پتانسیلهای برانگیخته شنوایی برنامه های کاربردی در مقابل تحقیقات بالینی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|36959||1997||13 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Psychiatry Research, Volume 69, Issues 2–3, 24 March 1997, Pages 183–195
Evidence of abnormal auditory evoked potentials (EPs) in patients suffering from schizophrenia has been accumulating. In spite of the magnitude of the EPs in schizophrenia literature, EPs have not been found to be clinically useful thus far. In this study we attempted to replicate the findings in a large sample of schizophrenia patients, and describe how auditory EPs may be used as supplemental tests in the differential diagnostic process. Five subject groups were formed; paranoid (PAR) and disorganized/undifferentiated (disorg/undiff) schizophrenics, schizoaffective (SA), bipolar, and a normal control group. All patients were stable on medications. Subjects underwent one EP recording session. Classification and regression trees (CART) based on EP amplitudes were used to classify subjects into subgroups. The optimal Bayes classification rule that minimizes the expected misclassification cost was then constructed for various misclassification cost functions. In a standard ‘Odd Ball’ paradigm the N100 amplitudes were significantly decreased in the disorg/undiff group than in the bipolar or normal subjects. The P200 amplitude was smaller in the PAR, disorg/undiff and the SA groups than in the normal controls. Both the disorg/undiff and the PAR groups had significantly lower P300 amplitudes than the normal controls. Classification rules used to classify subjects into normal or ill were sensitive to the relative cost of misclassifying a subject, as well as the prior clinical probability that this subject was ill. Our data largely agree with the existing literature showing abnormally decreased N100, P200, and P300 amplitudes in schizophrenic patients, particularly the disorg/undiff patients. We conclude that whether EP measures are clinically useful depends on the clinical situation. In particular, the prior probability of the diagnosis in question being present and the cost of misclassifying the patient are critical.