Encyclopedia of Autism Spectrum Disorders

Living Edition
| Editors: Fred R. Volkmar

Action Prediction in Autism

  • Tobias SchuwerkEmail author
  • Markus Paulus
Living reference work entry
DOI: https://doi.org/10.1007/978-1-4614-6435-8_102206-1


Action prediction is the inherent social cognitive ability to anticipate how another individual’s action will unfold over time. Such projections are essential for smooth reciprocal social interaction and involve the predictions of others’ action goals as well as the means they use to achieve their goals. In everyday life, humans constantly coordinate their actions with others. For example, having a cup of coffee at a café involves numerous joint actions, such as ordering the coffee when the waiter is attending, giving the cash and receiving the change, or holding up the cup so that the waiter can refill it with more coffee from the coffeepot. All these actions have to be sensitively attuned in order to successfully enjoy the cup of coffee without dropping money or spilling hot coffee on one’s pants.

Eye movements during the observation of another individual’s action reveal that, instead of passively following the movement trajectory of this action, humans proactively anticipate the end state as well as the trajectory of an ongoing action (Gredebäck and Falck-Ytter 2015). These gaze patterns are highly similar to eye movements during the performance of one’s own actions (Flanagan and Johansson 2003). Developmental psychological research showed that infants in their first year of life already produce such anticipatory eye movements when they observe another’s action (Falck-Ytter and von Hofsten 2006).

A substantial body of research suggests that action processing is altered in autism spectrum disorders (ASDs) and that this affects social interaction and communication abilities in individuals with ASD. Recently, it was proposed that impaired action prediction is a key factor that determines social interaction and communication skills in ASD. Yet, theoretical positions and empirical evidence remain controversial. This entry presents current knowledge on if and how action prediction is altered in ASD. Hence, the entry puts emphasis on more recent work on visual action anticipation (often assessed by eye tracking technology) and mentions older work on verbal action prediction only in passing.

Historical Background

Research on action prediction in ASD was mainly framed by two theoretical backgrounds. First, the theory of mind deficit hypothesis suggested that individuals with ASD have difficulties with the prediction of other’s actions, because they are impaired in attributing mental states, such as beliefs or desires, to others and themselves (see Frith 2012). This hypothesis is based on the premise that action predictions require the prior representation of another’s goal (e.g., the guest at the neighboring table also wants a coffee refill) and belief about that goal (e.g., the guest thinks that drawing the waiters attention to herself will get her to that goal; cf., Dennett 1989). Thus, if individuals with ASD have difficulties in attributing mental states such as goals and beliefs, they should also be impaired in predicting the corresponding action (e.g., the guest will lift her empty cup of coffee when the waiter passes).

Second, the broken mirror hypothesis suggested that the mirror neuron system is dysfunctional in ASD. The mirror neuron system is a network of brain regions that presumably matches observed actions with one’s own motor system and thereby enables action imitation and interpretation (Gallese et al. 2004). It was proposed that an ontogenetically early deficit in this neuronal mechanism of mapping another’s and one’s own actions in ASD leads to impaired imitation, understanding, and prediction of other’s actions (Oberman and Ramachandran 2007).

However, mounting evidence is incompatible with both hypotheses. Children and adults with ASD show typical imitation, understanding, and prediction of goal-directed actions in a variety of paradigms (e.g., Cusack et al. 2015; Falck-Ytter 2010; Marsh et al. 2015; Sebanz et al. 2005). On the other hand, studies using different paradigms suggest that in some situations individuals with ASD do have difficulties with accurate action predictions (e.g., von Hofsten et al. 2009; Vivanti et al. 2011). Together, these findings challenge the theory of mind deficit hypothesis and the broken mirror hypothesis because these frameworks cannot fully explain the nuanced characteristics of action prediction in ASD. Consequently, more refined theories that account for impaired and intact aspects of action prediction in ASD are required (Hamilton 2009).

Current Knowledge

Recent years have seen a surge of empirical work on various levels of action processing in ASD, ranging from the detection of biological motion to belief-based action prediction. These findings have in turn driven rapid advance in the development of new theories.

