Evoked compound action potential (ECAP) is extensively used as biomarker for neural activity in closed-loop spinal cord stimulation (SCS). However, accurate monitoring of neural activity is challenged by the low signal-to-noise ratio of ECAP recording, primarily due to the interference from stimulation-induced artifacts. This study proposes biphasic electrostimulation artifact model (BEAM) for ECAP extraction and assesses its efficacy using an enhanced quantification approach. The BEAM model is built based on the principle of additivity and incorporates boundary conditions that account for the equivalent electrical environment and temporal relationships. A bilinear growth curve model is introduced to calculate the E-score, enabling a comparison of ECAP extraction results with other methods. The proposed approach is extensively validated using clinical data from nine-month SCS therapy. BEAM method achieves the highest E-scores and reduces stimulation artifacts clinically without distorting information related to neural activity. BEAM accurately characterizes the morphology of artifacts generated by stimulation across various neurostimulation scenarios. Our model may lead to unique interpretation of neural activation in ECAP signals, evoked by electrical stimulation especially in closed-loop controlling strategies that aim to maintain consistent neural excitation.
Zhang et al. (Thu,) studied this question.