1
|
Wernisch L, Edwards T, Berthon A, Tessier-Lariviere O, Sarkans E, Stoukidi M, Fortier-Poisson P, Pinkney M, Thornton M, Hanley C, Lee S, Jennings J, Appleton B, Garsed P, Patterson B, Buttinger W, Gonshaw S, Jakopec M, Shunmugam S, Mamen J, Tukiainen A, Lajoie G, Armitage O, Hewage E. Online Bayesian optimization of vagus nerve stimulation. J Neural Eng 2024; 21:026019. [PMID: 38479016 DOI: 10.1088/1741-2552/ad33ae] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/13/2023] [Accepted: 03/13/2024] [Indexed: 04/04/2024]
Abstract
Objective.In bioelectronic medicine, neuromodulation therapies induce neural signals to the brain or organs, modifying their function. Stimulation devices capable of triggering exogenous neural signals using electrical waveforms require a complex and multi-dimensional parameter space to control such waveforms. Determining the best combination of parameters (waveform optimization or dosing) for treating a particular patient's illness is therefore challenging. Comprehensive parameter searching for an optimal stimulation effect is often infeasible in a clinical setting due to the size of the parameter space. Restricting this space, however, may lead to suboptimal therapeutic results, reduced responder rates, and adverse effects.Approach. As an alternative to a full parameter search, we present a flexible machine learning, data acquisition, and processing framework for optimizing neural stimulation parameters, requiring as few steps as possible using Bayesian optimization. This optimization builds a model of the neural and physiological responses to stimulations, enabling it to optimize stimulation parameters and provide estimates of the accuracy of the response model. The vagus nerve (VN) innervates, among other thoracic and visceral organs, the heart, thus controlling heart rate (HR), making it an ideal candidate for demonstrating the effectiveness of our approach.Main results.The efficacy of our optimization approach was first evaluated on simulated neural responses, then applied to VN stimulation intraoperatively in porcine subjects. Optimization converged quickly on parameters achieving target HRs and optimizing neural B-fiber activations despite high intersubject variability.Significance.An optimized stimulation waveform was achieved in real time with far fewer stimulations than required by alternative optimization strategies, thus minimizing exposure to side effects. Uncertainty estimates helped avoiding stimulations outside a safe range. Our approach shows that a complex set of neural stimulation parameters can be optimized in real-time for a patient to achieve a personalized precision dosing.
Collapse
Affiliation(s)
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - Guillaume Lajoie
- Université de Montréal and Mila-Quebec AI Institute, Montréal, Canada
| | | | | |
Collapse
|
2
|
Choinière L, Guay-Hottin R, Picard R, Lajoie G, Bonizzato M, Dancause N. Gaussian-process-based Bayesian optimization for neurostimulation interventions in rats. STAR Protoc 2024; 5:102885. [PMID: 38358881 PMCID: PMC10876592 DOI: 10.1016/j.xpro.2024.102885] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/14/2023] [Revised: 12/11/2023] [Accepted: 01/29/2024] [Indexed: 02/17/2024] Open
Abstract
Effective neural stimulation requires adequate parametrization. Gaussian-process (GP)-based Bayesian optimization (BO) offers a framework to discover optimal stimulation parameters in real time. Here, we first provide a general protocol to deploy this framework in neurostimulation interventions and follow by exemplifying its use in detail. Specifically, we describe the steps to implant rats with multi-channel electrode arrays in the hindlimb motor cortex. We then detail how to utilize the GP-BO algorithm to maximize evoked target movements, measured as electromyographic responses. For complete details on the use and execution of this protocol, please refer to Bonizzato and colleagues (2023).1.
Collapse
Affiliation(s)
- Léo Choinière
- Department of Neurosciences and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada.
| | - Rose Guay-Hottin
- Department of Neurosciences and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Department of Electrical Engineering and Institute of Biomedical Engineering, Polytechnique Montréal, Montreal, QC H3T 1J4, Canada; Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada
| | - Rémi Picard
- Department of Neurosciences and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Department of Electrical Engineering and Institute of Biomedical Engineering, Polytechnique Montréal, Montreal, QC H3T 1J4, Canada
| | - Guillaume Lajoie
- Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada; Mathematics and Statistics Department, Université de Montréal, Montreal, QC H3T 1J4, Canada
| | - Marco Bonizzato
- Department of Neurosciences and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Department of Electrical Engineering and Institute of Biomedical Engineering, Polytechnique Montréal, Montreal, QC H3T 1J4, Canada; Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada
| | - Numa Dancause
- Department of Neurosciences and Centre Interdisciplinaire de Recherche sur le Cerveau et l'Apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada.