In contrast to previous research (e.g., Blake et al. 2003; Klin and Jones 2008), several recent studies found no group difference between children, adolescents, and adults with ASD and typical participants in cognitive processes that are presumably required for successful action prediction (Cusack et al. 2015; Murphy et al. 2009; Saygin et al. 2010). For example, Cusack et al. (2015) reported intact detection of animacy, detection of biological motion, and discrimination of different types of actions in adolescents with ASD. Thus, deficient precursor mechanisms cannot serve as explanation for altered action prediction in ASD.

Moreover, the prediction of goals and means of a simple “pick-and-place” action was found to be unaffected in 5-year-old children with ASD (Falck-Ytter 2010). The children watched an agent reaching for objects on a table and placing it into a container at the other side of the table. Just as children from the comparison group, 5-year-olds with ASD visually anticipated the end state of these actions, i.e., grasping the corresponding ball and placing it in the container. In contrast, eye movements of children with and without ASD were reactive when the objects were moving self-propelled, i.e., their gaze followed the movement. This suggests that, when observing other agents, children with ASD proactively process the agent’s goal and predict according actions.

Yet, action prediction in ASD seems to reach its limits when it becomes necessary to consider another’s false beliefs about a certain situation. This was shown in explicit theory of mind tasks testing verbal action predictions (Baron-Cohen et al. 1985), and more recently using implicit theory of mind tasks, eye tracking versions of the classical explicit paradigms. In these implicit tasks, participants are familiarized with an agent’s goal to get an object by opening one of two doors. Individuals with and without ASD need only a few trials to generate predictive eye movements towards the door that is about to be opened. In the subsequent test trial, the agent falsely believes the object would be located behind door A, although it actually is either behind door B or it was removed completely from the scene. To accurately predict the agent’s action in this trial (that she will open door A), participants have to take into account that the agent’s upcoming action is based on her false belief that the object would still be behind this door. Children and adults with ASD systematically fail to correctly predict this false belief-based action (Schuwerk et al. 2015; Senju et al. 2010). However, it is important to note that, unlike previously thought, individuals with ASD are able to correctly predict the agent’s false belief-based action in a variety of explicit theory of mind tasks (Scheeren et al. 2013). One explanation for this finding is that individuals with ASD, especially those with good intellectual and language abilities, develop compensatory strategies to pass these tasks (cf., Livingstone and Happé 2017).

In sum, current mixed and partially inconclusive evidence on intact and impaired aspects of action prediction in ASD speaks against the ideas of a general action prediction deficit in ASD and that single links in the chain of action processing are broken. Individuals with ASD are in principle able to predict other’s actions. Rather, the finding that this ability is hampered in some contexts but not in others suggests that computations associated with successful action prediction work less effectively, and/or alternative cognitive routes are taken to get to an accurate action prediction (Livingstone and Happé 2017).

An appealing way to explain these altered cognitive processes that affect action prediction in ASD is offered by the theoretical framework of hierarchical predictive processing. In short, this currently prominent theory holds that we do not perceive the world by the unbiased interpretation of the information conveyed by sensory systems. Rather, we have a model of how the world should look like and our brain uses it to actively and optimally predict incoming sensory information. The flow of information is bidirectional: at each level within the postulated cognitive hierarchy, downward driving predictions and upward transmitted sensory information are compared. The part of sensory input that cannot be explained by the prediction results in a prediction error, which is passed upward so that adjusted and more accurate predictions can be generated. This theory, describing a cognitively very efficient way of making sense of our world, is able to explain cognitive information processing in a variety of domains, ranging from vision to social cognition (Clark 2013).

Related claims have been made in developmental psychology. Ruffman (2014) suggested a reduced ability for learning from statistical information in ASD. Given that action anticipation and social learning might rely on implicit statistical learning (Paulus 2014; Ruffman 2014), such a deficit could account for a variety of problems associated with ASD.

