Understanding the Development of Theory of Mind: Contributions from Neurobiology
Understanding the Development of Theory of Mind: Contributions from Neurobiology
When the term was first used by Premack and Woodruff (1978), theory of mind was meant to refer to the ability of attributing mental states to others in order to make inferences about their behaviour. Since that time, the definition of theory of mind has undergone considerable change and the conceptualization of theory of mind development has been largely debated amongst psychologists (Astington, 2006). In one camp, researchers have argued for a modular, or domain-specific, account of theory mind (Frith & Frith, 1999; Saxe & Powell, 2006). Within this view, theory of mind is an innate psychological structure that develops independently and has evolved specifically to deal with issues of human social functioning. Alternatively, other researchers have argued that theory of mind is an integrated construct that incorporates domain general abilities such as executive function and language (Milligan, Astington, & Dack, 2007; Apperly, Samson, Chiavarino, Bickerton, & Humphreys, 2007). According to this constructivist framework, the development of theory of mind is largely influenced by the concurrent development of associated cognitive abilities and is, therefore, more variable in its expression and underlying structure. Importantly, proponents of this latter approach do not believe that theory of mind understanding necessitates the development of a domain-specific mechanism (Stone & Gerrans, 2006). Thanks to recent advancements in the field of developmental neuroscience, a potential resolution to this debate might be found through neurological mappings of social cognitive functions. The purpose of this paper will be to review current neurological evidence in an effort to assess the validity of each account. Specifically, this paper will address whether (a) theory of mind understanding arises from the development of neurological structures that are specialized to respond to the processing of mental states, or (b) theory of mind understanding arises from the integration of multiple structures that are associated with the domain-general ability of metarepresentation.
Evidence in Support of a Modular Approach
Proponents of a modular account for theory of mind believe that a domain-specific mechanism is primarily responsible for the attribution and understanding of mental states. Evidence in support of this account has traditionally come from research examining the false belief understanding – i.e., understanding that the mental states of others can contradict a current state of reality – of typical and atypical populations (Leslie & Thaiss, 1992; Saxe & Baron-Cohen, 2006). Specifically, research has shown that individuals can have unique deficits in false belief understanding while still demonstrating a normal range of functioning on similar tasks that do not require the metarepresentation of mental states (Leslie & Thaiss, 1992). The existence of selective deficits in mental state understanding has consequently led certain researchers to hypothesize the existence of brain structures that are selectively engaged in the processing of mental states. Importantly, proponents of this view maintain that structures associated with theory of mind understanding should show unique patterns of activation during the processing of mental state representations and should be less active during processing of non-mental state representations (Saxe & Baron-Cohen, 2006).
Through the use of brain imaging technology, researchers have recently begun to test the modular account directly and identify brain regions that appear to be closely associated with theory of mind understanding. These regions include the right and left temporo-parietal junction and the posterior cingulate (Frith & Frith, 1999; Saxe & Powell, 2006; Sabbagh & Taylor, 2000). The most consistent findings to date, however, have been for the selective activation of the right temporo-parietal junction during the processing of false belief tasks (Perner, Aichhorn, Kronbichler, Staffen, & Laudurner, 2006; Saxe & Powell, 2006). In these studies, researchers have shown that activation of the right temporo-parietal junction may be selectively associated with the processing of mental-state perspectives but not with more domain-general types of metarepresentation. For example, in a study conducted by Perner and colleagues (2006), fMRI images were obtained for adult participants who read and answered questions related to three similar metarepresentation scenarios: false beliefs, false signs, and false photographs. Importantly, the scenarios were designed so that participants would only have to consider the mental state of others when answering a question related to a false belief (e.g., a tourist who incorrectly believes the location of a landmark after being given false directions) but not when interpreting a false sign (e.g., a road sign indicating the wrong direction) or a false photograph (e.g. a dated photograph of a scene that has since changed). Thus, the researchers were able to distinguish neural activation patterns that were uniquely associated with the metarepresentation of mental vs. non-mental states. In their results, Perner and colleagues found that while activation in the left temporo-parietal junction (TPJ-L) did not distinguish between processing of false beliefs and false signs, activation in the right temporo-parietal junction (TPJ-R) was significantly greater during the processing of false beliefs than during the processing of the other two scenerios. The authors consequently concluded that the TPJ-R may be uniquely associated with the metarepresentation of mental states and may therefore be a good candidate for a theory of mind “module”. Similar findings were also obtained by Saxe and Powell (2006), who found that the TPJ-R was selectively activated when participants were asked to read descriptions of a protagonist’s thoughts or beliefs but not when they read about a protagonist’s appearance or subjective physical feelings (e.g., hunger).
