Is there really no Theory of Mind deficit in autism? Part III: Do Theory of Mind tests fail to predict autistic traits?

In my first two posts on Gernsbacher and Yergeau’s 1029 paper, Empirical Failures of the Claim That Autistic People Lack a Theory of Mind, I critique (here and here) their claims that Theory of Mind (ToM) tests lack empirical validity: that the original test results with autistic subjects have failed to be replicated, and that the tests themselves fail to converge on a meaningful psychological construct.

Morton A. Gernsbacher, Professor of Psychology, University of Wisconsin

As the article proceeds, Gernsbacher and Yergeau argue that ToM tests also fail to predict both autistic traits and autism-related social phenomena. This argument, too, is problematic. The ToM tests the authors consider here are:

• The Strange Stories Test, which measures the ability to deduce the social/emotional reasons for characters’ behaviors in a narrative sequence

• The Animated Triangles Test, which measures the ability to ascribe emotions to self-propelled, interacting shapes in a short animation

• The Eyes Test, which measures the ability to recognize emotions from facial information that is restricted to a rectangular region around the eyes

• The Faux Pas Test, which measures the ability to detect social blunders

• The false-belief tests, which measures the ability to make inferences about the beliefs held by (or about) individuals who are missing key pieces of information

  • The Sally-Anne “unexpected location change”  test, in which Sally doesn’t witness Anne changing the location of her marble

  • The Smarties “unexpected contents” test, in which a candy box contains a pencil rather than candy

First, some of the studies Gernsbacher and Yergeau reference look only at non-autistic individuals. As discussed earlier, however, ToM tests appear to operate differently in non-autistic individuals: recall the lower levels of language development that non-autistic individuals need in order to pass them. This, together with the likelihood that social variables (including autistic traits) operate differently in non-autistic individuals  it’s not clear how relevant these studies are.

As for those studies that do involve individuals with autism, some of these pertain only to second-order ToM tests (e.g., “Where does Ann think that Sally will look for her marble?”). These tests, as I discussed earlier, lack the specificity of first-order ToM tests (particularly the implicit ToM tests that don’t involve language) in distinguishing individuals with autism. As for the first-order ToM tests (“Where will Sally look for her marble?”), some of the studies, on one hand, do show a failure of ToM tests to correlate with either individual autistic traits or with autism severity. On the other hand, however, these studies consistently show performance on first-order ToM tests, particularly the language-free implicit false-belief test, as significantly distinguishing of autism.

One of the issues here is that not all individual autistic traits—for example, restrictive/repetitive behaviors or social motivation—are directly related to what the different ToM tests tap into (i.e., cognitive perspective-taking, in the case of false-belief tests; or facial expression reading, in the case of the Eyes Test). Burnside et al. (2017), for example, found that performance on the implicit false-belief tasks and social motivation tasks were unrelated. Similarly, Salter et al. (2008) found that performance on the Animated Triangles test was not well correlated with standardized parent reports of behavior or with the quality of the child's interactions with an observer. Finally, Clemmensen et al. (2016) found that a battery of ToM tests, including first and second-order false-belief tests, as well as the ability to detect white lies, irony, double bluffs, and social faux pas, was not significantly correlated with a Social Responsiveness Scale, a comprehensive measure of autistic traits that includes restrictive/repetitive behaviors.

Another confound is the heterogeneity of the autism spectrum, with the possibility of various traits and factors combining in different ways in different subgroups. Lombardo et al., 2015, for example, found evidence for two discrete subgroups, one which showed significant difficulty with the Eyes test and another that did not—even though “[t]hese ASC [autism spectrum condition] subgroups are not systematically different across a range of other variables including … autistic traits... and autism symptom severity.” Gender, too, may play a role.  Baron-Cohen et al., 2015, for example, found better performance in in the Eyes test, along with a tight correlation between Eyes test performance and Autism Spectrum Quotient, in females with autism.

Another confound is executive function (EF). There is some evidence that EF skills contribute to performance on some of the false-belief tests and that autism involves EF deficits as well as ToM deficits. Lukito et al. (2017), for example, note an association between EF and ToM impairments and a correlation between EF skills and performance on the false-belief and Strange Stories tests. And a three-year study by Ozonoff & McEvoy (1994) found that

both executive function and theory of mind abilities are seriously deficient in autistic individuals, improve little with development, may never reach normal functioning levels, and appear to eventually hit a developmental ceiling.

