Latest / The Biological Psychiatry Podcast / White Matter Network Disruptions in Bipolar Disorder
Transcript
- Elvisha Dhamala: That's all for today's episode. Thank you to Dr. Nibulsi for joining us to share her work on white matter disruptions in bipolar disorder. For our listeners, you can find the full paper in Biological Psychiatry, and we'll include the link in our episode description. To learn more about our journals, please visit biologicalpsychiatryjournal.com. me discuss this work is Dr. Leila Nibulsi from the University of Southern California, first author of the paper in Biological Psychiatry. Dr. Nibulsi, welcome to the Biological Psychiatry podcast. Bipolar disorder disrupts mood, cognition, and daily functioning for millions of people worldwide. Decades of research have shown us that certain brain areas look different in people with the condition. But the brain doesn't work as a collection of isolated parts. It works as a network with white matter pathways acting as the cables or wires that connect brain regions into functional circuits. Understanding how these connections are altered in bipolar disorder has been difficult, largely because most studies You can find The Biological Psychiatry Podcast wherever you get your podcasts. now so you don't miss our next episode and follow us on our social media channels for highlights and updates. Thanks listening. ⁓ This is for informational purposes only and does not constitute medical or professional advice. ⁓ The expressed are those of the participants and do not necessarily reflect those of the Society of Biological Psychiatry, ⁓ Leila Nabulsi: Thank so much for having me. Elvisha Dhamala: we get into the study itself, can you tell us about Enigma? What is the Enigma Consortium? How does it work? And why is pooling brain imaging data from sites around the world so important for studying psychiatric disorders? From the Biological Psychiatry family of journals, is the Biological Psychiatry podcast, ⁓ we discuss the latest research on psychiatric disorders and their treatments. ⁓ you for joining us. I'm your host and Social Media Editor, Elvisha Dhamala. ⁓ have been too small or too variable in their methods to produce reliable answers. A new study from the Enigma Bipolar Disorder Working Group tackles this challenge. Leila Nabulsi: historically we know that neuroimaging studies were limited to small inconsistent methods, poor reproducibility. ⁓ So study would often produce conflicting results ⁓ that generalize. ⁓ So where ENIGMA kind of started. So ENIGMA stands Enhancing Neuroimaging Genetics Through Metanalysis. ⁓ It's a global founded and led in part by Professor Paul Thompson at the University of Southern Elvisha Dhamala: its family of journals, its editors ⁓ rights reserved, including those for text and data mining or use in AI systems. ⁓ Leila Nabulsi: California and it was created back in 2009 to directly address these limitations of single studies. ⁓ So the goal of Enigma is to coordinate cross-site analysis so that the data collected worldwide is analyzed in a consistent manner, reproducible way through shared analytical protocols, standardized quality control, distributed workflows. Up to today, I believe it consists of 85 working groups in total. Some of them are disorder-focused, so they span across conditions from neurological to neuropsychiatric conditions, ⁓ but there's a lot of methods focus groups these cover different imaging modalities, genetics, but also emerging AI approaches. ⁓ the way Enigma operates, it works in a modular way so that mythological advances developed in one area can be rapidly ⁓ across many disorders and data sets. ⁓ but why does this matter, right? ⁓ brain differences in psychiatry are usually subtle. ⁓ in smaller studies, ⁓ been a lot of them, obviously, but differences can really be hard to detect. So like I said, some of the study findings are often inconsistent, but only when you bring together thousands of participants across many sites internationally, we can start to improve statistical power. And that's when reliable reproducible signals can start to emerge. So we think also skills matters for generalizability of findings. So when an effect is consistent across diverse populations, age ranges, clinical settings from all around the world, then you can be a little bit more confident that it reflects the true biology of the disorder. Elvisha Dhamala: ⁓ And in this study, focused on bipolar disorder. So can you give us an overview of the disorder? What does it look like clinically? How common is it? And why is it so challenging to study and to treat? Leila Nabulsi: Yeah, so bipolar disorder is a chronic recurrent mood disorder. characterized by ⁓ alternating of mania or hypomania, which is a milder form of mania that doesn't reach the severity of a full manic episode. ⁓ Also and in some cases psychosis. There are different types of bipolar disorder. There's type one ⁓ that's defined by having at least one full manic episode, depressive episode, which ⁓ could psychotic features like hallucination delusions. ⁓ And then there's a type two bipolar disorder where we, ⁓ diagnosis, we need to see at least ⁓ hypomanic episode and one major depressive episode, but there's no full mania. It historically used to be considered a milder form of bipolar disorder, but ⁓ actually the individuals spend in major depressive episodes make it just as serious as the other ⁓ type. So bipolar disorder roughly 1 in 25 adults in the US. So it contributes substantially the global disability and carries one of the highest suicide rates among psychiatric disorders. Most neuroimaging ⁓ studies, including are done utymia, is a period of