View complete transcript
She is the senior author of recent paper published on Movement Disorders Journal titled, Temporal Signature of Task Specificity in Isolated Focal Laryngeal Welcome the Simonyan and many thanks your time.
[00:00:48] Prof. Kristina Simonyan: Thank you very much for the invitation. It's a pleasure to be here.
[00:00:51] Dr. Sarah Camargos: First of all, professor, tell us about your line of research.
[00:00:56] Prof. Kristina Simonyan: I have been interested in understanding the neural [00:01:00] signature of dystonia, which is, as you know movement disorder. And I have been working in this field for past 20 years or so, and I have been specifically interested in understanding the pathophysiology of task specific dystonias, including laryngeal dystonia, writer's cramp, musician's dystonia, and so on.
[00:01:26] Dr. Sarah Camargos: And did you define laryngeal dystonia as a task specific dystonia? Can you tell us about this?
[00:01:37] Prof. Kristina Simonyan: Sure, laryngeal dystonia is defined as a task specific dystonia because it specifically and predominantly affects voice during speaking, but not other behaviors such as whispering less during singing. Shouting, yelling are usually preserved and in fact patients report[00:02:00] no differences from normal baseline in their voice while they do those tasks.
Singing stands out a little bit differently from other vocal behaviors some patients do note that their singing is affected, but not to the degree as. speeches being affected. In a rare form of laryngeal dystonia. Singing is specifically affected, and this is a form that intersects both laryngeal dystonia and musician's dystonia, and symptoms appear in professional singers.
So the disorder is called singer's laryngeal dystonia. And it's a quite interesting phenomenology when speaking is usually not affected while singing is. These patients may develop additional impairments of their speech production as well, but [00:03:00] usually it starts with singing.
Fortunately, it's quite rare. However if the affected individual is a professional singer, this can be a career changing disorder.
As with any other dystonia, singer's dystonia can also progress into multifocal or segmental form and it can take the usual course as in other forms of dystonia, so while singing is a predominant behavior that is affected, it can also further affect speech and other body regions.
[00:03:38] Dr. Sarah Camargos: Very interesting. Could you please tell us about the methodology of your study?
[00:03:45] Prof. Kristina Simonyan: In this study, we included 51 subjects 26 of whom were patients with isolated focal laryngeal dystonia and 27 were healthy controls without any neurological or psychiatric [00:04:00] or voice and speech problems. Our population of patients was specifically selected also not to have any other, any genetic mutation, any psychiatric problems they had normal radiological reading their structural MRI they were all right handed.
Native English speakers, which is important as we try to capture the basic organization of neural networks. And importantly, patients were also post their botulinum toxin treatment at least three months after their last injection which was an important criterion, selection criterion for them to be fully symptomatic during study participation.
As they were asked to do both symptomatic and asymptomatic tasks during the experiment. So with this population, we acquired EEG data using [00:05:00] 128 channel EEG system which is a high density EEG system allowing us to capture high resolution signal. From brain regions of interest and as I mentioned earlier both patients and healthy individuals were asked to produce symptomatic for patients, symptomatic speech which consisted of preselected sentences and the sentences were loaded with vowels or voiceless consonants to help.
Evoking their voice symptoms. In addition to that, they were also asked to whisper the same sentences. And this presented an asymptomatic task as whispering is normal. And usually patients whisper instead of speaking out because in this way they overcome their symptoms. And the third,
[00:05:59] Dr. Sarah Camargos: [00:06:00] Now, a sensory trick for them. The whispering,
[00:06:06] Prof. Kristina Simonyan: Right. I'm not really sure if it falls into the sensory trick category. But there is somewhat different physiology, laryngeal vocal fold physiology involved with whispering versus speaking loudly, so we still don't understand quite well the difference between whispering and speaking, but this study and more studies are definitely needed to understand and How similar, not exactly the same, but similar behavior leads to completely different symptomatology.
[00:06:48] Dr. Sarah Camargos: Yes.
