A Clinical Feasibility Study of Spinal Evoked Compound Action Potential Estimation Methods.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Pain Management

Background:

• Spinal cord stimulation (SCS) treats chronic neuropathic pain.
• Evoked compound action potential (ECAP) sensing enhances SCS functionality.
• Accurate ECAP isolation from stimulation artifact (SA) is crucial for SCS control.

Purpose of the Study:

• To characterize the sensitivity of ECAP growth curve slope to neural response and SA contamination.
• To evaluate four spinal ECAP estimation methods using a novel performance measure (|S_resp/S_art|).

Main Methods:

• Collected 112 ECAP and artifact recordings from 14 human subjects.
• Applied high-pass (HP) filtering, artifact modeling (AM), or template correlation to reduce SA.
• Compared methods against each other and the standard P2-N1 method.
• Fit ECAP estimates to a growth curve to calculate neural response slope (S_resp) and artifact slope (S_art).

Main Results:

• The ratio of neural signal preservation to SA misclassification (|S_resp/S_art|) was superior with HP and AM schemes (p < 0.05).
• These methods demonstrated better performance in isolating neural signals compared to other schemes.

Conclusions:

• This is the first comprehensive assessment of spinal ECAP estimation schemes.
• Understanding ECAP estimation sensitivities is vital for closed-loop SCS systems.
• Misclassification of artifact as neural signal can lead to suboptimal SCS therapy adjustments.