Check out Colton Casto gorgeous preprint on cerebellar language functions - a true tour de force that I was lucky to be able to take part in.

**ecstatic** to share our ICLR 2025 paper: sparse components distinguish visual pathways & their alignment to neural networks, with Nancy Kanwisher @[email protected] and Meenakshi Khosla (openreview.net/forum?id=IqHeD…) 1/n

What are the organizing dimensions of language processing? We show that voxel responses are organized along 2 main axes: processing difficulty & meaning abstractness—revealing an interpretable, topographic representational basis for language processing shared across individuals.

Thrilled to announce the 2025 recipients of #KempnerInstitute Research Fellowships: Elom Amemastro, Ruojin Cai, David Clark, Alexandru Damian, William Dorrell, Mark Goldstein, Richard Hakim, Hadas Orgad, Gizem Ozdil, Gabriel Poesia, & Greta Tuckute! bit.ly/3IpzD5E

We’re thrilled to introduce the 2025 cohort of #KempnerInstitute Graduate Fellows! This year’s recipients include grad students enrolled across five Harvard University Ph.D. programs. Read more: bit.ly/3TYzZ5H Harvard SEAS, Harvard Griffin GSAS, Harvard Physics Program in Neuroscience at Harvard Medical School #AI #NeuroAI #ML

New paper with Jianing Mu & Preston Lab! We propose a simple model for human memory of narratives: we uniformly sample incoming information at a constant rate. This explains behavioral data much better than variable-rate sampling triggered by event segmentation or surprisal.

🚨 New Preprint 🚨 Targeting intracranial electrical stimulation (ES) to network regions defined within individuals causes network-level effects By Cyr et al. Q: Can we use individualized network maps from precision fMRI to modulate a targeted network via ES? A: Yes! 🧵:

New paper with Richard Antonello, Chandan Singh, Suna Guo, Gavin Mischler, Jianfeng Gao, & Nima Mesgarani: We use LLMs to generate VERY interpretable embeddings where each dimension corresponds to a scientific theory, & then use these embeddings to predict fMRI and ECoG. It WORKS!

Human auditory cortex integrates information in speech across absolute time (e.g., 200 ms), not phonemes, syllables, words, or any other time-varying speech structure: nature.com/articles/s4159…