The university library will be closed from December 20, 2025 to January 11, 2026. From January 12, 2026, regular opening hours will apply again. Exception: The main medical library and the central library will be open again from January 5, 2026. Further information

Result: MotorNet, a Python toolbox for controlling differentiable biomechanical effectors with artificial neural networks.

Title:
MotorNet, a Python toolbox for controlling differentiable biomechanical effectors with artificial neural networks.
Authors:
Codol O; Western Institute for Neuroscience, University of Western Ontario, Ontario, Canada.; Department of Psychology, University of Western Ontario, Ontario, Canada., Michaels JA; Western Institute for Neuroscience, University of Western Ontario, Ontario, Canada.; Department of Physiology & Pharmacology, Schulich School of Medicine & Dentistry, University of Western Ontario, Ontario, Canada.; Robarts Research Institute, University of Western Ontario, Ontario, Canada., Kashefi M; Western Institute for Neuroscience, University of Western Ontario, Ontario, Canada.; Department of Physiology & Pharmacology, Schulich School of Medicine & Dentistry, University of Western Ontario, Ontario, Canada.; Robarts Research Institute, University of Western Ontario, Ontario, Canada., Pruszynski JA; Western Institute for Neuroscience, University of Western Ontario, Ontario, Canada.; Department of Psychology, University of Western Ontario, Ontario, Canada.; Department of Physiology & Pharmacology, Schulich School of Medicine & Dentistry, University of Western Ontario, Ontario, Canada.; Robarts Research Institute, University of Western Ontario, Ontario, Canada., Gribble PL; Western Institute for Neuroscience, University of Western Ontario, Ontario, Canada.; Department of Psychology, University of Western Ontario, Ontario, Canada.; Department of Physiology & Pharmacology, Schulich School of Medicine & Dentistry, University of Western Ontario, Ontario, Canada.
Source:
ELife [Elife] 2024 Jul 30; Vol. 12. Date of Electronic Publication: 2024 Jul 30.
Publication Type:
Journal Article
Language:
English
Journal Info:
Publisher: eLife Sciences Publications, Ltd Country of Publication: England NLM ID: 101579614 Publication Model: Electronic Cited Medium: Internet ISSN: 2050-084X (Electronic) Linking ISSN: 2050084X NLM ISO Abbreviation: Elife Subsets: MEDLINE
Imprint Name(s):
Original Publication: Cambridge, UK : eLife Sciences Publications, Ltd., 2012-
MeSH Terms:
Neural Networks, Computer* , Software*, Biomechanical Phenomena ; Humans ; Algorithms
References:
IEEE Trans Biomed Eng. 2007 Nov;54(11):1940-50. (PMID: 18018689)
PLoS Comput Biol. 2016 Nov 4;12(11):e1005175. (PMID: 27814352)
J Neurophysiol. 1997 Jul;78(1):554-60. (PMID: 9242306)
J Neurosci. 1994 May;14(5 Pt 2):3208-24. (PMID: 8182467)
J Mot Behav. 1993 Sep;25(3):140-152. (PMID: 12581985)
Nat Neurosci. 2024 Jul;27(7):1349-1363. (PMID: 38982201)
Exp Brain Res. 2000 Dec;135(4):474-82. (PMID: 11156311)
J Biomech. 1995 May;28(5):513-25. (PMID: 7775488)
Elife. 2020 Dec 28;9:. (PMID: 33370235)
Curr Opin Neurobiol. 2020 Dec;65:59-69. (PMID: 33142111)
Math Biosci. 2020 Nov;329:108455. (PMID: 32835693)
J Biomech Eng. 2003 Feb;125(1):70-7. (PMID: 12661198)
Proc Natl Acad Sci U S A. 2020 Dec 15;117(50):32124-32135. (PMID: 33257539)
Nature. 2023 Nov;623(7988):765-771. (PMID: 37938772)
Nature. 2015 May 28;521(7553):436-44. (PMID: 26017442)
J Neurophysiol. 2010 Dec;104(6):2985-94. (PMID: 20884757)
Curr Opin Neurobiol. 2020 Dec;65:iii-v. (PMID: 33357763)
Nat Rev Neurosci. 2021 Jan;22(1):55-67. (PMID: 33199854)
Nat Neurosci. 2019 Apr;22(4):529-533. (PMID: 30742115)
Nat Neurosci. 2019 Nov;22(11):1761-1770. (PMID: 31659335)
Neural Comput. 2017 May;29(5):1229-1262. (PMID: 28333583)
Neuron. 2013 Jan 9;77(1):168-79. (PMID: 23312524)
Exp Brain Res. 2008 Mar;185(3):359-81. (PMID: 18251019)
Proc ASME Des Eng Tech Conf. 2013 Aug;2013:. (PMID: 25905111)
Exp Brain Res. 1995;107(1):125-36. (PMID: 8751070)
Nature. 2001 Sep 13;413(6852):161-5. (PMID: 11557980)
J Hum Kinet. 2021 Jan 29;76:9-33. (PMID: 33603922)
Eur J Neurosci. 2010 Sep;32(6):1049-57. (PMID: 20726884)
Biol Cybern. 1997 Mar;76(3):163-71. (PMID: 9151414)
Nonlinear Dyn. 2010 Oct 1;62(1):291-303. (PMID: 21170173)
Atten Percept Psychophys. 2019 Oct;81(7):2265-2287. (PMID: 31161495)
Curr Biol. 2020 Jun 8;30(11):R629-R632. (PMID: 32516607)
J Neurosci. 2014 Mar 26;34(13):4608-17. (PMID: 24672006)
J Neurophysiol. 2006 May;95(5):2898-912. (PMID: 16436480)
J Morphol. 2000 Sep;245(3):206-24. (PMID: 10972970)
Curr Opin Neurobiol. 2020 Dec;65:194-202. (PMID: 33334641)
Biol Cybern. 1999 Jul;81(1):39-60. (PMID: 10434390)
J Biomech Eng. 2013 Feb;135(2):021005. (PMID: 23445050)
J Neurosci. 2013 Jun 26;33(26):10898-909. (PMID: 23804109)
Curr Biol. 2020 Sep 21;30(18):R1025-R1030. (PMID: 32961152)
Elife. 2023 Feb 10;12:. (PMID: 36718990)
Nat Neurosci. 2004 Sep;7(9):907-15. (PMID: 15332089)
J Neurophysiol. 1997 Feb;77(2):826-52. (PMID: 9065853)
Nature. 2015 Feb 26;518(7540):529-33. (PMID: 25719670)
Curr Opin Neurobiol. 2021 Feb;66:250-257. (PMID: 33358629)
J Neurophysiol. 2007 Jun;97(6):4235-57. (PMID: 17376854)
PLoS Comput Biol. 2018 Jul 26;14(7):e1006223. (PMID: 30048444)
Nat Rev Neurosci. 2020 Jun;21(6):335-346. (PMID: 32303713)
Exp Brain Res. 1986;63(2):331-40. (PMID: 3758250)
Grant Information:
RGPIN/05458-2018 Natural Sciences and Engineering Research Council of Canada; PJT-175010 Canada CAPMC CIHR; RGPIN-2022-04421 Natural Sciences and Engineering Research Council of Canada; PJT-156241 Canada CAPMC CIHR
Contributed Indexing:
Keywords: biomechanical model; computational model; motor control; motor learning; muscle model; neural network; neuroscience; none
Entry Date(s):
Date Created: 20240730 Date Completed: 20240730 Latest Revision: 20240801
Update Code:
20250114
PubMed Central ID:
PMC11288629
DOI:
10.7554/eLife.88591
PMID:
39078880
Database:
MEDLINE

