PDF: Psychology and Neuroscience of Affective Disorders Lecture Transcript

Module: Psychology and Neuroscience of Affective Disorders Week 1: Introduction to affective disorders Topic 3: Emerging focuses in affective disorders (Part 2 of 4) Lecture transcript Dr Rebecca Strawbridge Department: Centre for Affective Disorders Slide 4 In this part, we discuss emerging resear...

Module: Psychology and Neuroscience of Affective Disorders Week 1: Introduction to affective disorders Topic 3: Emerging focuses in affective disorders (Part 2 of 4) Lecture transcript Dr Rebecca Strawbridge Department: Centre for Affective Disorders Slide 4 In this part, we discuss emerging research aiming to better understand the mechanisms related to affective disorder development and maintenance. Of course this is a broad field which has been investigated for decades, however, as you may remember from right at the start, we still lack understanding. Slide 5 I've already mentioned a lack of consistent results coming from different studies. A key reason for these factors is the heterogeneity of affective disorders. Here we focus on depression which is particularly challenging in this regard. A lot of the issues here are covered very nicely by the referenced paper by Eiko Fried. His research has pointed out that modern diagnostic criteria for depression allows more than 1,000 unique symptom combinations, but heterogeneity also exists in the realms of illness causes, comorbidities, longitudinal course, treatment response, and others. Despite this and despite some sub-types of depression going way back, almost all research treats depression as a unitary construct, also conceptualising illness severity based on an equal weighting of symptoms course that themselves differ © King’s College London 1 between different measures used. From this perspective, it's not that surprising that we don't find the same results coming from different even methodologically similar studies. Slide 6 So where is this going. There is now a lot of consensus that we need to discover subgroups that people can be categorised in and within which there is homogeneity. The NIMH RDoC initiative goes back some years now, but this is an ongoing collaborative initiative to do this. For example, let's look at people with anhedonia. Let's determine the severity of this and assess them together. As a result even though we are trying to split an illness like depression, because there is many features common to many disorders, this is actually encouraging trans diagnostic assessment to progress towards homogeneous sub-types. Sub-types are being sought both in biological and non-biological domains. A few approaches to sub-typing include a dimensional approach where subgroups are sought based on theory and existing knowledge. Data-driven sub grouping where groupings emerge based on the data itself, and combinations of the previous two. Network approaches can be helpful in delineating how different characteristics are related to one another in order to understand constructs more comprehensively. And finally, it's being increasingly considered important to be looking at different types of data together, so multimodal investigation. Slide 7 Computational modelling is actually one approach that could help disentangle heterogeneity, and more broadly improve mechanistic understanding of these illnesses now that we have more sophisticated technologies. In the past to build understanding in this area, we had to do a series of singular experiments based on previous theories. This meant that nuances got lost between complex connections and interactions resulting in reductionist models of understanding. With the progression of technology and understanding, we've recently been able to progress things in a more sophisticated manner. Computational models can do this by trying to better understand the brain by reproducing its processes mathematically. Slide 8 Computational models can be applied in attempting to understand the brain by reproducing its processes. Computational psychiatry aims to mathematically explain relationships between neurobiology, environment, and symptoms. It can use cognitive models in combination with computational power and well-designed experiments to unite different levels or facets rigorously. Generated models can be used to predict future progression of states. This might help to bring together previously discrepant findings and to avoid reductionist theories. It is a fast moving field still basically in its infancy, but work is now looking at developing more complete models of how effective disorders develop and perpetuate, and is trying to advance closer to having translational benefits to patients. Slide 9 A key example here is studies like that of Camacho et al. They were looking towards finding the right combination of treatments for people with depression by firstly creating computational models of neurobiological factors; that is, monoamines, cortisol, and 2 © King’s College London testosterone, and then using machine learning to train what would happen to these upon administration of different drugs. And they highlighted some medications that then could be useful for helping people with treatment-resistant depression. This is nice because it's putting computational psychiatry looking forward to clinical application, but I would like to stress some cautions here. First of all, if you show this study to many clinicians, you'll see a lot of suspicion. At the moment there is in general a big gap between