Elsevier

Brain Research

Volume 1232, 26 September 2008, Pages 173-184
Brain Research

Research Report
Neural correlates of self-face recognition: An effect-location meta-analysis

https://doi.org/10.1016/j.brainres.2008.07.010Get rights and content

Abstract

Recent evidence from neuropsychological patients with focalized lesions and functional brain imaging studies indicate that processing of self is distinguishable from processing of information about others (e.g., recognizing a familiar face). Here, we conduct an effect-location meta-analysis (Fox et al., 1998) of 9 functional neuroimaging studies of self-face recognition. The evidence provides support for a right-dominated, but largely bilaterally distributed model for self-face processing. Four areas are consistently activated: the left fusiform gyrus, bilateral middle and inferior frontal gyri, and right precuneus. The evidence is interpreted in light of a developing model of self-face recognition as part of a larger social cognitive stream of processing.

Introduction

That non-human species are able to recognize themselves in a mirror was first shown convincingly by Gallup (1970). Using the now classic mark test, he anesthetized chimpanzees and marked them with odorless dye. Their subsequent inspection of the marked areas while in the presence of a mirror, but not in absence of a mirror, demonstrated that chimpanzees could learn the capacity to recognize their own mirror self-reflection (Gallup, 1970).

Since then, mirror self-recognition has been convincingly demonstrated in bonobos and orangutans (Suarez and Gallup, 1981, Walraven et al., 1995), while reports that gorillas (Patterson and Cohn, 1994), bottlenose dolphins (Reiss and Marino, 2001) and Asian elephants (Plotnik et al., 2006) pass the test are more controversial. Gallup (1982) suggested that the ability of an animal to recognize itself reveals something important about the way an animal thinks. Gallup postulated that the ability to recognize yourself is essentially the ability to become the object of one's own attention. He hypothesized that this ability led to the capacity for introspection about oneself, and that this is the cognitive foundation of our ability to infer the mental states of others, referred to as theory of mind (see also, Keenan et al., 2003a, Keenan et al., 2003b). The model suggests that individuals construct mental models of themselves and use these to infer the mental experiences of others.

The species that exhibit mirror self-recognition have interesting commonalities. All but the orangutan live in highly complex social groups (Dunbar, 1998), and in the case of the primates, they have highly developed frontal lobes (Semendeferi et al., 1997a). Gallup (1998) suggested that since the frontal lobes are the most recently evolved part of the brain, they may be a required substrate for the capacity to engage in self-processing.

In the first neuropsychological investigations of self-recognition in humans, Preilowski (1977) (see also Sperry, 1982, Sperry et al., 1979) measured galvanic skin response in split-brain patients and demonstrated that the right cortical played a greater role in self-face recognition than the left hemisphere. In contrast, more recently Turk et al. (2002) worked with a split-brain patient and presented varying levels of morphed images of self and familiar-other faces separately to the right and left hemispheres. The patient's task was to determine whether a given image portrayed himself or a familiar other. They found that in both the self and other condition, the proportion of positive responses increased as the images approached completeness (i.e. the percentage of either self or other increased) when pictures were presented to either the left or the right hemisphere. The authors concluded that both hemispheres are capable of recognizing self and other. Furthermore, they also found that the proportion of positive responses in the “self” condition was significantly greater when the images were presented to the left than to the right hemisphere. The opposite pattern was obtained in the “other” condition. They suggested that the left hemisphere might have a special role to play in the “self-memory system” (see Gazzaniga, 1998).

