The WIN Mood Disorders Focus Day will take place on the 26th of March between 9am and 5pm in the Richard Doll building (note venue change). Lunch and refreshments will be provided.
Registration for the day is still open: oxfordxpsy.az1.qualtrics.com/jfe/form/SV_a4ZukVf3CJG9Y0t.
The day will involve data talks, short oral presentations, posters and a discussion session on how best to facilitate the translation of pre-clinical work so that we can ask clinically interesting questions.
Theme1: Physiology, big data and neural dynamics.
Chair: Matthew Rushworth
Serotonergic influences on emotional learning
The serotonin transporter regulates synaptic serotonin availability and drugs targeting the transporter (e.g. SSRIs) remain the first line treatment for anxiety and depression. Perhaps surprisingly, recent human genetic studies have failed to find convincing evidence implicating serotonin transporter-related genes on the aetiology of anxiety or depression. In contrast, studies in rodents have reliably demonstrated increased anxiety / depression-like phenotypes arising from genetic knock-out of the serotonin transporter, and reduced anxiety / depression-like phenotypes when the serotonin transporter is over-expressed. Here I will describe some of this animal research, focusing on how genetically-altering serotonin transporter expression or blocking the transporter with drugs influences emotional learning as well as activity in brain structures such as the amygdala and hippocampus. Collectively, these data support the idea that increasing serotonin availability enhances learning for emotionally-relevant events.
Relating markers of mental well-being to specific amygdala connections in humans
There has been increasing interest in using neuroimaging measures to predict psychiatric disorders. However, predictions usually rely on large numbers of brain connections, thus lacking anatomical specificity and limiting possibilities for targeted interventions, and there is large disorder heterogeneity. We addressed both challenges using resting-state functional MRI (rs-fMRI) and behavioural measures from the Human Connectome Project. First, we parcellated the amygdala, a key region implicated in mood disorders, into seven nuclei using rs-fMRI. Next a factor analysis on a large number of questionnaire scores provided four latent behaviours that captured, at sub-clinical levels, sub-problems frequently found in anxious-depressive individuals, such as negative emotions and sleep problems. Finally, for each latent behaviour, we identified the most predictive connections between individual amygdala nuclei and regions of interest e.g. in brainstem and medial prefrontal cortex. A small number of connections (<8) was in each case sufficient to predict the latent behaviour, providing unprecedented levels of specificity, in humans, for relating mental well-being to precise anatomical connections.
Theme 2: Dopamine, apathy, anhedonia and motivation.
Chair: Susannah Murphy
Action and reward interactions in mesolimbic dopamine
There is broad consensus that the activity of midbrain dopaminergic neurons and downstream dopamine release in the nucleus accumbens (NAc core) correlate with a reward prediction error. Yet there is also evidence that mesolimbic dopamine release and the activation of dopaminergic receptors on ventral striatal medium spiny neurons may play a causal role in the initiation of goal-directed action. To understand how action and reward interacts in the mesolimbic dopamine system, I have used fast-scan cyclic voltammetry and targeted pharmacological manipulations in conjunction with a novel behavioural task that varies both action requirements and reward size on offer. I will discuss findings that demonstrate a role for action initiation in canonical prediction error signalling, as well as the importance of dopamine D1-receptor activation in the NAc core for in cue-driven action.
Apathy in Small Vessel Disease: a multimodal investigation.
Clinical apathy is a common, debilitating syndrome that occurs across a multitude of conditions, including cerebrovascular small vessel disease (SVD), where it affects more than a third of patients. Despite its high prevalence, little is known about the mechanisms underlying apathy in SVD. Here we used an approach that combines relatively novel effort-based decision making tasks, that allow better behavioural phenotyping of the syndrome, with multimodal neuroimaging . Our preliminary findings suggest a decrease in reward sensitivity in SVD patients with clinical apathy compared to those without apathy. They were less inclined to invest effort for low rewards. Importantly, this effect was independent of depression, despite a significant correlation between apathy and depression in this group. Interestingly, when apathetic patients did choose to exert effort they took a significantly longer to make decisions that involve cost-benefit evaluation than their motivated counterparts. These findings in SVD are consistent with findings from investigations of apathy in other conditions, such as Parkinson’s disease, lending support to common underlying brain mechanisms across clinical conditions with different pathologies.
Constructing value in the medial prefrontal cortex
How do we know what is valuable and what is not? In most standard models, value is simply emitted by the environment and can be learned by a decision-maker through reinforcement. However, in many real-world scenarios value needs to be actively constructed from both this environmental feedback and one’s internal state, a process that is impaired in many psychiatric disorders. I will first show in two experiments how an area in the medial prefrontal cortex may support active value construction by tracking the accumulation of behaviourally relevant assets (the internal state) over time. This tracking may provide a simple proxy for reward: Value is encoded as the amount by which a choice rebalances the internal state. Finally, I will argue that a framework of active value construction may provide a better description of the specific impairments of patients with mood disorders.
Theme 3: Computational approaches to mood disorders.
Chair: Jacinta O’Shea
Generalisation of structural knowledge in the entorhinal cortex
Several mental conditions can be characterised in terms of faults of inference processes underlying the construction of mental models of the world. An important component of such models is the statistical structure of the world: the relationships between the different entities in the environment. Representing these relationships explicitly, i.e. in a manner that is divorced from the sensory particularities of the entities, is useful for generalising previous knowledge to new environments with the same relational structure – a key component of flexible behaviour. Theoretical and experimental work suggest that “grid cells”, originally discovered in the entorhinal cortex, might encode such an explicit representation of the structure of spatial tasks.
