Automated calibration of somatosensory stimulation using reinforcement learning.
AI
Automatic calibration
Electrical stimulation
Neuropathy
Neurostimulation
Reinforcement learning
Sensory feedback
TENS
Journal
Journal of neuroengineering and rehabilitation
ISSN: 1743-0003
Titre abrégé: J Neuroeng Rehabil
Pays: England
ID NLM: 101232233
Informations de publication
Date de publication:
26 09 2023
26 09 2023
Historique:
received:
17
02
2023
accepted:
13
09
2023
medline:
28
9
2023
pubmed:
27
9
2023
entrez:
26
9
2023
Statut:
epublish
Résumé
The identification of the electrical stimulation parameters for neuromodulation is a subject-specific and time-consuming procedure that presently mostly relies on the expertise of the user (e.g., clinician, experimenter, bioengineer). Since the parameters of stimulation change over time (due to displacement of electrodes, skin status, etc.), patients undergo recurrent, long calibration sessions, along with visits to the clinics, which are inefficient and expensive. To address this issue, we developed an automatized calibration system based on reinforcement learning (RL) allowing for accurate and efficient identification of the peripheral nerve stimulation parameters for somatosensory neuroprostheses. We developed an RL algorithm to automatically select neurostimulation parameters for restoring sensory feedback with transcutaneous electrical nerve stimulation (TENS). First, the algorithm was trained offline on a dataset comprising 49 subjects. Then, the neurostimulation was then integrated with a graphical user interface (GUI) to create an intuitive AI-based mapping platform enabling the user to autonomously perform the sensation characterization procedure. We assessed the algorithm against the performance of both experienced and naïve and of a brute force algorithm (BFA), on 15 nerves from five subjects. Then, we validated the AI-based platform on six neuropathic nerves affected by distal sensory loss. Our automatized approach demonstrated the ability to find the optimal values of neurostimulation achieving reliable and comfortable elicited sensations. When compared to alternatives, RL outperformed the naïve and BFA, significantly decreasing the time for mapping and the number of delivered stimulation trains, while improving the overall quality. Furthermore, the RL algorithm showed performance comparable to trained experimenters. Finally, we exploited it successfully for eliciting sensory feedback in neuropathic patients. Our findings demonstrated that the AI-based platform based on a RL algorithm can automatically and efficiently calibrate parameters for somatosensory nerve stimulation. This holds promise to avoid experts' employment in similar scenarios, thanks to the merging between AI and neurotech. Our RL algorithm has the potential to be used in other neuromodulation fields requiring a mapping process of the stimulation parameters. ClinicalTrial.gov (Identifier: NCT04217005).
Sections du résumé
BACKGROUND
The identification of the electrical stimulation parameters for neuromodulation is a subject-specific and time-consuming procedure that presently mostly relies on the expertise of the user (e.g., clinician, experimenter, bioengineer). Since the parameters of stimulation change over time (due to displacement of electrodes, skin status, etc.), patients undergo recurrent, long calibration sessions, along with visits to the clinics, which are inefficient and expensive. To address this issue, we developed an automatized calibration system based on reinforcement learning (RL) allowing for accurate and efficient identification of the peripheral nerve stimulation parameters for somatosensory neuroprostheses.
METHODS
We developed an RL algorithm to automatically select neurostimulation parameters for restoring sensory feedback with transcutaneous electrical nerve stimulation (TENS). First, the algorithm was trained offline on a dataset comprising 49 subjects. Then, the neurostimulation was then integrated with a graphical user interface (GUI) to create an intuitive AI-based mapping platform enabling the user to autonomously perform the sensation characterization procedure. We assessed the algorithm against the performance of both experienced and naïve and of a brute force algorithm (BFA), on 15 nerves from five subjects. Then, we validated the AI-based platform on six neuropathic nerves affected by distal sensory loss.
RESULTS
Our automatized approach demonstrated the ability to find the optimal values of neurostimulation achieving reliable and comfortable elicited sensations. When compared to alternatives, RL outperformed the naïve and BFA, significantly decreasing the time for mapping and the number of delivered stimulation trains, while improving the overall quality. Furthermore, the RL algorithm showed performance comparable to trained experimenters. Finally, we exploited it successfully for eliciting sensory feedback in neuropathic patients.
CONCLUSIONS
Our findings demonstrated that the AI-based platform based on a RL algorithm can automatically and efficiently calibrate parameters for somatosensory nerve stimulation. This holds promise to avoid experts' employment in similar scenarios, thanks to the merging between AI and neurotech. Our RL algorithm has the potential to be used in other neuromodulation fields requiring a mapping process of the stimulation parameters.
TRIAL REGISTRATION
ClinicalTrial.gov (Identifier: NCT04217005).
Identifiants
pubmed: 37752607
doi: 10.1186/s12984-023-01246-0
pii: 10.1186/s12984-023-01246-0
pmc: PMC10523674
doi:
Banques de données
ClinicalTrials.gov
['NCT04217005']
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
131Informations de copyright
© 2023. BioMed Central Ltd., part of Springer Nature.