The first attempts to investigate the neural correlates of self-recognition in non-patient subjects using modern neuroimaging and psychophysical techniques were conducted by Keenan et al., 1999, Keenan et al., 2000. They developed a reaction time task that measured hand-asymmetry when responding to one's own face as opposed to other faces that were either familiar or novel. They found a left-hand advantage (faster reaction times) when subjects were responding to self-faces, but not to other faces. A second experiment asked subjects to stop a movie of a morphed face that transitioned between a famous face and self-face when they thought that the image looked more like self than famous. Keenan et al. (2000) found that subjects stopped the movies sooner when responding with the left hand; i.e. they were more likely to stop the movie when it had less self in the morph (sensitive to self) with the left hand rather than with the right hand. Because of contralateral motor control, they interpreted their findings as reflecting right hemisphere dominance of self-face recognition (see also Platek and Gallup, 2002, Platek et al., 2003a, Platek et al., 2003b, Platek et al., 2004). Keenan et al. (2001) have gone on to show that the right hemisphere is implicated in self-face recognition by utilizing the WADA-sodium pentobarbital technique (Wada, 1949) to “knock out” each of the cortical hemispheres. Using patients with intractable epilepsy as subjects, they measured reactions to the self-face morphed with famous faces, and found that when the right hemisphere was anesthetized patients were more likely to report that the face was famous (i.e., not self), but when the left hemisphere was anesthetized they were more likely to report that the face looked like themselves.

Further evidence of right hemisphere localization of self-face recognition comes from case studies of delusional misidentification, particularly mirrored self-misidentification. Such patients typically show extensive right hemisphere damage (e.g., Feinberg 2000; (Breen et al., 2001, Feinberg and Keenan, 2005); Spangenberg et al. 1998). Recently, Platek et al. (2004;2006) and others (Devue et al., 2007, Sugiura et al., 2000, Sugiura et al., 2005) have also demonstrated right hemisphere activation when viewing one's own face.

Platek et al. (2004) measured BOLD signal response while subjects saw their own face, the face of a stranger, and the face of someone famous. We found that when comparing activation associated with viewing self-faces to viewing famous and novel faces the right hemisphere was selectively active; i.e. self-face recognition preferentially activates the right frontal lobes. Using personally familiar gender and age matched control faces, Platek et al. (2006) found a distributed bilateral network involved in self-face recognition that included right superior frontal gyrus, right inferior parietal lobe, bilateral medial frontal lobe, and left anterior middle temporal gyrus (see Fig. 1). This is the first study to employ rigorous gender and familiarity matched control faces.

Kircher et al. (2000 ";; 2001) also found a distributed bilateral network that included left prefrontal regions and right temporal lobe regions. However, in a reanalysis of Kircher et al.'s (2001) data, Keenan (personal communication) showed a slightly greater amount of right hemisphere activation in the self-face condition. These inconsistencies are striking and may be partially explained by differences in methodology (Gillihan and Farah, 2005).

Sugiura et al. (2005) contrasted self-faces with a prelearned novel face and two personally familiar faces and found activation in left fusiform gyrus, right frontal operculum, and right occipitotemporoparietal junction. The personally familiar faces activated regions in the right occipitotemporoparietal junction as well, suggesting that the right frontal operculum and left fusiform gyrus are critical regions involved in self-face recognition.

Uddin et al. (2005) presented subjects with varying levels of morphed images and found activation in right superior frontal and inferior parietal lobes. They suggest that this activation to “self” represents activation of self–other frontal–parietal mirror network. That is, they suggest that the way in which one discriminates self from other is by recruiting a mirror neuron like network that compares self to other.

More recently, Devue et al. (2007) contrasted self-face with a familiar colleague and found activation in the right inferior frontal lobe and right insula. Additionally, they found activation in anterior (right medial and frontal) regions when distinguishing self body from non-self body. They suggest that there is a posterior–anterior stream of processing whereby posterior regions serve as first level structural characterization of faces and bodies. They suggest that anterior regions serve to differentiate self from other at a higher level of processing and perhaps at an abstract level of knowledge about the self.

