Matthew Kempen Hello, my name’s Matt Kempen. I’m the Marketing Manager for ACAMH. This week we’re really excited to be releasing two podcasts to celebrate Eating Disorders Awareness Week. You’ll hear from two early career Researchers making great strides in their field. In an odd coincidence, both have left their native shores and flown to opposite sides of the Atlantic to conduct their research. Today, you’ll hear from Dr Michelle Sader of the University of Aberdeen in Scotland. Wednesday of this week, we’re releasing our chat with Dr Caitlin Lloyd of Columbia University in the USA, on “Anorexia Nervosa.” Let’s get on with the chat. Dr Clara Faria Hello, welcome to the In Conversation podcast series for the Association for Child and Adolescent Mental Health, or ACAMH for short. I’m Clara Faria, an ACAMH Young Person Ambassador, and in today’s episode, I have the pleasure to talk to Dr Michelle Sader. Dr Sader is a Postdoctoral Fellow at the University of Aberdeen School of Medicine. Her research focus on neurobiological mechanisms in eating disorders, particularly ARFID. Today, we will be discussing her newly published paper, “Neural Correlates of Children with Avoidant Restrictive Food Intake Disorder Symptoms A Large-Scale Neuroanatomical Analysis of a Paediatric Population,” which was published in JCPP. Welcome, Dr Sader. Thank you so much for being here today. Can you start with an introduction, giving an overview of what you do? Dr Michelle Sader Hello, Clara. Thank you so much. As mentioned, I’m currently a Postdoctoral Research Fellow at the University of Aberdeen in Scotland. I work as part of the Eating Disorders and Autism Collaborative, or EDAC for short, and this is a research network hosted across the UK by the University of Aberdeen, King’s College London and the University of Edinburgh. Our group’s work focuses on better conceptualising and understanding eating disorders in autistic people, and I’m one of our members on the team with a research background in medical imaging, particularly the use of MRI. The work we’re discussing today stems from my PhD, in which I worked on better understanding regions of the brain that are implicated in volitional or voluntary appetite control. I was fortunate enough to be able to focus on a special research interest of mine, which is avoidant/restrictive food intake disorder, or ARFID. Dr Clara Faria That sounds fantastic, and recently, ARFID gained more awareness in the media and in research, but it is still a relatively understudied eating disorder compared to the other types. Can you give us an overview of what is ARFID and what led you to research about it? Dr Michelle Sader ARFID is a feeding and eating disorder that was introduced in the 2013 edition of the DSM-5, and it’s primarily characterised by extreme restriction in dietary intake, and that’s to the point where a person’s weight, growth, nutritional intake or managing their daily life events are affected. And ARFID is clinically distinct from other eating disorders, such as anorexia or bulimia nervosa, and this is primarily due to the fact that the disorder is not driven by body shape or weight concerns. However, research has shown that the adverse psychological and physiological effects that are associated with ARFID can actually be very similar to those with anorexia, even to the extent where we’ve seen reductions in bone density that are similar across both ARFID and anorexia. The reasons behind an individual with ARFID restricting their intake can significantly vary, so there are three underlying drivers that are associated with dietary restriction or avoidance seen in ARFID. And these three drivers are referred to as sensory sensitivity that’s associated with food, a lack of interest associated with food, as well as a fear of averse consequences, such as choking or vomiting while eating. And these three drivers are quite new in the ARFID literature. They’ve only really formally come into play within the text revision of the DSM-5 that was published in 2022. Importantly, an individual does not present with ARFID if their restriction or avoidance is due to cultural practice that’s associated with mealtimes or if the restriction is secondary to a pre-existing mental or physical illness. But what’s also important to note is that ARFID can affect individuals of any weight range and of any age range. Compared to those with anorexia or bulimia, however, people with ARFID have been reported as more likely to be male. They’re more likely to develop the disorder at a younger age, and they may struggle with the dietary restriction or avoidance for a longer period of time, as well. Despite what we do know, there’s actually loads that we don’t know about ARFID, especially when it comes to the brain imaging research on the feeding and eating disorder. So, prior to our work, there was actually no large-scale structural imaging study investigating the brain regions, kind of, implicated or associated with ARFID symptoms. I personally think it’s really important for Researchers and Clinicians to understand the regions of the brain that may be associated with dietary restriction, and it could also lead towards potential implications for treatment, particularly personalised forms of treatment. I will absolutely admit that in the middle of my PhD, I had only just discovered ARFID and that was back in 2020/2021, and then, later that year, I actually got diagnosed with ARFID. Dr Clara Faria Oh. Dr Michelle Sader So, just a collision of worlds right