A proof-of-concept study applying machine learning methods to putative risk factors for eating disorders: results from the multi-centre European project on healthy eating.

BACKGROUND: Despite a wide range of proposed risk factors and theoretical models, prediction of eating disorder (ED) onset remains poor. This study undertook the first comparison of two machine learning (ML) approaches [penalised logistic regression (LASSO), and prediction rule ensembles (PREs)] to conventional logistic regression (LR) models to enhance prediction of ED onset and differential ED diagnoses from a range of putative risk factors. METHOD: Data were part of a European Project and comprised 1402 participants, 642 ED patients [52% with anorexia nervosa (AN) and 40% with bulimia nervosa (BN)] and 760 controls. The Cross-Cultural Risk Factor Questionnaire, which assesses retrospectively a range of sociocultural and psychological ED risk factors occurring before the age of 12 years (46 predictors in total), was used. RESULTS: All three statistical approaches had satisfactory model accuracy, with an average area under the curve (AUC) of 86% for predicting ED onset and 70% for predicting AN v. BN. Predictive performance was greatest for the two regression methods (LR and LASSO), although the PRE technique relied on fewer predictors with comparable accuracy. The individual risk factors differed depending on the outcome classification (EDs v. non-EDs and AN v. BN). CONCLUSIONS: Even though the conventional LR performed comparably to the ML approaches in terms of predictive accuracy, the ML methods produced more parsimonious predictive models. ML approaches offer a viable way to modify screening practices for ED risk that balance accuracy against participant burden.

Effects of Intranasal Oxytocin on the Interpretation and Expression of Emotions in Anorexia Nervosa.

Altered social-emotional functioning is considered to play an important role in the development and maintenance of anorexia nervosa (AN). Recently, there has been increasing interest in investigating the role of intranasal oxytocin in social-emotional processing. The present study aimed to investigate the effects of intranasal oxytocin on the interpretation and expression of emotions among people with AN. Thirty women with AN and 29 age-matched healthy women took part in the present study, which used a double-blind, placebo-controlled, cross-over design. The participants received a single dose of 40 IU of intranasal oxytocin in one session and a placebo spray in the other. Fifteen minutes after administration, the participants completed the Reading the Mind in the Eyes Test to assess the interpretation of complex emotions and mental states followed by a video task, which assessed expressions of facial affect when they were viewing humorous and sad film clips. The intranasal oxytocin did not significantly influence the expression or interpretation of emotions in the AN or healthy comparison groups. The AN group expressed significantly less positive emotion, spent more time looking away and reported experiencing a significantly more negative affect in response to the film clips. The finding that intranasal oxytocin had little to no effect on the interpretation or expression of emotions in either group supports the notion that the effects of oxytocin on social-emotional processing are not straightforward and may depend on individual and environmental differences, as well as the emotion being processed. Replication of these findings is necessary to explore the effect of timing on the effects of oxytocin before firm conclusions can be drawn. Nonetheless, these findings add to the steady accumulation of evidence that people with AN have reduced emotional expression and avoidance of emotionally provoking stimuli.

The set-shifting profiles of anorexia nervosa and autism spectrum disorder using the Wisconsin Card Sorting Test: a systematic review and meta-analysis.

Difficulties in set-shifting are commonly reported in both autism spectrum disorder (ASD) and anorexia nervosa (AN) populations. Despite this, it is not known whether this cognitive profile persists across different ages, or whether the profiles seen in ASD and AN are comparable. This systematic review and meta-analyses aimed to compare the set-shifting profiles, as measured by the Wisconsin Card Sorting Test (WCST) in adults and younger people with either ASD or AN, relative to healthy controls (HCs) and to statistically compare performance on the WCST between ASD and AN. In all, 24 studies on ASD and 22 studies on AN were identified. In ASD, there were significant differences between the clinical group and HCs, with the ASD group making significantly more perseverative errors, indicating greater difficulty in set-shifting [pooled effect size of d = 0.67, 95% confidence interval (CI) 0.53-0.81, p ⩽ 0.001]. This effect was consistent across the age span. For AN studies, there was a significant difference between adults with AN and HCs (d = 0.52, 95% CI 0.36-0.68, p ⩽ 0.001) but a non-significant effect in child studies (d = 0.25, 95% CI -0.05 to 0.55, z = 1.66, p = 0.096). Meta-regression indicated no effect of diagnosis (AN or ASD) on performance in adult studies but there was a non-significant trend (p = 0.053) towards children with ASD performing worse than children with AN. While difficulties with set-shifting appear to be stable in ASD, there may be differences between children and adults with AN, which warrant further investigation.

