Monday, November 26, 2012

Eating Disorders and the Not So Female Brain

Attempts to link eating disorders to a hypothetical "extreme female brain" involve some convoluted logic 
A well-known theory of psychological sex differences proposed that autism represents a manifestation of an ‘extreme male brain’ type. A recent paper has claimed that the opposite type, or ‘extreme female brain’ is manifested in eating disorder symptoms, such as anorexia and bulimia. The evidence provided in the paper for this notion seems rather mixed, especially considering that some of the results applied more clearly to males than females. Additionally, there is evidence that eating disorders and autism have certain features in common even though they are supposed to represent opposite brain types. Characterising certain conditions as extremely ‘male’ or ‘female’ based on gender stereotypes may actually create more confusion than real understanding.
Is there a distinct "female brain" type? (Image courtesy of Victor Habbick at

Autism and the ‘extreme male brain’
Simon Baron-Cohen[1] proposed that men and women have evolved different brain-types[2] specialising in either of two distinct information processing modes. The two modes are empathizing, considered to be the drive to understand other people’s thoughts and feelings in order to predict how they will behave, and systematizing, considered as the drive to understand the rules that govern a system in order to predict how the system works. Individuals have their own balance of how much they have developed either of these capacities and most people have reasonable ability in both. According to Baron-Cohen’s research, men tend to be more specialised in systematizing, whereas women tend to be more specialised in empathizing. Baron-Cohen applied this theory to understanding the nature of autism spectrum conditions including Asperger’s syndrome. People with autism tend to have social and communication difficulties due to impairments in their ability put themselves ‘in another person’s shoes’, that is, to imagine another person’s thoughts and feelings. On the other hand, autistic people often have particular strengths in understanding physical systems. For example, studies on autistic children have found that they tend to have an unusually good grasp of intuitive physics. Baron-Cohen therefore argued that autistic people have a profile of well develop systematizing and poorly developed empathizing. He referred to this profile as representing the ‘extreme male brain’ on the grounds that autism occurs many times more frequently in males than females.

Mr Spock: someone with an extreme male brain?
Baron-Cohen proposed that therefore an ‘extreme female brain’ type might exist. This profile would be associated with particular strength in the area of empathizing and difficulty in the area of systematizing. He argued that this profile should be more common in women than men but did not attempt to describe what such a condition might be like, although he suggested that people like this might get along well in modern society due to their well-developed people skills, as long as they could avoid dealing with modern technology. 
Evidence for an ‘extreme female brain’?
Bremser and Gallup (2012) proposed that eating disorders are a manifestation of an ‘extreme female brain’ (EFB) that is a ‘mirror image of autism’. They justify this on the grounds that eating disorders are much more prevalent in females than males and cite evidence linking eating disorders to sex hormones. Additionally, they claim that fear of negative evaluation and social anxiety play an important role in the development of eating disorders. They argue that fear of negative evaluation and social anxiety can be linked to high levels of empathizing. Therefore, high empathizing (a feature of the proposed EFB) could predispose a person to eating disorders by making them vulnerable to fear of negative evaluation and social anxiety.
Their argument for the connection between empathizing and social anxiety is that sociability can be represented on a continuum with social apathy (characteristic of autism) at one end and social anxiety at the high end. However, anxiety about social interaction indicates that one is lacking confidence in one’s social skills. According to Baron-Cohen though empathizing is supposed to be a particular strength of the EFB. I therefore find it puzzling that high levels of empathizing would be assumed to go hand in hand with inadequate social skills. Current theories suggest that social anxiety is associated with self-focused rumination (thinking about oneself too much during social interactions) which does not sound much like empathic concern for other people. Furthermore, their own results contradict their argument. They did find that fear of negative evaluation and social anxiety were positively related to disordered eating. However, self-reported empathizing was found to have no relation at all to fear of negative evaluation and to be negatively correlated with a measure of social anxiety. That is, people who were high in empathizing were actually low in social anxiety and vice versa.[3]
Restrictive eating is more common in women than men. (Image courtesy of sattva at

A reasonably detailed description of the Bremser and Gallup study appears at Christian Jarrett’s blog. (A briefer description can be found here, while a more critical view can be found here.) Briefly, the authors performed a series of four studies to test their hypothesis that eating disorder symptoms are associated with a pattern of high empathizing and low systematizing. They did find modest positive correlations between self-reported empathizing and disordered eating when gender was not considered, apparently supporting their hypothesis. However, when they looked at differences between males and females there were some puzzling findings. In study 2 females who were either high or low in disordered eating did not differ in their level of self-reported empathizing (see Figure 2, p. 471). However, males who were high in disordered eating were higher in self-reported empathizing than males low in disordered eating. An even more puzzling finding was evident when they examined the relationship between emotion recognition (a test of empathic ability) and disordered eating (see Figure 1). Females high on disordered eating scored somewhat higher than other participants on this task, although it was not really clear from the authors’ report if this difference was statistically significant. What was more striking was that males high in disordered eating actually scored noticeably lower than all other participants on the emotion recognition task. Now remember that the authors’ hypothesis was that high empathizing would be associated with disordered eating, yet males with disordered eating actually scored worse than everyone else on a test of empathic ability, even though their self-reported empathy was higher. This suggests to me that these males had a lack of insight into their actual ability to register another person’s emotions. Yet quite oddly the authors claim that this anomalous finding actually supports rather than refutes their hypothesis:

