Intraoperative electrocorticography for physiological research in movement disorders: principles and experience in 200 cases

Abstract

OBJECTIVE Contemporary theories of the pathophysiology of movement disorders emphasize abnormal oscillatory activity in basal ganglia-thalamocortical loops, but these have been studied in humans mainly using depth recordings. Recording from the surface of the cortex using electrocorticography (ECoG) provides a much higher amplitude signal than depth recordings, is less susceptible to deep brain stimulation (DBS) artifacts, and yields a surrogate measure of population spiking via "broadband gamma" (50-200 Hz) activity. Therefore, a technical approach to movement disorders surgery was developed that employs intraoperative ECoG as a research tool. METHODS One hundred eighty-eight patients undergoing DBS for the treatment of movement disorders were studied under an institutional review board-approved protocol. Through the standard bur hole exposure that is clinically indicated for DBS lead insertion, a strip electrode (6 or 28 contacts) was inserted to cover the primary motor or prefrontal cortical areas. Localization was confirmed by the reversal of the somatosensory evoked potential and intraoperative CT or 2D fluoroscopy. The ECoG potentials were recorded at rest and during a variety of tasks and analyzed offline in the frequency domain, focusing on activity between 3 and 200 Hz. Strips were removed prior to closure. Postoperative MRI was inspected for edema, signal change, or hematoma that could be related to the placement of the ECoG strip. RESULTS One hundred ninety-eight (99%) strips were successfully placed. Two ECoG placements were aborted due to resistance during the attempted passage of the electrode. Perioperative surgical complications occurred in 8 patients, including 5 hardware infections, 1 delayed chronic subdural hematoma requiring evacuation, 1 intraparenchymal hematoma, and 1 venous infarction distant from the site of the recording. None of these appeared to be directly related to the use of ECoG. CONCLUSIONS Intraoperative ECoG has long been used in neurosurgery for functional mapping and localization of seizure foci. As applied during DBS surgery, it has become an important research tool for understanding the brain networks in movement disorders and the mechanisms of therapeutic stimulation. In experienced hands, the technique appears to add minimal risk to surgery.

Keywords: DBS = deep brain stimulation; DLPFC = dorsolateral prefrontal cortex; ECoG = electrocorticography; FDA = Food and Drug Administration; LFP = local field potential; MER = microelectrode recording; PAC = phase-amplitude coupling; Parkinson's disease; brain oscillations; deep brain stimulation; electrocorticography; functional neurosurgery; movement disorders; primary motor cortex.

Figures

FIG. 1

Planning the placement of a motor cortex ECoG strip using surgical planning software (Stealth Framelink 5) on gadolinium enhanced T1-weighted MRI. Upper: Axial slice near the cortical surface showing the intended location (arrow) of the midpoint of the ECoG strip with respect to the central sulcus and “hand knob” of the precentral gyrus. Lower: Parasagittal slice passing through the intended DBS target (the subthalamic nucleus). The intended DBS trajectory (left line) is projected onto this image. A second trajectory plan (right line) is created in Framelink to connect the intended location of the midpoint of the ECoG strip (shown in Upper) with the planned DBS target and used to generate the ring and arc settings on the stereotactic frame that will guide the placement of a radio-opaque scalp marker over the intended ECoG target.

FIG. 2

EcoG placement and confirmation. A: Intraoperative photograph showing the insertion of a 28-channel ECoG strip (in this example, it is directed laterally to the DBS bur hole towards the dorsolateral prefrontal cortex). B: Intraoperative lateral fluoroscopy image showing the placement of a 6-contact ECoG strip with respect to the scalp marker over the intended motor cortex target (arrow) that is placed stereotactically as described in Fig. 1. The planning trajectory for motor cortex ECoG placement is superimposed on the lateral fluoroscopy image (dotted line), showing that Contact 5 of the ECoG strip overlies the precentral gyrus. C: Intraoperative CT (O-arm, Medtronic) fused with the preoperative MRI to show the gyral localization of each ECoG contact. In this example, Contact 4 (arrow) is over the precentral gyrus.

FIG. 3

Examples of the use of ECoG recordings in physiological research. All recordings are from Parkinson’s disease patients. A: ECoG potential recorded from M1 in a patient at rest (bipolar recording) in comparison with simultaneously recorded subthalamic nucleus (STN) LFP (monopolar recording from Contact 1 of a Medtronic model 3389 permanent DBS lead) without (left side) and with (right side) STN DBS (monopolar stimulation 1–3 V, 145 Hz, 60 μsec). The arrow marks the time of DBS initiation. B: Time-frequency color plot of primary motor ECoG frequency components before and after movement initiation (time = 0) in a Parkinson’s disease patient, showing the movement-related decrease in the beta band and increase in broadband gamma (over 50 Hz). C: Schematic illustration of PAC. Low- and high-frequency rhythms are extracted from the signal by filtering (top 2 lines). If the rhythms are coupled (red trace), high-frequency activity occurs at a preferred phase (shown here in the trough) of the low-frequency rhythm. D: Color plot showing the magnitude of coupling at many combinations of phase frequency and amplitude frequency in a 30-second M1 ECoG potential. Strong coupling is observed on the left between the 10- to 30-Hz beta rhythm and the amplitude of a broad range of gamma frequencies at 50 to 200 Hz, indicating that population spiking is strongly synchronized to the beta rhythm. This synchronization is reduced by therapeutic DBS (right plot) (bipolar stimulation settings: Contact 1–2+, 4 V, 144 Hz, 90 μs). E: Simultaneously recorded subthalamic nucleus spontaneous single-unit discharge (top) and M1 ECoG (middle). Spike-time averages (bottom) from 5 bipolar cortical electrode pairs arranged from posterior to anterior showing synchronization of STN spiking to the phase of a cortical 20-Hz oscillation that is localized to M1. Spike-time averages are calculated by averaging 1-second segments of ECoG that were centered on each of 2000 STN spikes (time 0).