Okay. And our last speaker is Professor Jean Régis, who will be speaking about pathological networks and essential tremor and the surgical implications of these. Thank you very much for your kind invitation. How does it work? Yes. Years ago when we started to perform radiosurgery, that was a few years after the introduction of DBS by Alim Ben-Abid, and that was clearly not an option for us to use radiosurgery. And surprisingly, in the 2000s, neurologists have started to ask us to perform radiosurgery in some of these patients. And we have seen, certainly, a very unexpected change in our practice, which is a dramatic increase in the demand for radiosurgery in these patients. This is a paper we published some years ago that was our first cohort of 50 patients who were blindly evaluated by the Grenoble Group, by Paul Crack, and by Marwan Harris in Queens Square for the radiological aspects. Demonstrating quite a good efficacy and safety of radiosurgery in this indication. And the consequence has been this dramatic increase of the practice of radiosurgery for tremor. We are doing more than 100 tremor radiosurgery per year, and we do something like 20 DBS for VIM in tremor. The targeting is remaining, basically, based on inter-individual targeting of your scheme with the technique you can see here. Even if we use the visualization on the diffusion of the preliminiscal radiation, this is basically still targeted inter-individually, without physiology, of course. The vast majority of the patients are displaying the same changes after radiosurgery. A few patients are a little bit hyporesponder. Some of us are hyporesponders. But there is quite a consistent answer radiologically of these patients. We don't have the physiology when you do a radiosurgery. We, of course, like people doing IFU, we are interested to develop more direct targeting. The microstructure of VIM is known not to be very different from the structures around, and that may be not the best direction to go. The connectivity is quite different. So it varies in different ways. The clustering is one direction. The fiber tracking is another direction. The group we are working with in the APFL have worked on the clustering based on diffusion imaging and higher resolution MRIs. And to make a long story short, we have a nice clustering of the thalamus, but usually what we extract is the ventrolateral group, which is including VIM, but is not enabling us to distinguish the VIM itself from the other structures. So we, like everybody, we think that the connectivity is better and certainly because there is this anterior-posterior organization with more negral and palatal afferent anteriorly and more sensory afferent posteriorly. And the VIM itself is more cerebellar afferences. For a long time, everybody thought that the dentate nucleus was traditionally the major contributor to cerebellar thalamic cortical projections. We do know that there is also interpositional fastidial thalamocortical connections. And this is a little bit more complex in terms of connectivity. There is a number of excellent works in the literature nowadays. This is the thesis of Akram showing how you can subdivise using tractography, the thalamic areas. But these techniques are coming with a number of limitations and that's still a work in progress. The second point of interest in this group of patients for us was to try to understand what can predict the outcome of this patient. What are the good predictors of outcome? And we have started a program we call GARP for Gamma Life Response Prediction. And we are screening a lot of information. Genetic sampling, phenotypes, DTI location, tremor phenotype, comorbidities, medications, radio biology. And surprisingly, we found some months ago in our pet studies strange information. Our pet guys came to us and they told us, we are able to predict the good responders of radiosurgery for tremor based on the occipital cluster. I was absolutely not believing this. That was making no sense to me. In spite of the fact that the data were quite strong, you can see that the sensitivity and the specificity of this occipital cluster, the capacity to predict the responders versus non-responders was quite high. So we started to work a little bit more on this and we took a group of patients. That's a group of 38 patients, very nicely screened before and after for the clinical aspects and the MRI aspects. And we were having 82% of responders in this group of patients. And here you see the cluster we found, which is in the occipital, mesial, and basal area, which is more or less corresponding to what we call the par-hypercampal place area, which is known to be involved in integrating movement in the scenery, visual scenery. In order to question this finding, we proposed to use the VBM. So we took all the MRIs of these patients and we performed VBM studies, voxel-based morphometry, comparing responders versus non-responders, responders and non-responders between time point and baseline versus one year. And we found, interestingly, exactly the same thing. We found a cluster in the occipital area, which is the same cluster we got in pets. So what is very interesting is that the only, in this study, the only predictor of quality of outcome for this patient, that was the quantity of gray matter in this right occipital associative cortex. All these patients were treated on the left VIM for right tremor. So we found the same thing. A cluster was different between the responders and the non-responders in the same area in VBM. And when we looked at the comparison between before and after, we found that the radiosurgery was changing the quantity of gray matter. And that was changing the quantity of gray matter differently in the responders and the non-responders. So we thought we, that was worth doing more research on this aspect of the involvement of occipital cluster. And we started to perform systematically resting state in this patient before and after. I don't have to explain here what is resting state. And we started by working in this material without applying any a priori hypothesis in this patient. And we found a different connectivity, cluster of connectivity in the occipital area. And the connectivity in this area was demonstrated in resting state to be increased in non-responders before radiosurgery. And this decreased in responders. And the amplitude of the decrease of the connectivity was correlating with the amplitude of the tremor improvements. And the QUEST score and the head tremor score at one year also were correlating to the decrease of the connectivity. We found also in this work another cluster, which was related to the response, which is an insular cluster we will find again later. So we tried to work differently. Now we will introduce a priori, which is to look at the connectivity with the area we treated. And we were defining this area using diffusion. We were extracting automatically the cluster corresponding to the ventrolateral area. And we were putting seed there. And when you look