there, academically and personally. I’m lack of interest, so I would actually rather do anything else. I will forget to eat, genuinely, every day. My partner’s a life saver and I’m not underestimating that term. He genuinely reminds me to eat, ‘cause I just get excited by other things and it’s incredibly ironic, because I always tell him, like, “I can’t have a meal right now, I’m focused on my ARFID research.” Dr Clara Faria Wow. Yeah, can I just say, I think it’s amazing all the work you’re doing in neuroimaging around ARFID, ‘cause as you said yourself, it’s such a new area of research in eating disorders. And listening to you talk about the three main drivers behind ARFID reminded me of a recent paper that was published by Dasha Nicholls and her group, in The Lancet, talking about the epidemiology of ARFID overall. And reading that, it just made me think there’s just so much, as you said, that we ‘don’t know’ yet. Dr Michelle Sader Absolutely. Dr Clara Faria And – yeah, and going to your paper, so as you said, you looked at neurobiological correlates of ARFID and it was the first study to do that. And most importantly, to our audience, it was also the first study that compared those findings with a control group. So, Michelle not only looked at people who had ARFID symptoms, but she also looked at people without ARFID symptoms and looked for neuroanatomical differences. So, can you tell us a little bit more about the rationale of your study? What were your hypotheses? What led you to use the dataset you did? Dr Michelle Sader So, I think pre-emptively, in terms of answering the question, what led me to use the dataset I did, was a slightly unfortunate, yet very fortunate, accident when it came to my PhD. So, to be perfectly frank, I did begin my PhD in November of 2019 and then, as we all know, about three or four months later, we all internationally got hit with a worldwide pandemic. So, that, of course, threw a huge spanner in the works of what my personal plans were for my PhD and what sort of recruitment or imaging modalities we were going to explore. On the other hand, however, it gave me such a phenomenal opportunity to work with large-scale datasets that were already rigorously and comprehensively built together. So, while I didn’t have the chance or opportunity, let’s say, to recruit my own sample, I had a phenomenal opportunity to conduct research on a standardised, well regarded, well validated, sample and that dataset was the Generation R cohort in the Netherlands, based in Rotterdam. And so, I was able to collaborate with them to conduct our study focusing on ARFID, but what was very interesting about our collaboration was that there were no standardised measurements to capture ARFID characteristics in that sample. What we ended up having to do is to create an index through the wide breadth of measurements that Generation R had collected, and we used certain measurements that really reflected the existing DSM-5 diagnostic criteria for ARFID, to create an ARFID Index. So, what’s really important on our end, as authors of the study and as Researchers, to clarify, is that we consider these children that we conducted the study with as presenting with ARFID symptoms. I’m sure we’ll get into it a bit later down the line in our conversation, but one of the things that I’d really like to explore, moving forward with the research, is formal diagnosis of ARFID using either a Clinician-based diagnosis or a validated questionnaire, such as the Nine Item ARFID Screen. But that’s the dataset that we ended up working with, and in all honesty, it was an absolute pleasure to work with Generation R, as well. It was a lovely coincidence, at the end of the day. But diving into our hypotheses, there’s very limited literature on the neurobiology associated with ARFID, as I’m sure you’ve read, as well. We really had to have a think about regions of the brain that may be implicated in the feeding and eating disorder. So, we anticipated that the structure of the following regions would be significantly different between children exhibiting versus not exhibiting ARFID symptoms. And these regions are the orbitofrontal cortex, which is associated with the processing of both reward and punishment, the anterior cingulate cortex, which is associated with conflict monitoring, the cerebellum, which traditionally, was largely ascribed to roles in motor co-ordination, but is now receiving increased attention for proposed roles in emotional regulation and appetite control. We also hypothesised the insula would be a region of interest and that’s due to its associated roles with discussed processing. And lastly, the superior, as well as the middle frontal gyri, which are regions of the brain that are very important for executive function, more specifically the processing of impulsivity and attentional resources, as well. So, those are loads of brain regions, but they were primarily driven by three distinct trains of thought. The first train of thought is the very limited, but existing, functional MRI studies on ARFID. The second train of thought that we worked with were existing neurobiological theories that were proposed by other ARFID experts, and our final train of thought were regions that were associated with other restrictive eating disorders and also, generalised appetite control. So, concerning the limited functional MRI studies on ARFID, there were two smaller studies that used brain scans to see how people with ARFID responded to images of high versus low