Taken together, the above studies lend strong support to the modular account of theory of mind. By demonstrating that the right temporo-parietal junction is selectively activated when an individual processes the thoughts or beliefs of another person, the current evidence suggests that certain regions of the human brain do develop to specialize in theory of mind understanding. Importantly, the fact that the TPJ-R does not show a similar pattern of activation during the processing of non-mental state representations further supports the unique specialization of this region.
Evidence in Support of a Constructivist Approach
From a constructivist framework, theory of mind understanding is a specialized cognitive construct that arises from the integration of various domain-general processes. As a result, proponents of this account do not believe that the development of theory of mind requires the functional specialization of brain regions (Stone & Gerrans, 2006). Instead, theorists in this camp believe that neurological structures associated with theory of mind understanding may also be recruited for the processing of non-mental states. As proposed by Stone and Gerrans (2006), theory of mind understanding does not necessitate a domain-specific mechanism because it can more parsimoniously be accounted for by the integration of related social functioning abilities.
To date, much of the evidence in support of this account has demonstrated that that theory of mind abilities are in fact associated with highly integrated networks of neural activation (Abu-Akel, 2003). Specifically, studies have shown that theory of mind understanding often requires the activation of structures that are not uniquely associated with false belief understanding (Abu-Akel, 2003; Apperly et al., 2007; Stone & Gerrans, 2006). For example, damage to the amygdala has been shown to be associated with impairments in social understanding (Abu-Akel, 2003). Activation of the orbitofrontal cortex and medial prefrontal cortex has also been associated with the processing of mental states (Baron-Cohen, Ring, Moriarty, Schmitz, Costa & Ell, 1994; Stuss, Gallup, & Alexander, 2001). Thus, regions of the brain that are typically associated with more general processing of social input also appear to be important for the understanding of theory of mind.
Recently, research with brain damaged patients has also brought into question the domain-specificity of certain neurological structures associated with theory of mind understanding. Specifically, a study conducted by Apperly and colleagues (2004) demonstrated that patients with damage to the left temporo-parietal junction (TPJ-L) were selectively impaired on false belief tasks. This evidence contradicts Perner and colleagues’ (2006) research which shows that activation of the right temporo-parietal junction, as opposed to the left, selectively differentiated between the processing of mental state and non-mental state representations. Moreover, in a follow-up study conducted by Apperly and colleagues (2007), results indicated that patients who had previously shown selective deficits in false belief understanding also showed deficits in reasoning about non-mental state representations (i.e., false photographs). This research, therefore, suggests that similar neural pathways are recruited for the metarepresentation of mental and non-mental states. Importantly, it also suggests that the TPJ-R may not be as uniquely associated with theory of mind understanding as others have argued.
In the end, although regions have been identified that are uniquely associated with mental state reasoning, most theory of mind tasks appear to require the integrative activation of several brain regions.
Summary of Evidence: Modular vs. Constructivist Accounts
In the current paper, neurological evidence in support of both modular and constructivist accounts of theory of mind were reviewed in an effort to resolve the structural nature of theory of mind development. Specifically, this paper sought to address whether the ability to reason about mental states is dependent on: (a) the development of highly specialized neural structures, or (b) the formation of pathways between structures that are associated with domain-general functions. Based on the current evidence, it appears as if there is strong support for the existence of structures that are specialized to deal with theory of mind understanding (Saxe & Powell, 2006; Perner et al., 2006). Although several studies have also shown that mental state reasoning is associated with an integrated set of domain-general structures (e.g., medial prefrontal cortex), evidence supporting the specialized role of the right temporo-parietal junction cannot be readily dismissed. As such, the current evidence seems to suggest that theory of mind understanding is dependent on both specialized and domain-general structures in the brain. What is currently less clear, however, is whether the specialization of neural structures in the adult brain provides unequivocal support for the modular development of theory of mind in young children. As will be discussed in the following section, there are many problems inherent in making this assumption. Based on recent models of neural development (Johnson, 2000; Karmiloff-Smith, 2007), it may also be possible to account for the development of domain-specific structures using a constructivist perspective. In the final section of this discussion, an alternate model of theory of mind development will be proposed based on the principles of probabilistic epigenesis – i.e., a developmental framework that emphasizes the dynamic interactions between function and structure (Johnson, 2000; Karmiloff-Smith, 2007).
