Edited by: Bert Timmermans, University Hospital Cologne, Germany
Reviewed by: Marc Slors, Radboud University Nijmegen, Netherlands; Leon De Bruin, Ruhr-University Bochum, Germany
*Correspondence: Vivian Bohl, Department of Philosophy, Institute of Philosophy and Semiotics, University of Tartu, Jakobi 2, Tartu 51003, Estonia. e-mail:
This is an open-access article distributed under the terms of the
Traditional
In the past three decades, intensive discussions on social cognition have taken place in philosophy and in empirical sciences. Until recently, the so called
The aim of this paper is to argue that ToM and interactionism ought not to be considered as mutually exclusive opponents. Instead, they should be integrated into a single comprehensive theoretical framework for understanding social cognition. In order to develop new research questions and hypotheses from the integrative ToM-interactionist framework, we draw upon dual process theories of social cognition that contrast two different types of social cognitive processing. The first type (labeled
The paper is divided into several parts. In section “Theory of Mind and Interactionism” we introduce the theoretical framework of the traditional ToM accounts and outline the interactionist alternative to ToM. Next, in section “ToM or Interactionism? Both!”, we argue that ToM and interactionism address different aspects of social cognition and should therefore not be considered as mutually exclusive paradigms, but rather as complementary. This is followed by a brief overview of the dual process accounts in section “The Dual Models of Cognition: Type 1 and Type 2 Processes”. Here we introduce the dual process theories of social cognition that have been proposed in social psychology and social neuroscience, and we draw parallels between the Type 1 and Type 2 processes on the one hand, and the interactionist and ToM accounts on the other. In “Joining Theory of Mind and Interactionism to Study Dual Processes” we propose that marrying ToM and interactionism for studying different types of social cognitive processes leads to building a new comprehensive theoretical framework that raises new interesting research questions and hypotheses. In “An Integrative Framework” we outline a new hypothesis, based on the integrative account, that Type 1 and Type 2 processes can have mutual effects upon each other during social interactions. Finally, in section “Theoretical Concerns Become Methodological Concerns”, we address some of the methodological issues that follow from this new approach and discuss how the integrative framework can guide experimental research on social interaction.
The framework of ToM encompasses two theories: TT and ST. There are several versions of both of them, as well as hybrid versions that combine the two. The central claim of ToM is that we understand other people's behavior by attributing mental states to them.
According to TT, our ability to attribute mental states to others as well as to ourselves relies on the use of a specific theory—folk psychology. Folk psychology postulates mental states (beliefs, intentions, desires, pains, fears, etc.) as theoretical entities that are bound together in commonsensical principles and generalizations, enabling us to explain and predict observable behavior in mental terms, i.e. to “mindread” or “mentalize”
According to ST, there is no need for mastering the folk-psychological theory, since we use our own minds as models for understanding other minds. We apprehend the mental states of others by simulating them with the help of our own cognitive mechanisms. There are various versions of ST. Simulation can either be taken to be a conscious use of imagination and inference, i.e., mentally putting oneself in the “shoes” of the other (Goldman,
As stated above, various authors have also argued in favor of hybrid versions that combine theorizing and simulation in order to overcome the problems that each of these theories face by themselves (e.g., Nichols and Stich,
Although TT and ST have been developing as rivals, they share several fundamental assumptions and methodological strategies (which is why we discuss them under the common label “theory of mind”)
According to interactionists, one of the core implications, no matter which version of TT, ST, or their hybrids we consider, is the denial of the possibility that we can directly grasp other people's mental states (see e.g., Reddy,
Another criticized assumption is that of “third-person mindreading” being essential for social cognition (see e.g., Gallagher,
A final common feature to all ToM approaches is that they focus on cognitive processes of individuals and pay little attention to other enabling factors of social cognition, such as context, environment, and embodiment. The aim of ToM is to explain what cognitive processes are responsible for human social cognition. This aim has been criticized as being excessively individualist and overly cognitivist because it underestimates the supra-individual
These common assumptions have had strong influence on the empirical methodology that has been used for testing the empirical hypotheses of ToM. For example, in the classical false-belief tasks, test subjects are asked to interpret others' behavior by merely looking at a scene as passive observers; they do not interact with the characters they observe. Interactionists have pointed out that observational mindreading tasks reveal little about the processes that are important for participatory social interactions
We will discuss some of the assumptions of ToM to a greater extent in the next section where we introduce the interactionist position as based on a critique against ToM.
