Eating Concerns and Compulsivity

This study uses a computational task to examine differences in adaptive learning to both rewards and punishments between three groups: those who have recovered from anorexia nervosa, those who score highly on the EAT-26 (Eating Attitudes Test - 26 item version; an eating disorder symptom scale), and healthy controls. This task also allows the examination of pupil response (thought to reflect norepinephrine activity) in response to expected and unexpected wins and losses.

This study involves using a novel computational task (the volatility task, designed by Dr Michael Browning) to examine differences in adaptive learning in terms of sensitivity to environmental change in those who are in eating disorder 'risk' groups (defined as those with a previous diagnosis of AN, and those who score highly in the EAT-26 for eating disorder symptoms. This study allows us to investigate whether or not these individuals are able to pick up key environmental statistics and adapt their behaviour accordingly. We hypothesise that those in eating disorder risk groups will show a deficit in this area, which might begin to explain why the cognitive phenotype of 'cognitive inflexibility' is found so commonly in these patients. Using pupillometry measures will also allow us to putatively form links between this behaviour and the norepinephrine system in these participants, as pupil dilation measures are thought to track environmental statistics of this kind. Additionally, this task allows us to identify whether there is a particular deficit in tracking and learning about positive or negative environmental information. We will be using standard clinical interviews and questionnaires to define the groups and to record key variables (e.g. mood information) within groups. This study will consist of a single visit, including these interviews and questionnaires, the volatility task with pupillometry measures, and the Wisconsin Card Sort Task, which we hope to use to demonstrate a baseline difference between groups on cognitive flexibility.

Three parallel groups: those who have recovered from anorexia nervosa; those who worry a lot about eating, shape and weight; and healthy control participants will be included, and all will complete some questionnaires and two tasks.

Those who have a past diagnosis of AN (defined by the Diagnostic and Statistical Manual of Mental Disorders-5 (DSM-5) criteria) but are currently recovered, as shown by BMI over 18.5 throughout the last 12 months (self-report and current weight measured). Defined as either 'fully recovered'and: score must be within the 'normal range' of the Eating Disorders Examination Questionnaire (EDE-Q) global mean scores for young women, below 20 on the EAT-26 and below 16 on the Clinical Impairment Assessment for Eating Disorders (CIA); or partially recovered where one or more of these scores may be above the above-mentioned cutoffs.

Those who score above 20 on the EAT-26, but who do not declare a former diagnosis of an eating disorder (though they may meet criteria for a current diagnosis during the Structured Clinical Interview for the DSM-5).

No history of or current diagnosis of any psychiatric disorder (especially eating disorders) which could impact study results.

Difference between eating disorder risk groups and healthy controls in extent to which learning rate difference between win-volatile and loss-volatile blocks changes.

Difference in relative inverse logit learning rate (alpha) for the volatile versus stable blocks between groups.

Whether there is a difference in the learning rate for different valence environmental information (positive vs. negative) across groups.

To compare changes in learning rate across blocks for reward vs. punishment information across groups.

Examine whether post-outcome pupil dilation tracks environmental volatility and outcome surprise to the same extent across groups.

Correlation between relative log learning rate (alpha) change between blocks and eating disorder symptom scores on the Eating Attitudes Questionnaire - 26 item version

The EAT-26 is a questionnaire which measures eating disorder symptoms. The total score will be used (summing of individual items). Lower scores represent lower presence of eating disorder symptoms.

Correlate relative log learning rate and beta size (an inverse temperature parameter) with perseverative errors on the Wisconsin Card Sort task across groups

Inclusion Criteria: Participant is willing and able to give informed consent for participation in the study. Females aged 18 to 45 years. BMI over 18.5 and has remained so for the last year. Participant is a fluent English speaker. Inclusion: recovered from anorexia Past formal diagnosis of AN (defined by DSM-5 criteria). Fully recovered: Score must be below 2.767 on the EDE-Q, below 16 on the CIA, and below 20 on the EAT-26 or partially recovered: scores may be above 2.767, 16 and 20 respectively. Inclusion: high scoring on EAT-26 Score above 20 on the EAT-26 questionnaire. Inclusion: healthy control Score below 2.767 on the EDE-Q, below 16 on the CIA, and below 20 on the EAT-26. Exclusion Criteria: Any current diagnosis of a psychiatric disorder which in the investigator's opinion could impact study results (e.g. significant depression, anxiety or OCD). Any current psychotropic medications. Current regular cigarette smoking of over 5 cigarettes per day. Recent use of illicit drugs. Alcohol intake which indicates an element of alcohol abuse; or unwillingness to refrain from drinking the night before the study visit. Exclusion for high scoring EAT-26 • A former formal diagnosis of an eating disorder. Note that in this group a current diagnosis of EDNOS will not be an exclusion criterion.