Multivariate predictive logistic regression model for schizophrenia.

Multiple logistic regression model: n = 174, percentage of concordant pairs = 90.5%, percentage of discordant pairs = 9.3%, percentage of tied pairs = 0.1%, adjusted generalized R2 = 0.61, Deviance goodness-of-fit test p = 0.985>0.05 (df = 169), Pearson goodness-of-fit test p = 0.534>0.05 (df = 169), and Hosmer and Lemeshow goodness-of-fit test p = 0.816>0.05 (df = 8). The estimated probability of having schizophrenia (i.e., the predicted value,) can be calculated using the following formula (FCz is the mean mismatch negativity amplitude at electrode FCz; Arithmetic, Block Design, and Performance IQ are test scores of WAIS-III): .