[00:06:49] Prof. Kristina Simonyan: So as whispering was a task related to their speaking, we also decided to look into an asymptomatic task [00:07:00] that is not related to their moderate symptoms, and for that we chose writing of the same sentences. And this went along this line of previous research that we conducted looking into dopaminergic abnormalities during asymptomatic speaking and asymptomatic writing in these patients and we did find at that time somewhat normal dopamine release that was also associated with abnormal reaction time in these patients. So three tasks. Symptomatic speaking, asymptomatic whispering, and asymptomatic writing. That were completed by all patients and healthy volunteers while we were recording their EEG signal.
So, in terms of data analysis we looked into different aspects and we [00:08:00] looked into, you know, when we conducted EEG signal pre processing, this was important to remove especially movements associated with orofacial motion that may affect or most likely to affect EEG signal.
And next we moved into analysis of spectral topography of cortical oscillation source localization because it was also important for us to understand not only which band. May or may not be involved, but also where this localizes in the brain. And then the final portion of the study data analysis included effective connectivity of statistically significant source activity to address the question of how these regions interact and how they may or may not influence activity going on within the network.
[00:08:58] Dr. Sarah Camargos: It was very interesting [00:09:00] and I saw that you, the patient was asked to write 50 times and to say the sentences, three sentences, a hundred times, each one of them, right?
[00:09:15] Prof. Kristina Simonyan: Oh, correct. So there were more sentences. The figure in the paper just shows an example of three sentences. Yes, there were more sentences. But yes, that's correct. Subjects repeated the sentences, they were cued with the visual stimulus and with auditory stimulus and they had four seconds to speak or to whisper and they had about five seconds to write the sentence.
And we did collect a hundred trials of each speaking and whispering, and 50 trials of writing. There was a lot of speaking and writing involved.
[00:09:57] Dr. Sarah Camargos: And a lot of data to analyze, [00:10:00] actually.
[00:10:01] Prof. Kristina Simonyan: Correct. Which has an advantage because we use a high density EEG system, but we also collected many data points, so that power the study appropriately, in addition, we had relatively large number of participants and this was an advantage over the previous studies.
First of all, there were not many studies that looked into temporal dynamics or temporal signature of speech or any other behaviors in patients with laryngeal dystonia. Two studies that have been reported to date both studies focus on analysis and data collection of a simple vowel.
So as we know from, you know clinical practice and from previous studies and patient's reports and so on vowels are not very symptomatic in this patient. So you really need to evoke [00:11:00] speech. So speech related voice breaks and effort and breathiness depending on the form of the dystonia.
[00:11:08] Dr. Sarah Camargos: Yes, and I love the quantification of the voice. You count the number of breaks in each sentence, right? This is a lot of work.
[00:11:22] Prof. Kristina Simonyan: Correct. So we conducted a quantitative analysis of voice samples. We quantified the number of breaks in each sentence and also on the scale, zero to a hundred. We quantify the additional symptoms as harshness or effort in abductor form of laryngeal dystonia and breathness in abductor form.
And then an average score of symptoms has been computed for each subject for each patient and then this score has been [00:12:00] used for running correlations between significant changes in the EEG signal to understand where we can or cannot link abnormalities in EEG signal to clinical symptomatology.
[00:12:15] Dr. Sarah Camargos: Yes, amazing. Really, really well done. methodology is amazing. And what did you find in your study? What did you think is most important in relation to your study?
[00:12:29] Prof. Kristina Simonyan: It was quite interesting to see that speech, symptomatic speech did show abnormalities in a very specific band, and this was gamma band, and in contrast to that, whispering did not show any significant abnormalities, and as a reminder, whisper is quite normal behavior there are. Using whispering voice because they don't have symptoms.
And so it was [00:13:00] interesting to find the confirmation in EEG signal, where we observed that whispering is close to the normal levels and the signal did not have, did not show in any band any significant differences from healthy controls. So on the other hand the asymptomatic task speaking showed more distributed pattern of abnormalities though it was localized to the gamma band, but source analysis further showed that there are a number of regions that are involved in these alterations.
Interestingly, these were mostly prefrontal and parietal and somatosensory areas. There was not much in the motor cortex which can also indirectly tell us that motions during speaking probably were [00:14:00] filtered out properly and were not the contributing factor for which reason we potentially did not see motocultural abnormalities.
