The neurobiological basis of empathy deficits in Autism Spectrum Conditions

Rebecka Ahl, University of Cambridge School of Clinical Medicine, Addenbrooke's Hospital, Hills Road, Cambridge, CB2 0SP

Homer once wrote: “taught by time, my heart has learn’d to glow for others’ good, and melt at others’ woe” [1]. While a poet attempts to describe empathy in verse, it is neurologically complex. It can be argued that empathy defines mankind and severe deficits are considered a key part of the social difficulties that is a defining feature of autism spectrum conditions (ASC). The autism spectrum is often viewed as a triad of difficulties. According to DSM-IV [2] ASC is an Axis I disorder and diagnostic criteria include markedly atypical development in: social interaction, communication and repertoire of activity and interests. ASC have a prevalence of approximately 1% [3], a male to female ratio of 4:1, and are usually diagnosed in early childhood [4]. This group is characterised by heterogeneity with considerable variability in intelligence and adaptive functioning. A diagnosis of an ASC has a recognised genetic basis and is a life-long disability for which no effective treatment exists. One major reason for this is the absence of an established neurobiological basis of both cause and course of the empathy deficits.

Empathy is regarded as the capacity to both predict the behaviour of agents by inferring mental states – the so-called ‘Theory of Mind’ (ToM) or ‘mentalizing’ - and respond with appropriate emotion. It is assumed that ‘mind-blindness’ [5], a cognitive failure in ToM, plays an essential role in empathy deficits, and that the capacity to ‘read’ other people’s minds is imperative for normal development of empathy. Typical observations in individuals with ASC include reduced capacity in reading emotions from pairs of eyes [6], from voices [7], and below average Empathy Quotients in both children [8] and adults [9]. Attempts to identify the neurobiological basis of empathy deficits are ongoing and research in functional neuroimaging and hormonal influence on brain development are two areas that may improve our understanding of the autism spectrum and subsequent detection, diagnosis and treatment.


Neural Network of Theory of Mind

Neuroimaging suggests a neural network of medial prefrontal cortex (MPFC), superior temporal sulcus (STS) and temporal pole adjacent to the amygdala, during mentalizing [10]. Observations in functional magnetic resonance imaging (fMRI) studies suggest further functional subdivision of the MPFC [11] and thus higher levels of specialization (Figure 1).

ToM predicts that neurobiological abnormalities will profoundly affect our empathic abilities. Case-control studies have corroborated the role of a mentalizing network when employing empathy. Castelli et al. [12] used positron emission tomography (PET), with individuals with ASC exhibiting reduced blood flow versus controls in MPFC, STS (at the temporoparietal junction) and basal temporal area (including fusiform gyrus and temporal pole) when performing a ToM-task. This supports the three-component ToM network. Figure 2 demonstrates these large activity differences between individuals with ASC and controls in core ToM brain areas, also identifying an additional fourth component (extrastriate regions of occipital cortex). This area showed typical increased activation in individuals with ASC, but reduced connectivity with STS compared to controls. Consequently, while the increased ‘mentalizing complexity’ of the ToM animations seems to be visually registered, no such information appears to reach the mentalizing network. In the ten adults with high-functioning autism or Asperger’s syndrome, atypical activity was accompanied by less accurate descriptions.

Fig 2: Reproduced from Castelli F, et al. Autism, Asperger Syndrome and brain mechanisms for the attribution of mental states to animated shapes, Brain, 2002, 125, 8, 1839-49, by permission of Oxford University Press

Mirror and Limbic Systems

Impairments in interpreting and responding to emotional faces suggest an interaction between atypical neural face processing and ‘mirroring mechanisms’ in empathy deficits. It follows that some developmental abnormality in proposed mirror systems (MS) of premotor cortex and inferior parietal cortex may be involved in the neurobiological basis of deficient empathy in ASC [13]. It is, however, questionable whether emotion imitation requires comprehension: MS may merely contribute to ToM and empathy via automatic simulation of observed actions. For instance, rudimentary forms of imitation are even seen in newborns [14]. Reduced spontaneous electromyography responses in facial muscles of both teenagers and adults with ASC during imitation of photographed faces support this, although it can be debated if this generalises to three-dimensional faces [15].

