Beitragstitel | Disorganized Gyrification Network Properties During the Transition to Psychosis |
---|---|
Beitragscode | P06 |
Autoren | |
Präsentationsform | Poster |
Themengebiete |
|
Abstract |
Background: There is urgent need to improve the limited prognostic accuracy of psychopathology-based classifications to predict the onset of psychosis in clinical high-risk (CHR) subjects for psychosis. However, as yet no reliable biological marker has been established to differentiate CHR subjects who will develop psychosis from those who will not. Methods: 44 healthy controls (HC), 63 at-risk mental state (ARMS) subjects without later transition to psychosis (ARMS-NT), 16 ARMS subjects with later transition (ARMS-T), and 38 antipsychotic-free patients with first-episode psychosis (FEP) underwent baseline magnetic resonance imaging (MRI) with follow-up assessment to determine the transition status of CHR subjects. Graph theory has been applied to construct gyrification connectomes and group differences in network measures were assessed using Functional Data Analysis. Extremely randomized trees with repeated, nested cross-validation was performed to differentiate ARMS-T from ARMS-NT individuals. Results: Small-worldness is reduced in both ARMS-T and FEP patients, secondary to reduced integration and increased segregation in both groups. Using the connectome properties as features, we obtained a high classification performance above 80%. Conclusion: Our findings indicate that gyrification-based connectomes might be a promising means to generate systems-based measures from anatomical data that improves individual prediction of psychosis transition in CHR subjects. |