Functional MRI Biomarkers of Addiction-Related Neuroplasticity in Neurological Disorders: A Comprehensive Review of Functional Connectivity, Brain Network Reorganization, and Clinical Applications

Abstract

Recent advances in resting-state functional MRI have enabled non-invasive characterization of large-scale functional networks, providing novel biomarkers for diagnosis, prognosis, and treatment monitoring.Functional magnetic resonance imaging (fMRI)-particularly resting-state functional connectivity (rsFC) and graphtheoretical network analysis-has become the principal noninvasive method for characterizing addiction-related neuroplasticity in humans.This narrative review synthesizes fMRI evidence on addiction-related network reorganization across substance use disorders, including nicotine, alcohol, opioids, cocaine and other stimulants, and cannabis, as well as addiction-adjacent phenomena arising within primary neurological disease, with particular emphasis on impulse control disorders in Parkinson's disease and substance-use vulnerability following traumatic brain injury.The review the rapidly evolving computational layer built on these findings, including graph neural networks, explainable artificial intelligence methods for biomarker attribution, multimodal foundation models pretrained on large unlabeled neuroimaging corpora, and federated learning and harmonization strategies designed to enable multi-site, privacy-preserving biomarker discovery.We also examine fMRI-guided neuromodulation, including repetitive transcranial magnetic stimulation, transcranial direct current stimulation, and deep brain stimulation, as both a mechanistic probe and a treatment strategy, and evaluate the reproducibility, generalizability, and clinical translation barriers that currently constrain this literature.We conclude with a precision-medicine roadmap emphasizing standardized acquisition and analysis pipelines, prospective longitudinal and multimodal designs, externally validated and explainable machine-learning biomarkers, and dedicated connectivity studies in understudied neurological-comorbidity populations.Realizing the clinical potential of fMRI biomarkers of addiction-related neuroplasticity will require moving from descriptive grouplevel neuroimaging findings toward reproducible, individually actionable, and mechanistically interpretable tools.