“The data from males are consistent with the idea that disordered eating is associated with the hyper-empathizing that characterizes the EFB type.”
They attribute the failure of these males to correctly identify emotions as due to “hyper-mentalizing associated with the EFB” manifesting as “faulty inferences about mental states”. So they found a pattern associated with a particular group of males and then decide that this is evidence of an extreme female brain, even though none of the females manifested this pattern. Furthermore, failure in a test of empathic ability is interpreted as due to ‘hyper-empathizing’. In other words, when people are too high in empathizing they over-analyse other people and therefore make mistaken attributions about what they are thinking and feeling. The authors go on to explain: “This may be because they are using their own experience to model the experience of others, and their bias to classify emotions with a negative bias may influence their attributions.” In other words, they project their own concerns onto other people rather than trying to understand others on their own terms. This does not sound much like empathy to me. It actually sounds more like autism. People with autism also have trouble imagining that other people feel differently from themselves.
The authors actually cite research findings that people with eating disorders often show impaired emotion recognition. This would imply a failure of empathizing, yet Bremser and Gallup argue that this is actually due to “hyper-mentalizing”. They also appear to ignore previous research findings indicating commonalities between eating disorders and autism. For example, research had found that autism spectrum disorders sometimes precede the development of eating disorders and that 16% of teenage sufferers of anorexia have been estimated to have autism (Oldershaw, Treasure, Hambrook, Tchanturia, & Schmidt, 2011). Additionally, autism and anorexia may coexist within the same family suggesting they could have a shared genetic basis.
The authors argue that errors in emotion recognition tasks may be due to either a deficit in understanding (as in Asperger’s syndrome) or to abnormal or excessive attribution of mental states associated with psychotic type mental processes (referred to in the paper as schizotypal traits). I think they actually made a valid point about this. There is evidence that schizotypal traits play a role in eating disorders, particularly anorexia. This might seem to justify their claim that failures of emotion recognition are related to “hyper-mentalizing”. However, their own results show that although schizotypal traits were related to disordered eating they were largely unrelated to empathizing (see Table 7). Therefore their claim that “hyper-mentalizing” (associated with schizotypal traits) is related to “hyper-empathizing” seems unwarranted.
There was a significant correlation between the schizotypal scale ‘constricted affect’ and empathizing but this was in the negative direction. That is, people who were high in empathizing tended to be less constricted in their emotional expression. The authors noted that males who were low in constricted affect (and therefore emotionally expressive) also scored higher on disordered eating. In females there was no such relationship. The authors once again try to claim that this supports their theory by arguing that emotional expressiveness is a ‘feminine’ trait that is also related to empathizing. So therefore the finding that emotionally expressive men were more eating disordered is evidence of a relationship between the EFB and disordered eating. So yet again, a relationship found in men, but not women, is taken as evidence of a female brain type.
As previously noted the EFB is supposed to be high in empathizing and also poor in systematizing. Therefore, the authors predicted that disordered eating would be associated with poor systematizing. The actual results they found were mixed. Self-reported systematizing and a test of intuitive physics were unrelated to disordered eating. However, a test of mental rotation was found to have a significant negative correlation with disordered eating, indicating that those who performed more poorly on the mental rotation task had more disordered eating. Research has found that men tend to perform much better on tests of mental rotation than women, although some scholars have claimed that this is due to the psychological effects of gender stereotypes rather than innate differences between men and women (Ortner & Sieverding, 2008).  Bremser and Gallup did not report results for each gender so we do not know if men and women had different patterns of results. The authors acknowledge that self-starvation associated with disordered eating can produce deficits in task performance. This might explain why mental rotation was poorer in people with disordered eating. On the other hand, there were no impairments in performance on the intuitive physics task, so the results are difficult to interpret. 
What Conclusions can be drawn
Bremser and Gallup proposed that disordered eating would be associated with a pattern of high empathizing and poor systematizing they refer to as the EFB. They did find that there were modest positive associations between self-reported empathizing and disordered eating. Also, one of their studies found a moderate negative association between a systematizing task (mental rotation) and disordered eating, although two other studies using different systematizing measures found no such relationship. However, closer inspection of their results found that, when gender differences were reported, the relationship between empathizing and disordered eating occurred in men but not women. Their findings would seem to indicate that a pattern of high self-reported empathizing, poor emotion recognition, and emotional reactivity is associated with disordered eating in males but not females. Yet they claim this as evidence for an extreme female brain. Perhaps they should call this the ‘stereotypically feminine brain that leads to eating disorders in men’. Not nearly as catchy I know. Furthermore, they use some rather convoluted reasoning to explain why deficits in emotion recognition (that is, failures in empathizing) found in eating disorders should be associated with high empathizing by invoking “hyper-mentalizing”. The latter could plausibly be a feature of fear of negative evaluation and schizotypal tendencies, yet their own results indicated that these were unrelated to empathizing. There is in fact evidence for at least some overlap between eating disorders and autism, even though the EFB is supposed to represent the opposite of an autistic condition. I get the impression that the authors of this study decided that because eating disorders are so strongly associated with females that they would make a good candidate for an EFB, so they decided to try to force the result to fit their theory.