Facial expression to emotional stimuli in non-psychotic disorders: A systematic review and meta-analysis.

Facial expression of emotion is crucial to social interaction and emotion regulation; therefore, altered facial expressivity can be a contributing factor in social isolation, difficulties with emotion regulation and a target for therapy. This article provides a systematic review and meta-analysis of the literature on automatic emotional facial expression in people with non-psychotic disorders compared to healthy comparison groups. Studies in the review used an emotionally salient visual induction method, and reported on automatic facial expression in response to congruent stimuli. A total of 39 studies show alterations in emotional facial expression across all included disorders, except anxiety disorders. In depression, decreases in facial expression are mainly evident for positive affect. In eating disorders, a meta-analysis showed decreased facial expressivity in response to positive and negative stimuli. Studies in autism partially support generally decreased facial expressivity in this group. The data included in this review point towards decreased facial emotional expressivity in individuals with different non-psychotic disorders. This is the first review to synthesise facial expression studies across clinical disorders.

Alterations in brain structure in adults with anorexia nervosa and the impact of illness duration.

BACKGROUND: Brain structure alterations have been reported in anorexia nervosa, but findings have been inconsistent. This may be due to inadequate sample size, sample heterogeneity or differences in methodology. METHOD: High resolution magnetic resonance images were acquired of 33 adult participants with anorexia nervosa and 33 healthy participants, the largest study sample to date, in order to assess whole-brain volume, ventricular cerebrospinal fluid, white matter and grey matter volume. Voxel-based morphometry was conducted to assess regional grey matter volume. Levels of depression, anxiety, obsessionality and eating disorder-related symptoms were measured and used to explore correlations with brain structure. RESULTS: Participants with anorexia nervosa had smaller brain volumes as well as a global decrease in grey matter volume with ventricular enlargement. Voxel-based morphometry revealed a decrease in grey matter volume spanning across the cerebellum, temporal, frontal and occipital lobes. A correlation was found between grey matter volume loss and duration of illness in the cerebellum and mesencephalon. No correlations were found with clinical measures. CONCLUSIONS: Findings are in accordance with several previous studies on brain structure and match functional studies that have assessed the symptomatology of anorexia nervosa, such as body image distortion and cognitive bias to food. The correlation with duration of illness supports the implication of cerebellar atrophy in the maintenance of low weight and disrupted eating behaviour and illustrates its role in the chronic phase of anorexia nervosa. The lack of other correlations suggests that these findings are not related to the presence of co-morbid disorders.

Functional MRI investigation of verbal working memory in adults with anorexia nervosa.

Literature regarding verbal working memory (vWM) in anorexia nervosa (AN) has been inconsistent due to a misunderstanding of the key components of vWM and introduction of confounding stimuli. Furthermore, there are no studies looking at how brain function in people with AN relates to vWM performance. The present study used functional magnetic resonance imaging (fMRI) with a letter n-back paradigm to study the effect of increasing vWM task difficulty on cortical functioning in the largest AN sample to date (n=31). Although the AN group had low BMI and higher anxious and depressive symptomology compared to age-matched controls (HC), there were no between-group differences in accuracy and speed at any task difficulty. fMRI data revealed no regions exhibiting significant differences in activation when groups were compared at each difficulty separately and no regions showing group x condition interaction. Although there was a trend towards lower accuracy as duration of illness increased, this was not correlated with activity in regions associated with vWM. These findings indicate that vWM in AN is as efficient and performed using the same cognitive strategy as HC, and that there may not be a need for therapies to pursue remediation of this particular neurocognitive faculty.