While standard ToM approaches rarely emphasize social interaction, it is the heart of the approach that we ascribe to “interactionists”: “[w]hat we call social cognition is first of all social interaction” (Gallagher,
The position of interactionists can be characterized as anti-individualist
Interactionism embraces insights from enactivism, which is the view that cognition first and foremost consists of actively relating with the environment (see e.g., McGann and De Jaegher,
While advocates of the ToM approach do not specify their concept of perception, they seem to have a rather narrow notion of it, given that one of their core assumptions is arguably that the minds of others are imperceptible. For interactionists, in contrast, perception is one of their core concepts, being rich both in content and function: perception does most of the work needed to understand and interact with others
What about the interactionist understanding of mind and mental states? It is not easy to give interactionist account of them, as these notions do not seem to inherently belong to interactionist vocabulary. However, interactionists often use these concepts in order to enter into a dialog with the proponents of ToM, or, more broadly, with analytic philosophers of mind and cognitive psychologists. In contrast to ToM, interactionists claim that the mental states of others are not things which are completely hidden away or of a solely theoretical nature. The upshot is that in most everyday social encounters there is little need to figure out, either explicitly or implicitly, either via folk psychological theory or via simulation, what the hidden mental states of others behind their behavior are (Zahavi,
According to ToM (especially TT) our epistemic attitude toward others in its presumed third-person observational format is similar to the attitude of scientists toward their research objects. Interactionists, in contrast, emphasize the importance of affect and engagement in interpersonal interactions, which is assumed to give access for the participants to information that is otherwise much less obvious (Reddy and Morris,
In general, interactionists have been rebelling against the monopoly of ToM over social cognition research. In doing so, some of them, that we refer to as
At this point, it is crucial to distinguish between the different aims of social cognition research envisioned by interactionists and proponents of ToM. Asking the general question of how to explain human social cognition is different from the more specific issue of what individual cognitive (neural and psychological) mechanisms appertain to social cognition. According to interactionists, the advocates of ToM make the mistake of equating these two questions. At the same time, radical interactionists seem to make a similar mistake by claiming that explaining social cognition is a matter of detecting and explaining the perceptual and supra-individual processes that constitute real life social interactions.
Some of the radical interactionists are also
Another vulnerable aspect of interactionism is that it tends to over-emphasize social interactions that are honest, smooth, and cooperative. It might well be the case that in situations where social interaction evolves smoothly in the direction that is agreeable for all participants, there is little need for the participants to bother with attributing mental states to others based on either simulation or inference. But it is questionable whether this is the case in situations of competition, disagreement, conflict, or obvious misunderstandings. It is surely possible that people sometimes observe other people from a purely third-person point of view and, moreover, attribute mental states to others to gain a better understanding of what they are up to.
Also, it is hardly possible to expand the characteristic of perceptibility to all mental states as it is obvious that some mental states (e.g., thoughts that are not expressed) are less perceivable than others (e.g., strong emotions), even by interactionist's standards. At times we do indeed find ourselves trying to figure out those more hidden aspects of others' mental lives. In addition, it is doubtful that interactionist approaches are able to account for many linguistic forms of interaction (e.g., those requiring Gricean assumptions about communicative intentions, see Grice,
To conclude the subject, a theory of social cognition should be able to account for both, the more direct forms of social interaction where no mindreading seems to be necessary, and the more detached and sophisticated forms of social understanding that require mindreading.
Recently, a small number of authors have provided different arguments in favor of merging elements of ToM and interactionism instead of preferring one account over the other (Gangopadhyay and Schilbach,
We have tried to demonstrate that while ToM is an individualist and a mentalizationist approach, interactionism is in many respects its antipode in preferring explanations in terms of embodiment, perception, and supra-individual processes. It seems, however, that ToM and interactionism may describe different aspects of social cognition. The important issue is therefore not asking whether ToM or interactionism is right, but rather asking what aspects of social cognition do these approaches capture and how are these different aspects related.
