View complete transcript
This idea was born out of a paper that Professor Edwards published last year in the Movement Disorders Journal with Anna Sadnicka, titled Between Nothing and Everything, Phenomenology in Movement Disorders, that explains the limitations of a phenotype equals disease approach to movement disorders. And we thought it could be interesting to have a conversation around this with experts in the field who have thought deeply about this and written about it as well.
So thank you both for joining us today.
[00:00:58] Prof. Mark Edwards: It's a pleasure. Thanks. Thanks for [00:01:00] having us.
[00:01:01] Dr. Sarah Lidstone: All right, so we can just get started. Perhaps it might be a good way to open the discussion by Professor Edwards. Did you want to share a little bit with us about the background of this paper and, what led to its development in terms of the ideas that you and Anna were thinking about and discussing at the time?
[00:01:18] Prof. Mark Edwards: Thanks very much, Sarah. So this was a paper that Anna Sadnicka and I have been thinking about for a long time. And I suppose it came from two places. One was that there had been a number of publications, I think particularly publications by Alberto Espay, but also many others, which had sort of highlighted the problem of classification of disease via phenotype, particularly when you're trying to develop biologically sensible treatments for diseases.
And it highlighted the issues that many of us are aware about, that disease is classified on the basis of clinical phenomenology, which was always the way we did things, when it comes down to genetics [00:02:00] and biology. You often see blurring of disease categories, so that you might see similar processes going on in phenotypically different disorders, and we can see that particularly in genetic disorders, where once a gene is described for something, you find that lots of different phenotypes can be related to the same genetic problem.
And that's a particular issue when you're trying to develop disease modifying treatments. And Alberto and others have written extensively about that in developing treatments for Parkinson's disease and other neurodegenerative disorders. So there was that going on, which is a sort of in a sense, anti phenotype way of thinking about things.
And then Anna and I were thinking about things from the other way around, that even regardless of what's going on at a disease level, that there might be Specific mechanisms that cause specific phenotypes. And we were thinking particularly about movement phenotypes that might be a bit agnostic or a bit unrelated [00:03:00] to diseases to auto a specific disease process.
So to put that another way, there are specific, but quite generic mechanisms. that produce specific kind of phenotypes, or specific kind of ways that movement go wrong. And that we need to think about those maybe a bit separately from all of the processes that might be going on, say, in a neurodegenerative disease.
And that if you don't do that, you run into problems, because if you just interrogate the system by doing, say, an fMRI study, or an electrophysiological study, or something like that. You might pick up abnormalities, but then there's a temptation to say that those abnormalities that you picked up must be mechanistically related to the movement problem you're seeing in front of you.
And if you start to think about that, that's not logical. Because those things that you picked up might be mechanistically related but they might just have something to do with the disease and not [00:04:00] really be part of the mechanism of the actual movement problem you're seeing. So that led us to talk about this importance of having a phenotype level understanding of what's going on as well as a disease level understanding.
And you should do both of these things in parallel because they might have different therapeutic implications. And different ways in which you develop treatment, so that's where it all I suppose came from that was the basis of the paper.
[00:04:26] Dr. Sarah Lidstone: And I love what you wrote about how the nervous system can break in a limited number of ways. And that analogy, I think, or that visualization really stuck with me. And the way you're describing it now, it's almost like you're saying that there's these kind of networks or these, these states that a disease process might pass through, right?
As it's progressing, according to its own biological parameters. So the example in the paper you also wrote was Huntington disease and how there's a great example of how multiple different phenotypes exist over the spectrum of somebody having Huntington's so [00:05:00] might begin first as chorea and then later on develop into a more of a hypokinetic syndrome, a parkInsonism.
[00:05:06] Prof. Mark Edwards: Exactly and I think that we were thinking about it again at a very maybe a rather sort of superficial level about the idea that we have a sort of you could think about a sort of sensory motor control system that's involved in the production of movement and that, that system can only fail in a set number of ways or maybe out of fail only in a set number of ways.
And we were thinking about the fact that in a system like that, when you perturb it, It's trying to get to another stable state, so it kind of resolves into a particular state, which we might then see as bradykinesia, or might see as chorea. And maybe there are actually much finer gradations within that if we had the ability to analyze the movement difficulty at a higher dimension than we can just by our own kind of clinical skills.
And that that's maybe an opportunity, [00:06:00] because if you could understand where the system is, maybe you could shift it somewhere else, even if you weren't doing anything to the underlying disease process. And we gave the example of maybe DBS being a technology which does that. So we were talking about in the paper about tremor, that if you have tremor due to Parkinson's, MS, whatever it might be, usually if you put a lead in the right place in the thalamus and stimulate, you usually get some resolution of the tremor, often quite profound resolution of the tremor.