The frontal lobes are developing rapidly between the first and third years of life (Thatcher, 1999, Semendeferi, 1999), which is the period in which children are also developing the capacities to represent self and other (see Amsterdam, 1972, Lewis, 2003). This has led to suggestions of a frontal lobe localization theory of self-awareness and theory of mind (Frith and Frith, 1998, Gallup, 1982, Keenan et al., 2000, Keenan et al., 2003a, Keenan et al., 2003b; Stuss and Anderson, 2004). Additionally, the frontal cortex/prefrontal cortex (PFC) appears to be the most recently evolved portion of the neocortex (Rakic, 1995) and is a highly intricate multi-modal information-processing center (Gibson, 2002). Neuropsychologists have known for decades that damage to the PFC in humans (and animals) will produce drastic changes in personality and behavior (Weinberger, 1993). It is interesting to note that among non-human primates, chimpanzees, one of three non-human species who show evidence of self-recognition and self-awareness, have the most developed frontal cortex (Semendeferi et al., 1997b). In contrast, the gorilla, the outlier species that has not clearly shown evidence for mirror self-recognition or theory of mind, appears to have the least developed frontal lobes (Semendeferi, 1999) and may also be the least lateralized (LeMay and Geschwind, 1975). The frontal lobes may thus be a necessary substrate for the capacity to engage in self-processing (Gallup, 1998).

However, several substrates besides the frontal lobes are involved in self-referential processing. For instance, recent evidence has shown that the inferior parietal lobes and left anterior temporal lobe may also be involved in self-face recognition (Platek et al., 2006). Similarly, the cortical midline structures such as the precuneus and cingulated gyrus appear to be involved in self-referent information processing and discriminating between self- and other-descriptive words/phrases (Fossati et al., 2003, Fossati et al., 2004, Kelley et al., 2002, Lou et al., 2004, Macrae et al., 2004, Northoff and Bermpohl, 2004, Seger et al., 2004). The medial parietal lobes, posterior cingulate and precuneus have also been associated with autobiographical memory retrieval (Maddock et al., 2001), engaging in self-generated actions and self-monitoring (Blakemore et al., 1998), self-reference in morphed faces (Platek et al., 2008), and discriminating between theory of mind stories and “physical” stories (Fletcher et al., 1995; see also Vogely et al., 2001). This again suggests a major role for midline cortical structures in the capacity to be self-aware. Keenan et al., 2003a, Keenan et al., 2003b) summarize a number of neuropsychological studies, which show both right hemispheric lateralization and localization of self-recognition in the prefrontal cortex (see also Feinberg, 2000). There is also the suggestion that self–other processing is subserved by processes associated with mirroring others' action—the so-called mirror neuron system (Gallese et al., 2004, Iacoboni, 2004, Decety and Chaminade, 2003), localized to the left inferior frontal lobe and left inferior parietal lobe (see also Uddin et al., 2005).

The evidence that individuals with autism, schizophrenia, and related, but milder disturbances such as Asperger's syndrome and schizotypy perform less well on tasks associated with self-processing further implicate the frontal lobes, since these conditions have been linked to deficits in frontal lobe functioning (Baron-Cohen et al., 1985, Baron-Cohen et al., 1997a, Corcoran et al., 1997, Doody et al., 1998, Frith and Corcoran, 1996, Langdon and Coltheart, 1999, Pickup and Frith, 2001, Ellis and Gunter, 1999, Mattay et al., 1999, Callicot et al., 2000, Craig et al., 2004 for discussions of frontal lobe functioning and autism or Asperger's, and Schizophrenia).

People who score high on measures of schizotypal personality show impairments in self-face recognition (Platek and Gallup, 2002) and self-descriptive adjective processing (Platek et al., 2003a, Platek et al., 2003b). There is mounting evidence that schizophrenia is associated with frontal lobe pathology (e.g., Frith, 1997) and individuals with schizophrenia show pronounced mental state attribution deficits (e.g., Frith and Corcoran, 1996, Besche et al., 1997, Sarfati et al., 1999). Individuals with schizophrenia often see their own reflections in mirrors as independently alive, alien, or sinister (Harrington et al., 1989). They also have been observed talking and laughing at their mirrored reflection as if it were another person (Rosenzweig and Shakow, 1937). In a now classic series of studies, Traub and Orbach, 1964, Orbach et al., 1966) showed that individuals with schizophrenia had difficulty with a task that involved self-referent mirror use. The task involved rectifying a distorted mirror image of the subject, or an inanimate object using controls on the mirror. Although schizophrenics were as good as the controls at “fixing” the image of the inanimate object (a door), unlike controls, they were unable to adjust the mirror to appropriately represent themselves. This suggests that schizophrenia involves not a deficit in mirror understanding, but a deficit in self-processing.