Interestingly, it seems that both ToM and interactionism aim to give alternative explanations to the same empirical findings (e.g., developing alternative accounts of the function of mirror neurons or interpreting developmental studies differently), but often they also refer to different empirical studies (cf. e.g., Goldman,
Many everyday tasks in our lives require high speed and effectiveness of processing on the one hand, and great flexibility together with executive control on the other. Social interaction is no exception. Empirical evidence demonstrates that these characteristics do not fit well together, so how can coping in these tasks be explained? We favor a dual process approach to social cognition
Several recent cognitive models for social cognition in psychology and social cognitive neuroscience contrast two different processing types labeled e.g., lower level vs. higher level (Apperly,
The existence of numerous differently labeled but analogous dual process accounts is partly the result of the fact that many of them have been proposed in narrow fields of research independently of each other (e.g., in cognitive psychology and in social psychology, see Evans,
Type 1 processes are typically described as fast, efficient, stimulus-driven, and relatively inflexible. They place hard constraints on what information is considered or how it is processed to gain high efficiency and speed at the expense of low flexibility. These processes are thought to be closely linked with affect and their outcome is usually experienced as perceptions or feelings. Several authors have also pointed out that Type 1 processes might be evolutionarily older and shared with other species. Type 2 refers to a type of information processing that is relatively slow and cognitively laborious, making the opposite trade-off by being relatively flexible. The Type 2 processes typically involve some combination of effort, intention, and awareness. The processes tend to interfere with one another, and their outcome is sometimes experienced as self-generated thoughts. The Type 2 processes are thought to be responsible for explicit reasoning and to have no direct link with emotion. As such, they are considered to be evolutionarily more recent and uniquely developed in humans. (See e.g., Chaiken and Trope,
The typically correlated characteristics listed above should not, however, be considered as necessarily co-occurring features (Stanovich and Toplak,
It is beyond the scope of this paper to give a fair review on the wealth of the dual process accounts and their critiques
Dual process accounts of social cognition have emerged independently in the social neurosciences. Recent neuroscientific research suggests that the understanding of others' intentions is supported by two neural systems that perform complementary roles;
Although the MNS is thought to underlie the understanding of fairly complex motor intentions, it also has its limitations. For instance, Rizzolatti and Sinigaglia claim that “[u]nderstanding the reasons behind an agent's motor intention requires additional inferential processes” (2010, p. 271). Whilst we can directly grasp that someone intends to pick up a book from the table, the MNS is not thought to provide us with an understanding of the reasons that underlie that motor intention (for example, the wish to read the book, to put it back on the shelf or to bring it back to the library).
Generally, this distinction in action understanding corresponds to the difference between
Our aim is not to give an exhaustive overview of all processes or brain areas involved in social cognition (for extensive reviews see Lieberman,
In this section we argue that there is a viable parallel to be drawn between the interactionism vs. ToM discussion and the Type 1 vs. Type 2 process models of cognition and that a promising way for enhancing research on social cognition is to consider interactionism as emphasizing those aspects of social cognition that pertain mostly to Type 1 processes, while considering ToM as emphasizing aspects that typically characterize Type 2 processes.
Interactionism stresses the perceptual and the affective dimension of social cognition. There is a clear parallel here to Type 1 processes. It is often stated in the dual process models that the outcome of Type 1 processes is usually experienced as perceptions, feelings, or emotions, whereas we have little conscious control over the processes that lead to these outcomes. The “intuitive mind” is considered to be shared across species and it is thought to be both phylogenetically and ontogenetically older and therefore more basic than the “reflective mind” (Evans,
On the other hand, standard accounts of ToM aim to find out what makes human social cognition distinctively human (see e.g., Saxe,
The connection between interactionism and the Type 1 processes may be more difficult to grasp. Earlier we criticized the tendency of the “radical interactionism” to prefer supra-individual descriptions over sub-personal explanations where actually both are needed. If interactionists lack sub-personal explanations for social cognition then the connection between Type 1 processes and interactionism is not obvious—after all, Type 1 processes are processes described on the sub-personal level. Most of the interactionists, however, do agree that a sub-personal story of the processes underlying social perception needs to be given as a part of social cognition theory. We argue that interactionism has potential to enrich our understanding of the Type 1 processes. Firstly, although the explanations given by ToM may seem plausible for explaining the functions of the Type 2 processes, the interactionist critique gives good reasons to doubt that the same explanations are suitable for the Type 1 processes. Secondly, interactionism provides detailed personal-level descriptions of social perception which can be used as proper scientific