Blair [16] argued for distinction between motoric empathy (dependent on MS) and emotional empathy (dependent on limbic and somatosensory areas). However, using fMRI, the direct relationship between MS and empathy in typically developing ten-year-olds has been studied [17]. This shows a neural circuit comprising MS (pars opercularis, ventral premotor cortex and rostral inferior parietal lobule), insula and amygdala during both observation and imitation of facial emotional expressions. Activity in these regions correlated with scores on the Interpersonal Reactivity Index of empathy. While being cautious about mistaking correlations for causes, this suggests a complex network of empathizing, where neural empathy could ultimately result from an interactive relationship between MS, brain areas feeding into the MS, and the limbic system. Although this is based purely on typically developing subjects, previously mentioned studies support atypical brain activity in both amygdala and MS in individuals with ASC. While most agree on the existence of a “social brain” it is often debated exactly what brain components are involved [18]. Figure 3 attempts to illustrate those brain areas believed to play a role in empathizing.

Fig.3 Reprinted by permission from Macmillan Publishers Ltd: Nature Reviews Neuroscience; Blakemore S. The social brain in adolescence. Nature Reviews Neuroscience. 2008: 1; 267-277, copyright 2008.


Foetal Testosterone

Observations of autistic traits early in childhood have led to examination of prenatal influences on brain development. As Figure 4 shows, there are distinctive sex differences in relation to empathy scores and foetal testosterone (FT) levels. Using amniocentesis in the second trimester of 193 women, Chapman et al. [19] established a negative correlation between FT and 6-8 year-old children’s scores on empathy measures (Empathy Quotient for children and ‘Reading the Mind in the Eyes’ task). This may suggest that individuals with ASC have abnormally high FT levels, though this has not been directly tested. Using medical questionnaires, a study by Ingudomnukul et al. [20] supports the involvement of abnormal FT in women with ASC and mothers of children with ASC compared to mothers of typically developing children. They found an increase in testosterone-related atypical behaviour (such as bisexuality, asexuality and tomboyism) and testosterone-related medical conditions (e.g. menstrual dysfunctions, hirsutism, breast and uterine cancers) in women with ASC and the broader autism phenotype. Consequently, hormonal abnormalities appear linked with ASC. However, between foetal factors and empathy outcome there are many possible interventions (including social experience).

Fig.4 Reproduced from Chapman, E. et al. Fetal testosterone and empathy: Evidence from the Empathy Quotient (EQ) and the “Reading the Mind in the Eyes” Test. Social Neuroscience. 2006: 1; 135-148, by permission of Taylor and Francis,

Empathy is thought to involve coordination of neural information from several distinct regions, and thus depend on ‘long-range’ connections. It has been proposed that typical male brains show reduced long-range connectivity tracts – the so-called ‘Extreme Male Brain’ (EMB) theory of autism [21]. This may consequently explain typical male behaviour – higher levels of systemizing and lower levels of empathizing than in typical females. Furthermore, ASC-behaviour involves even lower average empathizing scores and atypical neural responses during empathizing tasks. Thus, as FT has a strong impact on brain development, it is tempting to suggest that empathy deficits in ASC might arise from development of extreme typical male neuroanatomy, with shifted balance between local and long-range connectivity [22]. For instance, MRI demonstrates abnormal brain overgrowth in ASC, dominated by white matter hyperplasia, confined to local connectivity tracts [23]. The observed male preponderance in ASC could originate from innate sex differences in hormonal starting points.

Studying typically developing 4-year-old boys and girls, Knickmeyer et al. [24] reported results supporting the proposed relationship between FT levels (established by amniocentesis) and mentalizing abilities. They found that when describing two interacting animated triangles, boys demonstrated lower use of intentional propositions (negatively correlated with FT) and more use of neutral propositions (directly correlated with FT) than girls, indicating reduced emotional involvement. Additionally, 204 children took part in a longitudinal study of FT effects where FT was the “only significant predictor” of systemizing when boys and girls were analysed together [25]. The fact that gender was not in the final regression model suggests FT levels are more significant than gender on degree of systemizing demonstrated. Direct extrapolation to autism is unjustifiable since these studies tested typically developing children and not individuals with ASC, but these findings are relevant to EMB theory discussions.


