Elsevier

Neuropsychologia

Volume 40, Issue 8, 2002, Pages 1129-1138
Neuropsychologia

Reading the mind from eye gaze

https://doi.org/10.1016/S0028-3932(02)00008-8Get rights and content

Abstract

Baron-Cohen [Mindblindness: an essay on autism and theory of mind. Cambridge, MA: MIT Press, 1997] has suggested that the interpretation of gaze plays an important role in a normal functioning theory of mind (ToM) system. Consistent with this suggestion, functional imaging research has shown that both ToM tasks and eye gaze processing engage a similar region of the posterior superior temporal sulcus (STS). However, a second brain region associated with ToM, the medial prefrontal (MPF) cortex, has not been identified by previous eye gaze studies. We discuss the methodological issues that may account for the absence of MPF activation in these experiments and present a PET study that controls for these factors. Our experiment included three conditions in which the proportions of faces gazing at, and away from, the participant, were as follows: 100% direct [0% averted], 50% direct—50% averted, and 100% horizontally averted [0% direct]. Two control conditions were also included in which the faces’ gaze were averted down, or their eyes were closed. Contrasts comparing the gaze conditions with each of the control conditions revealed medial frontal involvement. Parametric analyses showed a significant linear relationship between increasing proportions of horizontally averted gaze and increased rCBF in the MPF cortex. The opposite parametric analysis (increasing proportions of direct gaze) was associated with increased rCBF in a number of areas including the superior and medial temporal gyri. Additional subtraction contrasts largely confirmed these patterns. Our results demonstrate a considerable degree of overlap between the medial frontal areas involved in eye gaze processing and theory of mind tasks.

Introduction

A recent functional imaging study by Hoffman and Haxby [19] has shown that matching faces for their direction of gaze engages different brain areas to matching the same faces for their identity. The gaze task was associated with a significantly larger signal in the left posterior superior temporal sulcus (STS) and intraparietal sulcus, whereas the facial identity task produced significantly larger bilateral signals in the fusiform gyri and inferior occipital gyri. The same study also found that averted gaze was associated with a higher left STS signal than direct gaze.

The above results replicate and extend the findings of two earlier neuroimaging studies. The first showed that viewing faces with direct and averted gaze produced an increase in the regional cerebral blood flow (rCBF) of areas adjacent to the posterior STS when compared with viewing faces with downward gaze [31]. In contrast to Hoffman and Haxby, however, Wicker et al. [31] found no reliable difference between their averted and direct gaze conditions. The second study showed posterior STS involvement when participants were shown a dynamic facial stimulus switching between averted and direct gaze [30], relative to when they were shown a static facial image looking directly at them. Interestingly, the results of none of these three studies bear much resemblance to an investigation of eye gaze processing by Kawashima et al. [22] that found no evidence of STS involvement. Instead, both direct and averted gaze conditions in the Kawashima et al. study produced amygdala activation, with the direct gaze condition showing significantly greater right amygdala activation than the averted gaze condition: Kawashima et al. attributed the amygdala involvement to the emotional component of processing another person’s gaze.

More recent research has examined the neural correlates of eye gaze in more detail. George et al. [15] showed that relative to averted gaze, direct gaze produced more metabolic activity in an area of the fusiform gyrus that has consistently been shown to respond to facial images. The authors suggested that this finding follows because (i) previous research has shown that attending to facial images in a display enhances the signal in the fusiform face area, and (ii) direct gaze can convey a number of meanings (e.g. threat, sexual attraction, interest, etc.), and hence it is important to pay careful attention to other aspects of the face (e.g. identity, facial expression, sex) in order to attribute the correct meaning.

Finally, Kampe et al. [21] demonstrated that certain neural correlates of gaze interact with the attractiveness of the face. They showed that the signal in an area identified as the ventral striatum increased with increasing facial attractiveness for faces displaying direct gaze, and decreased with increasing facial attractiveness for faces displaying averted gaze. The authors suggested that the ventral striatal signal reflected the reward value of making eye contact with an attractive individual.

Section snippets

The role of STS in the attribution of mental states to others

Wicker et al. [31] suggested that the STS activation they observed was related to the perception of both averted and direct gaze, whereas Puce et al. [30] concluded that this area was involved in processing the direction signalled by gaze. Hoffman and Haxby [19] on the other hand have suggested that this area of STS has a more general role in the perception of changeable aspects of faces (e.g. gaze, expression, lip speech, etc.) (see also [17]).

All of these suggestions are plausible

Participants

Nine female members of the MRC Cognition and Brain Sciences Unit participant panel (mean age=58.3, S.D.=2.9) participated in the experiment for payment. All participants were post-menopausal, right-handed, and had normal or corrected-to-normal vision. All had given written informed consent to participate in the study. Ethical approval for the study was obtained from the Cambridge Local Research Ethics Committee.

Materials

The stimulus materials consisted of colour photographs of 16 young models (eight men

Behavioural data

Participants’ mean RTs and percent errors for the five conditions (100% direct, 50% direct—50% horizontally averted, 100% horizontally averted, down, and closed) are summarised in Table 1. Errors were low indicating that participants were concentrating on the task. RTs and error scores (arcsin transformed) were each submitted to separate repeated measures ANOVAs examining the five conditions. Both produced significant effects of condition: RTs, F(4.24)=3.56, P<0.02; errors F(4.24)=2.87, P<0.05.

Discussion

Our results demonstrate that an area of medial frontal cortex (BA 8/9), previously associated with the ToM tasks, is indeed engaged by processing another person’s gaze. This concurs with the idea that gaze plays a central role in social communication, and hence, the attribution of mental states to others (i.e. ToM) [4]. From an evolutionary perspective this would seem to make good sense, in that survival depends on maintaining a constant degree of vigilance and an ability to second guess

Acknowledgements

Thanks to Brian Cox for preparing stimulus materials and graphics.

References (33)

  • K.A. Nichols et al.

    Eye gaze and the GSR

    Journal of Experimental Social Psychology

    (1971)
  • B. Wicker et al.

    Brain regions involved in the perception of gaze: a PET study

    Neuroimage

    (1998)
  • T. Ban et al.

    Cortico-cortical projections from the prefrontal cortex to the superior temporal sulcal (STS) area in the monkey studied by means of HRP method

    Archives Italiennes de Biologie

    (1991)
  • Baron-Cohen S. Mindblindness: an essay on autism and theory of mind. Cambridge, MA: MIT Press,...
  • M. Brett et al.

    Using the Talairach atlas with the MNI template

    NeuroImage

    (2001)
  • J. Driver et al.

    Gaze perception triggers reflexive visuospatial orienting

    Visual Cognition

    (1999)
  • Cited by (294)

    • Amygdala function in emotion, cognition, and behavior

      2022, Handbook of Clinical Neurology
    View all citing articles on Scopus
    View full text