mood ⁓ between episodes. It's easier to scan patients, But also what's important is that if we scan eutamic patients, then we can capture trait-like features of the disorder and not measure acute mood state effects. So only ⁓ 30 % of patients respond to ⁓ the standard first line agent, ⁓ ⁓ has been in use since I believe the 70s. But most patients require polypharmacy. So the concurrent use of multiple medications to manage their symptoms, ⁓ but many treatment resistant. ⁓ So there's treatment specific to manage the cognitive symptoms of bipolar disorder. ⁓ many symptoms can even arise after takes antidepressants. So adds to the complexity of management of bipolar disorder. ⁓ Elvisha Dhamala: so in this study, you're measuring these white matter pathways reconstructed from diffusion MRI. What exactly are these white matter pathways and what does it mean when they're disrupted? Leila Nabulsi: the brain has two main types of tissues. ⁓ The gray matter, which contains the cell bodies. And then we have the white matter, ⁓ which lies beneath and is made up of long nerve fibers wrapped in a fatty insulated sheet that is called myelin. If we think of gray matter as the city and white matter as the motorway connecting them, ⁓ motorways, the cities ⁓ just exist in isolation, so they wouldn't communicate. ⁓ that's what white matter pathways do. They're the brain's communication cables. So they carry electrical signals between regions ⁓ the speed and reliability of these communication depends on how well myelinated these fibers are. So how do we measure them? White matter pathways obviously can be seen directly with standard brain scans. We operate at the macro scale with MRI, but one of the acquisition protocols of ⁓ diffusion MRI, gives us an indirect window, ⁓ say, by tracking how water molecules can move through the brain. In well-organized fiber bundle, water moves along the direction of the fibers, like let's say flowing through a pipe. So the scanner picks up on this signal and produces the images that, as researchers, can go and investigate. ⁓ And so from images, ⁓ we can use sophisticated cutting edge algorithms to reconstruct the pathways in the brain. And that's a technique that we call tractography. And we can derive several measures that describe the organization, the density and microstructural properties of white matter connection. our study, we focused on two key measures, fractional anisotropy or FA. ⁓ measure reflects how well organized each fiber bundle is. ⁓ And then we also focus on the number of reconstructed streamlines between regions. And that gives us an estimate of the connection density. And so when we talk about white matter disruptions ⁓ bipolar disorder specifically, ⁓ we mean is that the organization of these pathways is altered. ⁓ example, if we see lower FAA, then we would assume there's a less coherent fiber bundle organization. And could reflect several changes ⁓ at microstructural level. what is important is the consequence of this change is that have less reliable communication between brain systems ⁓ and ⁓ in our study we found disruption not just at ⁓ individual but in the network level organization of connections across the whole brain ⁓ network. ⁓ Elvisha Dhamala: So in this study, as you mentioned, you're combining brain scans from 16 different sites which use different scanners and different protocols. How do you go about harmonizing all of that data to be able to study it together? Leila Nabulsi: Yeah, that's a very important question. That's one of the challenges of Enigma actually. ⁓ And for researchers who are analyzing and organizing the data, of course, it does take a village. ⁓ obviously different sites different scanner brands, ⁓ different protocols, acquisition All of these can introduce systematic differences that aren't related to the biology ⁓ of the disorder that what we want to study. And so that's what we call noise. So we collected raw scans from all the 16 sites. ⁓ They sent those scans to us to the central processing team here at USC. So an important step was to verify that each site's acquisition protocol was suitable ⁓ for the tractography-based connectome reconstruction that we wanted to do. All data was then processed centrally using standardized pipelines. Scanning site also was explicitly modeled in our statistical analysis to try to separate these between sites variability from the biological signal. Elvisha Dhamala: Two concepts that come up throughout your paper and your results are network integration and network segregation. Can you explain what those mean in the context of the brain and why the balance between integration and segregation matters so much? Leila Nabulsi: ⁓ yeah. So network neuroscience, emerged in the late, 2000. So it's quite recent. ⁓ It borrows from graph theory. That's a branch of mathematics they use to study everything from social networks to the internet, ⁓ but also to study the brain, not an isolated region, but as an integrated system of connections. So these concepts gave rise to ⁓ the concept of the human connectome. ⁓ Basically, that's a complete of how every region in the brain is connected to every other region. So these are complex concepts that we can break down with, I guess, a simple analogy. If we think ⁓ of ⁓ the global air traffic network, we have airports that could be brain regions ⁓ flight that are the connection between ⁓ So you wouldn't ⁓ assess that network by at each airport in isolation. You'd ask how efficiently can travel between two cities, whether certain hub airports carry a disproportionate share of traffic maybe and whether regions cluster connect well to the global system. So the brain works exactly