Spatial, temporal, and spectral relationships in event-related activity in the human electroencephalogram: Methods to define distinct phenotypes of psychosis
Abstract
Event-related potential paradigms are widely studied in the search for biological endophenotypes of schizophrenia. Abnormalities observed on two of these measures, P50 suppression and P300 generation, are considered some of the most reliable and robust effects in the schizophrenia literature, and appear to aggregate within affected families and high-risk groups. Evidence for similar findings in bipolar disorder, however, warrants more systematic evaluation before proposals regarding a genetic overlap with schizophrenia can be entertained. In the current study, the diagnostic efficiency of candidate endophenotypes for schizophrenia (SZ; n = 84) was tested in classification analyses including healthy normal (HN; n = 81) and bipolar disorder (BP; n = 50) comparison groups. P50 suppression and P300 amplitude values were scored according to conventional procedures. Spectral power and hemispheric asymmetry measures were used to assess evoked-response activity within functionally discrete frequency bands, and at electrode sites more proximal to bilateral temporal generators. Values were entered as predictors in multivariate regression models designed to optimize sensitivity and specificity of diagnostic classifications. Seventy-nine percent classification accuracy of SZ against HN, and 64% against BP, was achieved using a derived multivariate model. These values reflected a substantial improvement in diagnostic specificity over that achieved by standard P50 and P300 amplitude measures. A meaningful distinction in profiles was observed between BP, characterized by impairment in late stages of information processing (e.g., novelty detection, encoding) and SZ, characterized by a combination of these deficits with earlier, sensory-related, processing abnormalities. Interpretations of results regarding diagnostic overlap and differentiation of model parameters are discussed.
Degree
Ph.D.
Advisors
McGrew, Purdue University.
Subject Area
Clinical psychology
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.