Anatomical Data Driven Modeling of Evoked Compound Action Potentials Recordings During Spinal Cord Stimulation in a...

A spinal cord stimulation (SCS) approach has been developed that uses inactive electrode contacts to record epidural spinal recordings (ESRs) generated during SCS. ESRs contain evoked compound action potentials (ECAPs) which represent a quantitative measure of synchronous neural recruitment in the spinal cord. ECAPs may be utilized as a control signal for closed-loop stimulation and aid in optimal electrode placement and parameter selection. However, the various physiological and technical factors underlying the composition of these signals are difficult to obtain experimentally due to subject variability and sources of noise, which may limit the use of ECAPs in elucidating mechanisms of SCS-induced analgesia. Therefore, the goal of this study was to use computational modeling based on detailed anatomical imaging paired with preclinical physiological data to investigate the neuromodulatory effects of SCS.

We developed a computational model from an experimental data set containing imaging and ESRs from six swine. We coupled our finite element method model with multicompartment cable models to simulate the neural response to SCS. We then used a reciprocity-based approach to calculate model ECAP recordings.

Model ECAPs were dependent on stimulation parameters (ie, tonic stimulation waveform and configuration) and anatomical variations (ie, dorsal cerebrospinal fluid thickness and mediolateral lead location). Our modeling results indicate that the combined choice of stimulation waveform and stimulation configuration may result in action potential initiation at different locations, which, when recorded, gives rise to ECAPs with different morphologies and amplitudes, even for approximately the same level of underlying neural activation.

Our findings suggest that ECAP characteristics may not directly represent the level of neural recruitment to a stimulus and are highly dependent on stimulation parameters. Overall, the results of this study provide a mechanistic understanding of how various factors affect the composition of ECAP recordings and will help optimize the utility of ESRs in SCS.