the same way and network neuroscience ⁓ is a very cool tool that lets us ask the same questions but about the brain communication. So there are two important concepts in network neuroscience, integration and segregation like you mentioned. There's been several graph theory studies conducted so far, not just in bipolar disorder, but in other conditions. And these studies of the human connectome in general, people with no diagnosed disorder, have shown that the brain has a tendency to organize into two complementary properties, so the segregation and integration. And that the balance between these properties ⁓ is fundamental to sustain healthy brain function. So network segregation, refers to the brain's tendency to cluster into specialized structural communities. So regions that work together, group tightly, each dedicated to specific function like emotion, memory, ⁓ this clustering is important because it supports the efficiency with which information is processed locally ⁓ these communities. ⁓ so one would think that when segregation breaks down, the brain loses its ability to process information in a focused, organized way. And then there's network integration. So that process refers to how efficiently the information travel across the whole brain, ⁓ allowing the specialist communities to rapidly coordinate with one another. And so when integration breaks down, communication becomes less direct and less efficient. ⁓ Connecting these two properties are also a set of highly interconnected brain regions. These are known as hubs, like our airport hubs, maybe Atlanta could be one in the US. So this form a structural backbone that support this global neural communication and support ⁓ higher order cognitive functions. So one would think the damage to this core system creates a vulnerability in the network so the brain needs both segregation integration imbalance and when the balance is disrupted like we see in bipolar disorder ⁓ this is associated with alteration how the brain supports thoughts, emotion, behavior and how really each part of the brain communicates with one another. Elvisha Dhamala: So in your study, at the whole brain level, you found that individuals with bipolar disorder had lower network density, longer path lengths, lower efficiency, and higher betweenness centrality. What exactly does that mean in terms of integration and segregation and just how information flows in their brains overall? Leila Nabulsi: Yeah, So all these findings aren't isolated. They paint a coherent picture of what the brain network ⁓ looks like in bipolar disorder and it looks like it's less well connected, less efficient transferring information, more dependent on a small number of critical bottlenecks. ⁓ if we think back to the traffic analogy, if you imagine a network where fewer routes exist, ⁓ journeys will take longer. Fewer passengers can move efficiently across the system and everything funnels through just a handful of overloaded hub airports. So that's essentially what we're seeing in bipolar disorder at the level of the whole brain. ⁓ so if we want to break down each met, individually in more plain language, We found lower density relative to those individuals who ⁓ not diagnosed with the illness. So that means fewer connections are present across the whole brain network. So the wiring is sparser. So some regions that would normally communicate directly are less reliably connected. ⁓ found longer characteristic path length ⁓ bipolar disorder. So that means on average, it takes more steps to get from any one brain region to any others. So information has to travel via more intermediate regions rather than direct routes. It's like having to take three connecting flights instead of one direct. So we also found lower global efficiency. That's a direct consequence of the lower density and longer path length. So the brain is less efficient and integrating and distributing information across the whole system simultaneously. And then we found higher betweenness centrality. So that lets us think that there's a small number of hub regions in the brain that are now carrying a disproportionate share of the network traffic. So more journeys are being routed through fewer airports, making these hub critical bottlenecks. ⁓ Together, these findings do paint a picture of the brain network that is less well connected, less efficient at transferring information across the network overall, more reliant on a small number of critical hubs to maintain communication, so a more fragile architecture overall. So a network that depends heavily on a small number of hubs is also more vulnerable to disruption. So if one of these hubs is affected, we can think the consequences ripple across the whole system. Elvisha Dhamala: You also found two specific sub networks that were primarily disrupted in bipolar disorder, one involving the limbic and basal ganglia regions and another linking default mode salience and frontal limbic areas. Can you tell us a bit more about those specific networks and circuits overall and what roles they play? Leila Nabulsi: So the first subnetwork that you mentioned is the limbic-bisogangia circuit. ⁓ So we found altered white matter microstructure, ⁓ FA, predominantly between and within limbic regions ⁓ in basal ganglia nodes. The limbic system, we can think of it as the brain's emotional core, and the basal ganglia are central to reward, motivation, goal-directed behavior. So together, this circuit supports how we experience and regulate emotions. ⁓ how we respond to reward and threat. So the changes we see here makes us think that. why matter connection between these regions are altered in bipolar disorder, it