What does it sound like when you think? We all experience different types of private speech-like thoughts on a given day depending on what we’re thinking about. Many researchers have studied this phenomenon by prompting subjects to repeat a word or sentence to themselves (Hinke et al., 1993; Simons et al., 2010; Aleman et al., 2005), but other kinds of inner speech have been largely ignored. Alderson-Day et al. try to expand research on inner speech in their paper “The brain’s conversation with itself: neural substrates of dialogic inner speech.” The team investigates which parts of the brain are recruited to support both monologic (talking to oneself) and dialogic (imagining a conversation between two or more people) inner speech.

In addition to its inherent intrigue as a research topic, inner speech is also involved in self-awareness, past and future thinking, and emotional reflection (D’Argembeau et al., 2011; Morin et al., 2011). Some researchers and theorists, like Vygotsky (1987), believe that internalized verbal thought reflects external dialogues one experiences with others. On this theory, thought would often take the form of conversational inner speech. Naturally, this complex inner dialogue is importantly related to atypical cognition. For example, auditory verbal hallucinations (AVH) are thought to be instances of misattributed instances of inner speech in which the patient believes another being has produced the verbalization (Bentall, 1990; Fritch, 1992). Uncovering more about the social mechanisms underlying inner speech will be key in learning how it relates to AVH‌.

The authors of this paper explain that internal conversation is much more than just mimicking others’ voices. Dialogic inner speech likely would involve theory-of-mind (ToM) capacities as well as complex mental imagery of setting and spatial position. In other words, in order to imagine a conversation, you would likely need to take into account the different mental and visual perspectives involved. In order to separate these potential elements, the researchers chose three tasks; 1) a ToM task in which the subject is required to decipher different actors’ intentions, 2) a visual perspective switching task during which the subject is asked to imagine a scene for a particular point of view and then switch to a different view when prompted, and 3) the inner speech task which the subject is presented with a description of a scenario involving either dialogue or monologue and then asked to generate inner speech in the scenario. All of the subjects performed these tasks while in a functional magnetic resonance imaging (fMRI) machine, which allows researchers to observe blood flow during a given task.

Because of the complexity of dialogic inner speech, the researchers’ expectation was that dialogic inner speech would engage regions beyond those typically associated with language, otherwise known as the perisylvian language network. The perisylvian language network is in the left hemisphere, but many of the hypothesized areas of activation are in the right hemisphere, such as the the right IFG, right MTG, and right STG/STS. Their findings suggest that dialogic inner speech does indeed activate areas beyond the perisylvian language network, including the precuneous, posterior cingulate and right STG, more so than monologic inner speech. Activation of the right STG was more anterior than the left STG and slightly overlapped with areas thought to be involved in spatial processing which suggests that dialogic thought is not solely representing extra voice representations, but also seems to rely on extra spatial representations. Importantly, activation during ToM reasoning and that of dialogic speech overlapped significantly in the right posterior STG (Figure 1). This area has previously been linked to imagery of personal perspectives (Ruby and Decety, 2001)


Figure 1: Activation during dialogic inner speech task and ToM task. (a) Area of overlap in activation between the two tasks is highlighted in green. (b) Area activated during ToM task (yellow) covers anterior to posterior STG while areas activated during dialogic inner speech task (blue) are right STG, cingulate and frontal gyrus. Their overlap (green) is evident in the right posterior STG.

In addition to the three main tasks used in this experiment, the researchers also solicited self-reports after the scanning session. They were asked to report (1) how vividly they imagined the scenarios, (2) the vividness of any visual imagery they used during the task and (3) the everyday characteristics of their own inner speech, using the VISQ (McCarthy-Jones and Fernyhough, 2011). High ratings of vividness and dialogic inner speech (as a characteristic of their typical inner speech) correlated with greater activation in the  right MTG and cingulate gyrus, but did not correlate with greater activation of the left hemisphere activation (typically activated by language tasks). This suggests that covert articulation may not be necessary for particularly vivid imagination. Based on the areas that were activated in these individuals, one possible takeaway is that a greater tendency to engage in dialogic inner speech on a daily basis may indicate a greater tendency to recall and re-engage in past interactions as well as plan future interactions.

These findings also have important implications for psychopathology. As one might expect, inner speech and ToM play crucial rules in abnormal cognition, such as AVH. These hallucinations tend to manifest as if they were in someone else’s voice, but scientists have had a hard time explaining why that might be. In order to treat people with AVH, researchers must first figure out the mechanism behind the misattribution of these thoughts. This study is an important stepping stone for that discovery.

Written by Nora Bradford


Alderson-Day, B., Weis, S., McCarthy-Jones, S., Moseley, P., Smailes, D., & Fernyhough, C. (2016). The brain's conversation with itself: neural substrates of dialogic inner speech. Social cognitive and affective neuroscience, 11(1), 110–120.

Aleman A., Formisano E., Koppenhagen H., Hagoort P., de Haan E.H.F., Kahn R.S. (2005). The functional neuroanatomy of metrical stress evaluation of perceived and imagined spoken words. Cerebral Cortex, 15(2), 221–8. [PubMed] [Google Scholar]

Bentall R.P. (1990). The illusion of reality: a review and integration of psychological research on hallucinations. Psychological Bulletin, 107(1), 82–95. [PubMed] [Google Scholar]

Frith C. (1992). The Cognitive Neuropsychology of Schizophrenia. Lawrence Erlbaum: Hove. [Google Scholar]

Hinke R.M., Hu X., Stillman A.E., Kim S.G., Merkle H., Salmi R., et al.  (1993). Functional magnetic resonance imaging of Broca’s area during internal speech. Neuroreport, 4(6), 675–8. [PubMed] [Google Scholar]

Morin A. (2009). Inner speech. In Bayne T., Cleeremans A., Wilken P., editors. , editors. Oxford Companion to Consciousness, pp. 380–2, Oxford: OUP. [Google Scholar]

Morin A., Uttl B., Hamper B. (2011). Self-reported frequency, content, and functions of inner speech. Procedia - Social and Behavioral Sciences, 30, 1714–8. [Google Scholar]

Ruby P., Decety J. (2001). Effect of subjective perspective taking during simulation of action: a PET investigation of agency. Nature Neuroscience, 4(5), 546–50. [PubMed] [Google Scholar]

Simons C.J.P., Tracy D.K., Sanghera K.K., O’Daly O., Gilleen J., Dominguez M.-G., et al.  (2010). Functional magnetic resonance imaging of inner speech in schizophrenia. Biological Psychiatry, 67(3), 232–7. [PubMed] [Google Scholar]

Vygotsky L.S. (1987). Thinking and Speech. The Collected Works of Lev Vygotsky, Vol. 1, New York: Plenum Press. [Google Scholar]