| |
Collapse
|
3
|
Bonizzato M, Guay Hottin R, Côté SL, Massai E, Choinière L, Macar U, Laferrière S, Sirpal P, Quessy S, Lajoie G, Martinez M, Dancause N. Autonomous optimization of neuroprosthetic stimulation parameters that drive the motor cortex and spinal cord outputs in rats and monkeys. Cell Rep Med 2023; 4:101008. [PMID: 37044093 PMCID: PMC10140617 DOI: 10.1016/j.xcrm.2023.101008] [Citation(s) in RCA: 5] [Impact Index Per Article: 5.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/08/2021] [Revised: 02/16/2023] [Accepted: 03/20/2023] [Indexed: 04/14/2023]
Abstract
Neural stimulation can alleviate paralysis and sensory deficits. Novel high-density neural interfaces can enable refined and multipronged neurostimulation interventions. To achieve this, it is essential to develop algorithmic frameworks capable of handling optimization in large parameter spaces. Here, we leveraged an algorithmic class, Gaussian-process (GP)-based Bayesian optimization (BO), to solve this problem. We show that GP-BO efficiently explores the neurostimulation space, outperforming other search strategies after testing only a fraction of the possible combinations. Through a series of real-time multi-dimensional neurostimulation experiments, we demonstrate optimization across diverse biological targets (brain, spinal cord), animal models (rats, non-human primates), in healthy subjects, and in neuroprosthetic intervention after injury, for both immediate and continual learning over multiple sessions. GP-BO can embed and improve "prior" expert/clinical knowledge to dramatically enhance its performance. These results advocate for broader establishment of learning agents as structural elements of neuroprosthetic design, enabling personalization and maximization of therapeutic effectiveness.
Collapse
Affiliation(s)
- Marco Bonizzato
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Department of Electrical Engineering and Institute of Biomedical Engineering, Polytechnique Montréal, Montreal, QC H3T 1J4, Canada; CIUSSS du Nord-de-l'Île-de-Montréal, Montreal, QC H4J 1C5, Canada; Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada.
| | - Rose Guay Hottin
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Department of Electrical Engineering and Institute of Biomedical Engineering, Polytechnique Montréal, Montreal, QC H3T 1J4, Canada; Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada
| | - Sandrine L Côté
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada
| | - Elena Massai
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada
| | - Léo Choinière
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada
| | - Uzay Macar
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada
| | - Samuel Laferrière
- Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada; Computer Science Department, Université de Montréal, Montreal, QC H3T 1J4, Canada
| | - Parikshat Sirpal
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada
| | - Stephan Quessy
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada
| | - Guillaume Lajoie
- Mila - Québec AI Institute, Montreal, QC H2S 3H1, Canada; Mathematics and Statistics Department, Université de Montréal, Montreal, QC H3T 1J4, Canada
| | - Marina Martinez
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada; CIUSSS du Nord-de-l'Île-de-Montréal, Montreal, QC H4J 1C5, Canada
| | - Numa Dancause
- Department of Neurosciences and Centre interdisciplinaire de recherche sur le cerveau et l'apprentissage (CIRCA), Université de Montréal, Montreal, QC H3T 1J4, Canada.
| |
Collapse
|
4
|
Chang YC, Ahmed U, Jayaprakash N, Mughrabi I, Lin Q, Wu YC, Gerber M, Abbas A, Daytz A, Gabalski AH, Ashville J, Dokos S, Rieth L, Datta-Chaudhuri T, Tracey KJ, Guo T, Al-Abed Y, Zanos S. kHz-frequency electrical stimulation selectively activates small, unmyelinated vagus afferents. Brain Stimul 2022; 15:1389-1404. [PMID: 36241025 PMCID: PMC10164362 DOI: 10.1016/j.brs.2022.09.015] [Citation(s) in RCA: 6] [Impact Index Per Article: 3.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/24/2022] [Revised: 09/02/2022] [Accepted: 09/30/2022] [Indexed: 12/30/2022] Open
Abstract
BACKGROUND Vagal reflexes regulate homeostasis in visceral organs and systems through afferent and efferent neurons and nerve fibers. Small, unmyelinated, C-type afferents comprise over 80% of fibers in the vagus and form the sensory arc of autonomic reflexes of the gut, lungs, heart and vessels and the immune system. Selective bioelectronic activation of C-afferents could be used to mechanistically study and treat diseases of peripheral organs in which vagal reflexes are involved, but it has not been achieved. METHODS We stimulated the vagus in rats and mice using trains of kHz-frequency stimuli. Stimulation effects were assessed using neuronal c-Fos expression, physiological and nerve fiber responses, optogenetic and computational methods. RESULTS Intermittent kHz stimulation for 30 min activates specific motor and, preferentially, sensory vagus neurons in the brainstem. At sufficiently high frequencies (>5 kHz) and at intensities within a specific range (7-10 times activation threshold, T, in rats; 15-25 × T in mice), C-afferents are activated, whereas larger, A- and B-fibers, are blocked. This was determined by measuring fiber-specific acute physiological responses to kHz stimulus trains, and by assessing fiber excitability around kHz stimulus trains through compound action potentials evoked by probing pulses. Aspects of selective activation of C-afferents are explained in computational models of nerve fibers by how fiber size and myelin shape the response of sodium channels to kHz-frequency stimuli. CONCLUSION kHz stimulation is a neuromodulation strategy to robustly and selectively activate vagal C-afferents implicated in physiological homeostasis and disease, over larger vagal fibers.