could explain for some of the functional deficits we know that characterize the illness. And then the second sub-network we found was connection density, so number of ⁓ reconstructed streamlines, ⁓ spanning across the fault mode network nodes. This is the network, let's say, of the inner self. It kind of supports self-referential thinking, autobiographical memory, ⁓ but salience network nodes. ⁓ that's like the brain's relevance detector and frontal limbic connections as well. So it's quite like a wide network ⁓ linking prefrontal reasoning with the limbic emotional system. So these are three ⁓ important circuits within the sub network. And so ⁓ the circuit is disrupted, the brain may struggle to assess what is emotionally relevant. ⁓ It may affect the ability of an individual to distinguish internal from external states, difficulty in recruiting what we call top-down control, so the ability to pause and regulate on emotional responses and then acting on it when emotions escalate. And so these are conclusions, obviously, that would be great to test in the future using detailed behavioral and cognitive measures. anatomical location of our findings make us think we can pinpoint towards those cognitive deficits of bipolar disorder. ⁓ it's interesting to also note that within both of these two subnetworks, we found a mix of lower and higher connectivity. So this makes us think that there are both disruption, but also potentially compensatory processes that are happening. So when core pathways are compromised, then the brain recruits alternative routes to maintain communication ⁓ and to maintain a utamic state of the illness. Elvisha Dhamala: So along those lines, you also looked at specific clinical features and how they were related. you found that longer illness duration, later age of onset, and a history of psychosis ⁓ were with greater network disruption. What does that tell us about how the illness might be shaping the brain over time? ⁓ Leila Nabulsi: one of the most important thing we can do ⁓ in study like these and with the power of enigma and rich phenotypes is go beyond simply comparing bipolar disorder to controls. ⁓ So, and can ask within the bipolar disorder group, does this network disorganization vary with how severe or how long standing the illnesses? ⁓ If brain disruption tracks with illness burden, that's perhaps evident that these aren't just pre-existing traits that they may reflect something that accumulates or evolves within the disorder. So we found in relation to illness duration the longer someone had lived with bipolar disorder, ⁓ the more the brain's communication network appears to be affected overall. ⁓ And regionally, we'd reduce connectivity specifically between hippocampus to the amygdala ⁓ So regions that we know are central ⁓ to emotional memory, fear, regulation, and this is consistent with idea of cumulative disease burden, that maybe there are chronic inflammation and changes in the brain tissue happening over time, gradually affecting white matter pathways ⁓ that keep brain systems connected. And we also found that illness onset was associated with more pronounced alteration globally and in specific circuits connecting the cerebellum, the thalamus, ⁓ frontal limbic pathways as well, all networks are involved in emotion regulation. Usually earlier onset ⁓ is associated with greater clinical severity. ⁓ So more pronounced deficits seen at a onset of the disorder. ⁓ maybe are just reflecting differences in treatment timing, in the illness recognition, or maybe there are some underlying biological and clinical factors, perhaps genetic resilience allows an individual to tolerate illness burden before symptoms emerge. ⁓ We also found that individuals with a history of psychosis showed more fragmented and less efficient brain network, fewer connection, longer communication paths, and greater overreliance in a small number of hub regions. So this is consistent with prior work that shows that psychosis in bipolar disorder is associated with more severe white matter changes. we believe our findings just extend that picture to show that psychosis leaves a detectable mark on the brain's overall communication architecture. ⁓ Elvisha Dhamala: also looked at manic episodes, which are another characteristic aspect of bipolar disorder as you mentioned, ⁓ you found that they were associated with higher connectivity in a fronto-subcortical circuit. How do you interpret those findings? Leila Nabulsi: We found that a greater number of manic episodes was with a higher connectivity ⁓ between the anterior middle cortex and some subcortical regions ⁓ just below the thalamus. these are kind of like the opposite direction from all the findings in the study, but we believe one possible interpretation is that perhaps mania induces some sort of neuroplasticity in the brain. So having had repeated manic episodes, perhaps drives structural reorganization of this frontal subcortical circuit through the brain's reward and excitatory signaling systems, because those are like the neurotransmitter systems that are passing through the circuit that we found. ⁓ so... ⁓ think of it like if a road gets used so frequently, it gets So there's a structural of all this repeated activation. We did not record any connectivity change associated with having experienced a greater number of depressive episodes. I suppose this is a reminder that bipolar disorder is not just a single uniform process that perhaps different aspects of this illness can leave different biological signatures in the brain. Elvisha Dhamala: Earlier on, you introduced the idea of polypharmacy, which is that most people with bipolar disorder are taking