Oxytocin levels have also been associated with empathy in typical adults and general social impairments in ASC, but findings are conflicting. In a double-blind placebo-controlled study of 30 healthy men, Domes et al. [26] found that intranasal administration of a single dose of 24 IU oxytocin improved performance on ‘Reading the Mind in the Eyes’ test. There was a larger effect on the ability to infer mental states for more complex emotions. Consequently, they suggested that oxytocin injections may have modulated and improved neural networks for emotion recognition, and that individuals with ASC might show reduced levels. Their findings raise the question of whether oxytocin effects are short-term only or also developmental. Additionally, Andari et al. [27] reported that adults with ASC demonstrate “stronger interactions (than ASC-placebos and controls) with the most socially cooperative” player in a simulated ball game after oxytocin inhalation. Inhalations also led to self-reports of more trust and preference for the ‘good’ over the ‘bad’ player in the ball game, and increased gazing-times on eye regions in observed photographs. While this suggests a possible therapeutic role of oxytocin, they do not consider practical issues that might limit this approach (such as a short half-life).

Direct links between oxytocin and empathy deficits in ASC are limited but general findings support the role of oxytocin in social situations. For example, Wahl [28] hypothesised that high levels of administered oxytocin during labour induction may cause a down-regulation of oxytocin receptors in the foetal brain, affecting future development. Wahl’s also established a link between higher rates of pitocin (a synthetic oxytocin) induction in ASC-births than normal births. A permanent effect on early brain development is possible , for example if oxytocin receptive fields in the brain are labile during birth this could invite atypical density shifts in response to external oxytocin influence. Although oxytocin normally facilitates development of social behaviours, it may be that elevated levels at birth are harmful.

Gene-environment interactions

While the evidence for a genetic component in ASC is strong, environmental influences should not be excluded. Family conversation has been emphasised in development of ToM and abilities to understand false-belief and other mentalizing tasks in both deaf children of hearing parents and autistic children. Both groups show poor language skills and understanding in areas like make-believe and abstract ideas, possibly because they lack an “early conversational partner” [29]. Consequently, early absence of daily discussions about own and others’ mental states may profoundly affect typical brain development and should be considered in the biological discussion of empathy deficits in ASC.

Furthermore, environmental influences on the development of empathy deficits are important when considering treatment. Different approaches have been made to improve emotion recognition in ASC: Golan et al. [30] showed an animated series, “The Transporters”, daily to children with ASC. Results suggest improvements in both emotion recognition at three task-levels and emotional vocabulary (Figure 5). Although this supports environmental effects, small sample sizes and/or lack of application of improved skill to real situations make it hard to conclude whether these interventions improve ‘true’ neural empathy or simply task performance.


Limitations in Functional Neuroimaging

While functional neuroimaging studies provide the most consistent evidence of empathy deficits in the autism spectrum, there are limits. Reduced brain activity when inferring mental states and during emotion recognition has been reported in individuals with ASC versus controls in regions typically linked to the mentalizing network [12]. However, this hypoactivation pattern is neither proof of direct links between these regions and task performance, nor that these regions are essential for task accuracy. Perhaps there are different approaches to solving empathy-related tasks and different neural networks (e.g. internal simulation vs. memory retrieval from long-term stores) which would be all but impossible to standardise. Moreover, the common trend of reduced brain activity in ASC could, theoretically, simply mean that ASC-cohorts show increased neural efficiency. The data suggests an association between these regions, and impairments in recognition of emotions and mental states, but it does not prove a direct relationship.

What might strengthen the field is investigating the relationship between degree of hypoactivation and degree of mentalizing deficits. For example, Wang et al. [7] not only found MPFC activity inversely related to severity of emotion judgment deficits in individuals with ASC (supporting the importance of this region), they also provided evidence for the hypothesis of atypical hypoactivation in MPFC, with raised brain activity in individuals with versus controls when instructed to attend to facial expression and tone of voice.

Finally, it is commonly observed that severity of empathy deficits in high-functioning children and adolescents with ASC decrease in adulthood but nobody has established why. It would be interesting to conduct a longitudinal study investigating whether such improvements are purely psychological or if proposed neural circuits of empathy deficits (e.g. mirror systems and ‘ToM-areas’) show a proportional improvement in brain activity.

Missing Links

From neuroimaging studies of ASC, a pattern of brain regions linked to empathy deficits emerges, including ToM components, limbic structures and mirror activity areas. There are also strong connections between hormones (FT, oxytocin) and empathy deficits, and a proposed role for environmental factors. For example, Baron-Cohen’s neurological basis of the EMB theory [21] allows FT to influence the developing brain, the male to female ratio and observed sex differences in empathizing abilities of the general population.