Emotion recognition and emotional theory of mind in chronic fatigue syndrome.

BACKGROUND: Difficulties with social function have been reported in chronic fatigue syndrome (CFS), but underpinning factors are unknown. Emotion recognition, theory of mind (inference of another's mental state) and 'emotional' theory of mind (eToM) (inference of another's emotional state) are important social abilities, facilitating understanding of others. This study examined emotion recognition and eToM in CFS patients and their relationship to self-reported social function. METHODS: CFS patients (n = 45) and healthy controls (HCs; n = 50) completed tasks assessing emotion recognition, basic or advanced eToM (for self and other) and a self-report measure of social function. RESULTS: CFS participants were poorer than HCs at recognising emotion states in the faces of others and at inferring their own emotions. Lower scores on these tasks were associated with poorer self-reported daily and social function. CFS patients demonstrated good eToM and performance on these tasks did not relate to the level of social function. CONCLUSIONS: CFS patients do not have poor eToM, nor does eToM appear to be associated with social functioning in CFS. However, this group of patients experience difficulties in emotion recognition and inferring emotions in themselves and this may impact upon social function.

Neurocognition in bulimic eating disorders: a systematic review.

UNLABELLED: Van den Eynde F, Guillaume S, Broadbent H, Stahl D, Campbell IC, Schmidt U, Tchanturia K. Neurocognition in bulimic eating disorders: a systematic review. OBJECTIVE: The aim of this study was to review the literature on neurocognition comparing people with a bulimic eating disorder in the acute phase of the illness with healthy controls (HC). METHOD: The review follows the PRISMA (preferred reporting items for systematic reviews and meta-analysis) statement guidelines. Three databases (Medline, Web of Science, and Scopus) were searched combining the search terms 'bulimic disorder', 'bulimia nervosa (BN)', 'binge-eating disorder (BED)' with terms referring to cognitive domains (e.g. 'executive functions'). RESULTS: Thirty-seven studies on people with BN and four on people with BED were selected for review. Overall, sample sizes were relatively small [bulimic disorders: median and range 22 (12-83); HC: 27 (13-172)]. The diversity in methodology precluded a meta-analytical approach. People with a bulimic disorder did not present with a clear neurocognitive profile. Inclusion of salient, disorder-related stimuli (e.g. body weight/shape words) in the neurocognitive paradigms tended to generate differences between people with a bulimic disorder and HC. CONCLUSION: Neurocognition in bulimic eating disorders is under researched, and the available evidence is inconclusive. This review outlines strategies for further research in this area.

Insight in eating disorders: clinical and cognitive correlates.

BACKGROUND: The aim of this study was to explore the extent of lack of insight and its components in eating disorders (EDs) and to investigate the relationship between insight and clinical and cognitive characteristics in this group. METHOD: Seventy-five participants were enrolled in the study: 25 with anorexia nervosa (AN), 15 with bulimia nervosa (BN) and 35 healthy controls (HC). Insight was assessed with a modified version of the Schedule for the Assessment of Insight for EDs (SAI-ED) and multi-dimensional scaling (MDS) analysis was used to clarify the internal structure of the scale. Neuropsychological tests included the Trail Making Test (TMT), the Brixton Test and a Verbal Fluency Task. RESULTS: Only a subgroup of AN patients (24%) had severe impairment of insight. Patients with the restricting type of AN (AN-R) had poorer overall insight than patients with the binge-purge type of the disorder (AN-B/P). More of the ED patients displayed a deliberate denial of illness rather than a lack of awareness of the illness. A regression model revealed that only performance in part B of the TMT (TMT-B) was a moderate predictor of insight level. No association was found between insight and other cognitive or clinical variables. CONCLUSIONS: Impaired insight is a significant feature of some ED patients. Insight in EDs seems to be partially dependent on intact mental flexibility.

