Pallidotomy microelectrode targeting: neurophysiology-based target refinement.

Objective: Microelectrode recording can refine targeting for stereotactic radiofrequency lesioning of the globus pallidus to treat Parkinson's disease. Multiple intraoperative microelectrode recording/stimulating tracks are searched and assessed for neuronal activity, presence of tremor cells, visual responses, and responses to kinesthetic input. These physiological data are then correlated with atlas-based anatomic data to approximate electrode location. On the basis of these physiological properties, one or more tracks are selected for lesioning. This study analyzes the track physiological factors that seem most significant in determining the microelectrode recording track(s) that will be chosen for pallidal lesioning.

Methods: Thirty-six patients with Parkinson's disease underwent microelectrode-guided pallidotomy. Between one and five microelectrode recording tracks were made per patient. Usually, one (n = 23) or two (n = 12) of these tracks were lesioned. Electrode positions in the x (mediolateral) and y (anteroposterior) axes were recorded and related to track neurophysiological findings and final lesion location. The stereotactic location and sequence of microelectrode tracks were recorded and plotted to illustrate individual search patterns. These patterns were then compared with those noted in other patients. Neurophysiological data obtained from recording tracks were analyzed. A retrospective analysis of track electrophysiology was performed to determine the track characteristics that seemed most important in the surgeon's choice of the track to lesion. Track physiological properties included general cell spike amplitude, tremor synchronous neuronal firing, kinesthetically responsive neuronal firing, and optic track responses (either phosphenes reported by the patient during track microstimulation or neuronal firing in response to light stimulus into the patient's eyes). Orthogonally corrected postoperative magnetic resonance images were used to confirm the anatomic lesion locations.

Results: In patients who had a single mapped track lesioned, specific track electrophysiological characteristics identified the track that would be lesioned most of the time (20 of 24 patients). Tracks that exhibited a combination of tremor synchronous firing, joint kinesthesia, and visual responsivity were lesioned 17 (85%) of 20 times. Analysis of intraoperative electrode movement in the x and y axes indicated a significant subset of moves but did not result in microelectrode positioning closer to the subsequently lesioned track. Accuracy of initial electrode movement in the x and y axes was most highly correlated with a measure of first-track electrophysiological activity. The number of microelectrode recording tracks did not correlate with clinical outcome. Anatomic analysis, using postoperative magnetic resonance imaging, revealed that all lesions were placed in the globus pallidus. Most patients (35 of 36) improved after surgery.

Conclusion: The level of electrophysiological activity in the first track was the best predictive factor in determining whether the next microelectrode move would be closer to the ultimately lesioned track. The analysis of electrode track location and neurophysiological properties yields useful information regarding the effectiveness of microelectrode searching in the x and y axes. Within an institution, the application of this modeling method may increase the efficiency of the microelectrode refinement process.