Modeling of Recruitment Curves
Conditions
Keywords
TMS, SCS, SCI, Hierarchical Bayesian, Motor threshold, Transcranial Magnetic Stimulation, Spinal cord stimulation, Evoked Potentials
Brief summary
The purpose of this study is to better understand how electrical or magnetic stimulation affect the nervous system by optimizing the way researchers measure muscle responses. The relationship between stimulation intensity and muscle response is described by "neural recruitment curves," which are critical for monitoring the state of the nervous system during therapies like transcranial magnetic stimulation (TMS) and spinal cord stimulation (SCS). This study tests a new, real-time computational approach based on our previously developed methods (Hierarchical Bayesian models) to estimate these recruitment curves more efficiently. The primary goal is to use this model to dynamically guide the experiment, automatically selecting the optimal stimulation intensities to test. The investigators hypothesize that this optimized approach will accurately estimate the entire recruitment curve, or specific targets components of it like the motor threshold, using significantly fewer samples than standard methods. By reducing the number of measurements required, this approach aims to decrease experimental time and minimize participant burden, making future TMS and SCS therapies and experiments more feasible and efficient.
Detailed description
Transcranial magnetic stimulation and other types of neurostimulation play a crucial role in advancing the understanding and manipulation of neural activity for both research and therapeutic purposes. The proposed approach to sampling recruitment curves in real-time promises to significantly improve the efficiency and precision of experiments that use electrical or electromagnetic stimulation techniques, reducing the experimental burden for participants as well as experimenters. By enhancing experimental efficiency in multiple experimental settings and techniques, this research directly contributes to accelerating the translation of scientific discoveries into clinical applications. This study will benchmark the relative performance of different methods against each other by testing existing and proposed algorithms using neurostimulation in people, and comparing the resultant estimates in recruitment curve parameters, and the number of samples required to reach predefined tolerances on these parameters.
Interventions
OTHERAlgorithm: Uniform Sampling
Standard uniform distribution sampling used as a baseline comparison.
OTHERAlgorithm: hbMEP-adaptive algorithm (version 1)
An active sampling algorithm for recruitment curve estimation.
OTHERAlgorithm: hbMEP-adaptive algorithm (version 2)
An alternative active sampling algorithm for recruitment curve estimation.
OTHERML-PEST
Algorithm: Adaptive threshold hunting using the Parameter Estimation by Sequential Testing (PEST) algorithm.
DEVICEMagPro X100 Transcranial Magnetic Stimulation
The proposed algorithms will deliver stimulation by using this magnetic stimulation methodology.
DEVICEDigitimer DS8R Transcutaneous Electrical stimulation
The proposed algorithms will deliver stimulation by using this electrical stimulation methodology.
Sponsors
Columbia University
National Institute of Neurological Disorders and Stroke (NINDS)
Study design
Allocation
NA
Intervention model
SINGLE_GROUP
Primary purpose
BASIC_SCIENCE
Masking
NONE
Masking description
Participants are functionally masked to the specific interventions, as the stimulation parameters generated by the different algorithms are randomly interleaved pulse-by-pulse.
Intervention model description
This is a single-group, within-subject methodological study designed to compare different TMS sampling algorithms. All participants undergo multiple experiments in a single sessions. A single experiment will compare multiple sampling algorithms. Specifically, the neurostimulation pulses dictated by each active algorithm are interleaved in a randomized sequence. This interleaved design ensures that any time-dependent physiological variables impact the threshold and recruitment curve estimations of all tested algorithms equally.
Eligibility
Sex/Gender
ALL
Age
18 Years to 90 Years
Healthy volunteers
Yes
Inclusion criteria
* Healthy adult volunteers aged 18 years and older. * Able to understand study procedures and provide written informed consent.
Exclusion criteria
* 1\. History of adverse reaction to Transcranial Magnetic Stimulation (TMS) or non-invasive neurostimulation. * 2\. History of seizures, epilepsy, or family history of epilepsy. * 3\. History of stroke, brain injury, or illness causing brain injury. * 4\. History of head injury or neurosurgery. * 5\. History of neurological diseases, or central nervous system lesions. * 6\. Presence of metallic implants or foreign bodies in the head (outside of dental work/fillings). * 7\. Presence of implanted electronic or medical devices (e.g., cardiac pacemakers, medical pumps, implanted stimulators). * 8\. Current pregnancy or possibility of pregnancy. * 9\. Currently taking medications that alter cortical excitability or lower seizure threshold.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Number of stimuli to reach a pre-defined threshold error | Through completion of the study visit, 2-4 hours. | Number of stimuli required for the compared methods to reach a pre-defined error threshold relative to the ground truth, computed from recruitment curves fitted after sampling using aggregated data. |
| Number of stimuli to reach a pre-defined predictive curve error | Through completion of the study visit, 2-4 hours. | Number of stimuli required for the compared methods to reach a pre-defined error threshold relative to the ground truth, computed from recruitment curves fitted after sampling using aggregated data. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Mean absolute error in a given parameter (e.g. threshold, predictive curve, slope) for a given number of stimuli | Through completion of the study visit, 2-4 hours. | The error of the methods under comparison, with the ground truth computed from recruitment curves fitted subsequent to sampling using aggregated data. |
Countries
United States
Contacts
CONTACTJames R McIntosh, PhD
PRINCIPAL_INVESTIGATORJames R McIntosh, PhD
Columbia University
Outcome results
None listed