Rethinking Parkinson’s: The Somato-Cognitive Action Network (SCAN) hypothesis

What if it is a brain wide action network that integrates movement with arousal, autonomic physiology, and motivation? I am Michele Matarazzo the editor of the podcast, and today we will discuss a nature paper proposing exactly that. The article is titled Parkinson's Disease as a Somato Cognitive Action Network Disorder. Our guest is Professor Heshang Liu from the Changping Laboratory in Beijing, China. The last and corresponding author of the paper. Heshang, welcome to the podcast.

Prof. Hesheng Liu: Hi nice to be here. Thank you for the introduction.

Dr. Michele Matarazzo: Okay. So I would [00:01:00] like to start with the concept itself because it sounds elegant, but it can also sound a bit abstract. Many of us first learned the neuroanatomy through the classic penfield homonculus and until recent that was the mainstay. But recently a work published in Nature in, 2023, I believe this challenged the traditional somatotopic view of the motor cortex, and introduced the concept of the Somato cognitive Action Network or SCAN. Can you explain us what SCAN is and why might it be relevant for Parkinson's disease?

Prof. Hesheng Liu: Yes. SCAN is a newly discovered network that sits in between the effector motor regions. So in the past when Dr. Penfield identified the homunculus using cortico stimulation. Of course when they stimulate the effector area, they will see the response of the arms, the the legs or the tongue.

But there are regions sitting in [00:02:00] between those effect regions. But if you stimulate those regions, you don't see any obvious movement of these effectors. So it was missed 80 years ago when Dr. Penfield stimulated cortex. But using the advanced neuroimaging technology, now you can clearly see there are some isolated regions sitting there in the motor region.

So in the 2023 Nature paper when Gordon et al., identify those regions. They propose that these regions are related to the cognitive function, not not just the motor function. So they give the name as the Somato Cognitive Action Network. I think it's a very elegant finding

Dr. Michele Matarazzo: And just for everybody to understand this the current understanding is that this area integrates what happens from a motor and a cognitive standpoint, right?

Prof. Hesheng Liu: Yeah I think that's partially correct. It probably transmit some cognitive information to the motor circuit, to the [00:03:00] conventional motor circuit, or it has even more related functions for example to autonomic function. As well as some for example, the blood prep, low blood pressure as you will observe in the Parkinson's patients and the digest problems and sleep problems.

So they are probably all related to this SCAN network.

Dr. Michele Matarazzo: So now before getting to the paper itself and to your study, why did you think that this specific network was relevant to Parkinson's?

Prof. Hesheng Liu: Very good question. Back in 2016, we have been trying to identify the functional circuits responsible for Parkinson's. Which has been a the dream of a lot of researchers. Of course, we want to know, for this important functional disorder, what circuit is damaged or affected by the disease?

So we have been using functional imaging technology, trying to see this functional circuits. But to be honest, we didn't see anything new [00:04:00] until we saw this 2023 Nature Paper. Then it gave us a lot of inspiration because we have known that Parkinson's is not a very simple movement disorder, even though it's always named or defined as a movement disorder.

But you know that if there's a fire or it's an earthquake, the patients can immediately get their movement function back. They can run away from the scene. The cognitive input has a lot to do with their movement functions. So that's why I thought that Parkinson's probably, it's a broader question about the cognitive function and the motor function.

And when we saw the SCAN paper in 2023 that immediately resonates with a lot of questions. So now everything started to make sense. We immediately realized that this spot might be a broken hub in the Parkinson's patient's brain, which disrupted [00:05:00] the connection between the cognitive area and the motor area.

So then we started with this hypothesis. And then study those patients and then we saw, oh, yeah, that, that's exactly true.

Dr. Michele Matarazzo: Great. And now I want to make maybe the last question before we get deeper into the paper is that, do you think functional imaging is or specifically fMRI, is the best way of studying that? Or how does it compare with metabolic imaging, let's say for example, a FDG-PET?

Prof. Hesheng Liu: So there are actually different ways to study brain networks or the dysfunction areas of the patient's brain. Functional MRI has a lot of advantages over FDG PET in terms of the spatial resolution. And there are other approaches, for example, electrophysiology to study the neuro circuits in the past that, people have been recording neurophysiology signal during DBS stimulation.

So that's another way to infer the connectivity [00:06:00] between brain areas. However, I still believe functional MRI has the advantage of giving you a whole picture of the entire brain. And now especially with the MRI compatible DBS you can even study how the brain circuits respond to simulation.

Dr. Michele Matarazzo: Perfect. Thank you for this very nice introduction and overview of the original problem. Now let's get to the paper. The paper is quite dense in part also because it's a substantial translational effort. And you assembled a very large multimodal, multi interventional data set with I believe it's 863 participants across several cohorts, which is an amazing number. Now before we get to the interventions, what were the first key findings when you examine the resting state functional connectivity in Parkinson's disease compare with healthy controls? And also in general, when you analyze it what happens, l ooking at this SCAN specifically?

