Real-Time Artifact Suppression in Neuromodulation: A Model-Based Approach.

Closed-loop neuromodulation is a promising technique for treating neurological disorders such as Parkinson's disease and epilepsy. However, induced electrical stimulation artifacts remain one of the main challenges of closed-loop neuromodulation technology. Existing methods for removing these artifacts are not robust enough to restore the lost spectral and statistical characteristics of neural recordings and are not efficient enough for real-time applications. In this study, we propose a novel artifact reduction method based on the modeling of the brain signals that can remove stimulus artifacts from both cortical and subcortical recordings. We demonstrate the effectiveness of our method using data recorded from the human motor cortex and subthalamic nucleus. Our method restores the original neural recording by removing the stimulus-induced artifact and preserves the spectral features of the cortical and subcortical neuronal activities, which are essential for many types of closed-loop neuromodulation. Our method has the potential to enhance the performance of closed-loop brain stimulation by providing artifact-free neural signals and improving the efficiency of clinical treatments.Clinical Relevance- This study presents a novel artifact reduction method that enhances the accuracy of neural recordings during closed-loop neuromodulation. By effectively removing stimulation-induced artifacts while preserving critical features of neuronal activity, this approach enables more reliable neural signal interpretation. This method will enable neuromodulation therapies that utilize real-time monitoring of neurophysiologic data during stimulation for conditions such as Parkinson's disease and epilepsy, ultimately optimizing treatment outcomes and enhancing patient care.