Predictive validities of neurodevelopmental markers in schizophrenia and adolescents with risk behaviors

  • 甘 淑婷

Student thesis: Master's Thesis


Background The neurodevelopmental hypothesis proposes that schizophrenia is originated from aberrant brain development Minor physical anomalies (MPA) and neurological soft signs (NSS) are suggested as a biomarker associated with disruptions of fetal development Numerous studies report an increased frequency of neurodevelopmental impairments in schizophrenic individuals compared with controls Risk behaviors are those that can have negative effects on the overall development and well-being of youth There were limited researches on the relationship between MPA and risk behaviors which may be derived from abnormal neurodevelopment The aim of this study was to identify which speci?c MPA and NSS are more associated with schizophrenia and risk behaviors and to determine the optimal predictive value of MPA and NSS scores in schizophrenia and adolescents with risk behaviors Methods We developed a modified physical measurement scale composited of both qualitative and quantitative items and used Neurological Evaluation Scale (NES) to assess the NSS A self-administered questionnaire for the assessment of risk behaviors was also developed Group comparisons were conducted by using mixed-effect model among schizophrenia their first-degree relatives and stepwise logistic regression analysis between case and control of risk behavior The ROC curve was conducted to evaluate predictive accuracy sensitivity and specificity of these neurodevelopmental markers Results There were 210 patients with schizophrenia 109 nonpsychotic first-degree relatives 151 normal controls and a total of 118 adolescents aged 11-19 were recruited in this study The results of ROC curve analysis for schizophrenia were as follows: The analysis of MPA and NSS alone provided an accuracy of 75 8% and 80 6% respectively for schizophrenia vs controls The composite MPA and NSS provide a greater predictive validity in the model of patients with schizophrenia vs controls provided an accuracy of 83 8% The results of ROC curve analysis for risk behaviors were as follows: The model of violence vs controls provided an accuracy of 64 4 %; the model substance abuse vs controls provided an accuracy of 65 3%; the model of sex behavior vs controls provided an accuracy of 78 8%; the model of emotion vs controls provided an accuracy of 68 6% Conclusion We found that subjects were most accurately classified when MPA & NSS were considered as a composite neurodevelopmental marker rather than independently The MPA and NSS are more frequent in patients with schizophrenia their relatives and MPA were slightly different in adolescents with risk behaviors compared to controls which are consistent with the hypothesis of abnormal neurodevelopment in schizophrenia and adolescents with risk behaviors
Date of Award2014 Sep 3
Original languageEnglish
SupervisorSheng-Hsiang Lin (Supervisor)

Cite this