puc-header

Optimizing Neuroprosthetic Therapies via Autonomous Learning Agents

83 Pages Posted: 16 Sep 2021 Publication Status: Review Complete

See all articles by Marco Bonizzato

Marco Bonizzato

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA)

Sandrine L. Côté

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA)

Elena Massai

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA)

Rose Guay-Hottin

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA)

Samuel Laferrière

University of Montreal - Computer Science Department

Stephan Quessy

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA)

Guillaume Lajoie

Mila - Quebec AI Institute

Marina Martinez

Université de Montréal - Department of Neuroscience and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA); CIUSS-NÎM-Hôpital du Sacré-Coeur de Montréal

Numa Dancause

Université de Montréal - Department of Neurosciences and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA)

More...

Abstract

Neural stimulation can alleviate or even reverse paralysis and sensory deficits. Rapid technological advancements bring the possibility to develop complex and refined patterns of neurostimulation. However, multipronged interventions with high-density neural interfaces will require algorithmic frameworks to handle optimization in large parameter spaces. Here, we used an algorithmic class, Gaussian-Process (GP)-based Bayesian Optimization (BO), to solve this online problem. We show that GP-BO can efficiently explore the neurostimulation parameters’ space, exceeding extensive search performance after testing only a fraction of the possible combinations. It can quickly optimize multi-channel neurostimulation across diverse biological targets (brain and spinal cord), animal models (rats and non-human primates), in healthy and injured subjects. Moreover, since BO can embed and improve ‘prior’ expert/clinical knowledge, the performance can be dramatically enhanced even further. These results support broad establishment of learning agents as a structural part of neuroprosthetic design, enabling therapeutic personalization and maximization of intervention effectiveness.

Keywords: Machine-learning, Brain-machine interface, motor cortex, spinal cord injury, Bayesian optimization, neurostimulation

Suggested Citation

Bonizzato, Marco and Côté, Sandrine L. and Massai, Elena and Guay-Hottin, Rose and Laferrière, Samuel and Quessy, Stephan and Lajoie, Guillaume and Martinez, Marina and Martinez, Marina and Dancause, Numa, Optimizing Neuroprosthetic Therapies via Autonomous Learning Agents. Available at SSRN: https://ssrn.com/abstract=3925256 or http://dx.doi.org/10.2139/ssrn.3925256
This version of the paper has not been formally peer reviewed.

Marco Bonizzato

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA) ( email )

United States

Sandrine L. Côté

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA) ( email )

United States

Elena Massai

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA) ( email )

United States

Rose Guay-Hottin

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA) ( email )

United States

Samuel Laferrière

University of Montreal - Computer Science Department ( email )

United States

Stephan Quessy

University of Montreal - Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA) ( email )

United States

Guillaume Lajoie

Mila - Quebec AI Institute ( email )

Quebec
Canada

Marina Martinez

Université de Montréal - Department of Neuroscience and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA) ( email )

Quebec
Canada
5143437046 (Phone)
H3T 1J4 (Fax)

CIUSS-NÎM-Hôpital du Sacré-Coeur de Montréal ( email )

Canada

Numa Dancause (Contact Author)

Université de Montréal - Department of Neurosciences and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA) ( email )

Quebec
Canada

Click here to go to Cell.com

Paper statistics

Abstract Views
792
Downloads
67
PlumX Metrics