Hello, my name is Hiro Sparks, and I'm a fourth-year medical student at the David Geffen School of Medicine at UCLA. It is a great honor to share my group's research at this AANS 2020 virtual meeting. During this presentation, I will describe a framework for characterizing cortical oscillatory dynamics in Parkinson disease and essential tremor, when we record electrocorticography data from patients undergoing DBS surgery and co-localized electrodes to a common cortical surface. This enables spatial characterization of beta and gamma band power over the sensory motor cortices of ET and PD patients during rest and movement. Study of electrophysiological signatures in relation to the phenomenology of movement disorders has furthered our understanding of neural circuitry within the human brain. Movement disorders such as Parkinson disease can often be characterized by quantitative neurophysiological markers associated with disease state. One such indicator, the cortical beta oscillation, exhibits a variety of movement-related dynamics in healthy individuals and deviates from these normal dynamics in Parkinson disease patients. To this end, there's substantial evidence for the functional distinction of high and low beta bands, with high beta in particular corresponding to Parkinson disease-related network dysfunction. For example, our group has previously shown that high beta power is significantly greater in the primary and premotor region rather than the somatosensory region in Parkinson disease. Additionally, we showed that high beta gamma phase amplitude coupling correlates with PD symptom severity. Furthermore, we've shown that palatal deep brain stimulation, specifically modulates high beta gamma rather than low beta gamma cortical phase amplitude coupling. Given evidence for the functional distinction of beta band subdivisions, we sought to better characterize these signals in essential tremor through comparisons to Parkinson disease. One group has previously reported similar broadband beta power when comparing essential tremor to Parkinson disease, both during rest and movement. Meanwhile, a separate group has shown increased high beta power in Parkinson disease relative to ET, similar power in low beta band range during rest and movement. Deeper exploration of beta band dynamics in ET will further our understanding of neurophysiology that supports this disease state. Clarifying the dynamics of cortical ET electrophysiology is critical if these signatures are to be used to augment patient care in the future. This presentation can be organized into three objectives. Instruct the pipeline for co-localizing electrodes from a patient cohort onto a common cortical surface for spatial resolved analyses. We can capitulate core findings related to ET and PD cortical spectral power during rest and movement. More finely dissect beta band activity for ET and compared to known physiology of Parkinson disease. Here we show an overview of methodology for co-localizing electrodes. 3D coordinates for each subject's ECoG grid contacts were digitized using intraoperative fluoroscopy aligned with pre and post-op CT MRI imaging shown on the left. T1 MRI from each subject was aligned with the MNI standard brain using volume co-registration methods. The computed transformation matrix for each patient was then used to project 3D ECoG coordinates for each subject into MNI space shown on the right. Projection of electrophysiological data onto cortical source space was conducted using distance-based inverse exponential kernel shown in A, whose behavior is defined by piecewise function given in B. Computing kernel-based contributions from each ECoG channel to each grid point on the cortical MNI brain yields a transformation matrix C that can be used to view absolute electrode coverage of various cortical regions given all ECoG channels in a patient cohort. More importantly, it can be used to co-localize electrical activity from different patients onto the same cortical surface and conduct statistically robust spatial analysis. In healthy PD and ET patients, the initiation of movement is accompanied by desynchronization of cortical beta activity and synchronization of cortical gamma activity. We next sought to recapitulate these core findings in order to confirm the validity of our new framework for cortical electrode plotting and to confirm known changes in oscillatory dynamics in subjects with essential tremor and Parkinson's disease. To do this, we analyzed cortical oscillations and ECoG potentials recorded in subjects during DBS surgery. Patients were studied only in the off-medication state. Patients had an accelerometer glove attached to the index finger of the hand contralateral to the ECoG strip. Patients were instructed to alternate resting and finger tapping with the gloved hand for 30-second intervals. ECoG and accelerometer data were recorded for a two- to four-minute period of alternating finger tapping and rest. Accelerometer data from the tapping task was used to define trials of movement and subsequent rest. Spectral analyses were computed using each ECoG contact separately using the trial time limits applied to the ECoG data. Spatially resolved spectral analyses were then conducted by kernel-based projection of electrophysiological data onto the cortical surface using methods previously outlined. Cross-movement rest condition and PDET group comparisons were made with Mann-Whitney U tests for evaluation of significance. Overall, data was analyzed in 34 PDE patients contributing 212 electrodes and 14 ET patients contributing 136 electrodes. This totaled in 48 patients and 348 electrodes. Here we show the included cortical regions. Here we display ECoG channels across all PD subjects projected onto a standard cortical surface. Heat map indicates the relative channel count contribution to each vertex on