multiple medications. And in this study, you examined several different medication classes, also using something called the neuroscience-based nomenclature. Can you tell us a little bit more about that approach and what you found? Leila Nabulsi: The challenge in studies like ours ⁓ is separating each drug effects from others and from the illness itself. ⁓ So within Enigma, we had this wide pool of information, clinical information, medications. ⁓ And so we decided to tackle this issue by modeling current medication and illness severity together in order to identify those effects specifically and separate those effects from the illness effects. ⁓ so we chose to ⁓ medication under two complementary frameworks. So the conventional indication based categories, ⁓ then the neuroscience based nomenclature or MVN. ⁓ So the conventional categories group drugs by what they're prescribed for. ⁓ lithium, antipsychotics, antidepressants, anticonvulsants. ⁓ clinically intuitive, but neurobiologically, it's a bit blunt since drugs within the same category could act on very different mechanisms in the brain. we're looking at anatomy, so we want to be as precise as we can ⁓ effects that we're measuring. So ⁓ NBN group drugs instead by what they actually do to the brain. So instead of antidepressant, you'd call it a serotonin reuptake inhibitor. ⁓ it's kind of like a shift from what is the drug used for to how does this drug act on the brain. ⁓ So the most significant findings were seen for antidepressants and antipsychotics. So for antidepressants use, ⁓ particularly we found that the SSRIs, reuptake inhibitors, were associated with less communicating brain network, ⁓ weaker connection between two limbic regions, specifically the thalamus and the hippocampus. And then we found that a class of antipsychotics acting on multiple neurotransmitter system was linked to weaker connections specifically in the orbital frontal cortex. That's a region ⁓ in emotional decision-making. We believe take-home message from our medication analysis that treatment exposure is an important factor to consider when studying the biology of any disorder, really, but in particular bipolar disorder. People receiving certain medication may also differ in illness history or symptom severity. So ⁓ the effects of the illness from those of the medication is key in future study. And we do hope that our approach can encourage future studies to take these factors. into account. Elvisha Dhamala: Interestingly, you didn't find any significant associations with lithium use. What does that tell us? Leila Nabulsi: We found that lithium was actually the only ⁓ medication class that showed no significant with any of our network connectivity measures, ⁓ while antidepressants, antipsychotic, anticonvulsant, they all some sort of detectable, let's say, footprint on our network. So ⁓ one interpretation is ⁓ that rather altering the network, may be stabilizing it, preserving structural connectivity. to its baseline, ⁓ other medication perhaps leave a more measurable mark. ⁓ Elvisha Dhamala: So we take these findings together, what does that mean for the day-to-day experiences ⁓ of living with bipolar disorder? So in terms of their actual daily functioning ⁓ in regulation, decision-making, or cognitive function. Leila Nabulsi: Our study ⁓ measures brain structure, so symptoms directly. What we're saying is that the circuits we found disrupted are the same ones that are implicated in emotion regulation, cognition, reward by a large body of prior work. And since the vast majority of our participants were eutymic as scanning, we think this findings reflect the persistent trait-level biology of bipolar disorder, what's present in the brain when someone is clinically stable. But specifically what our findings may mean... for different aspects of this disorder is that, ⁓ for in ⁓ relation emotional regulation, the disruption within frontal circuits ⁓ in ⁓ the through, let's say, which the frontal cortex exerts this top down control over the limbic system, ⁓ when ⁓ well emotions may escalate more quickly, be harder to read back down. The salience network may further contribute to ⁓ Some stimuli that don't carry an emotional valence felt ⁓ intense or threatening. The reduced network efficiency and hub over-reliance offer another structural basis, I think, for the persistent cognitive difficulties we see in bipolar dysphoria, even during utymia. So sustained attention, working memory, cognitive flexibility, these all be affected. ⁓ And a default mode connectivity ⁓ could be with difficulty ⁓ disengaging from ⁓ We also basal ganga disruption. Those are central to reward processes. same motivation when their connections are altered, their ability to assess risk and reward, feel motivated, sustain goal-directed behavior, those could also be all compromised. picture if we can understand where symptoms map onto the brain and where... existing interventions map onto the same circuit, then we can begin to think about how to target them more precisely. ⁓ And I think this is the long-term promise of this kind of work. Elvisha Dhamala: If there's one key message you'd want clinicians, researchers, people affected by bipolar disorder, and their families and friends to take away from this work, what would that be? Leila Nabulsi: Bipolar disorder is a disorder of brain communication ⁓ and ⁓ being able that at a scale across the whole network the full course of the illness through collaborative efforts like ENIGMA is how we will eventually move psychiatry towards more objective biologically informed decision making ⁓ treatment that are truly matched ⁓ can really make a difference to the individual.