Critique of the Modular Account
While the evidence in support of a modular account of theory of mind is compelling, it is problematic to assume that the development of highly specialized neural structures is predetermined. In her framework of atypical neural development, Annette Karmiloff-Smith (2007) argues that by focussing on adult imaging data, researchers often fall into the trap of assuming the specialized neural functions arise from innate and localized neural development. Similarly, researchers may ignore the important functional changes that can occur in neural structures over time because most of these changes would have occurred by the time images are captured in adulthood. As such, structures that can appear domain-specific, or “modular”, in adulthood may in fact originate from less localized patterns of activation during childhood (Johnson, 2000; Karmiloff-Smith, 2007). This argument would certainly apply to the imaging studies reviewed in the current paper as all this research was conducted with adult populations. Although largely due to the limitations of current imaging technology, no study to date has examined the neural structures associated with children’s theory of mind understanding. Given that children begin to show sign of mental state awareness by 18-months (Frith and Frith, 1999) but do not typically acquire false belief understanding until the age of four-years (Astington, 2006; Saxe & Baron-Cohen, 2006), it is unlikely that the neural activation patterns of young children would appear as highly localized as those in adults.
Annette Karmiloff-Smith (2007) and Mark Johnson (2000) offer an alternate account of neural specialization that is based on the framework of probabilistic epigensis. Specifically, these researchers propose that the localization of psychological functions is determined by the dynamic interaction between genetics, neural maturation, and functional input. Early in development, few brain regions are specialized to process specific types of information. Instead, different regions or pathways, “compete” for the processing of different inputs and genetic biases ensure that certain neural pathways are better suited to process certain types of information (e.g., neurons in the occipital lobe may be biased to process visual stimuli). Through a process of natural selection, certain pathways eventually become more effective at processing specific types of input and less effective pathways are pruned, or cease to respond to the same input (Johnson, 2000; Karmiloff-Smith, 2007). Thus, neural patterns of activation become more selective and psychological functions become more efficient. What is unique about this approach is that it allows for the “partial operation” of psychological functions that have yet to develop more specialized patterns of activation (Johnson, 2000). Importantly, it also allows for highly specialized functions, such as theory of mind, to be temporarily processed by less specialized pathways in the brain (Karmiloff-Smith, 2007).
Applying Probabilistic Epigenesis to Theory of Mind Development
Using the frameworks proposed by Mark Johnson (2000) and Annette Karmiloff-Smith (2007), it is possible to propose an alternate account of theory of mind development that incorporates the evidence for both modular and constructivist approaches. First, this model would have to explain the discrepancies found the localization of mental state reasoning functions (Apperly, Samson, Chaivarino, & Humphreys, 2004; Perner et al., 2006). Second, this model would have to explain why theory of mind understanding would require both the specialization of neural structures and the recruitment of domain-general neural pathways.
With regards to the first point, this framework would predict that the localization of theory of mind functions would be open to a certain degree of variance due to the variability of individual social experiences. Thus, individuals who receive atypical social input, either because of atypical development or brain damage, may consequently develop distinct patterns of localization for the processing of theory of mind. Similarly, this framework would predict that individuals raised in an environment that demands increased amounts of mental-state reasoning (e.g., a daycare where children are engaged in vast amounts of social interaction) might show earlier patterns of localization as a result of their environmental input.
With regards to the second point, this model would predict that before psychological functions become fully localized in the brain, they can be temporarily executed by less specialized neural pathways. In the case of theory of mind, these less specialized structures may include regions of the limbic and paralimbic pathways that are associated with the processing of social input (Abu-Akel, 2003). As certain regions gain specialization in the processing of mental states, less specialized regions may still be recruited during the processing of mental states because they assist with the domain-general demands of a given task.
In the end, the application of probabilistic epigenesis to theory of mind understanding makes it possible to account for the development of domain-specific structures without having to make assumptions about the modularity of brain structures.