calorie food. So, one of these studies by Kerem et al. in 2022 found that teenagers with ARFID who had a higher body weight showed more activity in certain areas of the brain, primarily the orbitofrontal cortex and the interior insula. So, that evidence led us to anticipate that volumes of both of these regions would be increased in children with ARFID symptoms. There is also a paper by Professor Jennifer Thomas, which focuses on reviewing existing literature to determine if neurobiological features might be associated with ARFID. So, this review suggested that the different brain regions might play a role in ARFID depending on the underlying ARFID driver. And what I mean by that is, for instance, it was theorised that brain regions central to the processing of appetite, such as the insula, might be associated with a lack of interest ARFID driver. Alternatively, regions that might be associated with conflict monitoring and processing challenges, such as the interior cingulate cortex and the ventral prefrontal cortex, would also be associated with the aversive ARFID driver. And we expect that the volume in these regions would be increased in those with ARFID symptoms relative to those without. Finally, regarding the regions that were associated with other eating disorders or general appetite, our research, especially my PhD, focused on paying special attention to the cerebellum. And we proposed that the volume in this region would be smaller in children with, relative to without, ARFID symptoms, and this is due to the fact that research in this region has been associated with eating disorders, such as anorexia and obesity, as well as regular or unexplained weight gain. And the cerebellum has also been linked to the regulation of digestion on a physiological level. So, what we anticipate seeing is that in other eating disorders, there are also other regions that have been associated with disordered eating, and this involves the anterior cingulate cortex, middle frontal gyrus and the orbitofrontal cortex, and these differences have also been seen in those with anorexia. This, kind of, strengthens our existing hypotheses that these regions might be associated with ARFID in addition to the previously listed literature on the disorder. Ultimately, we worked with a lot of different information sources to determine regions that we anticipated would play a role in ARFID. Dr Clara Faria Thank you so much for such a comprehensive explanation of your rationale and I love to hear about the story about how, like, actually, you were not planning on using that dataset originally, but how COVID drove you to be creative and well, ended up doing a fantastic and super, like, landmark study for the ARFID research field. So, it’s really cool. I love those kind of stories. Dr Michelle Sader Thank you. Dr Clara Faria But going back to your paper. In your study, you looked at specific bain – brain regions that you have mentioned, and I was just wondering, can you tell us a little bit about your findings and about the differences you found looking at the MRIs of young people with ARFID-like symptoms? Dr Michelle Sader Yes, indeed. Our study provided significant strengths in that we were able to assess 121 children with symptoms, relative to 1,856 children without symptoms. Our findings showed that children with ARFID symptoms had a slightly thicker frontal cortex, a brain region, as mentioned, involved in executive function, as well as working memory, decision-making and impulse control. Differences between groups remained even after accounting for body mass index, meaning that across the weight spectrum, children with ARFID symptoms still present with that increased cortical thickness relative to children without ARFID symptoms. Interestingly, our other hypothesised brain regions, such as the insula, the anterior cingulate cortex, the orbitofrontal cortex and the cerebellum, were not found to be significantly different across our groups. And this suggests that brain regions, or at least these brain regions, might not play as important of a role, or an important role at all, in ARFID, relative to other feeding and eating disorders. Regarding our finding of cortical thickness, it is possible that this increased thickness of the superior frontal cortex may play a role in ARFID-related behaviours, such as food avoidance or difficulties engaging with mealtimes, but functional brain studies would really help us shed light on how the superior frontal cortex might play a role in the disorder symptomatology. However, our study was cross-sectional. This means that we only worked with the data at a singular timepoint, and as such, were unable to confidently state whether brain changes precede ARFID symptoms or whether ARFID symptoms, instead, precede the brain changes. Lastly, our findings focused on children, as well, and we do know that the brain undergoes rapid development between the ages of five to 11, particularly within the cortex and the outer layer of the brain. So, our study was conducted, again, in children aged ten and that might suggest that there might be developmental differences at play in these children, similar to other groups with neurodevelopmental differences, such as autistic children. Dr Clara Faria And just because you mentioned this, as well, one thing I found fascinating in the beginning is – was when you were talking about the similarities, especially regarding restricting behaviours, in ARFID – among ARFID and anorexia. And among all eating disorders, from