A Metric for Evaluating and Comparing Closed-Loop Deep Brain Stimulation Algorithms

Jeffrey Herron,Anca Velisar,Mahsa Malekmohammadi,Helen Bronte-Stewart,Howard Jay Chizeck

DOI: https://doi.org/10.48550/arXiv.1605.09312

2016-05-20

Abstract:Objective: Closed-loop deep brain stimulation (DBS) may improve current clinical DBS treatment for neurological movement disorders, but control algorithms may perform differently across patients. New metrics are needed for comparing and evaluating closed-loop algorithm performance that address the specific needs of closed-loop neuromodulation controllers. Approach: A metric is proposed for system performance that includes normalized terms that can be used to compare algorithm performance for a patient. This metric was evaluated using two closed-loop control algorithms that were tested in patients with Parkinson's Disease (PD) who experience rest tremor. Main Results: The metric's resulting balance between tremor treatment and power savings varied on a per patient and algorithm basis. This was expected given how each trial resulted in a variable reduction in stimulation power at the cost of additional tremor for the patient when compared to open-loop stimulation. Significance: The proposed metric will aid in clinical evaluation of new algorithms and provide a benchmark for future system designers. This will be important given the growing potential applications of dynamically adjusted neural stimulation. <a class="link-external link-http" href="http://ClinicalTrials.gov" rel="external noopener nofollow">this http URL</a> Identifier: NCT02384421.

Neurons and Cognition