MIMIC-MJX: Neuromechanical Emulation of Animal Behavior
Abstract
The primary output of the nervous system is movement and behavior. While recent advances have democratized pose tracking during complex behavior, kinematic trajectories alone provide only indirect access to the underlying control processes. Here we present MIMIC-MJX, a framework for learning biologically-plausible neural control policies from kinematics. MIMIC-MJX models the generative process of motor control by training neural controllers that learn to actuate biomechanically-realistic body models in physics simulation to reproduce real kinematic trajectories. We demonstrate that our implementation is accurate, fast, data-efficient, and generalizable to diverse animal body models. Policies trained with MIMIC-MJX can be utilized to both analyze neural control strategies and simulate behavioral experiments, illustrating its potential as an integrative modeling framework for neuroscience.
Keywords
Cite
@article{arxiv.2511.20532,
title = {MIMIC-MJX: Neuromechanical Emulation of Animal Behavior},
author = {Charles Y. Zhang and Yuanjia Yang and Aidan Sirbu and Elliott T. T. Abe and Emil Wärnberg and Eric J. Leonardis and Diego E. Aldarondo and Adam Lee and Aaditya Prasad and Jason Foat and Kaiwen Bian and Joshua Park and Rusham Bhatt and Hutton Saunders and Akira Nagamori and Ayesha R. Thanawalla and Kee Wui Huang and Fabian Plum and Hendrik K. Beck and Steven W. Flavell and David Labonte and Blake A. Richards and Bingni W. Brunton and Eiman Azim and Bence P. Ölveczky and Talmo D. Pereira},
journal= {arXiv preprint arXiv:2511.20532},
year = {2025}
}
Comments
Corrected LaTeX issues. Project page available at https://mimic-mjx.talmolab.org