In relation to the parallels outlined in the previous section, the question arises: what reasons are there to think that the Type 1 processes should be considered in terms of interactionism rather than in terms of implicit ToM? Implicit versions of ToM state that mindreading via theorizing or simulation takes place automatically and tacitly on the sub-personal level of processing. Some proponents of ToM have made a distinction between two kinds of mindreading which is similar to the differentiation between Type 1 and Type 2 processes. For example, Alvin Goldman (
Firstly, the concept of simulation involves some specific implications. For example, speaking of mirror neurons in terms of simulation implies that there is a primary and a secondary function for the mirror neurons. It is usually assumed that the primary function for mirror neurons is to contribute to the person's own actions and the secondary function is to simulate the actions of another “off-line”. However, it is possible that the activation of mirror neurons is not person-specific, so it has only one function; namely, to represent (a type of) an action as such. Information about who is the agent of the action may be “tagged” to this information in a later stage of processing. (see De Vignemont and Fourneret,
Secondly, even if the use of the concept “simulation” were a good way for speaking about the mirror neurons, this does not automatically guarantee the success of any form of ST in a strict sense. ST traditionally implies much more than simply using the concept of simulation for some sub-personal processes that matter for social cognition. In principle, one can loosely use the concept of simulation for the mirroring processes without being a proponent of any form of ST in a strict sense
Finally, even if the function of mirror neurons were best described by ST, the study of the role of these neurons for social interaction still requires a wider interactionist methodology. In addition to the observational experiments that have been dominating the field within the ToM framework, mirror neurons and other cognitive and affective processes need to be studied in real life situations and in interactive contexts, where interactionists have argued that specific environmental and contextual factors may be important for social cognition.
Overall, we argue that it is highly unlikely that all aspects of social cognition can be explained in the framework of ToM, which focuses on the issue of how we attribute mental states to others. It is more parsimonious and evolutionarily more plausible to think that not all aspects of human social cognition rely on mental state ascriptions and to consider the capability for mental state ascriptions to have evolved from other, more primitive social abilities that we share with our evolutionary ancestors. ToM was initially proposed as an account for deliberate reasoning about other individual's mental states. As it became clear that people do not constantly reason about others' mental states, which made explicit versions of ToM unlikely to be correct as general theories of social cognition, most of the mindreading was hypothesized to happen implicitly. Although it is possible that over time mindreading becomes habitual and carries on automatically on an implicit level, making implicit mindreading responsible for all aspects of human social cognition is an evolutionary non-starter, as it leaves a huge gap between human and animal social cognition.
All these trains of thought suggest that one should be at least careful to associate mirror neurons with the ST exclusively and to consider ToM as an approach that is able to explain all aspects of human social cognition. Alternative approaches to Type 1 processes of social cognition must be considered, especially those that interactionism has a potential to offer.
To sum up, our basic claim is that there is a coarse-grained mapping between the Type 1 processes and interactionism on the one hand, and the Type 2 processes and ToM on the other. We do not expect this mapping to be clean and perfect—undoubtedly ToM in passing mentions aspects of social cognition that can be related to Type 1 processes and vice versa—but in general the tendencies are rather clear. On the example of how dual processes operate in a single brain and complement each other, one can think of ToM and interactionism as complementary theories of social cognition in general. The aim of proposing the new integrative framework is, however, not merely to dissolve the ToM vs. interactionism opposition, but to generate new research questions and hypotheses that would help to gain new knowledge on social cognition and especially on social interaction. In the following section, we outline a hypothesis on the relation of Type 1 and Type 2 processes.
