induces and exacerbates schizophrenic symptoms in healthy individuals and patients, respectively. However, it remains unclear how ketamine affects neural oscillations across frequency bands and the extent to which changes in cortical firing are independent from ketamine’s cardiovascular effects. The present study used 1-norm support vector machine (SVM) as a classification approach to determine sources of psychophysiological signal, considering both EEG and ECG, characterizing primary effects of ketamine administration. EEG and ECG recordings of 11 participants were collected before (saline) and after the ketamine (0.5 mg/kg) intravenous infusion The 1-norm SVM model built on EEG and ECG in combination successfully differentiated the two sessions and identified prominent EEG features associated with neuronal oscillation alternation between sessions, which includes beta, delta frequency bands from both F3 and F4. However, SVM models based on EEG or ECG, taken alone, performed only slightly above chance. 1-norm SVM models suggest that machine learning methods could successfully distinguish saline and ketamine sessions, but only when both EEG and ECG single trial data were considered. Our findings show concurrent cardiac and brain effects of ketamine that may resemble alterations in the prodromal phase of schizophrenia.
Recommended Citation
Yang, Xiao, "1-norm Support Vector Machine on Single-trial EEG and ECG Data to Identify Neural Oscillatory Features in the Ketamine Model for Schizophrenia" (2019). Master's Theses. 1382.
https://digitalcommons.lib.uconn.edu/gs_theses/1382
Major Advisor
Chi-Ming Chen