
Preprints with The Lancet is part of SSRN´s First Look, a place where journals identify content of interest prior to publication. Authors have opted in at submission to The Lancet family of journals to post their preprints on Preprints with The Lancet. The usual SSRN checks and a Lancet-specific check for appropriateness and transparency have been applied. Preprints available here are not Lancet publications or necessarily under review with a Lancet journal. These preprints are early stage research papers that have not been peer-reviewed. The findings should not be used for clinical or public health decision making and should not be presented to a lay audience without highlighting that they are preliminary and have not been peer-reviewed. For more information on this collaboration, see the comments published in The Lancet about the trial period, and our decision to make this a permanent offering, or visit The Lancet´s FAQ page, and for any feedback please contact preprints@lancet.com.
Multivariate Classification Based on Large-Scale Brain Networks During Early Abstinence Predicted Relapse Among Male Detoxified Alcohol- Dependent Patients
19 Pages Posted: 14 Mar 2022
More...Abstract
Objective: Identification of biomarkers to predict relapse of alcohol-dependent (AD) patients remains remarkably challenging. The current study aimed to identify neurobiological features based on connectivity of brain networks that may be predictive of relapse after detoxification.
Methods: Sixty-six male AD patients in the early-abstinence stage after hospitalized detoxification underwent resting-state functional magnetic resonance imaging. We used both traditional analysis (ANOVA) and multivariate pattern analysis with a relevance vector machine (RVM) on large-scale brain networks with a novel-created dimensionality of 2,178 (242 nodes × 9 networks) connections to predict a relapse during the 6-month follow-up.
Results: During a 6-month follow-up, 38 (57.6%) of the AD patients drank again. Although ANOVA did not reveal any significant indicators based on connectome analysis for differentiation of relapsing vs. abstinent patients, RVM analysis yielded a model with high predictive performance (area under ROC curve =0.912). The accuracy of the RVM model documented by leave-one-out cross-validation was 0.833. The most heavily weighted connections for the relapse classification were among the limbic network (LIM), visual network (VIS), dorsal attention network, subcortical network (Sub) and cerebellum network. In addition, the intra-connections of VIS with LIM ( r = 0.44, p = 0.005) and Sub ( r = -0.46, p =0.004) were significantly associated with severity of relapse.
Conclusions: Application of RVM model in this study provided an elegant and efficient way to capture key information about neuroimaging biomarkers for AD prognosis, which indicates that deficits in networks of visual attention and the affective cognition behavior system could be predictors of relapse respect to AD patients.
Funding: The study was partly supported by grants from the National Natural Science Foundation (grant no. 81571305), the Department of Science and Technology of Sichuan provincial government (grant no. 2019YFS0153), the “1.3.5” project for disciplines of excellence, West China Hospital, Sichuan University (grant no. 2019HXFH026), and the Introduction Project of Suzhou Clinical Expert Team (grant no. SZYJTD201715).
Declaration of Interest: The authors declare no conflicts of interest.
Ethical Approval: This longitudinal study was approved by the Ethics Committee of West China Hospital of Sichuan University in 2016 (NO. 22). Informed written consent was obtained from every participant in this study
Keywords: alcohol dependence, relapse, predictor, rest-functional magnetic resonance imaging (rs-fMRI), relevance vector machine, multivariate pattern analysis
Suggested Citation: Suggested Citation