Recent findings from adult imaging studies have provided strong evidence for the localization of mental state reasoning in the brain. Although the existence of highly specialized structures may provide some support for modular accounts of theory of mind development, recent models of neural development raise some doubts to this conclusion. Alternatively, a model of theory of mind development based on the principles of probabilistic epigenesis may offer a more comprehensive explanation, not description of the current neurological evidence. Specifically, this model can account for potential discrepancies in the functional localization of mental state reasoning. In addition, this model can account for the simultaneous recruitment of highly specialized and domain-general pathways during the processing of mental state representations.
This was a very fine piece of work. I’m impressed! Your review of the evidence was clear and articulate and you framed the debate with insight as well as precision. You captured the gist of the argument with exquisite balance, as well as economy of writing, and you spelled out each side’s contentions precisely as well as concisely. But then you actually introduced a fresh vantage point which was capable of resolving the discrepancies in the data, by applying an intrinsically developmental formula to the problem. You showed how specialization might emerge, rather than being programmed in, through selectionist processes, so that specialized structures might show up against a background of more domain-general structures. This argument could have used further refinement and development, but the point of it was completely valid and justified. And your use of the Johnson model to tackle this particular problem was absolutely original, as far as I know. In fact, this degree of original problem solving in a domain which is not your home turf is really something to be proud of. It is probably the only approach that could actually resolve the debate without undermining either side. This was a rare treat in a term paper. Excellent job!
Abu-Akel, A. (2003). A neurobiological mapping of theory of mind. Brain Research Reviews, 43, 29-40.
Apperly, I.A., Samson, D., Chiavarino, C., Bickerton, W.L., & Humphreys, G.W. (2007). Testing the domain-specificity of a theory of mind deficit in brain-injured patients: Evidence for consistent performance on non-verbal, “reality-unknown” false belief and false photograph tasks. Cognition, 103, 300-321.
Apperly, I.A., Samson, D., Chiavarno, C., & Humphreys, G.W. (2004). Frontal and left temporo-parietal constributions to theory of mind: neuropsychological evidence from a false belief task with reduced language and executive demands. Journal of Cognitive Neuroscience, 16, 1773-1784.
Astington, J.W. (2006). The developmental interdependence of theory of mind and language. In N.J. Enfield & S.C. Levinson (Eds.). Roots of human society: Culture, cognition, and interaction (pp. 179-206). New York, NY: Berg.
Baron-Cohen, S., Ring, H.A., Moriarty, J., Schmitz, B., Costa, D., & Ell, P. (1994). Recognition of mental state terms. Clinical findings in children with autism and a functional neuroimaging study of normal adults. British Journal of Psychiatry, 165, 640-649.
Frith, C.D., & Frith, U. (1999). Interacting minds: a biological basis. Science, 286, 1692-1695.
Johnson, M.H. (2000). Functional brain development in infants: Elements of an interactive specialization framework. Child Development, 71(1), 75-81.
Karmiloff-Smith, A. (2007). Atypical epigenesis. Developmental Science, 10(1), 84-88.
Leslie, A., & Thaiss, L. (1992). Domain specificity in conceptual development. Cognition, 43, 225-251.
Milligan, K., Astington, J.W., & Dack, L.A. (2007). Language and theory of mind: Meta-analysis of the relation between language ability and false-belief understanding. Child Development, 78(2), 622-646.
Perner, J., Aichhorn, M., Kronbichler, M., Staffen, W., & Ladurner, G. (2006). Thinking of mental and other representations: The roles of left and right temporo-parietal junction. Social Neuroscience, 1(3), 245-258.
Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a ‘theory of mind’? Behavioural and Brain Sciences, 4, 515-526.
Sabbagh, M.A., & Taylor, M. (2000). Neural Correlates of theory-of-mind reasoning: An event related potential study. Psychological Science, 11(1), 46-50.
Saxe, R., & Baron-Cohen, S. (2006). Editorial: The neuroscience of theory of mind. Social Neuroscience, 1(3), 1-9.
Saxe, R., & Powell, L.J. (2006). It’s the thought that counts. Psychological Science, 17(8), 692-699.
Stone, V.E., & Gerrans, P. (2006). What’s domain-specific about theory of mind? Social Neuroscience, 1(3), 309-319.
Stuss, D.T., Gallup, G.G., & Alexander, M.P. (2002). The frontal lobes are necessary for ‘theory of mind’. Brain, 124, 279-286.