Pellicano and Burr (2012) suggested that these predictions of sensory information are attenuated in ASD. Thus, perception is less biased by prior expectations about sensory information. Assuming that strong expectations help the cognitive system to reduce the complexity of sensory input, attenuated expectations result in an overburdening stream of relatively unfiltered incoming information that has to be processed. This fits well with the clinical observation of sensory sensitivities and repetitive behavior patterns. Within this framework, the latter can be viewed as a way to reduce the need to process unpredictable external information by creating expectable and consistent sensory stimulation.

In the case of action processing, this means that individuals with ASD are affected in the ability to generate action predictions based on prior experience and current observations. This presumably not only affects the control of one’s own actions but also the prediction of other people’s actions (Sinha et al. 2014). Indeed, individuals with ASD show deficits in motor coordination like action preparation or action planning (Fournier et al. 2010). And also when watching interactions of others, children and adults with ASD are less likely to predict their actions (Chambon et al. 2017; von der Lühe et al. 2016; von Hofsten et al. 2009). This form of interpersonal action prediction is crucial for smooth interactive turn-taking.

Moreover, there is evidence that even very young children with ASD show less anticipation of other’s actions. For example, Brisson et al. (2012) retrospectively analyzed home videos of spoon-feeding situations of children around 5 months of age who have been diagnosed with ASD later. In contrast to a control group, they showed less anticipatory mouth opening when the caregiver moved the spoon towards the infant’s mouth. Interestingly, typically developing infants who displayed low-anticipation rates, improved rapidly. Although an increase in accurate anticipations was also observed in infants later diagnosed with ASD, they seemed to learn more slowly from experience in such feeding situations.

Also later in life, the ability to exploit past experience to generate action predictions seems to be affected in ASD. For example, adults and 10-year-old children with ASD showed altered action predictions in a task that elicited visual action anticipations of an agent who repeatedly produced one of two possible actions to get to its goal (Schuwerk et al. 2016). The participants with ASD not only showed overall lower rates of action predictions, they also profited less than the respective comparison groups from frequency information. Adults and children from the comparison groups rapidly used the observation that the agent repeatedly acted the same way to predict that, in the same situation, it will produce the same action again. However, participants with ASD showed less improvement in accurate action predictions over time, suggesting that this form of statistical learning is affected in ASD.

In sum, there is growing evidence that the ability to effectively build expectations of another’s upcoming actions based on previously observed actions under the same situational constraints is impaired in ASD. This might not affect the anticipation of simple actions with only one plausible outcome (Falck-Ytter 2010), or actions that follow certain rules (e.g., building a tower with colored pieces following the rule “alternate colors”; Vivanti et al. 2011). But, when options for an action become more complex or additional social cues or beliefs are involved, accurate action prediction might become challenging (Senju et al. 2010; Vivanti et al. 2011). It is important to note that the ability to learn from experience to generate accurate action predictions is not absent in ASD (Chambon et al. 2017; Schuwerk et al. 2015). Yet, it seems that this form of learning from experience works less efficiently in individuals with ASD.

Future Directions

Impaired hierarchical predictive processing is a promising theoretical account, which is able to elucidate a variety of empirical findings on altered action prediction ASD. However, more evidence is needed to firmly conclude that deficient predictive processing is at the core of observed social interaction deficits in ASD. Moreover, it is unclear whether a predictive processing deficit in action prediction sufficiently explains the entire range of symptoms of impaired social interaction and communication. In other words, is navigating the social world challenging for individuals with ASD merely because it is so unpredictable, or do other factors, for example, the motivation to engage with the social world, also play a role (Chevallier et al. 2012)? Further, if it were the case that decelerated learning from experience with past actions underlies altered action processing in ASD, this could be targeted by interventions that, for example, offer additional opportunities to learn from experience, or help to elaborate more explicit and rule-based strategies to predict other’s actions. It is up to future research to investigate if these are viable routes to modulate altered action prediction is ASD.

