Matthew Kempen Hello, my name’s Matt Kempen. I’m the Marketing Manager for ACAMH, and welcome to the second podcast we’re releasing this week to celebrate Eating Disorders Awareness Week. Today, you’ll hear from Dr Caitlin Lloyd of Columbia University in the USA on anorexia nervosa. Earlier this week, we released a podcast with Dr Michelle Sader of the University of Aberdeen in Scotland, on the topic of ARFID, so do please check that out. As well as these podcasts, there’s a host of other three resources, such as talks, video abstracts, blogs and a topic guide on eating disorders. So, do please visit www.a-c-a-m-h.org, and now, on with the podcast. Dr Clara Faria Hello, I’m Clara Faria, an ACAMH Young Person Ambassador and in today’s episode, I have the pleasure to talk to Dr Caitlin Lloyd. Dr Lloyd is a Postdoctoral Fellow at the Columbia Center for Eating Disorders. Her research focuses on neurobiological mechanisms in eating disorders. Today, we will be discussing her paper, “Food Choice and Neural Reward Systems in Adolescents with Anorexia Nervosa and Atypical Anorexia Nervosa,” which was published in the JCPP. Welcome, Dr Lloyd, thank you so much for being here today. Can you start with an introduction, giving a short overview of what you do? Dr Caitlin Lloyd Hi, well, it’s really great to be here and I’m really excited to discuss my research with you today. So, as you mentioned, I’m a Postdoctoral Research Fellow at Columbia in the Eating Disorders Research Unit and I have a longstanding interest in the neurobiology of eating disorders and have primarily been studying, I guess, for almost ten years now, the neural mechanisms of anorexia nervosa. And so, in our lab, what we do is really try to look at the brain differences that might be present between patients with eating disorders, such as anorexia nervosa, and individuals without eating disorders, both in terms of their structure and function. And then, we spend some time trying to really characterise the core disturbances in eating behaviour and use tasks that capture those disturbances and then, link those back to the underlying brain mechanisms, using technologies like functional MRI. Dr Clara Faria That all sounds fascinating. I know sometimes, you know, looking at neurobiological mechanisms behind certain diseases can help fully lead to new targets for drug intervention. And focusing on your paper, you looked at certain brain regions and you hypothesised that their activation patterns would be different in adolescents with anorexia nervosa, compared to healthy eatings. Can you tell us a little bit more about the rationale of the paper and why it doesn’t matter to investigate neurobiological mechanisms of eating disorders? Dr Caitlin Lloyd I guess, first of all, there’s really good evidence that eating disorders involve disturbances in the brain or, sort of, brain differences, which is, kind of, in contrast to perhaps what was thought a bit earlier in time, which was that these illnesses were really, primarily, kind of, sociocultural in nature. And we also know that eating disorders, and so, for example, anorexia nervosa, they’re, kind of, characterised behaviourally by disturbances in decision-making. So, for example, there’s a persistent restriction of food intake, a persistent selection of low-fat foods over higher fat alternatives, that is really coming at the expense of, kind of, physical and mental health. And so – and we also know that decision-making is driven by the brain and so, it stands to reason that maybe there are differences in underlying brain circuits that might be contributing to these disturbances in eating behaviour that we’re seeing in these patients. And the reason we wanted to look at adolescents is because we know that eating disorders typically emerge, and particularly anorexia, around, sort of, mid to late adolescence, so in the, kind of, 15 to 19-year-old period. And so, we think it’s really important to try and capture differences in the brain that might be driving restrictive eating, earlier on in illness, when we’re able to parse apart the mechanisms that might be contributing to the restrictive eating, versus changes in the brain that might have occurred because of a longstanding period of illness, so the effects of malnutrition, for example, which we know do change, sort of, structure and potentially function of the brain, as well. And our particular paper was looking, really, at reward systems, so the role of, kind of, heightened activation within reward circuits when a person is deciding what to eat. And that’s really based on the hypothesis that perhaps earlier on in illness, so when people first start engaging in restrictive eating behaviour, in – relatively shortly after their initial diagnosis of anorexia nervosa, that the thing that could be driving that behaviour is the reward system, and so, this idea that maybe it’s become rewarding for a person to engage in restrictive eating. Perhaps that’s because they’re getting reinforcement from people around them in terms of the weight loss being a good thing. Maybe it’s because it’s making them feel good in some way, or alleviating, kind of, negative emotions, like anxiety. And so, our hypothesis that we wanted to test