The socio-emotional processing stream in Anorexia Nervosa.

The significance of socio-emotional factors in development and maintenance of Anorexia Nervosa (AN) has been noted, but the literature is poorly integrated without clear models guiding research or treatment. This systematic review retrieved experimental studies of social-cognitive or affective processing in AN and categorised them using Ochsner's "Social-Emotional Processing Stream." Ochsner's "Processing Stream", based on healthy data, comprises five constructs: (1) acquisition of and (2) recognition and response to social-affective stimuli, (3) low-level and (4) high-level mental state inference and (5) context-sensitive emotion regulation. Thirty-seven experimental studies in Anorexia Nervosa were identified, mapping on to four of the five constructs (not Construct 3). A meta-analysis of nine affect recognition studies was conducted. AN patients demonstrated impairments in each of the four domains with preliminary reports that some difficulties are trait-like, and others ameliorate following recovery. Socio-emotional data was integrated with previous reports of neural abnormalities to generate an AN specific model of socio-emotional processing. Additional research is required for further definition and to translate experimental findings into clinical practice.

Emotional functioning in eating disorders: attentional bias, emotion recognition and emotion regulation.

BACKGROUND: Interpersonal processes, anxiety and emotion regulation difficulties form a key part of conceptual models of eating disorders (EDs), such as anorexia nervosa (AN) and bulimia nervosa (BN), but the experimental findings to support this are limited. METHOD: The Reading the Mind in the Eyes task, the Difficulties in Emotion Regulation Scale (DERS) and a computerized pictorial (angry and neutral faces) Stroop task were administered to 190 women [50 with AN, 50 with BN and 90 healthy controls (HCs)]. RESULTS: Those with an ED showed attentional biases to faces in general (medium effect), but specifically to angry faces over neutral faces (large effect) compared to HCs. The ED group also reported significantly higher emotion regulation difficulties (large effect) than HCs. There was a small difference between the ED and HC groups for the emotion recognition task (small-medium effect), particularly in the restricting AN (RAN) group. Depression and attentional bias to faces significantly predicted emotion regulation difficulties in a regression model. CONCLUSIONS: The data provide support for conceptualizations of EDs that emphasize the role of emotional functioning in the development and maintenance of EDs. Further research will concentrate on exploring whether these findings are state or trait features of EDs.

Cognitive remediation therapy (CRT) for anorexia in group format: a pilot study.

Cognitive Remediation Therapy (CRT) was designed to address the inflexible and detail focused thinking styles frequently reported in individuals with Anorexia Nervosa (AN). This paper reports the development and acceptability of a pilot of CRT for AN in group format. Four group sessions were designed and piloted with 30 patients in an eating disorder service. The voluntary drop out rate was 20%. Outcome measures targeting patients' self-reported cognitive flexibility, self esteem and motivation to change were completed in the first and last sessions. Statistically significant gains were found in self reported ability to change (p=0.03). Both patients and group facilitators found the group acceptable, useful and a positive experience.

Multidimensional self reports as a measure of characteristics in people with eating disorders.

This study used multidimensional self report assessments to measure perfectionism, impulsivity and obsessive compulsive characteristics in females with anorexia nervosa (AN), bulimia nervosa (BN) and in matched healthy controls (HC). The Frost Multidimensional Perfectionism Scale (FMPS), Barrett Impulsivity Scale (BIS) and Obsessive Compulsive Inventory-Revised (OCI-R) scale were completed by 107 participants (AN=30, BN=26, HC=51), in parallel with clinical measures. Results show that people with AN have the highest scores on the dimensions of the FMPS as well as on the overall score; the AN and BN groups have the highest scores on the dimensions and on the overall score of the OCI-R; on the BIS, the AN and BN groups have the highest scores on the attention subscale, but there are no group differences on the overall BIS scores. In relation to the FMPS, the global score, and the subscales 'concern over mistakes' and 'doubts about actions' are all highly correlated with both eating pathology (Eating Disorder Examination Questionnaire, EDE-Q) and low global functioning (Structured Clinical Interview for DSM IV, SCID). The subscale 'obsessing' on the OCI-R shows a strong correlation with eating pathology. The overall score and also the subscales of the BIS do not show strong correlations with eating pathology or poor global functioning. In conclusion, therapies should seek to address these specific areas which are highly correlated with eating disorder pathology.

