Classification of schizophrenia spectrum disorder using machine learning and functional connectivity: reconsidering the clinical application

Abstract Background Early identification of Schizophrenia Spectrum Disorder (SSD) is crucial for effective intervention and prognosis improvement. Previous neuroimaging-based classifications have primarily focused on chronic, medicated SSD cohorts. However, the question remains whether brain metrics...

Full description

Saved in:
Bibliographic Details
Main Authors: Chao Li, Ji Chen, Mengshi Dong, Hao Yan, Feng Chen, Ning Mao, Shuai Wang, Xiaozhu Liu, Yanqing Tang, Fei Wang, Jie Qin
Format: Article
Language:English
Published: BMC 2025-04-01
Series:BMC Psychiatry
Subjects:
Online Access:https://doi.org/10.1186/s12888-025-06817-0
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850181101372309504
author Chao Li
Ji Chen
Mengshi Dong
Hao Yan
Feng Chen
Ning Mao
Shuai Wang
Xiaozhu Liu
Yanqing Tang
Fei Wang
Jie Qin
author_facet Chao Li
Ji Chen
Mengshi Dong
Hao Yan
Feng Chen
Ning Mao
Shuai Wang
Xiaozhu Liu
Yanqing Tang
Fei Wang
Jie Qin
author_sort Chao Li
collection DOAJ
description Abstract Background Early identification of Schizophrenia Spectrum Disorder (SSD) is crucial for effective intervention and prognosis improvement. Previous neuroimaging-based classifications have primarily focused on chronic, medicated SSD cohorts. However, the question remains whether brain metrics identified in these populations can serve as trait biomarkers for early-stage SSD. This study investigates whether functional connectivity features identified in chronic, medicated SSD patients could be generalized to early-stage SSD. Methods Data were collected from 502 SSD patients and 575 healthy controls (HCs) across four medical institutions. Resting-state functional connectivity (FC) features were used to train a Support Vector Machine (SVM) classifier on individuals with medicated chronic SSD and HCs from three sites. The remaining site, comprising both chronic medicated and first-episode unmedicated SSD patients, was used for independent validation. A univariable analysis examined the association between medication dosage or illness duration and FC. Results The classifier achieved 69% accuracy (p = 0.002), 63% sensitivity, 75% specificity, 0.75 area under the receiver operating characteristic curve, 69% F1-score, 72% positive predictive rate, and 67% negative predictive rate, when tested on an independent dataset. Subgroup analysis showed 71% sensitivity (p = 0.04) for chronic medicated SSD, but poor generalization to first-episode unmedicated SSD (sensitivity = 48%, p = 0.44). Univariable analysis revealed a significant association between FC and medication usage, but not disease duration. Conclusions Classifiers developed on chronic medicated SSD may predominantly capture state features of chronicity and medication, overshadowing potential SSD traits. This partially explains the current classifiers’ non-generalizability across SSD patients with different clinical states, underscoring the need for models that can enhance the early detection of schizophrenia neural pathology.
format Article
id doaj-art-f3115b2f066f4ca1ab7ae98d96bcfe4b
institution OA Journals
issn 1471-244X
language English
publishDate 2025-04-01
publisher BMC
record_format Article
series BMC Psychiatry
spelling doaj-art-f3115b2f066f4ca1ab7ae98d96bcfe4b2025-08-20T02:17:58ZengBMCBMC Psychiatry1471-244X2025-04-0125111110.1186/s12888-025-06817-0Classification of schizophrenia spectrum disorder using machine learning and functional connectivity: reconsidering the clinical applicationChao Li0Ji Chen1Mengshi Dong2Hao Yan3Feng Chen4Ning Mao5Shuai Wang6Xiaozhu Liu7Yanqing Tang8Fei Wang9Jie Qin10Department of Radiology, The Third Affiliated Hospital of Sun Yat-Sen UniversityCenter for Brain Health and Brain Technology, Global Institute of Future Technology, Institute of Psychology and Behavioral Science, Shanghai Jiao Tong UniversityDepartment of Radiology, The Third Affiliated Hospital of Sun Yat-Sen UniversityPeking University Sixth Hospital, Institute of Mental HealthDepartment of Radiology, The Third Affiliated Hospital of Sun Yat-Sen UniversityYantai Yuhuangding Hospital, Qingdao UniversitySchool of Psychology, Shandong Second Medical UniversityEmergency and Critical Care Medical Center, Beijing Shijitan Hospital, Capital Medical UniversityDepartment of Psychiatry, Shengjing Hospital of China Medical UniversityEarly Intervention Unit, Department of Psychiatry, The Affiliated Brain Hospital of Nanjing