at the outcome, what we found is again that we were having a correlation with the occipital area and the outcome of this patient. So the connectivity before was predictive of the outcome. And the connectivity was changing after radiosurgery, especially in those who were responding to radiosurgery. So interestingly, when we started to send papers, we were several times rejected with comments like, oh, the offers went fishing. There is no other thing outside the cerebellopharmocortical axe in essential tremor. I found it not very serious scientifically. That's not because we don't know something that this thing doesn't exist. And I found it very poor in terms of comment of a high-ranked journal whatsoever. Few months after, we got this paper in Brain. That's a paper from a group who looked at the task MRI. And they found a cluster in the occipital area. So that was exactly the same thing. That was confirmation of our data. So that was an opportunity for us to write a letter to the editor. And this is this letter to the editor we wrote in Brain. What you can see on the left is the correlation between the connectivity and the improvement of the tremor. And you can see here the functional connectivity in those who were not responding and those who were responding. And you see that the connectivity is increased in the non-responders and is decreased in the responders. So we started to work again on this data. Now using what we call a group-level independent component analysis strategy. And we looked first at the head tremor. And we found that we were able to predict those who will respond, head tremor, which is not frequently responding to surgery in tremor. But we were able, based on the bilateral thalamus and limbic system and SMA connectivity, we were able to predict the responders in terms of head tremor reduction. And then we looked globally at the improvement of the tremor using this methodology looking at 20 major large-scale brain networks. And we found three networks with the interconnectivity statistically significantly predicting the outcome. The first was M1 interconnected with the inferior olivary nucleus. The second was bilateral thalamus interconnection with motor cerebellum lobulus V2. And the third was the anterior default mode network interconnected with Bronman area 10. For all, more positive pre-therapeutic interconnectivity was correlating with higher drop-in points in the respective scores. And these are the correlations for the three networks. So we looked at these data now comparing to healthy controls and looking at the changes after radiosurgery. And we found, still using this technique of the 20 large-scale networks, and we found that changes were displayed with reorganization of the dorsal attention and silence network after radiosurgery. And you see here the healthy, the baseline, and the one-year. And you see the healthy control versus the one-year for the three different clusters here. Let's make it a little bit faster in the interest of time. So just to summarize, that's very interesting out-of-the-box data. The more we are digging, the more we are finding a pertinence of a specific occipital area in patients with tremor. Of course, this technique we have been using have a number of limitations. We don't have a lot of subjects. You can discuss the use of the resting state. But whatever we do, whatever we use PETs, VBM, resting states, whatever the limit of each, all these techniques are showing the same thing. There is something strange in the occipital area which is correlating with the tremor, which is correlating with the result and outcome. And there is also other networks we have identified which are also of interest, like the involvement of the insula and I have no clue why and how the insula is involved in this story. But that's a lot of food to think, I will say. So just to summarize and to conclude, nowadays what we are working on is we are trying to use this data in the context of what we call the virtual brain. So we are working with a group of mathematicians who have developed a model, individual model of the brain network. They are taking the connectome of the patient and they are trying to modelize pathologies in the brain. And we are adding functional information on this base of connectivity informations. And we have started to use it first for epilepsy and now we are trying to use it to modelize this abnormal network of the central tremor based on all this data. And that's a very exciting work, but that's just work in progress. We, whatever you use, radio surgery or you use IFU, we need to improve our capacity to target. Although, without direct targeting, our results are quite good. There is certainly a margin for improvement by better individualized targeting. Interestingly, it's searching for biomarkers of response. We have found some very interesting new data about the abnormal network of the tremor. I think radio surgery and certainly IFU are especially interesting compared to DBS because we can work without the artifact of the lead. You can work before and after and you can follow in time the changes. We don't have the brain shift induced by the DBS when we do DBS. And that's something which is interesting in this kind of techniques. The great difficulty when we try to think about all this data we have nowadays is about the nature of what we observe. When you observe a pathological network, is it the tremor? Is it the effort of the patient to compensate the tremor? We don't know. Some of these things are different depending on the duration of the tremor, the history of the tremor. Some of this observation can be consequences of the tremor. Some can be the effort to control the tremor. We know the essential tremor is certainly a group of different diseases and maybe the tremor are different if these diseases are different. So there is still a lot of work to do. And if I'm taking a step backward, globally we have now a very interesting experience with the use of radio surgery. Twenty years ago, I was personally completely convinced that regional techniques were completely dead in movement disorders. And now we have this major increase of demand, both from neurologists and patients. And when I'm speaking about neurologists, I'm speaking about neurologists who are involved in highly specialized centers where there is DBS. So they know perfectly the advantages and disadvantages of each technique. And this interest is still sustained after 10 years referring us patients. So this is with long follow-up of our patients. So that's interesting to see how these things are changing rapidly. Thank you for your attention. So if I can just add something. This is the team. This is a course we're organizing in Marseille about functional radio surgery. And this is Ludwig Zlinzow. This is the congress you're organizing in Denver on behalf of the SSFN, which will be very much dedicated to psychiatry for those who are interested. Thank you.

radiosurgery

essential tremor

neurologists

efficacy and safety

gray matter