However, there is no substantial evidence to explain, using the EMB theory the neural pattern proposed by neuroimaging. Evaluating whether high FT levels sway the brain’s connectivity balance towards short-range connectivity tracts disturbing the neural circuitry of empathy (thought to rely on long-range connectivity tracts) necessitates a better definition of what neural empathy really is. Problems are raised already in this statement because how short are short-range connectivity tracts? Without a definition, how do we investigate them? There is also no reason to infer a causal relationship between FT and an ASC-diagnosis. FT levels have been negatively correlated with both EQ scores [19] and mentalizing abilities [24]. While this may demonstrate a plausible link between psychiatric traits and FT, it does not allow the inference that testosterone is connected to the ASC-diagnosis itself.

Symptomatological Overlap

Empathy deficits are not specific to ASC. They appear in several psychiatric conditions including schizophrenia, conduct disorder and psychopathy [2]. Since such different disorders include the same cognitive deficit, one must question the nature of empathy itself – could empathy consist of several distinct aspects, such as cognitive (thought) or affective (emotion). Distinguishing such components psychologically may allow neurological differentiation, which could help us better understand empathy as a personality trait, its biological basis in disorders, and whether empathy deficits form important parts of symptomatology.

It has been proposed, and to some extent shown, that psychopaths commonly demonstrate imbalances of the empathy spectrum [16]. For example, they may recognise others’ happiness but not their fear, while individuals with ASC seem to experience emotions but be less able to recognise and analyse them. Investigating the neurobiological basis underlying these differences may aid our understanding and may help guide future diagnostic criteria for ASC, which are currently inadequate.


Empathy is considered a personality trait, and if one were to test the ability of the general population to be empathic by use of a test, a wide range of scores would be seen. Comparing the test scores of controls and individuals with ASC, however, demonstrate a clear divide with controls obtaining significantly higher scores. A lot of effort is now placed on identifying those neural circuits involved in creating this observed divide, such that therapeutic tools can be aimed directly on the source of the problem. While some intervention studies currently show improved task performance, it is uncertain if this can be generalised to real social situations, and whether similar success is demonstrated in behaviours related to emotion recognition ability (i.e. outcome measure should be an independent task that can be linked to original test). Perhaps future research would benefit from neuropsychological assessments where degree of cognitive capacity advancement by individuals with ASC is compared to any increases in brain activity of, or connectivity between, brain areas thought to underlie neural empathy. Consequently, the challenge for the future is to extend our knowledge of the neurobiological basis of empathy deficits to improve diagnostics and intervention.


1. Homer. The Odyssey. USA: Popular Classics Publishing; 2012.

2. APA. DSM-IV Diagnostic and Statistical Manual of Mental Disorders. 4th edition. Washington DC: American Psychiatric Association; 1994.

3. Baird, G. et al. Prevalence of disorders of the autism spectrum in a population cohort of children in South Thames: the Special Needs and Autism Project (SNAP). Lancet. 2006: 368; 210-215. doi:10.1016/S0140-6736(06)69041-7

4. Kring, A.M. et al. Abnormal Psychology 10th edition. USA: John Wiley & Sons; 2007.

5. Baron-Cohen, S. Mindblindness: An Essay on Autism and Theory of Mind. Cambridge MA: MIT Press; 1995.

6. Baron-Cohen, S. et al. The ‘Reading the mind in the eyes’ test revised version: A study with normal adults, and adults with Asperger Syndrome or High-Functioning autism. Journal of Child Psychology and Psychiatry. 2001: 42(2); 241-252. doi:10.1111/1469-7610.00715

7. Wang, A.T. et al. Reading Affects in the Face and Voice: Neural Correlates of Interpreting Communicative Intent in Children and Adolescents With Autism Spectrum Disorders. Archives of General Psychiatry. 2007: 64(6); 698-708. doi:10.1001/archpsyc.64.6.698

8. Auyeung, B. et al. The Children’s Empathy Quotient (EQ-C) and Systemizing Quotient (SQ-C): sex differences in typical development and in autism spectrum conditions. Journal of Autism and Developmental Disorders. 2009: 39(11); 1509-1521. doi:10.1007/s10803-009-0772-x

9. Baron-Cohen, S. et al. The Empathy Quotient: An Investigation of Adults with Asperger Syndrome or High Functioning Autism, and Normal Sex Differences. Journal of Autism and Developmental Disorders. 2004: 34(2); 163-175. doi:10.1023/B:JADD.0000022607.19833.00