was really, is there differential activation within that reward system at the time that somebody is making a decision about what to eat, in patients with anorexia nervosa, relative to somebody without the illness? In this study, we were really able to look at people relatively early on in illness, which is quite unusual, and the patient participants were within their first year of diagnosis, which is quite an unusual thing for studies of anorexia nervosa. Dr Clara Faria Yeah, and I think the fact that you used patients who were early on in their illness just adds to the value of that study. So, in your study, Caitlin, you used, as you said, the Food Choice Task to test those neural mechanisms of choice among adolescents, and can you tell a little bit to your audience what this test consists of? Dr Caitlin Lloyd So, the Food Choice Task and – or variants of it, has been used in several different studies of eating behaviour, and it’s well validated, really, to capture somebody’s food preferences, as well as to be able to index those brain mechanisms that are driving those food preferences, as well. And so, what happens in this task, and it has three different phases, and in the first one, the participants rates 76 different foods and they’re presented individually. So, on one trial, you might see a picture of a cake and the next trial, you might see a picture of some carrots, and the person rates how tasty they think each of those food is on a scale of one to five, one being not tasty and five being very tasty. And then, the participant does exactly the same thing, but for healthiness. So, they’re then rating each of the foods how healthy they think they are, again on a scale of one to five. And this allows us to then identify one food that the participant has rated as, kind of, a neutral, so a three, on both tastiness and healthiness, and that food we then select as that person’s reference item. And then, in the third phase of the task, they complete a series of choices, so actually 75 trials, where they choose whether or not they would rather eat their reference item food or each of the other 75 food items. So, for example, if their reference item was crackers, they would then choose between on trial one, crackers and cake, on trial two, maybe crackers and carrots. And what’s really important is that they know that one of these trials is going to be selected at random and so, they will actually be asked to eat the food that they’ve selected on that trial, which makes the task then incentive compatible and makes it more likely that somebody is responding really in line with their true preferences. What it also allows us to do is because half of these foods are, sort of, high fat, where, you know, at least 30% of the total calories is derived from fat, and half of those are low-fat foods, we can identify the extent to which somebody is willing to eat a high-fat food on the task and the proportion of high-fat choices that they make. So, how often they select those high-fat foods over the reference item and then, similarly for the low-fat foods, how often they’re selecting those over the reference item. And this correlates very well with eating behaviour in patients who have anorexia nervosa, so that is the more restrictive eating somebody displays in a meal setting, or, kind of, self-reported restricted, correlates very strongly with the extent to which they’ll select high-fat foods on the task. So, the more restrictive their eating, the less likely they are to select those high-fat foods, and so, it’s a really good indicator of restrictive eating behaviour. And that’s really important, because when we administer this task, then, in the scanner, in an MRI scanner, we’re able to see the brain activation that’s linked to the decisions that somebody is making, so that’s underlying that choice behaviour, and to see, in particular, what brain circuits might be driving, sort of, a more restrictive pattern of food choice at the time of decision-making. Dr Clara Faria That’s super interesting and thanks for such a comprehensive explanation. And in your paper, in your study, you – your hypothesis was that you would observe more activity in this region called nucleus accumbens of the brain in adolescents with anorexia nervosa, compared with controls. Well, considering all this rationale of the study which you just explain to us, I wondered, can you tell us about the headline findings, and did the findings confirm your initial hypothesis? Dr Caitlin Lloyd The hypothesis was, I guess, on the basis that we think there might be, sort of, heightened activation within reward systems that’s driving restrictive eating, and the nucleus accumbens is part of this reward circuit. It’s very well-known to be implicated in, kind of, reward-driven or motivated behaviour, and so, we were expecting that this might be more active at the time of food choice and so, more driving those choices of patients with anorexia nervosa, relative to healthy controls. We actually did not find evidence to support that hypothesis. So, the nucleus accumbens was engaged during food choice in both the patients and the controls, but it seemed to be associated with choice preference, so how much somebody wanted something. And