Okay. Thank you so much. All right. Our last talk is by Dr. Vibhor Krishna, Distinct Cortical Regions Mediate Improvements in Individual Motor Symptoms After STN, DBS, and Parkinson's Disease. Good evening, everyone. Thank you for the organizers for giving us the opportunity to share some of the work. And Dr. DeSalis kind of set the background for this talk a little bit. In my lab, we are interested in studying the neuroimaging correlates of effective stimulation in patients undergoing DBS and other neuromodulation procedures. These are my sources for funding. The first two are related to this talk, and the others are not. So today, I'm going to talk to you about three major aspects of this research. First, the study hypothesis and how we came about to form this hypothesis based on the clinical observations that during DBS adjustment, some symptoms get better more than others. We looked at three main symptoms of Parkinson's disease, rigidity, bradykinesia, and tremor, and compared that to muscle contractions. We used probabilistic tractography of the whole cortical ribbon, and then looked at – and then I'll discuss with you the results that we found that there are unique cortical areas that are associated with each individual symptom improvement after STN, DBS. And then finally, I would share with you why we are passionate about this research and how it can impact patient care, especially in the field of shaping the stimulation with directional leads and also making the titration more efficient. So as we all know, Parkinson's disease has three major motor symptoms, rigidity, bradykinesia, and tremor. STN, DBS is shown to improve all three symptoms, but sometimes during DBS adjustment acutely, some symptoms get better more than others. A intuitive way of thinking about this is a picture of an electrode that I'm showing. At contact zero, when you stimulate at 0.5 volt, we do not observe any clinical improvement. But as you move from 0.5 to 1.5 volt, the stimulation predominantly improves tremor. We reasoned that a change in stimulation volume between 0.5 to 1.5 volt incorporates more axonal pathways in the region and may lead to the incorporation of cortical areas that may be involved in efficacy of STN, DBS. So we initially looked at this acute stimulation data, and using probabilistic tactography, we're able to see that between the two stimulation changes, there were more cortical areas that were incorporated. So we hypothesized that stimulation-induced distinct motor improvements after STN, DBS are mediated by unique cortical regions. So for each symptom improvement, there may be unique cortical areas that are involved and mediating this effect. We studied this in a cohort of 24 Parkinson's disease patients who had STN, DBS. And as I said before, we studied three symptoms that are of interest for therapeutic efficacy of STN, DBS, rigidity, bradykinesia, and tremor. And the observations were 52, 30, and 84 in each of these domains. And then we compared that to muscle contraction induced by STN, DBS, and the number of observations was 124. All patients had structural T1, diffusion-weighted imaging, and postoperative CT for localization of these electrodes. And we used an integrated pipeline using three different softwares for three different functions. So we used LEAD-DBS for calculating the volumes of tissue activation. And then used FSL for probabilistic tractography and group-level analysis in each symptom. And finally used AFNI for comparison across different symptoms and clustering and visualization. Just a little bit more about the group-level analysis. For each symptom, we did a permutation test with 5,000 permutations. And this p-value was 0.05. For comparison between different symptoms, we used mixed effect modeling and also corrected the results for false discovery rate. Finally, for clustering and visualization, we used a multimodality atlas produced by the Human Connectome Project group recently published in Nature. So first result looking at cortical areas that were associated with STN, DBS. Again, as shown by several other groups before, we identified that parts of the superior frontal gyrus and parts of Broadman area 9, medially supplementary motor area and the medial prefrontal cortex were shared across all three motor symptoms. So these were the regions shown in green that were shared across rigidity, bradykinesia, and tremor. The regions shown in yellow were shared only by two symptoms. For example, primary motor cortex, shown here, was shared between tremor and bradykinesia. Parts of the supplementary motor area were shared between tremor and rigidity, while again parts of Broadman area 9 were shared between tremor and rigidity again. Interestingly, there were some cortical areas that were unique to just one symptom improvement. For example, the premotor cortex was unique to tremor control only, while parts of the ventral prefrontal cortex were only associated with bradykinesia. Finally, looking at the stimulation-induced motor contraction, parts of the primary motor cortex and ventral premotor area were uniquely associated with motor contraction. So was the frontopolar region as well. Given the resolution of tractography at 2 millimeters, we then started to analyze whether in these cortical regions that were shared across different symptoms, were there any clusters that were uniquely associated with each symptom. And so I'm showing you an example of a high-resolution comparison between bradykinesia and tremor. And the regions that we of interest are primary motor cortex, superior frontal gyrus. Comparing the distribution of significant voxels between the two symptom improvements, we can see that there are voxels which are adjacent to each other that were associated uniquely with bradykinesia and tremor within the cortical regions that they shared. So the strengths of the study mainly include the sample size, 290 stimulation pairs that we analyzed, a data-driven approach using the whole cortical mask, and using probabilistic tractography, which is not limited by the concerns for crossing fibers. The major limitations of this analysis are a retrospective analysis of a prospective collected data. The imaging analysis was only restricted to the cortex, so we don't know what were the changes and what were the associations in the subcortical regions. We did not correct for free water contamination that we are starting to do in future studies. And finally, the functional relevance of these findings is unknown because this is just a probabilistic tractography study and we don't have functional studies. So finally, I want to talk to you about what are the implications of this kind of work for patient care in future. So when we looked at the parcellation of the subthalamic area based on the cortical masks that you just saw, we started to see that different parts of subthalamic region were associated with each symptom improvement. So the symptom improvement of bradykinesia is shown with golden-yellow color, which was more anterior and dorsal, while the violet color is improvement in rigidity, which was ventral and posterior, and the improvement in tremor was more medial. Finally, the contraction side effect was posterior and lateral. So this kind of work has implications for shaping the electrical field with the use of directional leads. Another implication of this work could be making stimulation titration efficient. So Dr. Okun has previously published results that patients who were deemed failures of DBS, one-third of those patients could be salvaged by doing methodical stimulation adjustment again. And each stimulation session, at least the initial titration sessions, can take somewhere between 60 to 90 minutes. So we started to look into whether this kind of an approach can make stimulation titration more efficient. So looking at the preferential connectivity of implanted DBS electrodes, we trained a classifier to identify electrodes that were preferentially connected to effect-producing cortical areas versus side effect, and then tested the classifier and looked at the results to validate it in a separate cohort. We could achieve a classification accuracy of up to 90 percent using this approach. So in conclusion, I discussed with you the research where we had clinical observations that some symptoms improved more than others during DBS adjustment. And using probabilistic tractography, we were able to identify cortical regions that are uniquely associated with symptom improvement, some that are shared across all three, some that are shared by two symptoms. And finally, there are areas that are unique to each symptom improvement. Even within shared cortical areas, there are voxels that are adjacent and are associated with single symptom improvement. And the future implications of this research for shaping of the electrical field and making stimulation titration more efficient. I'll take any questions if you have. I want to thank all my collaborators and the people in my lab who worked hard on this project. So, the question that you have is, at an individual level, can we resolve some of these things? Yeah. It's different. There is an increasable difference between area 1 and area 2. So, we haven't looked at that individual level analysis yet, but that's a great idea. So, it's not like individual? I don't have the answer to this question just yet. So, what we can do for individual difference, we could look at each electrode at individual level and classify these into, that could be more effective for tremor improvement or bradykinesia improvement overall. Thank you. Thank you. Thanks a lot to everybody, all the speakers. Very interesting work, and thanks to you guys for sticking around to the end. And this concludes the session. Thank you.

neuroimaging correlates

deep brain stimulation

Parkinson's disease

cortical regions

motor symptoms