Lune P Bellec #neuroAI preprint alert: studying the convergence of multimodal AI features with brain activity during movie watching. Notably identifies brain areas where multimodal features outperform unimodal stimuli representations. And it uses @cneuromod.ca 's movie10 dataset :) https://arxiv.org/abs/2505.20027
arXiv.orgMulti-modal brain encoding models for multi-modal stimuliDespite participants engaging in unimodal stimuli, such as watching images or silent videos, recent work has demonstrated that multi-modal Transformer models can predict visual brain activity impressively well, even with incongruent modality representations. This raises the question of how accurately these multi-modal models can predict brain activity when participants are engaged in multi-modal stimuli. As these models grow increasingly popular, their use in studying neural activity provides insights into how our brains respond to such multi-modal naturalistic stimuli, i.e., where it separates and integrates information across modalities through a hierarchy of early sensory regions to higher cognition. We investigate this question by using multiple unimodal and two types of multi-modal models-cross-modal and jointly pretrained-to determine which type of model is more relevant to fMRI brain activity when participants are engaged in watching movies. We observe that both types of multi-modal models show improved alignment in several language and visual regions. This study also helps in identifying which brain regions process unimodal versus multi-modal information. We further investigate the contribution of each modality to multi-modal alignment by carefully removing unimodal features one by one from multi-modal representations, and find that there is additional information beyond the unimodal embeddings that is processed in the visual and language regions. Based on this investigation, we find that while for cross-modal models, their brain alignment is partially attributed to the video modality; for jointly pretrained models, it is partially attributed to both the video and audio modalities. This serves as a strong motivation for the neuroscience community to investigate the interpretability of these models for deepening our understanding of multi-modal information processing in brain.
Wed 12 Feb 2025 
2-3pm GMT
Details and registration: 
Exploring secrets of human vision today at 
When: Wednesday 9th of January 2025 at 12 noon.
Read our findings: 
Sharing is highly appreciated!
) c’est parti!
Submissions for 
TA applications will be extended by one week! 
Apply here: 
so happy this is out! I wasn’t sure how the ML models would perform, given the frequency shift in monkey (and human) SWRs, and the larger post-ripple wave compared to those in mice. But, a few models rose to the top straightaway and performed even better with modest “top up” training! 
Now I’m curious how the 
both senior authors and one lead author were 