Prof. Hesheng Liu: So the key finding is the important [00:07:00] subcortical structures that we know are related to Parkinson's. For example, the STN/GPi and substantia nigra. These nuclei we all know that these structures are important for Parkinson's and we want to know how these nuclei connect to the cortical areas.

When we look at the resting state MRI, we saw that all these structures show a very strong abnormal connectivity to the SCAN area in the cortex. Compared to the healthy subjects Parkinson's showed a abnormal hyperconnectivity, that means overly connected. So that's the very surprising observation.

Dr. Michele Matarazzo: So basically you're suggesting that all these therapeutic targets that we've been using for years, the STN, the GPI, the thalamus, and so on they were modulating the SCAN all along. We just didn't know it right.

Prof. Hesheng Liu: Yes. In the past we [00:08:00] probably assumed that these subcortical structures are connected to the motor region. So we took for granted that these regions are part of the motor circuit. We assume they are connected to the M1, but we never systematically explore where these subcortical structures connect and project to. We guess that maybe project to the hand area, foot area, or tongue area.

But it turns out none of them is true. All project to the SCAN area.

Dr. Michele Matarazzo: And now if I may ask, maybe there might be some of our listeners that are not expert in neuroimaging or in neurophysiology. Can you clarify what you mean by hyperconnectivity?

Prof. Hesheng Liu: That means that the signal between the cortico SCAN area and the signal in those subcortical structures they are super correlated. Their correlation is very high, which is abnormal because they should be more independent. But something happened that caused some super synchronization between these [00:09:00] deep nuclei and the cortical area.

Dr. Michele Matarazzo: Perfect. Now, another important thing that you mention in the paper, is that there are many disorders that involve both motor and cognitive changes. Now you addressed this directly by including other neurological cohorts and conditions such as essential tremor and other conditions. What did this comparison show?

Prof. Hesheng Liu: We had a question whether the hyperconnectivity between SCAN and these subcortical structures. It's a common feature for all movement disorders. For example, dystonia, essential tremor, and ALS. With these questions in mind, we study these patients group and compare them with the healthy participants.

Surprisingly, the hyperconnectivity we observed in the Parkinson patients was not observed in ALS patients. Not in the essential tremor patients. Not in the dystonia patients. So that suggests that this [00:10:00] hyperconnectivity may have some specificity for Parkinson's. But I'm not saying that only Parkinson's can bring this hyperconnectivity.

But at least among these several movement disorders, we studied Parkinson's is the only disorder that showed this very strong hyperconnectivity.

Dr. Michele Matarazzo: And would you expect to find the same hyperconnectivity with other types of Parkinsonism?

Prof. Hesheng Liu: That's a good question. We are actively exploring with other subtypes of Parkinson syndrome, MSA or PSP. So that's an ongoing project.

Dr. Michele Matarazzo: I look forward to the results. Now let's let's move to the treatment effects you did not stop to at defining an imaging phenotype that you examined how it changes with therapy. Across levodopa, multiple neuromodulation approaches, the central messages that, effective treatments reduce this SCAN to subcortex hyperconnectivity. What were the key interventions you studied, [00:11:00] and what gave you confidence that this pattern is robust rather than a one-off finding?

Prof. Hesheng Liu: Yeah, I think all these treatments converge onto this hyperconnectivity, which is remarkable. We first study DBS because we collected a huge amount of data from 14 patients who underwent DBS treatment. We scanned them for many hours in the MRI. For example, for each patient before the DBS implantation, we collected the baseline data for for one day.

And then after the implantation, we scan the patient for two days continuously. Of course, only during the daytime. When the patients are feeling comfortable, we will put the patient in the scanner and then stimulate them using different parameters. For example, sometimes we stimulate them with 60 hertz stimulation.

Sometimes 90 hertz, sometimes 130 hertz. Sometimes we turn it on and turn it off. Sometimes we stimulate them continuously. For [00:12:00] each patient we schedule them for a two day scannin session. And we repeat this five times in one year. So that means we scan each patient for about 10 days after implantation.

So we know exactly how DBS modulates the functional activity in each patient's brain. We saw almost the identical of highly similar phenomena across all individuals. So that's why we have so much confidence that when we stimulate the STN, we saw the reduction of these connectivity between SCAN and STN.

So that's the the key information that a very well accepted treatment. It looks like it took effect by reducing this hyperconnectivity. And later we also try to see if the medication, which is the most common treatment for all the patients. We saw the same thing, that it still reduce the hyperconnectivity.