Image courtesy of Ventrilock at

Williams Syndrome: a better EFB?
So if eating disorders are not a very good candidate for a manifestation of an EFB is there something that is? The most logical candidate I am aware of is a rare condition known as Williams Syndrome. This condition is associated with extreme friendliness and sociability, and high levels of empathy (Klein-Tasman & Mervis, 2003), as well as subnormal IQ scores and difficulty understanding how a whole is made up of its parts, although language skills are generally highly developed (Farran & Jarrold, 2003). Williams Syndrome has even been referred to at times as the ‘anti-autism’ syndrome. For example, while people with autism show a disinterest in looking at faces, people with Williams Syndrome are fascinated by them (Riby & Hancock, 2008). Williams Syndrome neatly fits the profile of Baron-Cohen’s proposed EFB in that it combines strength in empathizing and difficulty in systematizing and seems more like a true ‘mirror image’ of autism. However, Williams Syndrome apparently occurs equally in males and females. Therefore, even though it meets most criteria for an EFB it does not seem to be particularly female. Anorexia is a predominantly female disorder yet it has actually been linked with autism, a supposedly extreme male disorder. Perhaps this shows that stereotyping certain conditions as extremely ‘male’ or extremely ‘female’ is actually misleading and hinders understanding them. Just because men tend to have more interest in systematizing than women does not necessarily make it a male province. Nor should empathizing be seen as a particularly female one. Furthermore, describing systematizing as a male domain and empathizing as a female one might have the effect of alienating member of the opposite sex from activities associated with each one and of needlessly reinforcing limiting gender stereotypes.

[1] Yes, he actually is related to Sacha Baron-Cohen of ‘Borat’ fame.
[2] The term ‘brain type’ seems a misnomer as the theory is based on observations about personality and behaviour rather than direct studies of the brain. However, to maintain consistency with the existing literature the term ‘brain’ will continue to be used in this article.
[3] See results for Study 3, Table 7. 
This post has previously appeared on my blog at Psychology Today Unique - Like Everybody Else.

Follow me on Facebook, Google Plus, or Twitter.

© Scott McGreal. Please do not reproduce without permission. Brief excerpts may be quoted as long as a link to the original article is provided.


Bremser JA, & Gallup GG Jr (2012). From one extreme to the other: negative evaluation anxiety and disordered eating as candidates for the extreme female brain. Evolutionary psychology : an international journal of evolutionary approaches to psychology and behavior, 10 (3), 457-86 PMID: 22947672
Farran, E. K., & Jarrold, C. (2003). Visuospatial Cognition in Williams Syndrome: Reviewing and Accounting for the Strengths and Weaknesses in Performance. Developmental Neuropsychology, 23(1-2), 173-200. doi: 10.1080/87565641.2003.9651891
Klein-Tasman, B. P., & Mervis, C. B. (2003). Distinctive Personality Characteristics of 8-, 9-, and 10-Year-Olds With Williams Syndrome. Developmental Neuropsychology, 23(1-2), 269-290. doi: 10.1080/87565641.2003.9651895
Oldershaw, A., Treasure, J., Hambrook, D., Tchanturia, K., & Schmidt, U. (2011). Is anorexia nervosa a version of autism spectrum disorders? European Eating Disorders Review, 19(6), 462-474. doi: 10.1002/erv.1069
Ortner, T. M., & Sieverding, M. (2008). Where are the Gender Differences? Male Priming Boosts Spatial Skills in Women. Sex Roles, 59(3-4), 274-281. doi: 10.1007/s11199-008-9448-9
Riby, D. M., & Hancock, P. J. B. (2008). Viewing it differently: Social scene perception in Williams syndrome and Autism. Neuropsychologia, 46(11), 2855-2860. doi: 10.1016/j.neuropsychologia.2008.05.003