10. Frith, U. et al. Development and neurophysiology of mentalizing. Philosophical Transactions of the Royal Society London B Biological Sciences. 2003: 358; 459-473. doi:10.1098/rstb.2002.1218

11. Amodio, D.M. et al. Meeting of minds: the medial frontal cortex and social cognition. Nature Reviews Neuroscience. 2006: 7; 268-277. doi:10.1038/nrn1884

12. Castelli, F. et al. Austism, Asperger syndrome and brain mechanisms for the attribution of mental states to animated shapes. Brain. 2002: 125(8); 1839-1849. doi:10.1093/brain/awf189

13. Williams, J.H.G. Self-Other Relations in Social Development and Autism: Multiple Roles for Mirror Neurons and Other Brain Bases. Autism Research. 2008: 1(2); 73-90. doi:10.1002/aur.15

14. Meltzoff, A.N. et al. Imitation of facial and manual gestures by human neonates. Science. 1977: 198; 75-78. doi:10.1126/science.897687

15. McIntosh, D.N. et al. When the social mirror breaks: deficits in automatic, but not voluntary, mimicry of emotional facial expressions in autism. Developmental Science. 2006: 9; 295-302. doi:10.1111/j.1467-7687.2006.00492.x

16. Blair, R.J.R. Responding to the emotions of others: dissociating forms of empathy through the study of typical and psychiatric populations. Consciousness and Cognition. 2005: 14(4); 698-718. doi:10.1016/j.concog.2005.06.004

17. Pfeifer, J.H. et al. Mirroring others’ emotions relates to empathy and interpersonal competence in children. NeuroImage. 2008: 39(4); 2076-2085. doi:10.1016/j.neuroimage.2007.10.032

18. Blakemore S. The social brain in adolescence. Nature Reviews Neuroscience. 2008: 1; 267-277. doi:10.1038/nrn2353

19. Chapman, E. et al. Fetal testosterone and empathy: Evidence from the Empathy Quotient (EQ) and the “Reading the Mind in the Eyes” Test. Social Neuroscience. 2006: 1; 135-148. doi:10.1080/17470910600992239

20. Ingudomnukul, E. et al. Elevated rates of testosterone-related disorders in women with autism spectrum conditions. Hormones and Behavior. 2006: 51(5); 597-604. doi:10.1016/j.yhbeh.2007.02.001

21. Baron-Cohen, S. The Essential Difference. London: Penguin Books; 2003.

22. Baron-Cohen, S. et al. Sex Differences in the Brain: Implications for Explaining Autism. Science. 2005: 310; 819-823. doi:10.1126/science.1115455

23. Lewis, J.D. et al. Growth-related neural reorganization and the autism phenotype: a test of the hypothesis that altered brain growth leads to altered connectivity. Deveopmental Science. 2008: 11; 135-155.doi: 10.1111/j.1467-7687.2007.00634.x

24. Knickmeyer, R. et al. Fetal testosterone and empathy. Hormones and Behaviour. 2006: 49(3); 282-292. doi:10.1016/j.yhbeh.2005.08.010

25. Auyeung, B. et al. Foetal testosterone and the child systemizing quotient. European Journal of Endocrinology. 2006: 155; S123-S130. doi:10.1530/eje.1.02260

26. Domes, G. et al. Oxytocin Improves “Mind-Reading” in Humans. Biological Psychiatry. 2006: 61(6); 731-733. doi:10.1016/j.biopsych.2006.07.015

27. Andari, E. et al. Promoting social behaviour with oxytocin in high-functioning autism spectrum disorders. Proceedings of the National Academy of the United States of America. 2010: 107; 4389-4394. doi:10.1073/pnas.0910249107

28. Wahl, R.U.R. Could oxytocin administration during labor contribute to autism and related behavioral disorders? – A look at the literature. Medical Hypotheses. 2004: 63; 456-460.doi:10.1016/j.mehy.2004.03.008

29. Peterson, C.C. et al. Insights into Theory of Mind from Deafness and Autism. Mind & Language. 2000: 15; 123-145.doi:10.1111/1468-0017.00126

30. Golan, O. et al. Enhancing Emotion Recognition in Children with Autism Spectrum Conditions: An Intervention Using Animated Vehicles with Real Emotional Faces. Journal of Autism and Developmental Disorders. 2010: 40; 269-279.doi:10.1007/s10803-009-0862-9