So that's an example of a disease agnostic phenotype level treatment, in theory and maybe if we understood more about the processes that generate other kinds of abnormal movement, we might be able to do that more effectively for other conditions.
[00:06:43] Prof. Mark Hallett: Right, let me sort of jump in here in relation to the first concept that you brought up, which I think is an extremely important one. And one that I think is now getting to be much more prominent all through the field of movement disorders. [00:07:00] And that is to think of any individual disease in two ways, one as the phenotype and the other as the etiology.
And there isn't a one to one match between them as you pointed out, and each of them can have its own type of treatment that is relevant for it. And it is generally the case that if you can treat the etiology, the underlying problem, that's the best thing to do. But often that's hard or it may be tricky and there may be phenotypic relevant treatments that are also there.
And that's certainly done in functional movement disorders where one has occupational therapy and physical therapy direct to the specific phenotype and that can be helpful for the patient too. Although in the long run. One might need to have the etiological [00:08:00] treatment, but I think the idea of thinking about any patient in two ways, the phenotype and the etiology, is absolutely critical.
And I think that that has been the way that we now sort of formally think about dystonia, formally think about tremor. These are the sort of MDS guidelines now in terms of thinking about it and to move functional disorders into the same way of thinking is really very important, I think,
[00:08:31] Prof. Mark Edwards: Obviously I agree and I think that one of the things that Anna and I were trying to think about in the paper is that, this applies to also papers definitely that I've been involved in the past, so things looking at pathophysiology of movement disorder, say dystonia that there's a real temptation, I think, to blur that parallel track of etiology and phenotype mechanism when it comes to talking about research findings.
So, you know, [00:09:00] you find an electrophysiological abnormality in dystonia, so dystonia is a good example. You find something around reduced intracortical inhibition or changes in reciprocal inhibition, and you take that finding, and then you say something like dystonia so the phenotype, is a problem of lack of inhibition.
And that's something which is tempting to do, but it's also potentially not true. Because you've picked up something correctly in terms of an abnormality in inhibition, but whether or not that's mechanistically related to the expression of the phenotype dystonia is a whole other question.
Maybe it's something relating to a disease process which is important to know about, but wouldn't be the way you would try and mechanistically change the phenotype, if you see what I mean. So it needs people to keep that in mind, I think, when thinking about and reporting and acting on the results of physiological investigations in [00:10:00] people with movement disorders and other conditions.
[00:10:02] Prof. Mark Hallett: Right? So I would basically agree and point out that I think what you're talking about is sort of a third way of thinking about a disease. You and I both started out by talking about phenotype and etiology, but actually in the middle is the pathophysiology, and that's really what you were talking about there.
And we know not as much as we should yet , and I guess you're pointing that out about how you get from etiology to phenotype. It is via the pathophysiology and which pathophysiological processes are going to be important. In any particular translation. We still need a lot of work in sorting that all out.
Just in terms of the general concept of how one goes about thinking about disease the pathophysiological axis is also an important one. That would be a third axis. So you have the etiology, the pathophysiology, and [00:11:00] the phenotype. We're not quite there yet, completely, in terms of understanding that pathophysiological translation, so that doesn't appear very formally yet in much of our thinking.
You don't know whether your process is going to be relevant or not. Yet it will eventually work out. I'm optimistic that we will know eventually, but that isn't anything that we're completely sure about at the present time,
[00:11:31] Dr. Sarah Lidstone: And if we're thinking about risks of a reductionist approach to different types of phenotypes, which you included in your paper mark of the, you know, six, seven different cardinal types of movement disorders, we are also overlooking very relevant non motor disorders. symptoms that also would contribute to phenotype as well.
And that's absolutely the case in functional movement disorder, which we've written about where we're kind of very interested in the role, for example, of the [00:12:00] autonomic nervous system and how that can show up in an examination setting or psychiatric symptoms or psychological factors even. In Parkinson's, constipation, REM sleep behavior, and sort of these other types of other phenotypes and by focusing too much on the movement, are we also missing the forest for the trees and other different ways of defining what phenotype might be.
[00:12:21] Prof. Mark Edwards: Yeah, I mean, taking the example of a functional movement disorder, once you start scratching the surface, there's so much complexity there. And I think that's one of the dangers, I suppose, of the change that's happened in talking about functional neurological disorder and functional movement disorder and encouraging neurologists in particular to become more comfortable with making that diagnosis, with explaining it to patients, with trying to be involved in treatment. All of which I agree with, that there's a danger sometimes of missing the complexity. And I think that you see that often as soon as something gets a disorder label. People just say, oh, that's [00:13:00] FMD, and like that's job done. But there's so many different levels at which you can then try to understand what's keeping symptoms going, the range of symptoms that are present. You can think of multiple different diagnoses, often overlapping, as you're saying, different systems being autonomic system of our intraception so, all of this, it's about trying to keep that all in mind at the same time.
