Hidden semi-Markov models in the computerized decoding of microelectrode recording data for deep brain stimulator placement.

Abstract

To describe an approach to the analysis of deep brain stimulation (DBS) of the subthalamic nucleus (STN) using a hidden semi-Markov model (HsMM) and early results of the analysis of microelectrode recordings for STN DBS. The author simulated the anatomy and electrophysiology of STN DBS and built a seven-state model to compare Hidden Markov model (HMM) and HsMM approaches. Accuracy of these competing models was similar for correctly identifying brain nuclei; however, HsMMs showed superior specificity in detecting microelectrode passes traversing the STN. Further clinical work must be done; however, based on these data, HsMMs may be best suited to computer-assisted anatomic delineation for DBS.