Multimodal Recurrent Ensembles for Predicting Brain Responses to Naturalistic Movies (Algonauts 2025)

TL;DR

This paper introduces a hierarchical multimodal recurrent ensemble that predicts brain responses to movies by integrating visual, auditory, and semantic data, achieving competitive results in the Algonauts 2025 challenge.

Contribution

It presents a novel multimodal recurrent model that effectively encodes temporal dynamics across multiple modalities for brain response prediction.

Findings

Achieved third place in Algonauts 2025 challenge

Highest single-parcel peak score among participants

Strong performance on challenging subject data

Abstract

Accurately predicting distributed cortical responses to naturalistic stimuli requires models that integrate visual, auditory and semantic information over time. We present a hierarchical multimodal recurrent ensemble that maps pretrained video, audio, and language embeddings to fMRI time series recorded while four subjects watched almost 80 hours of movies provided by the Algonauts 2025 challenge. Modality-specific bidirectional RNNs encode temporal dynamics; their hidden states are fused and passed to a second recurrent layer, and lightweight subject-specific heads output responses for 1000 cortical parcels. Training relies on a composite MSE-correlation loss and a curriculum that gradually shifts emphasis from early sensory to late association regions. Averaging 100 model variants further boosts robustness. The resulting system ranked third on the competition leaderboard, achieving an overall Pearson r = 0.2094 and the highest single-parcel peak score (mean r = 0.63) among all participants, with particularly strong gains for the most challenging subject (Subject 5). The approach establishes a simple, extensible baseline for future multimodal brain-encoding benchmarks.