what that means is it was more active if somebody wanted the food more or had a higher choice preference for that food, and that was the same in the patients and the controls. So, it didn’t seem to be differentially engaged, and the extent to which it was engaged was also not related to restrictive eating or the selection of high-fat foods in the task. So, contrary to our hypothesis, it didn’t seem to be associated with restrictive eating in our patients with anorexia nervosa, so not in line with the hypothesis. We did also look at another region, which was the anterior caudate, and the engagement of the anterior caudate is thought to reflect a, kind of, shift from perhaps reward-driven eating behaviour to a more habit-based eating behaviour in adults. So, that is whilst a person might initially engage in restrictive eating because it’s rewarding, with repetition over time, that behaviour might come to depend no longer on reward-based circuits, but actually, on those circuits more implicated in habitual automatic behaviour. And therefore, the restrictive eating in anorexia nervosa might actually be, sort of, more automatic and less thoughtful and goal directed. Which might make it then really hard for a patient who’s very motivated to recover to, kind of, eat in a way that’s consistent with recovery and weight gain, in particular. And so, having found this anterior caudate region to – which is associated with, sort of, automatic behaviours, to be more engaged during food choice in adults, we also wanted to see whether this was engaged, sort of, earlier in the course of illness in this adolescent sample. And whilst we didn’t find statistically significant differences between engagement of this region in patients versus controls in our adolescent group, we did see that the more this region was engaged in the patients, the more restrictive their food choices, so the less willing they were to eat high-fat foods. Which highlights to us, again, the potential role of this region in restrictive eating in anorexia nervosa. Dr Clara Faria No, that’s fascinating and I think that also aligns with what we know clinically about anorexia nervosa. Like, it’s well-known that after five years of illness, the rates of recovery are so much lower, and it’s really interesting to see how that relates to the neurobiology of illness, having, like, as you just explained beautifully, more consolidated circuits that are associated, like, with habit formation in patients who are adults and will likely have more years of illness. Well, going back to your study findings, well, as you just mentioned, and that’s actually something I think it’s fascinating about your study, is that you had a null finding. So, your results did not confirm your initial hypothesis, and you’ve just explained to us about the caudate, the anterior caudate, and you just mentioned that in your study, you found that the activation of that region was associated with more restrictive choices, but it was not more activated in adolescents with anorexia. And I was just wondering, how did you make sense of these findings? You just explained a little bit, but do you think that you might – like, the anterior caudate might be, then, a marker of, maybe illness severity? Dr Caitlin Lloyd So, that’s a great question and I think we’re, kind of, still trying to make sense of it, too. But I – firstly, I think one important thing to note is that we – having not found, kind of, statistically significant differences in patients versus controls, we then used a slightly different method, so a Bayesian statistical method, which allows us, really, to test whether or not the null hypothesis, so that is there being really no difference between patients and controls, is true. Which is something you can’t do with standard statistical tests, just because of the way in which the theoretical basis of the test is working. And what that showed for the nucleus accumbens was that there was good evidence that we should be accepting the null hypothesis, so there really being no differences between patients and controls in activation of that reward region. In contrast, in the anterior caudate, there actually wasn’t strong evidence for us being able to accept the null hypothesis, suggesting that actually, a bit more data was needed for us really to be able to rule out no difference in that region. And so, that probably points to there being, as you’re, kind of, thinking about perhaps there being individual differences in the patient group themselves, which is well supported by our data, too. So, there was variability in the activation of the caudate in patients with anorexia nervosa, some of whom were showing, sort of, engagement during choices and some of whom were not. In terms of what that might mean, it could mean, of course, yes, that maybe it’s related to duration of illness, so the longer that the illness continues, maybe there is a shift to a different type of circuit. Alternatively, it might mean that there are differences when somebody is initially diagnosed, that some people were showing, kind of, a more dependence on that caudate region than others, and maybe that’s prognostic. So, meaning that people who engage the caudate more at this baseline