NEUROBIOLOGICAL OVERLAPS AND DIFFERENCES IN METHAMPHETAMINE-INDUCED PSYCHOSIS AND SCHIZOPHRENIA: A RESTING-STATE FMRI AND MACHINE LEARNING STUDY

Abstract Background Methamphetamine-induced psychosis (MIP) is characterized by transient episodes, but a substantial portion persists long-term (Wearne &Cornish 2018). While symptom similarities exist between MIP and schizophrenia (Sz), their neural correlates remain uncertain. Furthermore, research into the neural basis underlying the presence and duration of MIP psychosis has been limited. Aims & Objectives This study aimed to investigate whether there would be differences in discriminatory ability when classifying methamphetamine (METH) users with or without psychosis and Sz using Sz classifier derived from resting-state fMRI. We also aimed to explore potential differences in the presence and duration of psychosis in METH users. Method Utilizing the machine learning algorithms (Yamashita et al. 2020), we constructed an Sz classifier using whole-brain functional connectivity from resting-state fMRI data sourced from a Japanese database with 78 Sz patients and 340 healthy individuals (Tanaka et al. 2021). The classifier was tested using independent datasets. For the Taiwan dataset, which consisted of 35 Sz patients, 106 regular METH users and 45 healthy individuals, we applied this classifier to differentiate fMRI data of Sz patients and METH users from healthy individuals. METH users were grouped as MNP (no psychosis, 30 participants), MBP (psychosis lasting 7-30 days, 52 participants), and MPP (psychosis >30 days, 24 participants). Discriminatory abilities of these groups were then compared. Results The Sz classifier's performance ranged from an AUC of 0.77 to 0.86, indicating reliable discriminatory performance. In the Taiwan dataset, Sz exhibited the highest AUC (0.90), followed closely by MPP (AUC: 0.82). Both MBP and MNP displayed values around 0.6. These patterns were consistent in accuracy. Discussion & Conclusion Both Sz and MPP demonstrated high discriminatory capacities, in contrast to MBP and MNP, which showed less distinction. This suggests that persistent MIP shares neurobiological characteristics more similar to Sz, while transient MIP shares more similar to METH users without psychosis and healthy individuals. The slight differences in performance between Sz and MPP might shed light on their respective pathophysiology. Our method provides a framework for categorizing METH users by their psychosis manifestation and duration, highlighting shared and distinct neurobiological aspects of MIP and Sz. References Wearne, T. A., &Cornish, J. L. (2018). A Comparison of Methamphetamine-Induced Psychosis and Schizophrenia: A Review of Positive, Negative, and Cognitive Symptomatology. Frontiers in psychiatry, 9, 491. Yamashita, A., Sakai, Y., Yamada, T., Yahata, N., Kunimatsu, A., Okada, N., Itahashi, T., Hashimoto, R., Mizuta, H., Ichikawa, N., Takamura, M., Okada, G., Yamagata, H., Harada, K., Matsuo, K., Tanaka, S. C., Kawato, M., Kasai, K., Kato, N., Takahashi, H. &Imamizu, H. (2020). Generalizable brain network markers of major depressive disorder across multiple imaging sites. PLoS biology, 18(12), e3000966. Tanaka, S. C., Yamashita, A., Yahata, N., Itahashi, T., Lisi, G., Yamada, T., Ichikawa, N., Takamura, M., Yoshihara, Y., Kunimatsu, A., Okada, N., Hashimoto, R., Okada, G., Sakai, Y., Morimoto, J., Narumoto, J., Shimada, Y., Mano, H., Yoshida, W., Seymour, B. &Imamizu, H. (2021). A multi-site, multi-disorder resting- state magnetic resonance image database. Scientific data, 8(1), 227.