Research map / Vision-language and multimodal / Human and embodied modeling
OmniMouse: Scaling properties of multi-modal, multi-task Brain Models on 150B Neural Tokens
Konstantin Friedrich Willeke, Polina Turishcheva, Alex Gilbert, Goirik Chakrabarty, Hasan Atakan Bedel, Paul G. Fahey, Yongrong Qiu, Marissa A. Weis, Michaela Vystrčilová, Taliah Muhammad, Lydia Ntanavara, Rachel E Froebe and 9 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| scale model from 80M to 300M parameters (ablation) | OmniMouse · 323 sessions forecasting | single-trial correlation: 0.25 -> 0.25 | No clear effect |
| scale model from 80M to 300M parameters (ablation) | OmniMouse · 323 sessions behavior decoding average | single-trial correlation: 0.77 -> 0.76 | No clear effect |
| scale model from 1M to 80M parameters (ablation) | OmniMouse · 323 sessions forecasting | single-trial correlation: 0.18 -> 0.25 | Improved |
| scale model from 1M to 80M parameters (ablation) | OmniMouse · 323 sessions behavior decoding average | single-trial correlation: 0.68 -> 0.77 | Improved |
| scale training data from 8 to 323 sessions (5M model) (ablation) | OmniMouse-5M · forecasting | single-trial correlation: 0.18 -> 0.22 | Improved |
| scale training data from 8 to 323 sessions (5M model) (ablation) | OmniMouse-5M · forecasting+stimulus | single-trial correlation: 0.25 -> 0.32 | Improved |
| scale training data from 8 to 323 sessions (5M model) (ablation) | OmniMouse-5M · population+stimulus | single-trial correlation: 0.27 -> 0.35 | Improved |
| scale training data from 8 to 323 sessions (5M model) (ablation) | OmniMouse-5M · behavior decoding average | single-trial correlation: 0.59 -> 0.69 | Improved |
| add visual stimulus input to forecasting (80M) (ablation) | OmniMouse-80M · 323 sessions forecasting | single-trial correlation: 0.25 -> 0.34 | Improved |
| train on 323 sessions instead of 8 sessions, versus baselines (ablation) | OmniMouse · running speed decoding | single-trial correlation: 0.44 -> 0.55 | Mixed |
| use OmniMouse over MtM baseline (data-matched) | OmniMouse-5M · 8 sessions forecasting | single-trial correlation: 0.12 -> 0.18 | Improved |
| use OmniMouse over Latent Model (data-matched) | OmniMouse-5M · 8 sessions forecasting+stimulus | single-trial correlation: 0.18 -> 0.25 | Improved |
Closest papers by meaning
- Neural Encoding and Decoding at Scale (ICML 2025 · similarity 0.90)
- Retrospective for the Dynamic Sensorium Competition for predicting large-scale mouse primary visual cortex activity from videos (NeurIPS 2024 · similarity 0.90)
- Scaling Laws for Task-Optimized Models of the Primate Visual Ventral Stream (ICML 2025 · similarity 0.90)
- Neuroformer: Multimodal and Multitask Generative Pretraining for Brain Data (ICLR 2024 · similarity 0.89)
- Towards a "Universal Translator" for Neural Dynamics at Single-Cell, Single-Spike Resolution (NeurIPS 2024 · similarity 0.88)
- Scaling Laws and Tradeoffs in Recurrent Networks of Expressive Neurons (NeurIPS 2026 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
