The Algonauts Project 2025 challenge assesses computational models' ability to predict fMRI-recorded brain activity during exposure to multimodal naturalistic movies, leveraging a curated subset of the CNeuroMod dataset. This dataset includes:
Training data:
- 80 hours per subject of fMRI responses
- Episodes from Seasons 1-6 of Friends
- Three feature films
- One documentary
Evaluation data:
- Season 7 of Friends (in-distribution test)
- Novel movies with distinct narrative/content features (out-of-distribution generalization test)
The challenge emphasizes OOD generalization as the primary selection criterion for winning models, pushing beyond traditional within-distribution performance metrics. This structure enables direct comparison between biological and artificial intelligence systems' capacity to handle complex, real-world sensory inputs.
Gifford AT, Bersch D, St-Laurent M, Pinsard B, Boyle J, Bellec L, Oliva A, Roig G, Cichy RM. 2025. The Algonauts Project 2025 Challenge: How the Human Brain Makes Sense of Multimodal Movies. arXiv preprint, arXiv:2501.00504. DOI: https://doi.org/10.48550/arXiv.2501.00504
Boyle J, Pinsard B, Borghesani V, Paugam F, DuPre E, Bellec P. 2023. The Courtois NeuroMod project: quality assessment of the initial data release (2020). 2023 Conference on Cognitive Computational Neuroscience.