a Clinician point of view, ARFID has very particular characteristics and does this also translate on the – at the neurobiological level then, do you – your finding of increased cortical thickness, is it – is exclusive to ARFID? I was just wanting to ask if the differences you’ve found are also present in other eating disorders, based on your knowledge of the literature in neurobiological findings in eating disorders? Dr Michelle Sader That’s an excellent question, especially concerning that a portion of our hypotheses were, at the very least, partially based on the literature from other restrictive eating disorders. So, there’s certainly a huge proportion of literature that’s identified differential structure, particularly smaller volumes of the superior frontal lobe and the generalised frontal lobe. However, some studies have focused on cortical thickness in this region, identifying thicker layers of the cortex in a manner similar to what our study has identified in children with ARFID. However, we do need more research to outright confirm these findings, particularly as a large proportion of these studies with anorexic participants are conducted in adolescence and adulthood, while our study was conducted in children. Some studies, particularly functional studies, have identified altered deactivation of the superior frontal gyri, therefore, less communication in those with bulimia nervosa, which has been associated with levels of conflict resolution and self-regulation in this group. We have made similar suggestions regarding the role of the superior frontal gyrus in ARFID, but of course, far more research has to be conducted to confirm or validate these associations. But interestingly, our findings have also strengthened the concept that ARFID is diagnostically distinct from other eating disorders. For instance, we identified no differences in the structure of the anterior cingulate cortex, associated with anorexia. The insula, which is associated with both anorexia and bulimia, as well as the cerebellum, which is associated with anorexia at one end of the BMI spectrum, as well as obesity on another end of that BMI spectrum. However, my thinking is we might be able to identify more specific neural correlates of ARFID symptoms by investigating these specific underlying drivers of ARFID in the future. Dr Clara Faria Yeah, that sounds fascinating, and I’m just asking myself, as you said, because you looked at children aged ten, and we know that the brain goes through so much change and plasticity in those earlier years, that if maybe you took patients, for example, that have been living with ARFID for a while, I don’t know, maybe you will find something different. But yeah, it’s, like, really fascinating research, and going back to your paper, so in the discussion section and also in our chat, you mentioned that there is a co-occurringness of ARFID and other neurodevelopmental disorders, such as autism, and also there has been extensive research published regarding neuroanatomical correlates of autism. So, I wanted to ask you, do autism and ARFID share neuroanatomical similarities based on your findings, or is this an area of research that you intend to explore in the future, as well? Dr Michelle Sader Again, another excellent question. So, the short version of this answer and the simple version of this answer is yes. The long answer requires a lot more nuance and a lot more reflection. So, one thing that I think is very important to state is that navigating the world of neuroimaging literature and autism research can be very tricky. So, firstly, the neuroimaging literature conducted with autistic participants can be quite inconsistent. As part of my own PhD, I previously conducted a meta-analysis with more than 50 papers, with all authors identifying regions of interest associated with autism. Yet, combining all of this data led to no particular region or biomarker that was associated with autistic groups. So, that raises the question as to how we’re conducting autism neuroimaging literature in the first place. A second thing I’ve noted is that I’ve recently been working with loads of autistic Researchers to reconsider the way that we approach neuroimaging literature in autistic people, and this has really raised a question that has stuck with me for quite some time now. Since so many autistic individuals struggle with the sensory environment and the noise associated with an MRI, and thus find the MRI as an imaging modality, inaccessible, it raises the question as to whether we’re really scanning a representative sample of autistic people, if only a certain part of the autistic community is able to engage with this form of research. So, those are two sticklers that have really been circulating in my mind as I’ve been, kind of, deconstructing the findings in our own study. Despite the lack of consensus, however, regarding neuroanatomical regions associated with autistic individuals, there are many studies, indeed, that have identified structural differences, particularly increased cortical thickness of the middle and superior frontal lobes in autistic people relative to non-autistic people. So, this does create a very straightforward association between autism, as well as ARFID, but again, loads more research needs to be conducted and perhaps even a reformulation of the way we approach autism research in neuroimaging modalities. What I will say is that the development of the brain in autistic, relative to non-autistic, individuals, differs