See Also

References and Reading

  1. Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a “theory of mind”? Cognition, 21(1), 37–46.CrossRefPubMedGoogle Scholar
  2. Blake, R., Turner, L. M., Smoski, M. J., Pozdol, S. L., & Stone, W. L. (2003). Visual recognition of biological motion is impaired in children with autism. Psychological Science, 14(2), 151–157.CrossRefPubMedGoogle Scholar
  3. Brisson, J., Warreyn, P., Serres, J., Foussier, S., & Adrien-Louis, J. (2012). Motor anticipation failure in infants with autism: a retrospective analysis of feeding situations. Autism, 16(4), 420–429.CrossRefPubMedGoogle Scholar
  4. Chambon, V., Farrer, C., Pacherie, E., Jacquet, P. O., Leboyer, M., & Zalla, T. (2017). Reduced sensitivity to social priors during action prediction in adults with autism spectrum disorders. Cognition, 160, 17–26.CrossRefPubMedGoogle Scholar
  5. Chevallier, C., Kohls, G., Troiani, V., Brodkin, E. S., & Schultz, R. T. (2012). The social motivation theory of autism. Trends in Cognitive Sciences, 16(4), 231–239.CrossRefPubMedPubMedCentralGoogle Scholar
  6. Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36(3), 181–204.CrossRefPubMedGoogle Scholar
  7. Cusack, J. P., Williams, J. H., & Neri, P. (2015). Action perception is intact in autism spectrum disorder. Journal of Neuroscience, 35(5), 1849–1857.CrossRefPubMedPubMedCentralGoogle Scholar
  8. Dennett, D. C. (1989). The intentional stance. Cambridge, MA: MIT press.Google Scholar
  9. Falck-Ytter, T. (2010). Young children with autism spectrum disorder use predictive eye movements in action observation. Biology Letters, 6(3), 375–378.CrossRefPubMedGoogle Scholar
  10. Falck-Ytter, T., & von Hofsten, C. (2006). Infants predict other people’s action goals. Nature Neuroscience, 9(7), 878.CrossRefPubMedGoogle Scholar
  11. Fournier, K. A., Hass, C. J., Naik, S. K., Lodha, N., & Cauraugh, J. H. (2010). Motor coordination in autism spectrum disorders: a synthesis and meta-analysis. Journal of Autism and Developmental Disorders, 40(10), 1227–1240.CrossRefPubMedGoogle Scholar
  12. Frith, U. (2012). Why we need cognitive explanations of autism. The Quarterly Journal of Experimental Psychology, 65(11), 2073–2092.CrossRefPubMedGoogle Scholar
  13. Gallese, V., Keysers, C., & Rizzolatti, G. (2004). A unifying view of the basis of social cognition. Trends in Cognitive Sciences, 8, 396–403.CrossRefPubMedGoogle Scholar
  14. Gredebäck, G., & Falck-Ytter, T. (2015). Eye movements during action observation. Perspectives on Psychological Science, 10(5), 591–598.CrossRefPubMedPubMedCentralGoogle Scholar
  15. Hamilton, A. D. C. (2009). Research review: Goals, intentions and mental states: Challenges for theories of autism. Journal of Child Psychology and Psychiatry, 50(8), 881–892.CrossRefPubMedGoogle Scholar
  16. Klin, A., & Jones, W. (2008). Altered face scanning and impaired recognition of biological motion in a 15-month-old infant with autism. Developmental Science, 11(1), 40–46.CrossRefPubMedGoogle Scholar
  17. Livingston, L. A., & Happé, F. (2017). Conceptualising compensation in neurodevelopmental disorders: Reflections from autism spectrum disorder. Neuroscience & Biobehavioral Reviews, 80, 729–742.CrossRefGoogle Scholar
  18. Marsh, L. E., Pearson, A., Ropar, D., & Hamilton, A. D. C. (2015). Predictive gaze during observation of irrational actions in adults with autism spectrum conditions. Journal of Autism and Developmental Disorders, 45(1), 245–261.CrossRefPubMedGoogle Scholar