And that's a really tricky thing to do. And I think we need to do that across patients that we're seeing, not just people with functional movement disorders, but across movement disorders generally. That you're trying to kind of picture the person in front of you with all of these multiple different levels at the same time.
Because that's the best way of trying to understand what's the most important set of interventions to do at a particular time. It's going to help you with the sort of sub classification of patients and thinking about sort of different patient groups where there may be different [00:14:00] etiologies and pathophysiologies going on and different treatments certainly might be needed.
[00:14:05] Prof. Mark Hallett: Right, I would agree too that functional disorders are really complex and all those other features that you mentioned Sarah, phenomic and psychiatric and psychological factors, they're all very important. And I suppose we should turn the tables on you as the moderator and ask you to discuss your own paper in this regard where you have tried to cluster patients on the basis of both their motoric and their other features, because I think that's an important way of thinking about these patients too.
[00:14:42] Dr. Sarah Lidstone: That's totally fair. It's a hot topic and it's my turn for the hot seat, but we were thinking about this in this very similar way that Mark has described with Anna at the beginning of this podcast. And this was born out of us, me with different disciplines, observing a patient together at once [00:15:00] forces the patient to be at the center.
And it forces a different way of analyzing. the presentation and the problems. And we also, knew anecdotally, but we didn't have data at the time that supported this is that the motoric phenotype absolutely changes over time in these patients. So, six months later they come back and they have weakness when they initially presented with tremor or episodic attacks.
And so we were recognizing that this was happening and it forced us to say, well, what are we missing? Is there anything that we can see in this deeper level that itself might be amenable to pattern recognition in the same way that we pattern recognize tremor and dystonia or gait disorders, for example?
So we tried to do a data driven approach as best as we could, again, very subjectively relying only on our clinical experience seeing these patients, but identify what kept coming up over and over again, and then running [00:16:00] a model statistically to see if anything precipitated out and what we found actually was that what was more important than the sub individual phenomenologies in functional movement disorder was whether or not the motor symptoms were occurring in an episodic or a constant fashion and we're still trying to define what this separation actually is because it's a combination of the patient's own perception of the constancy of their symptoms.
There were some patients as you know who very much say my gait is abnormal all the time. It's never normal or the weakness is present all of the time. There's never any periods where it's improved or it's better compared to patients with very distinct attacks of symptoms that are truly episodic.
And they can say to you, there's many times my tremor isn't present, or I don't have these dystonic attacks or these jerks, for example. And actually, this was done very pragmatically because those two different subpopulations of patients respond to different types of treatment. [00:17:00] So, it's very hard, for example, for a physiotherapist to treat episodic attacks because unless the patient's actually having them in front of them in the gym or the treatment center, they're not going to be able to do anything therapeutically with them.
And so this is what we found. We found there are these two clusters of episodic patients that tend to be more hyperkinetic. They have jerks, they have mixed jerks, they have tremor, but they also have very visible signs of somatic anxiety, hyperarousal, and history of trauma. And the more constant patients, look and feel very different.
They have weakness, gait disorders, fixed dystonia. They tend to have avoidance, both emotional and activity avoidance, and they have higher counts of symptoms. And so, there's an example of where the motoric phenotype is a place to start, but it incompletely describes what we're actually seeing. And, what's really fascinating just to kind of, to close this long discussion from my part is what's happened now is if there's a patient who presents with a gait disorder, for example, a constant gait disorder [00:18:00] that looks a certain way. So generally, this is a bit of an effortful gait. It's very slow. I wouldn't describe it as a kind of more kind of classic hyperkinetic gait disorder. That's often associated with depression. And so now we use the motoric phenotype to delve deeper into the more psychological features and actually can use that diagnostically to help with treatment.
So it goes both directions I guess is what we're learning.
[00:18:25] Prof. Mark Hallett: Right. So actually that goes back to the theme that we were talking about earlier if one thinks about the phenotype and the etiology one of the reasons for doing good phenotyping is because it may give you hints about the etiology. So patient comes into your office, you see them looking in a particular way.
It isn't completely irrelevant what they look like in terms of the etiology. There can be some hints in terms of what the etiology might be, which will help you develop [00:19:00] the best therapeutic program, which is basically what you were saying. So, I think that it is important to get good phenotyping as well as good etiology and there will be some links there that will be helpful and in the end our job is to help the patient, so we need to think about therapies in any way that we can.