Comparing views of patients, parents, and clinicians on emotions in anorexia: a qualitative study.

Patients with anorexia nervosa (AN) may experience difficulties in emotional processing that can adversely affect treatment and maintenance of the illness. Focus groups or questionnaires were undertaken with patients with AN, parents and clinicians, with the aim to explore the most salient issues pertaining to emotions and social cognition in AN. Qualitative thematic analysis was used to analyse the data. Seven primary themes were identified showing congruence across groups: 'emotional awareness and understanding'; 'emotional intolerance'; 'emotional avoidance'; 'emotional expression and negative beliefs'; 'extreme emotional responses'; 'social interactions and relationships'; and 'lack of empathy'. Clinical and empirical implications are discussed.

Lifetime course of eating disorders: design and validity testing of a new strategy to define the eating disorders phenotype.

BACKGROUND: Aetiological studies of eating disorders would benefit from a solution to the problem of instability of eating disorder symptoms. We present an approach to defining an eating disorders phenotype based on the retrospective assessment of lifetime eating disorders symptoms to define a lifetime pattern of illness. We further validate this approach by testing the most common lifetime categories for differences in the prevalence of specific childhood personality traits. METHOD: Ninety-seven females participated in this study, 35 with a current diagnosis of restricting anorexia nervosa, 32 with binge/purging subtype of anorexia nervosa and 30 with bulimia nervosa. Subjects were interviewed by a newly developed EATATE Lifetime Diagnostic Interview for a retrospective assessment of the lifetime course of eating disorders symptoms and childhood traits reflecting obsessive-compulsive personality. RESULTS: The data illustrate the extensive instability of the eating disorders diagnosis. Four most common lifetime diagnostic categories were identified that significantly differ in the prevalence of childhood traits. Perfectionism and rigidity were more common in groups with a longer duration of underweight status, longer episodes of severe food restriction, excessive exercising, and shorter duration of binge eating. CONCLUSIONS: The assessment of lifetime symptoms may produce a more accurate definition of the eating disorders phenotype. Obsessive-compulsive traits in childhood may moderate the course producing longer periods of underweight status. These findings may have important implications for nosology, treatment and future aetiological studies of eating disorders.

Brain-derived neurotrophic factor (BDNF) and set-shifting in currently ill and recovered anorexia nervosa (AN) patients.

BACKGROUND: Studies of patients with anorexia nervosa (AN) have shown that they do not perform well in set-shifting tasks but little is known about the neurobiological correlates of this aspect of executive function. The aim of this study was to measure serum brain-derived neurotrophic factor (BDNF) and to establish whether set-shifting difficulties are present in people with current AN and in those recovered from AN, and whether serum BDNF concentrations are correlated with set-shifting ability. METHOD: Serum BDNF concentrations were measured in 29 women with current AN (AN group), 18 women who had recovered from AN (ANRec group) and 28 age-matched healthy controls (HC group). Set-shifting was measured using the Wisconsin Card Sorting Test (WCST). Eating-related psychopathology and depressive, anxiety and obsessive-compulsive symptomatology were evaluated using the Eating Disorder Examination Questionnaire (EDEQ), the Hospital Anxiety and Depression Scale (HADS), and the Maudsley Obsessive-Compulsive Inventory (MOCI) respectively. RESULTS: Serum BDNF concentrations (mean+/-s.d.) were significantly lower in the AN group (11.7+/-4.9 ng/ml) compared to the HC group (15.1+/-5.5 ng/ml, p=0.04) and also compared to the ANRec group (17.6+/-4.8 ng/ml, p=0.001). The AN group made significantly more errors (total and perseverative) in the WCST relative to the HC group. There was no significant correlation between serum BDNF concentrations and performance on the WCST. CONCLUSIONS: Serum BDNF may be a biological marker for eating-related psychopathology and of recovery in AN. Longitudinal studies are needed to explore possible associations between serum BDNF concentrations, illness and recovery and neuropsychological traits.

