Massively Parallel Imitation Learning of Mouse Forelimb Musculoskeletal Reaching Dynamics
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908677621940224 |
|---|---|
| author | Leonardis, Eric Nagamori, Akira Thanawalla, Ayesha Yang, Yuanjia Park, Joshua Saunders, Hutton Azim, Eiman Pereira, Talmo |
| author_facet | Leonardis, Eric Nagamori, Akira Thanawalla, Ayesha Yang, Yuanjia Park, Joshua Saunders, Hutton Azim, Eiman Pereira, Talmo |
| contents | The brain has evolved to effectively control the body, and in order to understand the relationship we need to model the sensorimotor transformations underlying embodied control. As part of a coordinated effort, we are developing a general-purpose platform for behavior-driven simulation modeling high fidelity behavioral dynamics, biomechanics, and neural circuit architectures underlying embodied control. We present a pipeline for taking kinematics data from the neuroscience lab and creating a pipeline for recapitulating those natural movements in a biomechanical model. We implement a imitation learning framework to perform a dexterous forelimb reaching task with a musculoskeletal model in a simulated physics environment. The mouse arm model is currently training at faster than 1 million training steps per second due to GPU acceleration with JAX and Mujoco-MJX. We present results that indicate that adding naturalistic constraints on energy and velocity lead to simulated musculoskeletal activity that better predict real EMG signals. This work provides evidence to suggest that energy and control constraints are critical to modeling musculoskeletal motor control. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_21848 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Massively Parallel Imitation Learning of Mouse Forelimb Musculoskeletal Reaching Dynamics Leonardis, Eric Nagamori, Akira Thanawalla, Ayesha Yang, Yuanjia Park, Joshua Saunders, Hutton Azim, Eiman Pereira, Talmo Machine Learning Robotics Neurons and Cognition Quantitative Methods 68T05, 37N25, 37M1, 37N35, 70Q05, 37N40, I.2.6; I.5.1; G.1.10 The brain has evolved to effectively control the body, and in order to understand the relationship we need to model the sensorimotor transformations underlying embodied control. As part of a coordinated effort, we are developing a general-purpose platform for behavior-driven simulation modeling high fidelity behavioral dynamics, biomechanics, and neural circuit architectures underlying embodied control. We present a pipeline for taking kinematics data from the neuroscience lab and creating a pipeline for recapitulating those natural movements in a biomechanical model. We implement a imitation learning framework to perform a dexterous forelimb reaching task with a musculoskeletal model in a simulated physics environment. The mouse arm model is currently training at faster than 1 million training steps per second due to GPU acceleration with JAX and Mujoco-MJX. We present results that indicate that adding naturalistic constraints on energy and velocity lead to simulated musculoskeletal activity that better predict real EMG signals. This work provides evidence to suggest that energy and control constraints are critical to modeling musculoskeletal motor control. |
| title | Massively Parallel Imitation Learning of Mouse Forelimb Musculoskeletal Reaching Dynamics |
| topic | Machine Learning Robotics Neurons and Cognition Quantitative Methods 68T05, 37N25, 37M1, 37N35, 70Q05, 37N40, I.2.6; I.5.1; G.1.10 |
| url | https://arxiv.org/abs/2511.21848 |