In this talk, I will present results from an fMRI study showing that the entorhinal cortex also encodes an explicit representation of the relational structure of a very different task: a stimulus-outcome decision making task with two possible correlation structures between the reward probabilities associated with stimuli. This suggests a common coding framework for task structure across a wide variety of domains.
Keeping track: environmental evaluation in human decision making
Many real-world choices such as whether to accept a job or an invitation to dinner involve weighing up the merits of an option against an expectation about whether any better options are likely to materialise any time soon. Making such a decision requires having an estimate as to how prosperous our environment is. A process of belief updating allows us to integrate new evidence about our environment as it gets better or worse and adapt choices accordingly. I will show some unpublished behavioural data collected using an online gamified version of a classic problem from foraging theory: the prey selection task. Human participants were asked to decide whether to accept or reject sequentially encountered stimuli. Choice data was best described by a computational model in which global reward rate estimates were scaled up and down according to separate learning rates. This caused estimates to update sluggishly when the environment deteriorated (causing preferences for options to perseverate) but quickly when it improved (causing preferences to change). Autonomic systems recruited under threat are believed to index the overall quality of one’s environment. In the second part of the talk, I will present findings from a recently published study conducted under controlled conditions (participants in the lab) and in the real world (firefighters on call) revealing that sensitivity to negative information is enhanced under threat. This flexibility in how beliefs are updated may provide a mechanism via which individuals are able to adapt their behaviour in response to the level of global threat present in their environment.
Large behavioural data and psychiatry research online – developing and using novel paradigms
In experimental psychiatric research in the lab, we often want to understand behavioural and neural differences between patients and controls or the effects of treatments. Ideally, we would like to use the most sensitive paradigms. However, how do we establish what is the promising paradigm before running the lab study? In this talk, I will propose the use of behavioural data collected over the internet to address this question. I will use a recent example to illustrate our pipeline where I collected a large online data set (n=400) with a wide array of psychiatric disorders and personality measures. Using such online data and a computational model I was able to measure different aspects relating to distinct features of real-life behaviour: planning for long-term goals, working towards these goals while adaptively checking that they are still worthwhile, pre-emptively avoiding situations in which one knows that ones biases will hinder achieving one’s goals. I could then relate distinct aspects of behaviour to distinct psychiatric symptoms, controlling for comorbidities.
Theme 4: Inflammation and mood.
Chair: Mark Walton
Daniel C. Anthony
The behavioural consequences of inflammation in the brain.
In animal models, peripheral inflammatory disease induces the expression of proinflammatory cytokines in the brain, and is associated with stereotypical sickness behaviours that are similar to the behaviours exhibited by individuals suffering from major depressive disorder (MDD). For example, aggression, hyperactivity, impulsivity, helplessness and anhedonia are all signs of depressive-like disorders in humans and are often reported to be present in animal models of depression induced by inflammatory challenges. Chronically stressed mice display increased levels of anxiety, an increased propensity to float in the forced swim test, and demonstrate hyperactivity under stressful lighting conditions. These changes are associated with elevated expression of tumour necrosis factor alpha (TNF) and the 5-HT transporter (SERT) in the pre-frontal cortex and further supports the view that sickness behaviours and mood disorders may be closely related at a molecular level. Interestingly, other atypical stressors such as a high-fat diet also induce the same behaviours as chronic stress and provoke cytokine expression in the brain. However, the relationship is not as simple as it seems; it has become clear that both stress and inflammatory challenges induce distinct molecular and behavioural changes in the brains of rodents and the behaviours do not summate. The combination of a sub clinical inflammatory response with chronic mild stress exacerbates depressive behaviours, but inhibits aggressive behaviours. Understanding the interplay between stress, infection, and diet on the production of cytokines in the brain is essential if inflammation-targeted therapy is to become part of the therapeutic repertoire to treat MDD.
Riccardo De Giorgi
Fantastic treatments and where to find them: exploring the links between resistant depression and inflammation
Are depression and inflammation interconnected conditions? If so, what treatment strategies could be developed to break this vicious circle?
This talk will briefly cover the neurobiological underpinnings of the links between resistant depression and inflammation, consider their clinical implications, and discuss potential research avenues in this area.
Changes in emotional processing as a putative biomarker of antidepressant response
Among the most important research topics in depression is identification of treatment response biomarkers. Currently, often many attempts are necessary to find an effective treatment, which translates into delays in response and an increased burden of depression. If response biomarkers are available, an informed decision could be made about treatment most likely to lead to a symptomatic improvement in an individual patient in the shortest possible time. Functional neuroimaging has played a key role in elucidating neural substrates of treatment response and has been a useful tool in the search for biomarkers of antidepressant response.
One of the key features of depression is negative affective bias. Antidepressant treatments have been shown to modulate its behavioural and neural correlates early in the course of treatment, before improvement in symptoms is noticeable. It was hypothesised that these early changes in emotional processing may play a crucial role in symptomatic improvement seen later in the course of treatment.
This talk will focus on the data exploring early changes in processing of emotional information at the neural level as a putative biomarker of clinical response to antidepressant treatments, and will explore the link between mechanisms of action of antidepressant drugs and response prediction.