Dr. Michele Matarazzo: That's very interesting. And another [00:13:00] thing to discuss is the importance of the beta band, which is an a known electrophysiological marker for Parkinson's or for Parkinsonism? How does it relate to this hyperconnectivity?

Prof. Hesheng Liu: That's a very good question. We have recorded EEG signal from the DBS electrodes. So we know that when we stimulate the brain and the beta band activity changes. And there are some previous studies showing a hyper synchronization between the cortex and the subcortical regions in the beta band.

We believe that this desynchronization of the beta band is probably related to the reduced connectivity we saw in fMRI. But we need more direct evidence.

Dr. Michele Matarazzo: Okay. Now, you gave even a step further. So we discussed already the study of this hyperconnectivity as a biomarker. Then how it changes with the treatment. But you went a step further doing TMS trial. So you randomized [00:14:00] 36 patients with Parkinson's to intermittent theta burst stimulation over a personalized SCAN target or over the effector specific motor target. And apparently targeting SCAN appear to produce a substantially greater motor benefit.

Can you walk us through what you stimulated and how the target was individualized and what changed from the clinical standpoint?

Prof. Hesheng Liu: Yeah. So, we study 36 patients. We randomized that into two groups. So one group we run the functional MRI scan, and then we using a network approach. So we have been developing functional network population algorithms in the past decade. So now we have a sophisticated way to map out those regions.

So for each patient we can identify their SCAN area and the effector area. So for the SCAN treatment group, we will use TMS to stimulate [00:15:00] the middle piece of the SCAN. So if you look at the 2023 nature paper the original SCAN paper there are three spots in the precentral gyrus.

So the top one is oftentimes sitting in the sulci. So it's not very easily accessible using TMS but the middle one is relatively easy to be accessible by TMS. So we chose that one. And for the control group, we stimulate the effector area either on top of this middle SCAN spot or below the SCAN spot.

So the distance between the M1 target and the SCAN target is probably about one centimeter. But the effect is drastically different. The efficacy is about 2.5 times better when you stimulate SCAN compared to stimulating the effector M1. That tells us even in the motor area, the TMS [00:16:00] treatment should consider precise targeting.

If you can identify the correct SCAN target, you get a lot of benefit. But if you are missed by one centimeter you are off and then you are stimulating the effector motor region that efficacy is much lower.

Dr. Michele Matarazzo: And of course you also state clearly that this was a relatively small single center study, but just to make sure that the audience know this was a blinded trial, right.

Prof. Hesheng Liu: So the the patients and the staff, they did not know which patient is receiving a SCAN treatment, which one is receiving the M1?

Dr. Michele Matarazzo: Now, if you were designing the next trial with this, what would you prioritize patient selection outcomes such as gait or freezing or longer follow up? What would you study?

Prof. Hesheng Liu: So right now I have several follow up studies going on. One study is exactly like you described that we want to focus on gait and freezing. These symptoms are [00:17:00] very hard to treat compared to tremor. So these symptoms cannot be controlled by medication in a lot of patients.

DBS cannot control it, but in our patients we, even with the small sample we saw that we have very good effect on this gait problem. So now we want to select some patients with particular problems in the gait and freezing. So that's one study. And the other one is that we wanted to focus on some non-motor symptoms.

For example the sleep problems and the digestion and the low blood pressure. So we wanna see how SCAN treatment can improve those symptoms.

Dr. Michele Matarazzo: So you're trying also to target symptoms that are very difficult to manage often with the options that we currently have.

Prof. Hesheng Liu: I think that's a very strong need for this. And of course we want to do a multicenter study to corroborate our findings.

Dr. Michele Matarazzo: Perfect. Now for the focus ultrasound community and myself for one, we [00:18:00] do a lot of focus ultrasound and you also examined VIM thalamotomy and you introduced very practical concept. How close the lesion is to the patient's specific thalamic representation of SCAN. Now you studied a cohort of 10 patients.

Can you tell me what happened there?

Prof. Hesheng Liu: This is a very early study of the MRI guided focused ultrasound lesion study. So we lesion the VIM, this thalamic area in this tremor dominant patients. And we saw some improvements in some patients, and some patients are not responding very well. So after almost three years now we have a strong hypothesis that this SCAN circuit is the core circuit for Parkinson's.

We look at the original lesion targets see if there's anything we can find and we observe that if you lesion the area that's connected to the cortico SCAN, [00:19:00] then you see a very strong response in those patients. But if you are more off than the efficacy is worse.

So it looks like it's a strong correlation between the distance to the SCAN sweet spot. And the behavioral response. So I would guess is that in the future, maybe we could do a functional MRI scan for the patients before the focus ultrasound lesion, and then identify the sweet spot in the thalamus and then we lesion in that area that may improve the efficacy.