the cortical surface. For analyses and visualization, channel coverage was thresholded to only include regions of cortex with at least 1.5 effective contributing electrodes in order to minimize high-variance fringe effects from the boundaries of the covered cortical area. Here we display relative channel coverage of ET subjects. Here we show cortical surface projection of the difference between rest and movement beta power across all subjects for electrodes localized to broadening areas 1 through 6. Movement-related desynchronization of beta appeared maximally over the pre-central gyrus in the region of the hand knob, with spread across a larger region including the sensory motor and premotor gyri. Here we plot beta power during rest and movement for PD and ET. In PD, we see a significant decrease in beta power with movement. For the ET group, there are several possibilities for the lack of significant beta desynchronization with movement. However, this is most likely due to washout effects from including all regional electrodes instead of isolating activity to the hand knob region of the pre-central gyrus. Median desynchronization magnitude greater than 1 decibel is consistent with prior literature and would also suggest the lack of condition differences due to noise. In both PD and ET, we confirm a significant movement-associated gamma synchronization in all regions. In line with prior studies, we found broadband gamma power to be significantly greater in PD relative to ET during both movement shown on the left and rest shown on the right. Crowell et al. previously showed this relationship with anatomical specificity to M1. While we didn't test for anatomic specificity, here we show that when taken as an average over several cortical regions, gamma activity remains significantly higher in Parkinson's disease. When taken as a single band, beta power is not significantly different between PD and ET during movement shown on the left or at rest shown on the right. These findings are consistent with prior reports by Rowland et al. Given mounting evidence for the functional segregation of the beta subbands with respect to movement and movement disorders, we sought to further refine the results of ET-PD beta band comparisons. When the beta band is divided, ET demonstrates greater low beta power during rest. PD demonstrates greater high beta power during rest. ET demonstrates greater low beta power during movement. And PD demonstrates greater high beta power during movement. Our results are consistent with prior studies which show PD to be associated with an M1 spectral peak in the high beta range. By contrast, the spectral power peak in ET occurs in the low beta range. Additionally, we report a new finding that low beta power is significantly greater in ET relative to PD. Magnitude change in high beta and low beta during rest and movement were not different across disease groups. These studies are consistent with broadband beta findings reported by Rowland et al. More extensive evaluation of the power spectra indeed reveals a crossover at the 17 to 20 hertz range, supporting the finding that lower frequency power, low beta and even alpha shown here, are consistently higher in ET when compared to PD before transitioning to the PD-dominant spectral power in the high beta range. Note that the power spectral density for each ECoG channel was normalized to integrated power from 3 hertz to 200 hertz prior to averaging shown above. Our methodological pipeline for electrode co-registration and cortical surface projection is similar to a variety of other workflows published by labs working in the human neuroelectrophysiological research space. It is able to recapitulate standard findings of known motor electrophysiological activity present over the perirowlandic gyri. It will be utilized by our group for conducting subsequent patient group level analyses. Increased beta dominates cortical and subcortical network nodes in Parkinson disease. Our result complements literature which suggests the greater power within the high beta subband is more specific biomarker for Parkinson disease than elevated broadband beta power. In conjunction with our findings of diffuse cortical broadband gamma increase in Parkinson disease relative to ET, it is likely that the observed high beta increase in power is part of a larger broadband increase extending into the gamma ranges. Thus our findings help to refine the beta synchrony hypothesis of Parkinson disease. That is pathologically increased high beta synchrony may represent a limited portion in the spectrum of oscillatory derangements at the cortical level. Essential tremor is thought to arise from the cerebellophilic system rather than the basal ganglia. Cerebellophilic abnormalities could alter oscillatory activity in cortical areas so it is possible that elevated low beta frequency power seen in patients with essential tremor is itself pathological. Alternatively previous electrocorticography studies have shown similar M1's power spectral peaks in the alpha and low beta frequencies in subjects with without movement disorders. Thus it is also possible that the pathological relative low beta suppression Parkinson disease would explain the difference seen between the two groups. Another possible explanation for the degree of elevated low beta in ET compared to PD may be due to our full power spectrum normalization method with pathological suppression of power and other frequency bands driving up the observed low beta amplitude. More extensive analyses not limited to beta band will be conducted to ensure that this is not the case. Thanks to all those who viewed this presentation and a special thanks to all members of the UCLA Neurosurgical Brain Mapping and Restoration Lab.

Parkinson's disease

essential tremor

electrocorticography data

beta band power

gamma band power