Further information

Artificial neural networks (ANNs) are a powerful class of computational models for unravelling neural mechanisms of brain function. However, for neural control of movement, they currently must be integrated with software simulating biomechanical effectors, leading to limiting impracticalities: (1) researchers must rely on two different platforms and (2) biomechanical effectors are not generally differentiable, constraining researchers to reinforcement learning algorithms despite the existence and potential biological relevance of faster training methods. To address these limitations, we developed MotorNet, an open-source Python toolbox for creating arbitrarily complex, differentiable, and biomechanically realistic effectors that can be trained on user-defined motor tasks using ANNs. MotorNet is designed to meet several goals: ease of installation, ease of use, a high-level user-friendly application programming interface, and a modular architecture to allow for flexibility in model building. MotorNet requires no dependencies outside Python, making it easy to get started with. For instance, it allows training ANNs on typically used motor control models such as a two joint, six muscle, planar arm within minutes on a typical desktop computer. MotorNet is built on PyTorch and therefore can implement any network architecture that is possible using the PyTorch framework. Consequently, it will immediately benefit from advances in artificial intelligence through PyTorch updates. Finally, it is open source, enabling users to create and share their own improvements, such as new effector and network architectures or custom task designs. MotorNet's focus on higher-order model and task design will alleviate overhead cost to initiate computational projects for new researchers by providing a standalone, ready-to-go framework, and speed up efforts of established computational teams by enabling a focus on concepts and ideas over implementation.
(© 2023, Codol et al.)

OC, JM, MK, PG No competing interests declared, JP Reviewing editor, eLife