Additionally, Frith (1992 ";) and Blakemore et al. (2000) have shown that individuals with schizophrenia occasionally report seeing nothing in mirrored reflection (negative autoscopy) and are also unable to realize when behaviors such as speech are self-initiated. Frith (1992) has conceptualized schizophrenia as a disorder of mental states. He suggests that certain psychotic symptoms associated with schizophrenia may impair the ability to reason about other people's mental states. This inability can then lead to social withdrawal, inappropriate social behaviors and affective blunting (Frith, 1992). The literature has been fairly consistent in showing that currently symptomatic patients with schizophrenia perform worse on theory of mind tasks than non-psychiatric and psychiatric controls as well as patients in remission (Garety and Freeman, 1999). For instance, Corcoran et al. (1995) found that patients with schizophrenia perform poorer on a simple social inference task than normal and non-psychotic psychiatric control groups. Brunet et al. (2003) used PET to investigate brain activation during a nonverbal theory of mind task. Results indicated that while controls showed significant cerebral activation in the right prefrontal cortex, these activations were not found in the schizophrenia group.

We recently investigated the hypothesis that self-processing and mental state attribution are part of a shared behavioral and neurocognitive network that is impaired in schizophrenia-spectrum disorders (Irani et al., 2007). We found that unaffected first degree relatives of patients with schizophrenia take longer than patients to recognize their own face, but are more accurate in making self vs. other judgments. This was related to the level of schizotypy reported by the family members. Additionally, patients were more likely to misattribute self to unfamiliar faces; i.e. when they made errors at classifying a novel face, they were more likely to indicate that the face was self.

In children suffering from autistic-spectrum disorders, mirror self-recognition is developmentally delayed and sometimes even absent (Spiker and Ricks, 1984). A series of studies demonstrated that only between 50–70% of children with autism could pass a modified version of the Gallup mark test (Dawson and McKissick, 1984, Spiker and Ricks, 1984). Those that eventually pass the test, do so at a developmentally delayed age of 5–7 years when most children pass between 18–24 months of age. Similarly to schizophrenics, children with autism perform poorly on theory of mind tasks (Baron-Cohen et al., 1985, Baron-Cohen et al., 1986, Baron-Cohen and Cross, 1995, Baron-Cohen, 1997, Baron-Cohen et al., 1997b, Baron-Cohen et al., 1999, Baron-Cohen et al., 2001).

Patients with damage to the frontal cortex are not only impaired in their ability to recognize their own faces, but they show corollary deficits in self-evaluation and autobiographical memory (e.g., Keenan et al., 2003a, Keenan et al., 2003b). Additionally, when patients have the anterior portion of their temporal lobe removed in order to treat intractable seizures associated with epilepsy, this impairs the ability to accurately recognize their own face (slower reaction time and more errors) compared to personally familiar and famous faces. This finding, while inconsistent with a simple (right) hemispheric model for self-processing, supports our recent fMRI data (Platek et al., 2006) that showed a larger, distributed network for self-processing that involves the anterior middle temporal gyrus (Platek et al., 2008).

Thus, the data from these brain disorders supports the idea that self-recognition is impaired by deficiencies in frontal lobe processing that may be localized to specific regions (e.g., superior or inferior frontal gyri, as well as anterior temporal regions).

