A multimodal neuroimaging classifier for alcohol dependence

Abstract

With progress in magnetic resonance imaging technology and a broader dissemination of state-of-theart imaging facilities, the acquisition of multiple neuroimaging modalities is becoming increasingly feasible. One particular hope associated with multimodal neuroimaging is the development of reliable data-driven diagnostic classifiers for psychiatric disorders, yet previous studies have often failed to find a benefit of combining multiple modalities. As a psychiatric disorder with established neurobiological effects at several levels of description, alcohol dependence is particularly well-suited for multimodal classification. To this aim, we developed a multimodal classification scheme and applied it to a rich neuroimaging battery (structural, functional task-based and functional resting-state data) collected in a matched sample of alcohol-dependent patients (N = 119) and controls (N = 97). We found that our classification scheme yielded 79.3% diagnostic accuracy, which outperformed the strongest individual modality-grey-matter density-by 2.7%. We found that this moderate benefit of multimodal classification depended on a number of critical design choices: a procedure to select optimal modalityspecific classifiers, a fine-grained ensemble prediction based on cross-modal weight matrices and continuous classifier decision values. We conclude that the combination of multiple neuroimaging modalities is able to moderately improve the accuracy of machine-learning-based diagnostic classification in alcohol dependence.

White matter diffusion estimates in obsessive-compulsive disorder across 1,653 individuals: Machine learning findings from the ENIGMA OCD Working Group

“…Our model based on white matter features showed improved classification performance compared with the grey matter morphometry model in adults and pediatric samples, though a direct comparison may not be warranted due to different machine learning pipelines and different subsamples used in this study. Future studies should determine whether multi-modal machine learning using structural and functional MRI can increase classification accuracy (Calhoun & Sui, 2016;Kuo et al 2021;Guggenmos et al 2020;Menon & Krishnamurthy, 2021).…”

Modern views of machine learning for precision psychiatry

“…Right: feature importance maps of functional neuroimaging modalities. 14 (B) Unsupervised learning. Left: whole-brain functional-connectivity matrix averaged across all subjects.…”

Promises and pitfalls of deep neural networks in neuroimaging-based psychiatric research

“…For alcohol use disorder (AUD), the most prominent SUD, sMRI and fMRI studies have revealed a number of neurobiological correlates including enlarged ventricles, grey and white matter loss in frontal and reward-related brain areas, as well as altered functional connectiv-ity in the amygdala and nucleus accumbens [232,233,234,235,236]. Classical machine learning models have been employed to identify AUD or predict alcohol consumption / binge drinking on different kinds of data including demographics, history of life events, personality traits, cognition, candidate genes, as well as brain structure and function [237,238,239,240,174]. For sMRI data, it has been shown that a computer-based classification approach performed better than a blinded radiologist in diagnosing alcohol dependence based on regional grey matter (74% to 66%) and predicting future alcohol consumption [237].…”