computational and clinical experts. For example, computational researchers often cite diagnostic problems related to depression, whereas clinicians cannot see how a computer could possibly diagnose or treat a person better than all of the years of clinical education and experiences they've accumulated. There will be challenges in bringing these closer together. Slide 10 Moving forward, we can expect to see studies using more advanced ways of mapping environment and peripheral biological interactions with brain computations. These will provide better tools to move the field closer to having translational implications. For example, a good model of what happens to monoamine neurotransmitters in the brain alongside cognitive processes could be used to predict how a person will respond to specific monoaminergic antidepressants based on which receptors they target in combination with an individual's pre-medication function. As things progress, we also hope to get closer to knowing which models work for whom. So what is transdiagnostic and what is disorder specific? Computational psychiatry could advance at an increasing rate as we start to close the gaps between things we already know with the things we don't yet know. Slide 11 We move on now to biological marker research in disentangling the mechanisms of affective disorders. This is not a new field by any means, and other sessions in this course focus on the past, present, and future of a range of biological domains. To summarise, the first biological theories of depression were related to monoamine neurotransmitters in the brain, prominently serotonin but others as well. These persisted in the forefront for a long time before neuroimaging started taking off, and neuro anatomical models were catching up finding initially structural and later functional differences in the brain of depressed people. At a similar time; that is around three decades ago now, neuroendocrine theories also became popular finding hormonal associations with affective disorders focused mainly on an over secretion of the stress hormone cortisol, and then dysregulation of the HPA axis which is responsible for regulating related hormonal functions. More recently, neuroplasticity was found to be important. For example, finding reduced cellular production in the brain and peripherally. And this is one area still popular today alongside other biological work including that surrounding immune and inflammatory responses, which generally appear overactive in people with affective disorders. Genetics is here at the bottom, but this work has been ongoing also for a long time. Initially focused on specific candidate genes and more recently focusing on polygenic genome-wide measures. This is a really brief whistle stop tour of the biological history because there will be a lot more focus that you receive on individual biological domains in later lectures in this course. 3 © King’s College London Slide 12 Now that we know that so many biological systems are involved in affective disorders even within diagnoses, although similar between unipolar and bipolar disorders, we need to integrate data to more fully understand mechanistic pathways but also to subgroup participants, assess larger samples, which is becoming more possible through consortia, obtaining harmonised assessments, and further explore new areas that have emerged as important. Particularly worth noting here are small biomarkers from microbiomics and metabolomics as well as new advances in genetics. On this topic it's important to say that a lot of the 21st century neurobiological work is starting to focus more on translation towards the bedside. Pharmacogenetics studies have sequenced people's genomes and used this data in combination with pharmacological understanding of how specific medications work to suggest which treatments a patient might best respond to. We will talk more about this type of study goal later, but a word of warning here is that even though this might sound like a fruitful path, we don't see a consistent improvement in response from attempting to personalise treatment in this way from the studies we have so far. Slide 13 The hope then here is that in combination with what has already been covered in this section, future neurobiological work can help to disentangle the heterogeneity of mood disorders, be helped by technology, and in doing so help elucidate causal mechanisms and bring the field closer to translational benefits for patients. Slide 14 To summarise this section on mechanisms, we still have a long way to go to understand mood disorders. We're not short of research pathways with potential here to establish the mechanisms of both unipolar and bipolar disorders, but there are currently large explanatory gaps in our knowledge of the pathological mechanisms of mood disorders. We're very short of integration and really explaining the relationships between them. So research will need to very well model connections between the different players discussed in this section: biological, cognitive, clinical, and treatment effects in coherent and cohesive overarching theoretical frameworks, and all while accounting for the heterogeneity of patients within currently defined diagnostic categories. And even if we solve the vast heterogeneity issue, this will likely remain a challenge for some time because of the obstacles we've covered so far. However, this is critical and the implications for preventing and fully treating these illnesses are clearly huge. 4 © King’s College London