But on the other hand our findings have shown and confirmed recent studies from my group and others that have highlighted the role of prefrontal parietal network and connectivity being abnormal in dystonia and specifically in TASK specific dystonias. So with this study, we again showed that this prefrontal parietal axis is abnormal.
In addition, we showed that actually parietal areas influence the network, and how prefrontal and parietal regions interact, and with the directionality of abnormalities. And these findings were quite interesting because we [00:15:00] can start thinking about using these regions as potential biomarkers and potentially thinking about developing diagnostic and therapeutic opportunities targeting these regions.
Another interesting finding was that writing showed abnormality in beta band, and as I mentioned earlier, we did find abnormalities in dopaminergic system during writing though it's not affected asymptomatic motor task, but it also aligns with the thinking that there is more global motor deficiency in these patients, not only with laryngeal dystonia, but also with other forms of dystonia and some asymptomatic tasks may in fact be affected as subclinical features of the [00:16:00] disorder.
[00:16:01] Dr. Sarah Camargos: Wow, and speaking of therapeutic targets do you think there is some room for modulation?
[00:16:11] Prof. Kristina Simonyan: I think so. I think we can start thinking about potentially modulating these regions moving further away from basal ganglia and motor cortex, which were targets for decades but in terms of understanding pathophysiology and probing different types of neurostimulation. Unfortunately, literature shows that these were likely not very good targets as.
Well, for example, several studies using TMS or TDCS did not really report short or [00:17:00] long term symptom improvement. And I'm not talking about laryngeal dystonia, I'm talking about dystonias in general. There are very few studies that show transient improvement in symptoms after stimulating motor cortex.
So I think studies like this and others from other groups have laid foundation for starting to explore other targets outside of the motor cortex and non invasive neurostimulation would probably be the first step to define which regions may or may not show the most therapeutic effect and how this therapeutic effect can be maintained over these months.
Hopefully years. And long term, it would be great to start thinking about [00:18:00] some invasive neurosimulation, like DBS targets and so on. As you know, GPI is the most commonly used DBS target for dystonia patient, but there is no reason to believe that this has to be the only target for this disorder.
So more precise studies are needed for further defining these regions and expanding on research in this paper and from other groups and thinking about exploring other targets than GPI, which works great for some patients, it's very small percentage of patients with dystonia who are candidates for DBS.
[00:18:48] Dr. Sarah Camargos: And what do you think about your next steps in your search?
[00:18:54] Prof. Kristina Simonyan: We have, we have been exploring a number of different projects. We are [00:19:00] looking more into understanding the pathophysiology of laryngeal dystonia and in comparison with other forms of dystonia and other related neurological disorders such as tremor, for example we have been working on genetics of laryngeal dystonia and other forms of test specific dystonia one relatively large area of research is the development of objective diagnostic platforms and we have been successful in developing so called DystoniaNet .
Deep learning algorithm and platform that has shown 98 percent accuracy in identifying dystonia based on a single MRI . So in the past few years, we have been working on further testing DystoniaNet [00:20:00] that is currently in clinical trial. And hopefully can be transitioned into clinical use.
And we have been also working on developing extensions of the platform. Recently published paper showed that dystonia BotoxNet, as we call it, is quite accurate in predicting the response before treatment is administered. So and we have a number of other projects for extending the use of the DystoniaNet.
And our most recent addition to our research portfolio is developing of some novel, therapeutic tools and the therapeutic opportunities for patients with laryngeal dystonia and other forms of dystonia. We have developed a brain computer interface where we use the task specificity of laryngeal [00:21:00] dystonia and hopefully we can extend this as well to other forms of task specific dystonias where patients undergo a week long training to relearn their speaking based on their close to normal whispering.
So we have some preliminary results showing that it's actually working. Some patients have been giving quite positive feedback and we're reporting that they use similar strategies at home. So hopefully at the end of this clinical trial, we can have some specific and you know, positive outcomes for this new tool for rehabilitation of these patients.
[00:21:49] Dr. Sarah Camargos: Wow this is amazing.
I'm waiting eagerly to see your outcomes of this study. Professor, thank you very much your [00:22:00] time us. Your work is amazing. Thank you.
[00:22:04] Prof. Kristina Simonyan: Thank you very much. Thanks for having me.