Different dual process accounts for social understanding have been proposed by several authors over the past 10 years (e.g., Lieberman et al.,
Some authors have suggested that the MNS and ToMS constitute independent networks and have complementary roles in social cognition (e.g., Saxe and Wexler,
Furthermore, van Overwalle and Baetens (
A more common hypothesis in dual process models in general is that the Type 1 processes inform and support the Type 2 processes (e.g., Blakemore and Decety,
In line with this idea, some authors have suggested that the Type 1 processes associated with the MNS are a prerequisite for the ToM processes (e.g., Ohnishi et al.,
More recently, evidence has emerged in separate fields of research to support the idea that higher order (i.e., Type 2) representations regulate or bias perceptual (i.e., Type 1) processes. Several studies have already shown that higher order intentions can directly affect visual perception (Allport,
Taken together, the social neuroscience literature suggests that intentions are processed (and understood) at different levels of abstraction by different types of processes and that there are reciprocal interactions between these types of processing. Inspired by the predictive coding theory of the MNS (Kilner et al.,
Thus, although we fully agree that the “dynamic interaction between [distal] and [proximal] intentions modulates the processing of the observed actions of other people” (Ondobaka et al.,
From this integrative account, two major conclusions for the interdisciplinary study of social cognition follow. Firstly, it follows that although it is possible to study processes emphasized by ToM and interactionism in isolation from each other, if we want to know how social cognition functions in real-life, we should rather study how these different processes influence one another. Secondly, in order to study the interaction between Type 2 (ToM) and Type 1 (interactionism) processes, we need to study actual social interactions and how these unfold over time. Figure
Originally the field of social psychology was aimed at understanding how the behavior of people was influenced by the actual or imagined presence of others (Allport,
The interactionist critique addressing ToM is linked to a discussion over what empirical methods should be adopted for doing research on social cognition. Drawing on the discussions over the importance of “third-person” mindreading versus “second-person” engagement, interactionists insist that there is a need to overcome the “methodological solipsism” (Macmurray,
An interactive aspect of communication is in fact already inherent in standard “third-person” experiments. For example, before an experiment officially begins, the experimenter interacts with the test subject in order to explain what will happen and what is expected from the test subject. As Gallagher (
Such examples illustrate the interactionist claim that engaged social interactions are phenomenologically and cognitively different from passive observations of others. Although the suggestions for adopting a “second-person” methodology (experimenters interacting with test subjects) may seem too radical or not sufficiently rigorous to count as scientific, a less problematic way to improve the methodology is to find new ways for studying subjects while they are actively interacting during experiments rather than testing single persons who passively receive social stimuli. We hypothesize that studying social interactions in this manner will also allow us to learn more about the interaction between Type 1 and Type 2 processes. On many occasions, the division of labor between cognitive and social psychology has been drawn too sharply; both disciplines would gain from closer cooperation.
Within the neurosciences, social interactions have mostly been studied using game theoretical paradigms (for review see Rilling and Sanfey,
More recently, several research groups have developed novel experimental paradigms that are specifically aimed at revealing the neural correlates of the processes underlying social interactions. One such study, designed by Schilbach and colleagues, involved an interactive gaze following paradigm (Schilbach et al.,
The richness of social interactions that make controlled experiments so challenging also provides numerous new variables for experimental manipulation. Some examples include: the temporal dynamics of social interaction, degrees of coordination between participators, the nature, and history of the relation between participants (strangers/acquaintances/friends, etc.), the directness or mediation of contact, behavioral factors (body postures, gestures, eye-contact, etc.), experiential characteristics, or the mode of interaction (cooperative vs. competitive). For instance, cooperative vs. competitive social interactions are phenomenologically very different: cooperative interactions often feel effortless, whereas competition can be much more mentally taxing. To what extent do the relative contributions of Type 2 and Type 1 processes in cooperative vs. competitive contexts differ? What occurs when there is a shift from a cooperative to an antagonistic encounter? Or how might one's reputation (as a cooperator or non-cooperator) affect basic motor coordination processes? Although some of these questions can be answered using methods that have been traditionally used in the social cognitive neuroscience, the analyses of social interaction dynamics may require new methods, for example those used in dynamical systems analyses (De Jaegher,
Another challenge is to understand the putative roles of the ToMS and MNS processes in social interactions, and importantly the interaction between these processes. As we have pointed out, previous imaging studies have revealed the involvement of both systems, but have provided little insight in their possible interaction. Here we think that connectivity analyses
Finally, developmental studies provide unique opportunities to see how the MNS and ToMS interact in ways that are not typical for adults when they are fully matured (De Haan and Gunnar,
In this paper, we have characterized two rival approaches in the current social cognition research: the traditional
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The idea for the paper emerged from a series of meetings with other members of the interdisciplinary research group
1We use the concepts
2The term
3This implication does not hold for Gordon's simulation theory, however. See Gordon (
4It does not mean that the only alternative to the unobservability thesis is behaviorism. Interactionists simply deny that there is a sharp boundary between observable behavior and mental states, which makes mental states completely “hidden” from perception. It is, however, debatable, whether giving up the unobservability thesis would change anything substantial for ToM, as the question of “what are the sub-personal mechanisms that enable us to perceive the mental states of others?” would still remain. Some proponents of ToM have even stated that their view is compatible with the idea of observability of mental states (e.g., Carruthers,
5Supra-individual factors are factors which either lie outside of an individual or are not reducible to processes inside of an individual, such as e.g., physical and social environment, context, temporal dynamics of the situation, etc. If pressed on this point, proponents of ToM would not deny that supra-individual factors play a role in social understanding, but according to interactionism, ToM has downplayed the importance of supra-individual factors.