They note that “[t]he similar developmental trajectories of executive function and theory of mind performance… suggest that these skills may be related and interdependent. “

While many of the above studies cited by Gernsbacher and Yergeau partially support their claim that ToM tests don’t tightly correlate with autism symptomology/severity, other studies they cite fail to support this claim. Ronald et al., (2006), for example, found that autistic traits, particularly communication impairments, significantly predict ToM performance. Lukito et al. (2017), meanwhile, found that impairments in EF abilities were specifically associated with ADHD symptoms while impaired ToM was specifically associated with ASD symptoms. Jones et al. (2018) found that ToM ability was associated with both social communication symptoms and restrictive/repetitive behaviors and that “in adolescence ToM ability, but not EF, is directly related to autistic symptom expression.” Dziobek et al. (2006) found that all the ToM tests they looked at distinguished the autistic and non-autistic groups in their studies, though their new test, the MASC (in which participants viewed short films and answered questions about actors’ mental states) was the most successful of all the tests in distinguishing autism.

In short:

• Some studies find that ToM tests do not predict autism-related traits, but others reach the opposite conclusion.

• Autism-related traits are heterogeneous, ranging from restrictive-repetitive behaviors to social motivation, and some of these may be more correlated with ToM test skills than others.

• Since different ToM tests measure different aspects of ToM (see my previous post), different tests may correlate (or fail to correlate) with different autism-related traits.

• ·Autistic individuals are a heterogeneous group, and, in different individuals, different ToM tests may predict, or fail to predict, different autism-related traits.

There is, in other words, no convincing evidence of a general failure of ToM tests to predict autism-related traits, and some convincing evidence that specific ToM tests do predict specific autism-related traits.

REFERENCES:

Burnside, K., Wright, K., & Poulin-Dubois, D. (2017). Social motivation and implicit theory of mind in children with autism spectrum disorder. Autism Research, 10, 1834–1844. https://doi.org/10.1002/aur.1836

Clemmensen, L., Bartels-Velthuis, A. A., Jespersen, R. A., van Os, J., Blijd-Hoogewys, E. M. A., Ankerstrøm, L., . . . Jepsen, J. R. M. (2016). A psychometric evaluation of the Danish version of the theory of mind storybook for 8–14-year-old children. Frontiers in Psychology, 7, Article 330. https://doi.org/10.3389/fpsyg.2016.00330

Dziobek, I., Fleck, S., Kalbe, E., Rogers, K., Hassenstab, J., Brand, M., . . . Convit, A. (2006). Introducing MASC: A movie for the assessment of social cognition. Journal of Autism and Developmental Disorders, 36, 623–636. https://doi.org/10.1007/s10803-006-0107-0

Gernsbacher, M. A., & Yergeau, M. (2019). Empirical Failures of the Claim That Autistic People Lack a Theory of Mind. Archives of scientific psychology, 7(1), 102–118. https://doi.org/10.1037/arc0000067

Jones, C. R. G., Simonoff, E., Baird, G., Pickles, A., Marsden, A. J. S., Tregay, J., & Charman, T. (2018). The association between theory of mind, executive function, and the symptoms of autism spectrum disorder. Autism Research, 11, 95–109. https://doi.org/10.1002/aur.1873

Lukito, S., Jones, C. R. G., Pickles, A., Baird, G., Happé, F., Charman, T., & Simonoff, E. (2017). Specificity of executive function and theory of mind performance in relation to attention-deficit/hyperactivity symptoms in autism spectrum disorders. Molecular Autism, 8, Article 60. https://doi.org/10.1186/s13229-017-0177-1

Ozonoff, S., & McEvoy, R. E. (1994). A longitudinal study of executive function and theory of mind development in autism. Development and Psychopathology, 6, 415–431. https://doi.org/10.1017/S0954579400006027

Ronald, A., Viding, E., Happé, F., & Plomin, R. (2006). Individual differences in theory of mind ability in middle childhood and links with verbal ability and autistic traits: A twin study. Social Neuroscience, 1, 412–425. https://doi.org/10.1080/17470910601068088

Salter, G., Seigal, A., Claxton, M., Lawrence, K., & Skuse, D. (2008). Can autistic children read the mind of an animated triangle?Autism, 12, 349–371. https://doi.org/10.1177/1362361308091654