Epidural Spinal Cord Recordings (ESRs): Sources of Artifact in Stimulation Evoked Compound Action Potentials
Ashlesha Deshmukh,Megan L Settell,Kevin L Cheng,Bruce Knudsen,James K Trevathan,Maria LaLuzerne,Stephan Blanz,Aaron Skubal,Nishant Verma,Ben Romanauski,Meagan Brucker-Hahn,Danny Lam,Igor Lavrov,Aaron Suminski,Douglas Weber,Lee Fisher,Scott Lempka,Andrew Shoffstall,Hyunjoo Park,Erika Ross Ellison,Mingming Zhang,Kip Ludwig
DOI: https://doi.org/10.1101/2024.05.17.594739
2024-07-12
Abstract:Introduction: Evoked compound action potentials (ECAPs) measured using epidural spinal recordings (ESRs) during epidural spinal cord stimulation (SCS) can help elucidate fundamental mechanisms for the treatment of pain, as well as inform closed-loop control of SCS. Previous studies have used ECAPs to characterize the neural response to various neuromodulation therapies and have demonstrated that ECAPs are highly prone to multiple sources of artifact, including post-stimulus pulse capacitive artifact, electromyography (EMG) bleed-through, and motion artifact resulting from disturbance of the electrode/tissue interface during normal behavior. However, a thorough characterization has yet to be performed for how these sources of artifact may contaminate recordings within the temporal window commonly used to determine activation of A-beta fibers in a large animal model. Methods: We characterized the sources of artifacts that can contaminate the recording of ECAPs in an epidural SCS swine model using the Abbott Octrode™ lead. Muscle paralytics were administered to block muscle activation preventing EMG from contaminating the recorded ECAPs. Concurrent EMG recordings of the longissimus, a long muscle of the back, were used to confirm a 2-4 millisecond (ms) latency source of EMG bleed-through that frequently contaminated the A-beta temporal window. Additionally, we obtained recordings approximately 5-10 minutes post-mortem after clear evoked A-beta and associated EMG responses ceased to characterize the representation of stimulation artifact across the array. Results: Spinal ECAP recordings can be contaminated by capacitive artifact, short latency EMG from nearby long muscles of the back, and motion artifact from multiple sources. In many cases, the capacitive artifact can appear nearly identical in duration and waveshape to evoked A-beta responses. These sources of EMG can have phase shifts across the electrode array, very similar to the phase shift anticipated by propagation of an evoked A-beta fiber response across the array. This short latency EMG is often evident at currents similar to those needed to activate A-beta fibers associated with the treatment of pain. Changes in cerebrospinal fluid between the cord and dura, and motion induced during breathing created a cyclic oscillation in all evoked components of the recorded ECAP signal. Conclusion: Careful controls must be implemented to accurately separate neural signal from the sources of artifact in spinal cord ECAPs. To address this, we suggest experimental procedures and associated reporting requirements necessary to disambiguate the underlying neural response from these confounds. These data are important to better understand the conceptual framework for recorded ESRs, with components such as ECAPs, EMG responses and artifacts, and have important implications for closed-loop control algorithms to account for transient motion such as postural changes and cough.
Neuroscience