Medical UniversityDepartment of Radiology, The Third Affiliated Hospital of Sun Yat-Sen UniversityAbstract Background Early identification of Schizophrenia Spectrum Disorder (SSD) is crucial for effective intervention and prognosis improvement. Previous neuroimaging-based classifications have primarily focused on chronic, medicated SSD cohorts. However, the question remains whether brain metrics identified in these populations can serve as trait biomarkers for early-stage SSD. This study investigates whether functional connectivity features identified in chronic, medicated SSD patients could be generalized to early-stage SSD. Methods Data were collected from 502 SSD patients and 575 healthy controls (HCs) across four medical institutions. Resting-state functional connectivity (FC) features were used to train a Support Vector Machine (SVM) classifier on individuals with medicated chronic SSD and HCs from three sites. The remaining site, comprising both chronic medicated and first-episode unmedicated SSD patients, was used for independent validation. A univariable analysis examined the association between medication dosage or illness duration and FC. Results The classifier achieved 69% accuracy (p = 0.002), 63% sensitivity, 75% specificity, 0.75 area under the receiver operating characteristic curve, 69% F1-score, 72% positive predictive rate, and 67% negative predictive rate, when tested on an independent dataset. Subgroup analysis showed 71% sensitivity (p = 0.04) for chronic medicated SSD, but poor generalization to first-episode unmedicated SSD (sensitivity = 48%, p = 0.44). Univariable analysis revealed a significant association between FC and medication usage, but not disease duration. Conclusions Classifiers developed on chronic medicated SSD may predominantly capture state features of chronicity and medication, overshadowing potential SSD traits. This partially explains the current classifiers’ non-generalizability across SSD patients with different clinical states, underscoring the need for models that can enhance the early detection of schizophrenia neural pathology.https://doi.org/10.1186/s12888-025-06817-0Schizophrenia spectrum disorderSchizophreniaMachine learningFunctional connectivityClassification
spellingShingle Chao Li
Ji Chen
Mengshi Dong
Hao Yan
Feng Chen
Ning Mao
Shuai Wang
Xiaozhu Liu
Yanqing Tang
Fei Wang
Jie Qin
Classification of schizophrenia spectrum disorder using machine learning and functional connectivity: reconsidering the clinical application
BMC Psychiatry
Schizophrenia spectrum disorder
Schizophrenia
Machine learning
Functional connectivity
Classification
title Classification of schizophrenia spectrum disorder using machine learning and functional connectivity: reconsidering the clinical application
title_full Classification of schizophrenia spectrum disorder using machine learning and functional connectivity: reconsidering the clinical application
title_fullStr Classification of schizophrenia spectrum disorder using machine learning and functional connectivity: reconsidering the clinical application
title_full_unstemmed Classification of schizophrenia spectrum disorder using machine learning and functional connectivity: reconsidering the clinical application
title_short Classification of schizophrenia spectrum disorder using machine learning and functional connectivity: reconsidering the clinical application
title_sort classification of schizophrenia spectrum disorder using machine learning and functional connectivity reconsidering the clinical application
topic Schizophrenia spectrum disorder
Schizophrenia
Machine learning
Functional connectivity
Classification
url https://doi.org/10.1186/s12888-025-06817-0
work_keys_str_mv AT chaoli classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT jichen classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT mengshidong classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT haoyan classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT fengchen classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT ningmao classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT shuaiwang classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT xiaozhuliu classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT yanqingtang classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT feiwang classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication
AT jieqin classificationofschizophreniaspectrumdisorderusingmachinelearningandfunctionalconnectivityreconsideringtheclinicalapplication