6The methodological situation has started to change, presumably due to interactionist criticism. For instance, new “second-person” versions of the standard false-belief task have recently been developed where experimenters interact with test subjects in pragmatic contexts (see e.g., the study of Knudsen and Liszkowski,
7We distinguish between “radical” and “modest” anti-individualism. While “radical anti-individualism” is the view that once social cognition is explained on a supra-individual level, explanations on other levels become superfluous, “modest anti-individualism” insists that explanations on personal and sub-personal levels of explanation should be complemented by explanations on supra-individual level. On our account, all versions of interactionism are at least modestly anti-individualist, whereas some versions tend to be radical in their anti-individualism.
8See McGann and De Jaegher (
9Some proponents of ToM, especially simulationists (e.g. Goldman,
10Less radical interactionists admit that in some cases (e.g., if the behavior of the other puzzles us) we may rely on simulation or theorizing, so they do not insist on ruling out the mindreading paradigm completely (see e.g., Gallagher,
11In cognitive science it is a commonplace to distinguish between personal and sub-personal levels of explanation. This distinction goes back to Dennett (
12We are grateful to one of the reviewers for bringing this to our attention.
13Dual process accounts have been proposed by a great number of authors in a variety of cognitive domains, such as artificial intelligence, general reasoning, judgment, number cognition, memory, and social cognition. Here the discussion is restricted to dual process models of social cognition. In social psychology, the earliest models of dual processing date back to the 1980s; see e.g., Chaiken,
14We want to highlight one issue that is often raised by critics of dual process models: the fact that when we learn certain skills (e.g., learn to drive), we seem to move from deliberative Type 2 processes toward more automatic Type 1 processes. This can be seen as a problem for dual process accounts, because it suggests either a gradient between Type 1 and Type 2 processes or a magical translation from Type 1 to Type 2 processes. The issue is beyond the scope of the current paper, but because of its importance, we address it briefly by drawing on the richer field of the neuroscience of learning. Numerous studies have shown that there are different learning mechanisms operating in parallel, and these mechanisms are associated with different brain systems (for an overview, see Balleine et al.,
15Note that we use the processes associated with the MNS and ToMS as examples of Type 1 and Type 2 processes that are involved in social cognition, these are not meant to be considered exhaustive. For instance, other Type 1 non-MNS processes may exist, such as lower level perceptual processes.
16Chimpanzees and some other primates may be a possible exception. The issue of whether higher non-human animals are capable of some primitive form of mindreading has been a controversial matter since the end of 1970's and is far from being settled (cf. Premack and Woodruff,
17The same kind of argument can be used against TT. For example, claiming that social perception is inferential does not guarantee the success of TT.
18As one of our reviewers pointed out, one might argue that the two points about the concept of simulation could be met simply by adjusting one's definition of simulation. Our point is, however, that the interactionist critique of ST is not merely verbal—it is not reducible to the issue of whether one should use the term “simulation” or refrain from using it. Adjusting the definition of simulation to the extent that it would meet the above criticism would make the concept of simulation vacuous and would therefore rather weaken than strengthen the position of ST.
19The arrows on Figure
20Connectivity analyses are statistical methods that are aimed at understanding how the activity in one brain area (X) is related to another brain area (Y). One simple way of analyzing a possible relationship is to look at the correlation between time-series of areas X and Y. This is somewhat comparable of comparing how the index of different financial markets changes over time, and how the rises and falls of those indexes are correlated. In neuroimaging experiments, connectivity analyses are often aimed at testing whether the correlation between brain areas is affected by psychological states. Such analyses could for instance test whether there is a stronger connectivity between areas X and Y when someone observes a congruent or an incongruent action.
21DCM and GC are tools for testing different models of “causal” relationships between activities in multiple brain areas. As with other connectivity analyses, this involves the statistical analyses between time-series of different brain areas. The main idea is that a “causal” relation is assumed between X an Y when it is possible to predict changes in the time series of brain area Y based on what happens earlier in the time-series of X.