or earlier on in illness, are more likely to have a longer illness duration, potentially. And what’s nice about this study, actually, is that this is something we’re able to test because it was a longitudinal examination, such that the patients are being followed up one year and then two years after their initial scan. They’re having a repeat scan and they’re also redoing the Food Choice Task, and we’re also clinically evaluating them to be able to see, are these baseline differences in brain activity actually predicting the longer-term outcome? Dr Clara Faria Well, thank you so much, Caitlin, for sharing so much about your study. It is really fascinating, and I think when your longitudinal study is published, it will answer some very pressing questions on the neurobiology of anorexia. Still on your study, so you mentioned that you looked both at patients with anorexia nervosa and atypical anorexia nervosa, and I was just wondering, do you think that might have been, you know, one potential explanation why you didn’t find any differences in the region you’re looking for, considering that in atypical anorexia nervosa, we don’t have the severe weight loss that sometimes we see in patients with typical anorexia nervosa? Dr Caitlin Lloyd Yes, this is a great question that did come up somewhat in the review process as to whether some of those differences that we were not seeing was maybe because our sample was including individuals with atypical anorexia nervosa. To address this question, we rerun the analyses not including those individuals, so excluding the individuals with atypical anorexia nervosa, and we actually found the same thing. So, there’s still no evidence for the group difference. Again, not – also not being able to accept the null hypothesis for there being no group difference in that caudate region, but it’s – and similarly, with the nucleus accumbens, there was, again, no difference in that region when we just include individuals with anorexia nervosa. And I should say, as well, the associations between brain and behaviour, so the activation of the caudate being, kind of, negatively correlated, the high-fat food choices, were also still seen when we included only those individuals with anorexia nervosa. Dr Clara Faria Oh, thank you so much for that additional explanation, and to wrap up, I would like to ask you, where would you like to see your field go next? Is there any particular area of neurobiological mechanisms of eating disorders we – you think we need more research on? Dr Caitlin Lloyd Yeah, so I think it’s a really important time for the field, as, kind of, there’s a huge growth in the availability of tools for neuroscience that are allowing us to, kind of, visualise the brain in different ways, both structurally and functionally. And also, there’s an emergence of, kind of, new paradigms and statistical methods that really allow us to better test specific hypotheses about the mechanisms of illness. And I think there’s so many different valuable lines of potential study. I spend a lot of time thinking about this, but I think one that I’m particularly excited about is really trying to understand how somebody uses different information, both about food, so tastiness, healthiness, caloric content, as well as about future goals, which may or may not involved recovery, and how they’re able to pull in all of that information to construct value representations that we know guide choice. And I think the reason I’m particularly interested in that is because, you know, in some cases, there’s not differences in the way in which inf – value information is represented in the brain, so this activation that’s associated with choice, but the value that’s ascribed to different food choices, let’s say, it is quite markedly different between patients and healthy controls. And what this approach might allow us to do is to better understand those circuits that not only are contributing to different choices, but are underlying different affective responses to food, like eating-related anxiety, for example. And then, I think, sort of, building onto that, we might be able to look at how do we change both the information and processes that are being used to construct those value representations, and are there differences in how modifiable those processes might be in patients and healthy controls? The idea being, really, just to try and expand our set of evidence-based targets for treatment that can allow us to quantify the impact that our treatments might be having on mechanisms and hopefully, to improve the outcomes of treatments for our patients, as well. Dr Clara Faria Well, that’s really fascinating, and thank you so much, Dr Lloyd, for sharing your research with us today. And for more details on Caitlin’s work, please visit the ACAMH website, which is www.acamh.org. Also, we’re going to link Caitlin’s profile on Columbia here on the page of this podcast, if you want to check out more of her work. ACAMH is spelt A-C-A-M-H. Do keep an eye out for other podcasts, and don’t forget to follow us on your preferred streaming platform, let us know if you enjoy the podcast, with a rating or a review, and do share with friends and colleagues. Thank you so much for the listen.