and these differences seem similar to those with ARFID. So, in childhood, the cortex of autistic individuals is much thicker than non-autistic individuals, which is exactly what we found in our children with ARFID symptoms relative to those without. However, the cortex then becomes much thinner in autistic, relative to non-autistic, adults, as they transition from childhood to adulthood. So, again, we need to conduct further research to determine if developmental patterns of the cortex in autistic people might also be reflected in those with ARFID. Dr Clara Faria Wow, that sounds fascinating, and I hadn’t realised – what you just mentioned is really interesting. And I know it, like, also stems from a way bigger discussion and very important discussion about how we are including autistic people in research and how we are co-producing with them and how we still have to improve loads on that field. But yeah, like, an MRI is an incredibly uncomfortable experience for anyone, let alone for an autistic person with sensory limitations, so yeah, that’s actually a great point. You’re – we might be selecting – I mean, I particularly don’t work with MRIs. I work with epidemiology and secondary data analysis, but yeah, oh, my God, that’s a hu – but that’s a huge point you’ve just made there and it’s… Dr Michelle Sader Absolutely. Dr Clara Faria Yeah. Dr Michelle Sader And what I will say, is I’ll admit this myself, that I have undergone an MRI for volunteering in peers research, I will say I absolutely fell asleep. So, it is a very variable experience. It can be very uncomfortable for some people, but also right at home for others, strangely enough. Dr Clara Faria And it takes such a long time, as well, ‘cause it’s – I mean, especially if we’re talking about – I mean, that’s way too technical, but if we’re talking about MRIs for research, usually they use, like, a sev – that’s all your area, right? But I know that sometimes the guys use a very high Tes – like, a seven Tesla scan or something like that, and it takes a huge time to acquire the image. I’ve – the only time I’ve done an MRI was as a volunteer, as well, for, like, a research study and, yeah, and I remember being there thinking, oh, my God, it takes a long time. Dr Michelle Sader Absolutely. Dr Clara Faria Yeah, well, thanks for fascinating answer, and you gave us just so much food for thought. And to wrap up, I wanted to ask you, where would you like to see your field go next, and is there any particular area of ARFID research that you think we need more studies on? Dr Michelle Sader You’ve asked a golden question. So, when it comes to future insights in the field, there is such an extensive list of possibilities that we could explore. So, from my perspective, this study serves as us simply getting our foot in the door when it comes to ARFID research, rather than sitting back and patting ourselves on the back. So, first and foremost, I think it’s incredibly important to conduct future research that reflects that diverse nature of ARFID symptoms. I think that the three underlying drivers of ARFID, as presented by the DSM-5, can manifest incredibly differently and might even comprise of different communities or populations. There is a very real possibility that we’ve only captured brain regions associated with all three drivers, but that there may be very different or unique drivers associated. There is a very real possibility that we’ve only captured brain regions associated with all three drivers, but that there may be very different or unique regions associated with these respective drivers. Importantly, ARFID does occur across the weight and the age spectrum. So, I think we could really improve our understanding of ARFID by tackling whether changes in the brain are unique according to weight or BMI, as well as across adults and children. Importantly, longitudinal studies investigating brain structure across human development would allow us to determine whether ARFID symptoms precede, or rather, are consequent to, changes in brain structure. There’s even potential to understand the permanence of ARFID and our ability to provide useful interventions. There are many children that are able to outgrow ARFID symptoms, but there is a group of people in which their symptoms of ARFID are lifelong, persistent and very difficult to overcome. One thing that is important to acknowledge, as with many fields of research, is to highlight under-represented groups in future ARFID research. So, in many reviews and meta-analyses that I have come across, the majority of ARFID research is predominantly conducted with white participants. So, as Researchers and Academics, we need to continue to do our best in promoting ethical and representative sampling of our populations and participants, to hopefully reduce this divide. Dr Clara Faria Well, and on that very cheerful note and inspiring note, as well, I would like to thank you so much, Dr Sader, for sharing your research with us today. And if you want to know more about Dr Sader’s work, including her paper, you can visit ACAMH website, which is www.acamh.org and we will also link her profile and the University of Aberdeen in the page of this podcast. So, if you want to have a look at her other research work, you can also go there. ACAMH is spelt A-C-A-M-H. Do keep an eye out for other podcasts in the In Conversation series, and don’t forget to follow us on your preferred streaming platform and do share with friends and colleagues. Thank you so much for the listen.