[00:19:22] Prof. Mark Edwards: I was just thinking about this wider sort of phenotype. As you say, you want to broaden out, from the motor disorder through to non motor psychological and psychiatric symptoms and so on. And I also have been really struck by this clustering of sort of neurodevelopmental differences which may not really meet threshold for anything quite now but they're also really quite common in a group of patients and I'm talking about people who have traits which are seen in people with autism talking about people with hypermobility, so [00:20:00] joint hypermobility, not necessarily of something of a very severe type, but they are, clearly outside the typical normal range of joint flexibility, this idea of autonomic instabilities of people with easy fainting, et cetera.
Often these are things which didn't cause the person, any problems in particular, but they're there as a cluster of background factors which seem to make people more vulnerable than to developing a functional disorder later on. And then once the functional disorder is developed, These things often seem to flare up a little bit or add in additional symptoms.
And I suppose it's just another example of how there's this richness and complexity there that really wide phenotypic characterization can help with. But as you said, Sarah, often that needs multiple people involved. So people from other disciplines who have different perspectives and can give you insight into things that you wouldn't really have seen otherwise.
So there's still this beauty [00:21:00] and I think often neurologists are attracted into neurology by phenotype. There's this peculiar, even though neurological disorders are incredibly unpleasant, there's this peculiar beauty in neurological signs and it's to say that that's still relevant, and that those clinical skills and developing clinical skills is really relevant from a science point of view and from the point of view of trying to help patients day to day.
[00:21:26] Dr. Sarah Lidstone: Well it truly is an art and a science isn't it?
[00:21:29] Prof. Mark Hallett: True. Actually, if we have a little more time to go back to one of the other themes that we talked about earlier in terms of the different phenotypic manifestations, I must say that one of the things that I'm most commonly asked about by patients and by other doctors is why does a particular patient take a particular phenotype or have a particular phenotype? Given what we're been saying that there may be some non specificity in it, in [00:22:00] relation to the etiology, why do some people take a particular phenotype? And that's a very interesting question and I think that my sort of stock answer to that has been that people manifest what they know or what they have seen. So that the brain sort of learns a particular type of behavior. And then that can be brought out in this particular context. But I think there's a lot more to learn here.
[00:22:29] Dr. Sarah Lidstone: Yeah, and that speaks into, you know, Mark, you've written about predictive processing quite a lot and extensively around that. I think that that's one element of it. I also think trauma is another. If you're thinking about kind of the dissociative, especially when it comes to FND, at least the dissociative presentations feel distinct to me.
I think that maybe speaks to a deeper autonomic nervous system mechanism that is, that is happening there. But there is still so much more work to be done.
So I'm loving this discussion. I wish [00:23:00] we could go on for longer. But I would like to ask about the future of phenomenology and phenotyping. AI is emerging as being such a powerful tool, and medicine, I think, for a lot of very good reasons, are very reluctant to be using AI routinely, but do you think AI, or other types of technologies, are there different ways, or, in terms of complex data sets, and looking for patterns, do you think there might be a role in the future for phenotyping using AI?
[00:23:31] Prof. Mark Edwards: I certainly think that whether or not it's AI, but better kinematic measurement of the actual movement problems that we see, is going to give us much more what happens already and will continue to give us much more than we can see just from the human eye, even from someone who's seen lots and lots of movement disorders before.
And I think that, you think about our definitions of movement disorders, their collections of words descriptive words and there's often [00:24:00] this sort of debate about, is it this, is it that? And I think that we may well get more insights into that from better use of technology in phenotyping to give us a finer grain to look at things.
[00:24:11] Prof. Mark Hallett: Yeah, I think that going forward AI is going to be moving into medicine in general more and more all the time. And I've been giving advice to younger neurologists now that the thing to learn is AI because that's where the new advances are often going to be. The data that we accumulate on patients is increasing massively, and the number of etiological possibilities is increasing massively, and it's going to be hard to keep everything in mind at any one time.
I mean, obviously this is true in the disorders that are overtly genetic where there are hundreds of genes that can cause particular phenotypes that can be quite [00:25:00] similar. So I think that AI is going to play a major role, and all the data needs to be put in.
What's going to be relevant or not, we can't be sure about so I would certainly do good phenotyping that would be part of this.
[00:25:16] Dr. Sarah Lidstone: Well, thank you very much, Mark and Mark, for joining me for this conversation. It's extremely thought provoking. I propose we periodically revisit this topic as we learn more and more. But thank you both so much for your time.
[00:25:29] Prof. Mark Edwards: Thank you very much.
[00:25:30] Prof. Mark Hallett: You're very welcome.
[00:26:00]