Grant: Discovery Grants Program - Individual
École Polytechnique de Montréal
$38,000.00
Autonomous learning agents for intelligent neurostimulation
EF 2024-25
Résumé
The nervous system communicates via electrical signals. Electrical neurostimulation is obtained by positioning electrodes in contact with the brain, spinal cord or nerves and delivering stimuli that will modulate their activity. This powerful technique allows causal investigation of neural circuits, enabling neuroscientific discovery. It also constitutes the biophysical foundation of a class of medical interventions.
Neurostimulation always requires precise adjustment of several stimulation parameters, such as the spatial location of the stimulus, the timing, the frequency of stimulus delivery. Even in the most cutting-edge applications, stimulation tuning has been almost exclusively handled manually.
The lack of an algorithmic framework for control and optimization of neurostimulation has hindered scientific discovery. We lack protocols to interrogate and modulate neural circuits with multiple parallel inputs, which would be invaluable for studying multimodal sensorimotor integration, cognitive processes, and for developing more effective neuroprosthetic systems.
The LONG TERM OBJECTIVE of my research program is to provide the cornerstone demonstrations for the next generation of intelligent neurostimulation technology, tailored and personalized to each subject by autonomous learning agents. My working hypothesis is that real-time decision algorithms, combining learning and action, provide superior understanding and control of neural circuitries.
SPECIFIC OBJECTIVES:
In Obj1 we will establish continual learning strategies to handle non-stationarities typical of neural electrode arrays, track modifications occurring over weeks and co-adapt with changes naturally occurring in the brain.
In Obj2 we will design novel Bayesian strategies to allow multi-target, parallel neurostimulation optimization, learn to integrate multi-input-multi-output objectives for more powerful and diverse control of motor output.
In Obj3, we will develop autonomous learning agents capable to learn sequential `patterned' stimulation, rather than being limited to optimize single stimuli individually, for composite control of movement patterns.
Each of the Specific Objectives combines two research modalities: 1) theoretical innovations in optimization and 2) in vivo demonstrations of real-time neurostimulation guidance, to optimize previously intractable stimulation patterns.
In this program I will train 4 graduate and 5 undergraduate students and 1 postdoc with interdisciplinary backgrounds in engineering, AI and neuroscience. This pluralist environment is characterized by rare models and oriented toward discovery and collaboration. The trainees will develop profiles that are highly sought in academia and by a flourishing industry.
This work will greatly contribute to neuroscientific discovery by investigating the multimodal integration underlying sensorimotor function, and by offering breakthrough methods to create advanced neuroprosthetic technologies.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RGPIN-2023-04370
Name nom
Bonizzato, Marco
Department département
Génie Électrique
Organization id
64
Institution établissement
École Polytechnique de Montréal
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
38000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1502
Committee name en
Biological Systems and Functions
Committee name fr
Systèmes et fonctions biologiques
Area of application code
1206
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Engineering
Area of application fr
Génie
Research subject code
2509
Research subject group en
Electrical and electronic engineering
Research subject group fr
Génie électrique et électronique
Research subject en
Control systems
Research subject fr
Asservissements
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Autonomous learning agents for intelligent neurostimulation
Keywords
neurostimulation; neural interfaces; bayesian optimization; neuromodulation; learning agents; black-box optimization; control of uncertain systems; cortical control of movement
Application summary
The nervous system communicates via electrical signals. Electrical neurostimulation is obtained by positioning electrodes in contact with the brain, spinal cord or nerves and delivering stimuli that will modulate their activity. This powerful technique allows causal investigation of neural circuits, enabling neuroscientific discovery. It also constitutes the biophysical foundation of a class of medical interventions.
Neurostimulation always requires precise adjustment of several stimulation parameters, such as the spatial location of the stimulus, the timing, the frequency of stimulus delivery. Even in the most cutting-edge applications, stimulation tuning has been almost exclusively handled manually.
The lack of an algorithmic framework for control and optimization of neurostimulation has hindered scientific discovery. We lack protocols to interrogate and modulate neural circuits with multiple parallel inputs, which would be invaluable for studying multimodal sensorimotor integration, cognitive processes, and for developing more effective neuroprosthetic systems.
The LONG TERM OBJECTIVE of my research program is to provide the cornerstone demonstrations for the next generation of intelligent neurostimulation technology, tailored and personalized to each subject by autonomous learning agents. My working hypothesis is that real-time decision algorithms, combining learning and action, provide superior understanding and control of neural circuitries.
SPECIFIC OBJECTIVES:
In Obj1 we will establish continual learning strategies to handle non-stationarities typical of neural electrode arrays, track modifications occurring over weeks and co-adapt with changes naturally occurring in the brain.
In Obj2 we will design novel Bayesian strategies to allow multi-target, parallel neurostimulation optimization, learn to integrate multi-input-multi-output objectives for more powerful and diverse control of motor output.
In Obj3, we will develop autonomous learning agents capable to learn sequential `patterned' stimulation, rather than being limited to optimize single stimuli individually, for composite control of movement patterns.
Each of the Specific Objectives combines two research modalities: 1) theoretical innovations in optimization and 2) in vivo demonstrations of real-time neurostimulation guidance, to optimize previously intractable stimulation patterns.
In this program I will train 4 graduate and 5 undergraduate students and 1 postdoc with interdisciplinary backgrounds in engineering, AI and neuroscience. This pluralist environment is characterized by rare models and oriented toward discovery and collaboration. The trainees will develop profiles that are highly sought in academia and by a flourishing industry.
This work will greatly contribute to neuroscientific discovery by investigating the multimodal integration underlying sensorimotor function, and by offering breakthrough methods to create advanced neuroprosthetic technologies.