Movie encoding improves ADHD detection from brain scans during movies
MovieSTAGE: Scene, Transition, and Global Encoding for Movie-fMRI ADHD Classification
Machine Learning
Summary
Detecting ADHD using brain scans taken while people watch movies can be tricky because the brain activity changes with the movie’s story. The authors created a method called MovieSTAGE that looks at brain activity during different scenes, the changes between scenes, and the whole movie’s patterns. This method did better at telling apart people with ADHD from those without it compared to simpler approaches. Their results suggest that paying attention to how the brain reacts to different parts of a movie helps identify ADHD-related differences.
What this means in practice
- •For clinical neuroimaging teams: Use movie-based brain scan analysis with event-aligned models to better classify ADHD subtypes in patient data.
- •For medical imaging software developers: Integrate multiscale movie-fMRI encoding methods to improve brain disorder diagnostic tools involving dynamic stimuli.
Tested on one dataset.
Authors
Boseong Kim, Haejun Chung, Ikbeom Jang
Abstract
Naturalistic movie-fMRI provides a shared, temporally structured probe of brain dynamics, yet predictive models commonly rely on whole-run functional connectivity (FC) or temporally generic representations that are not aligned with narrative events. We introduce MovieSTAGE (Scene, Transition, and Global Encoding), a multiscale framework that combines hypergraph-structured FC-profile organization within scenes, unsigned FC-profile differences across adjacent scenes, and whole-movie FC. We evaluated 260 participants from the CMI-HBN Despicable Me cohort on case-control, ADHD-subtype, and three-class classification using 10 repetitions of stratified five-fold cross-validation, complete out-of-fold (OOF) predictions, and paired subject-cluster bootstrap and permutation tests. MovieSTAGE achieved AUROCs of 0.69, 0.73, and 0.75 and balanced accuracies of 67.6%, 69.8%, and 58.3%, respectively, yielding the highest mean point estimates among the evaluated methods. On the three-class task, the full model outperformed all two-branch variants, the HGNN scene encoder outperformed MLP, GAT, and BNT alternatives under matched settings, and the human-annotated partition outperformed duration-matched random and fixed-count GSBS controls. These controlled results support incremental predictive value from event-aligned scene and transition representations when combined with whole-movie FC in this cohort. Post-hoc model-derived analyses generated network-level hypotheses involving frontoparietal and default-mode systems.