Collapse
Affiliation(s)
- Yao-Chuan Chang
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Umair Ahmed
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Naveen Jayaprakash
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Ibrahim Mughrabi
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Qihang Lin
- Graduate School of Biomedical Engineering, University of New South Wales, Sydney, 2052, Australia
| | - Yi-Chen Wu
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Michael Gerber
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Adam Abbas
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Anna Daytz
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Arielle H Gabalski
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Jason Ashville
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Socrates Dokos
- Graduate School of Biomedical Engineering, University of New South Wales, Sydney, 2052, Australia
| | - Loren Rieth
- Department of Mechanical and Aerospace Engineering, West Virginia University, Morgantown, WV, 26506, United States
| | - Timir Datta-Chaudhuri
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Kevin J Tracey
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Tianruo Guo
- Graduate School of Biomedical Engineering, University of New South Wales, Sydney, 2052, Australia
| | - Yousef Al-Abed
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States
| | - Stavros Zanos
- Institute of Bioelectronic Medicine, Feinstein Institutes for Medical Research, Manhasset, NY, 11030, United States; Zucker School of Medicine at Hofstra/Northwell, Hempstead, NY, United States.
| |
Collapse
|
5
|
Govindarajan LN, Calvert JS, Parker SR, Jung M, Darie R, Miranda P, Shaaya E, Borton DA, Serre T. Fast inference of spinal neuromodulation for motor control using amortized neural networks. J Neural Eng 2022; 19:10.1088/1741-2552/ac9646. [PMID: 36174534 PMCID: PMC9668352 DOI: 10.1088/1741-2552/ac9646] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/31/2022] [Accepted: 09/29/2022] [Indexed: 11/12/2022]
Abstract
Objective.Epidural electrical stimulation (EES) has emerged as an approach to restore motor function following spinal cord injury (SCI). However, identifying optimal EES parameters presents a significant challenge due to the complex and stochastic nature of muscle control and the combinatorial explosion of possible parameter configurations. Here, we describe a machine-learning approach that leverages modern deep neural networks to learn bidirectional mappings between the space of permissible EES parameters and target motor outputs.Approach.We collected data from four sheep implanted with two 24-contact EES electrode arrays on the lumbosacral spinal cord. Muscle activity was recorded from four bilateral hindlimb electromyography (EMG) sensors. We introduce a general learning framework to identify EES parameters capable of generating desired patterns of EMG activity. Specifically, we first amortize spinal sensorimotor computations in a forward neural network model that learns to predict motor outputs based on EES parameters. Then, we employ a second neural network as an inverse model, which reuses the amortized knowledge learned by the forward model to guide the selection of EES parameters.Main results.We found that neural networks can functionally approximate spinal sensorimotor computations by accurately predicting EMG outputs based on EES parameters. The generalization capability of the forward model critically benefited our inverse model. We successfully identified novel EES parameters, in under 20 min, capable of producing desired target EMG recruitment duringin vivotesting. Furthermore, we discovered potential functional redundancies within the spinal sensorimotor networks by identifying unique EES parameters that result in similar motor outcomes. Together, these results suggest that our framework is well-suited to probe spinal circuitry and control muscle recruitment in a completely data-driven manner.Significance.We successfully identify novel EES parameters within minutes, capable of producing desired EMG recruitment. Our approach is data-driven, subject-agnostic, automated, and orders of magnitude faster than manual approaches.
Collapse
Affiliation(s)
- Lakshmi Narasimhan Govindarajan
- Cognitive, Linguistic & Psychological Sciences, Brown University, Providence RI USA
- Carney Institute for Brain Science, Brown University, Providence RI USA
| | | | | | - Minju Jung
- Cognitive, Linguistic & Psychological Sciences, Brown University, Providence RI USA
- Carney Institute for Brain Science, Brown University, Providence RI USA
| | - Radu Darie
- School of Engineering, Brown University, Providence RI USA
| | | | - Elias Shaaya
- Department of Neurosurgery, Brown University and Rhode Island Hospital, Providence RI USA
| | - David A. Borton
- Carney Institute for Brain Science, Brown University, Providence RI USA
- School of Engineering, Brown University, Providence RI USA
- Center for Neurorestoration and Neurotechnology, Department of Veterans Affairs, Providence RI USA
| | - Thomas Serre
- Cognitive, Linguistic & Psychological Sciences, Brown University, Providence RI USA
- Carney Institute for Brain Science, Brown University, Providence RI USA
| |
Collapse
|