  19. Murphy, P., Brady, N., Fitzgerald, M., & Troje, N. F. (2009). No evidence for impaired perception of biological motion in adults with autistic spectrum disorders. Neuropsychologia, 47(14), 3225–3235.CrossRefPubMedGoogle Scholar
  20. Oberman, L. M., & Ramachandran, V. S. (2007). The simulating social mind: The role of the mirror neuron system and simulation in the social and communicative deficits of autism spectrum disorders. Psychological Bulletin, 133, 310–327.CrossRefPubMedGoogle Scholar
  21. Paulus, M. (2014). How and why do infants imitate? An ideomotor approach to social and imitative learning in infancy (and beyond). Psychonomic Bulletin & Review, 21, 1139–1156.CrossRefGoogle Scholar
  22. Pellicano, E., & Burr, D. (2012). When the world becomes ‘too real’: a Bayesian explanation of autistic perception. Trends in Cognitive Sciences, 16(10), 504–510.CrossRefPubMedGoogle Scholar
  23. Ruffman, T. (2014). To belief or not belief: Children’s theory of mind. Developmental Review, 34, 265–293.CrossRefGoogle Scholar
  24. Saygin, A. P., Cook, J., & Blakemore, S. J. (2010). Unaffected perceptual thresholds for biological and non-biological form-from-motion perception in autism spectrum conditions. PloS one, 5(10), e13491.CrossRefPubMedPubMedCentralGoogle Scholar
  25. Scheeren, A. M., de Rosnay, M., Koot, H. M., & Begeer, S. (2013). Rethinking theory of mind in high-functioning autism spectrum disorder. Journal of Child Psychology and Psychiatry, 54(6), 628–635.CrossRefPubMedGoogle Scholar
  26. Schuwerk, T., Vuori, M., & Sodian, B. (2015). Implicit and explicit theory of mind reasoning in autism spectrum disorders: the impact of experience. Autism, 19(4), 459–468.CrossRefPubMedGoogle Scholar
  27. Schuwerk, T., Sodian, B., & Paulus, M. (2016). Cognitive mechanisms underlying action prediction in children and adults with autism spectrum condition. Journal of Autism and Developmental Disorders, 46(12), 3623–3639.CrossRefPubMedGoogle Scholar
  28. Sebanz, N., Knoblich, G., Stumpf, L., & Prinz, W. (2005). Far from action-blind: Representation of others’ actions in individuals with autism. Cognitive Neuropsychology, 22(3–4), 433–454.CrossRefPubMedGoogle Scholar
  29. Senju, A., Southgate, V., Miura, Y., Matsui, T., Hasegawa, T., Tojo, Y., et al. (2010). Absence of spontaneous action anticipation by false belief attribution in children with autism spectrum disorder. Development and Psychopathology, 22(2), 353–360.CrossRefPubMedGoogle Scholar
  30. Sinha, P., Kjelgaard, M. M., Gandhi, T. K., Tsourides, K., Cardinaux, A. L., Pantazis, D., et al. (2014). Autism as a disorder of prediction. Proceedings of the National Academy of Sciences, 111(42), 15,220–15,225.CrossRefGoogle Scholar
  31. Vivanti, G., McCormick, C., Young, G. S., Abucayan, F., Hatt, N., Nadig, A., et al. (2011). Intact and impaired mechanisms of action understanding in autism. Developmental Psychology, 47(3), 841–856.CrossRefPubMedPubMedCentralGoogle Scholar
  32. von der Lühe, T., Manera, V., Barisic, I., Becchio, C., Vogeley, K., & Schilbach, L. (2016). Interpersonal predictive coding, not action perception, is impaired in autism. Philosophical Transactions of the Royal Society B, 371(1693), 20,150,373.CrossRefGoogle Scholar
  33. von Hofsten, C., Uhlig, H., Adell, M., & Kochukhova, O. (2009). How children with autism look at events. Research in Autism Spectrum Disorders, 3(2), 556–569.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media LLC 2018

Authors and Affiliations

  1. 1.Department PsychologyLudwig-Maximilians-Universität MünchenMunichGermany