shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026@steven-j-west.bsky.social Olivier Winter @tyrellturing.bsky.social @glajoie.bsky.social @colehurwitz.bsky.social @mehdiazabou.bsky.social Matthew R. Whiteway Liam Paninski @evadyer.bsky.social 010
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026Laura Freitas-Silva Fei Hu Julia M. Huntenburg @anupkhanal.bsky.social Ines Laranjeira Petrina Lau Guido T. Meijer @natemiska.bsky.social @jpnoel.bsky.social Alejandro Pan-Vazquez Georg Raiser @rossant.bsky.social @karolinazsocha.bsky.social @anne-urai.bsky.social Miles J. Wells 110
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026@vinam.bsky.social Keshav Balaji @div1115.bsky.social @nandahkrishna.bsky.social Jingyun Xiao @yizi-zhang.bsky.social @ximengmao.bsky.social Wenrui Ma @hanyu42.bsky.social Daniel Birman @nbonacchi.bsky.social Gaelle A. Chapuis @joanafacatarino.bsky.social Felicia Davatolhagh Mayo Faulkner 110
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026A huge thank you to all of our team and collaborators for their hard work in making this benchmark possible! And a special thanks to the @intlbrainlab.bsky.social for the extraordinary shared dataset and resources behind this effort.shivashriganesh.bsky.socialshivashriganesh.bsky.social 110
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026🔎 Check out the project page: brainbench-org.github.io/ibl-bwb-proj... 🏆 Submit a model and view the BrainWideBench Leaderboard: bwb.iblcore.org 💻 Access the codebase: github.com/brainbench-o... And shoutout to torchbrain.org as the tool we used to process recordings into a standardized format.brainbench-org.github.ioIBL BrainWideBenchBenchmarking neural decoding on brain-wide recordings from the International Brain Laboratory. 110
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026🤝 Shared datasets, shared software, and shared benchmarks are essential if we want to measure progress consistently and move toward more general-purpose models of neural activity. We’d love for you to explore the benchmark and put your models to the test! 110
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026🌎 We invite the community to submit and evaluate their own models on BrainWideBench! Our goal is to build an ecosystem where researchers can focus efforts on developing new modeling ideas, and spend less time rebuilding evaluation infrastructures. 120
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026We evaluated 6 large-scale pretrained baseline models across multiple settings, and rigorously tuned 18 single-session baselines for individual tasks spanning the three task suites. 120
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026When developing new methods, researchers often have to rerun existing baselines, reproduce preprocessing pipelines, and evaluate multiple models across multiple tasks. In many cases, that evaluation effort can take as much time as, or even more than, developing the new model itself! 😵💫 120
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026and three complementary task suites: - Behavior: decoding movements and task variables - Neural dynamics: predicting masked or future neural activity - Anatomy: recovering brain-region identity from neural activity 110
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026A major bottleneck in modeling neural data today is rigorous and standardized evaluation. Modelling techniques need to be tested under multiple regimes and downstream tasks. Our benchmark spans: 🐭 139 mice | 276 brain regions | 452 recording sessions | 830,000+ unit-hours before filtering 120
shivashriganesh.bsky.social @shivashriganesh.bsky.social · 21/09/2026We’re excited to share BrainWideBench, a new benchmark for evaluating how models of neural activity generalize across animals and tasks! 📄 Check out the paper here! arxiv.org/abs/2609.22064 Special shoutout to @alx-adr.bsky.social, who contributed equally as co-first authors! 13516