Dr. Michele Matarazzo: So the broader implication is that we may actually gradually move from purely anatomical coordinates towards circuit defined targets in a personalized manner, and still within the thalamus, but guided by the single patient functional connectome, right?

Prof. Hesheng Liu: Yes. So that's the idea, I think because thalamus is a big structure. Now if you do the lesion in a conventional way, you are doing it in a trial and error fashion. But now if you can use the functional circuit to guide you that you may see a [00:20:00] better outcome.

Dr. Michele Matarazzo: Perfect. Now let me get to a couple of important questions and maybe more challenging questions. First of all, do you think hyperconnectivity is a cause, a compensatory response, or a downstream marker of other pathophysiological changes that are happening in Parkinson's? And what type of evidence would falsify or support the idea of SCAN as a core mechanism rather than just a marker?

Prof. Hesheng Liu: So my take of this is a cause instead of a compensatory mechanism. The reason is that if it is a compensatory mechanism, then the patients gets higher connectivity. It could, have an improved motor function. So that's what compensation means.

But we don't see those. Actually we saw that if you treat the patients with DBS or other treatments the improvement of the behavioral function is correlated with the [00:21:00] reduction of the hyperconnectivity, they go in the same direction.

So instead of the opposite, right? So that's why I think it's more likely a cause. 

Dr. Michele Matarazzo: May I add another question on that, if this hyperconnectivity is the cause of what happens in Parkinson's disease and of the disease itself, maybe, don't you think that all this treatment that modify and reduce this hyperconnectivity, don't you think they should have a disease modifying effect on the disease.

Prof. Hesheng Liu: Yeah. Yes. That's a very good question. I think that we all believe that dopamine neuron loss is the cause the accumulation of this alpha synuclein . So I believe that these microscopic changes in the brain that will lead to the macroscopic changes of the functional circuit. However, reducing the connectivity itself, whether this mean that disease modifying process or not, I think it's not directly [00:22:00] proving that reducing the connectivity is modifying disease. It probably controlling the symptoms. However, stimulating the SCAN, I personally believe it's modifying the disease process.

I have some evidence for that, not in this paper, but I have several things I wanna mention so that my first patient that I treated several years ago we treat this patient repeatedly every six months we treat her for two weeks, and we saw that the disease process is modified.

The patient her Parkinson's facial expression is gone. And her medication is reduced by 60 to 70%, which is very unusual. Usually after several years, you see the patients are taking more and more medication rather than less and less, and the patient's motor function is dramatically improved. So when I first treated her several years ago, she couldn't walk [00:23:00] independently, but now she's dancing almost every day in her dancing class. And she even could travel to Europe. So you can see that it's reversed.

And then we also performed some animal studies. We saw this effects very clear in our mouse model, if you stimulate SCAN that you see that even after you stop the stimulation, the animal's behavioral function is continues to improve. That means you don't need to do continuous stimulation like DBS and you can still see the improvement of the motor function.

But DBS, when you stimulate these subcortical structures, you don't see much disease modifying effect. So even after several years of DBS treatment, if you turn off the simulator the symptoms come back almost immediately, right? So it's a very different effect we observed in the cortex stimulation and the subcortical stimulation.

Dr. Michele Matarazzo: This is very fascinating. Of course, we will [00:24:00] love to see more of data of this and all of these ideas are very fascinating. Let's see what data you'll show us in the future. I will get to maybe one of the last questions, which is, you already told me that you are planning and you're doing already other studies, specifically looking at different manifestations of Parkinson's. But if you had to guess, which symptom domains do you expect to be most tightly linked to SCAN?

Prof. Hesheng Liu: I think it's probably the gait and the autonomic functions. So that also goes very well with the original proposal of the function of SCAN in the 2023 paper. Gordon et al., proposed that this networks is probably modulating the autonomic functions.

And so I believe that's that's the right direction to go.

Dr. Michele Matarazzo: Now the last question, if a movement disorders clinician takes one message from this paper, what should it be?

Prof. Hesheng Liu: So, parkinson's in the past it was defined based on the [00:25:00] symptoms. It's a movement disorder, but now we know the functional underpinning of this disease. So now the definition has been advanced to a SCAN disorders because we know exactly what functional circuit is broken in this disease.

Dr. Michele Matarazzo: Perfect. And I think it's a very timely message because we're discussing a lot lately about switching Parkinson's disease from a clinical definition to a biological definition, and then putting a network definition in the game here makes it all much more complex, but also interesting. Now Heshang thank you very much. This paper offers a genuine conceptual reframe of Parkinson's, not simply as a set of symptoms, but as you were mentioning, as a network level disorder of action, with SCAN hyperconnectivity as a unifying signature and a potential target that could make both invasive and noninvasive neuromodulation more precise.

Thank you very much for joining me.

Prof. Hesheng Liu: My pleasure. Thank you.

 [00:26:00]