Références
iScience. 2022 Mar 21;25(4):104129
pubmed: 35391829
Brain Stimul. 2019 Nov - Dec;12(6):1484-1489
pubmed: 31289013
J Neural Eng. 2007 Jun;4(2):L14-21
pubmed: 17409470
Sci Rep. 2017 Sep 7;7(1):10930
pubmed: 28883640
Med Eng Phys. 2003 Jan;25(1):75-8
pubmed: 12485788
Nat Rev Neurol. 2014 May;10(5):261-70
pubmed: 24709892
J Neural Eng. 2015 Dec;12(6):066002
pubmed: 26401550
Sci Transl Med. 2019 Oct 2;11(512):
pubmed: 31578244
Science. 2020 Oct 16;370(6514):290-291
pubmed: 33060348
Brain Stimul. 2016 May-Jun;9(3):438-452
pubmed: 26968805
J Neural Eng. 2020 Jun 12;17(3):036020
pubmed: 32348977
J Neurol Neurosurg Psychiatry. 2019 Jul;90(7):833-836
pubmed: 30100550
Cochrane Database Syst Rev. 2017 Sep 14;9:CD011976
pubmed: 28905362
Neurol Res. 2002 Jul;24(5):443-52
pubmed: 12117312
J Neural Eng. 2022 Jan 24;19(1):
pubmed: 34915454
Sci Robot. 2019 Jul 24;4(32):
pubmed: 33137773
Nat Med. 2019 Sep;25(9):1356-1363
pubmed: 31501600
Annu Rev Biomed Eng. 2008;10:275-309
pubmed: 18429704
Neuron. 2018 Oct 10;100(1):37-45.e7
pubmed: 30244887
J Neuroeng Rehabil. 2011 Dec 13;8:66
pubmed: 22165907
Science. 2021 Aug 6;373(6555):634-635
pubmed: 34353946
Diabetes Metab Syndr Obes. 2020 Nov 18;13:4447-4455
pubmed: 33239897
Nat Mater. 2021 Jul;20(7):925-939
pubmed: 33859381
Nat Biomed Eng. 2023 Apr;7(4):473-485
pubmed: 34059810
IEEE Trans Neural Syst Rehabil Eng. 2020 Jun;28(6):1452-1460
pubmed: 32286996
Nat Biomed Eng. 2023 Apr;7(4):443-455
pubmed: 33230305
Front Med Technol. 2021 Mar 09;3:619280
pubmed: 35047903
J Neural Eng. 2018 Aug;15(4):046002
pubmed: 29551756
Elife. 2018 Apr 10;7:
pubmed: 29633714
IEEE Trans Neural Syst Rehabil Eng. 2019 May;27(5):876-886
pubmed: 30951470
Curr Biol. 2021 Mar 8;31(5):1065-1071.e4
pubmed: 33417885
Neurology. 2017 Oct 31;89(18):1894-1903
pubmed: 28986411
Ann Neurol. 2019 Jan;85(1):137-154
pubmed: 30474259
Annu Rev Neurosci. 2006;29:229-57
pubmed: 16776585
Sci Transl Med. 2014 Feb 5;6(222):222ra19
pubmed: 24500407
Arch Phys Med Rehabil. 2006 Jun;87(6):853-9
pubmed: 16731222
J Neural Eng. 2018 Oct;15(5):056002
pubmed: 29855427
IEEE Trans Neural Syst Rehabil Eng. 2018 Feb;26(2):428-440
pubmed: 28500007
IEEE Trans Biomed Eng. 2004 Jan;51(1):176-81
pubmed: 14723507
Brain. 1996 Apr;119 ( Pt 2):439-47
pubmed: 8800939
Eur J Anaesthesiol. 2013 Jul;30(7):435-40
pubmed: 23624746
Elife. 2020 Jul 21;9:
pubmed: 32691733
Brain Stimul. 2016 May-Jun;9(3):425-437
pubmed: 26968806
N Engl J Med. 2020 Apr 30;382(18):1732-1738
pubmed: 32348644
Int J Neural Syst. 2017 Nov;27(7):1750012
pubmed: 28030999
IEEE Trans Neural Syst Rehabil Eng. 2022;30:1310-1320
pubmed: 35533165
Nature. 2015 Feb 26;518(7540):529-33
pubmed: 25719670
Europace. 2009 Feb;11(2):142-3
pubmed: 19109362
Front Neurosci. 2016 Sep 16;10:425
pubmed: 27695397
Sci Adv. 2021 Apr 21;7(17):
pubmed: 33883127
Exp Neurol. 2020 Jun;328:113274
pubmed: 32145251
J Diabetes Sci Technol. 2017 Jul;11(4):693-701
pubmed: 28627217
J Neuroeng Rehabil. 2021 May 21;18(1):83
pubmed: 34020662
Neurorehabil Neural Repair. 2008 Nov-Dec;22(6):706-14
pubmed: 18971385