A pilot study exploring Machiavellianism in anorexia nervosa.

Machiavellianism has not previously been systematically explored in the eating disorders. This personality dimension involves the social cognitive processes required to identify others' intentions and the use of strategies to deceive and manipulate others. Individuals suffering with Anorexia Nervosa (AN) often try to conceal their disorder through the use of deceptive strategies. In this study we explore Machiavellianism in 22 female AN patients and 41 female healthy controls (HCs). AN patients and HCs did not differ significantly in their self-reported Machiavellianism, but both groups did score highly in this domain. Possible explanations for these results are discussed.

Central coherence in eating disorders: a systematic review.

BACKGROUND: This review systematically appraised the research evidence for local versus global information processing to test the hypothesis that people with eating disorders (ED) had weak central coherence. METHOD: Searches on Medline, EMBASE, PsycINFO and ISI Web of Science databases were conducted in November 2006 and subsequently updated in September 2007. Each search was conducted in two steps: (1) neuropsychological tasks measuring central coherence and (2) words related to cognitive functioning in eating disorders. Data were summarized in a meta-analysis if the number of studies for a given test was >5. RESULTS: Data were extracted from 16 studies. Meta-analyses were conducted for four tasks obtaining moderate effect sizes. The majority of studies found global processing difficulties across the ED spectrum. The results are less clear regarding local processing. CONCLUSIONS: People with ED have difficulties in global processing. It is less certain as to whether they have superior local processing. Currently, there is insufficient evidence to refute the weak central coherence hypothesis.

The inability to ignore: distractibility in women with restricting anorexia nervosa.

BACKGROUND: Attentional difficulties reported in individuals with anorexia nervosa (AN) may be due to preferential processing of disease-salient stimuli at a pre-attentive or at a conscious level or to a general problem in attention. Attentional difficulties may be associated with duration of illness. METHOD: Female participants with AN (restricting subtype; n=24) and healthy comparison women (n=24) were randomly allocated to subliminal or supraliminal exposure to visual stimuli (food, neutral and aversive images) while performing the 1-back and 2-back working-memory tasks. RESULTS: Participants with AN made fewer errors than the healthy comparison group in the subliminal condition but significantly more errors in the supraliminal condition [condition x group interaction, F(1, 44)=6.82, p<0.01]: this was irrespective of stimulus type (food, neutral and aversive) and task (1-back or 2-back). The total number of errors made correlated positively with the duration of the AN for both the 1-back task (rs=0.46, p<0.05) and for the 2-back task (rs=0.53, p<0.01). CONCLUSIONS: Decreased ability to concentrate in the presence of explicit distracters is a feature of AN and is associated with longer duration of illness. This phenomenon could be addressed in psychological interventions.

Can cognitive exercises help treat anorexia nervosa?

Cognitive remediation therapy (CRT) is used as an intervention for people with brain lesions and psychosis. This case report demonstrates the possible benefits of introducing CRT into treatment packages for anorexia nervosa (AN). In our previous work, we reported that people with AN demonstrate inflexibility in cognitive set-shifting tasks. Weight gain alone does not improve the neuropsychological profile in set-shifting tasks. This case report illustrates how training programmes can address problems in cognitive rigidity. We acknowledge the limitations of case studies, however, this is a starting point in exploring the possibilities of introducing CRT as part of the treatment of AN.