Computational neuroimaging strategies for single patient predictions

• Published in: NeuroImage (Orlando, Fla.) Vol. 145; no. Pt B; pp. 180 - 199
• Main Authors: Stephan, K.E., Schlagenhauf, F., Huys, Q.J.M., Raman, S., Aponte, E.A., Brodersen, K.H., Rigoux, L., Moran, R.J., Daunizeau, J., Dolan, R.J., Friston, K.J., Heinz, A.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 15.01.2017; Elsevier Limited
• Subjects: Attention deficit hyperactivity disorder; Bayesian; Brain Diseases - diagnostic imaging; Classification; Clustering; Computational psychiatry; Decision making; Disease; EEG; fMRI; Generative embedding; Generative model; Humans; Medical imaging; Mental Disorders - diagnostic imaging; Model comparison; Model evidence; Model selection; Models, Theoretical; Neuroimaging - methods; Patients; Physiology; Schizophrenia; Translational neuromodeling; fMRI; Computational psychiatry; Generative embedding; Model selection; EEG; Bayesian; Classification; Translational neuromodeling; Model comparison; Clustering; Generative model; Model evidence
• ISSN: 1053-8119; 1095-9572; 1095-9572
• DOI: 10.1016/j.neuroimage.2016.06.038

Neuroimaging increasingly exploits machine learning techniques in an attempt to achieve clinically relevant single-subject predictions. An alternative to machine learning, which tries to establish predictive links between features of the observed data and clinical variables, is the deployment of computational models for inferring on the (patho)physiological and cognitive mechanisms that generate behavioural and neuroimaging responses. This paper discusses the rationale behind a computational approach to neuroimaging-based single-subject inference, focusing on its potential for characterising disease mechanisms in individual subjects and mapping these characterisations to clinical predictions. Following an overview of two main approaches – Bayesian model selection and generative embedding – which can link computational models to individual predictions, we review how these methods accommodate heterogeneity in psychiatric and neurological spectrum disorders, help avoid erroneous interpretations of neuroimaging data, and establish a link between a mechanistic, model-based approach and the statistical perspectives afforded by machine learning. •Reviews computational neuroimaging strategies for single patient predictions.•Generative models for inferring individual disease mechanisms in psychiatry and neurology.•Mapping inferred mechanisms to clinical predictions by Bayesian model selection and•generative embedding.•Links a mechanistic model-based approach to statistical perspectives by machine learning.