Section snippets

Results

As can be seen in Fig. 2 and Table 1, Table 2, several brain substrates have been associated with self-face recognition when contrasted with other face conditions. Overall, twenty-six brain regions were implicated in self-face recognition. Left hemisphere (M = 1.19) substrates were reported to be activated less than right hemisphere substrates (M = 1.92; t(25) =  2.35, p < .05) (total number of activations reported in LH = 31 versus RH = 50). Below we discuss those regions that were activated in five or

Discussion

Investigating self-referential processing continues to be a high priority in social and affective neuroscience. Intrinsic importance attaches to the notion of the self to nearly all cognitive processing. Further interest derives from the hypothesized role of self-referential processing in the development of theory of mind (Gallup, 1982, Keenan et al., 2003a, Keenan et al., 2003b). The evidence to support this idea encompasses neuroanatomical and behavioral data from non-human primates, as well

Experimental procedures

This meta-analysis examines neuroimaging studies of self-face recognition. The studies use similar paradigms, and all contrast self-face with another face of varying degrees of familiarity. The meta-analysis is of the “effect-location” type (see Fox et al., 1998). The aim in this type of analysis is not to investigate effect size, but rather to localize effects. All studies use hemodynamic methods, either fMRI (n = 8 studies) or PET (n = 1 study).

Acknowledgment

S.M.P. is supported by a grant from the Pioneer Fund, Inc.

References (109)

  • FeinbergT.E. et al.

    Where in the brain is the self?

    Conscious. Cogn.

    (2005)
  • FletcherP.C. et al.

    Other minds in the brain: a functional imaging study of “theory of mind” in story comprehension

    Cognition

    (1995)
  • FossatiP. et al.

    Distributed self in episodic memory: neural correlates of successful retrieval of self-encoded positive and negative personality traits

    NeuroImage

    (2004)
  • FoxP.T. et al.

    Beyond the single study: function/location metanalysis in cognitive neuroimaging

    Curr. Opin. Neurobiol.

    (1998)
  • GalleseV. et al.

    A unifying view of the basis of social cognition

    Trends Cogn. Sci.

    (2004)
  • GobbiniM.I. et al.

    Neural response to the visual familiarity of faces

    Brain Res. Bull.

    (2006)
  • GuX. et al.

    Attention and reality constraints on the neural processes of empathy for pain

    NeuroImage

    (2007)
  • HarringtonA. et al.

    Disordered recognition and perception of human faces in acute schizophrenia and experimental psychosis

    Compr. Psychiatry

    (1989)
  • KeenanJ.P. et al.

    Left hand advantage in a self-face recognition task

    Neuropsychologia

    (1999)
  • KeenanJ.P. et al.

    Hand response differences in a self-face identification task

    Neuropsychologia

    (2000)
  • KircherT. et al.

    Towards a functional neuroanatomy of self-processing: effects of faces and words

    Cogn. Brain Res.

    (2000)
  • KircherT.T.J. et al.

    The neural correlates of intentional and incidental self processing

    Neuropsychologia

    (2002)
  • LangdonR. et al.

    Mentalising, schizotypy, and schizophrenia

    Cognition

    (1999)
  • MaddockR.J. et al.

    Remembering familiar people: the posterior cingulate cortex and autobiographical memory retrieval

    Neuroscience

    (2001)
  • MazoyerB. et al.

    Cortical networks for working memory and executive functions sustain the conscious resting state in man

    Brain Res. Bull.

    (2001)
  • NorthoffG. et al.

    Cortical midline structures and the self

    Trends Cogn. Sci.

    (2004)
  • NorthoffG. et al.

    Self-referential processing in our brain—a meta-analysis of imaging studies on the self

    NeuroImage

    (2006)
  • PlatekS.M. et al.

    A left-hand advantage for self-description and the effects of schizotypal personality traits

    Schizophr. Res.

    (2003)
  • PlatekS.M. et al.

    Contagious yawning: the role of self-awareness and mental state attribution

    Cogn. Brain Res.

    (2003)
  • PlatekS.M. et al.

    Where am I? The neurological correlates of self and other

    Cogn. Brain Res.

    (2004)
  • PlatekS.M. et al.

    The neural basis for facial resemblance

    Neurosci. Lett.

    (2008)
  • RakicP.

    A small step for the cell, a giant leap for mankind: a hypothesis of neocortical expansion during evolution

    Trends Neurosci.

    (1995)
  • SarfatiY. et al.

    Investigating theory of mind in schizophrenia: influence of verbalization in disorganized and non-disorganized patients

    Schizophr. Res.

    (1999)
  • SegerC.A. et al.

    Cortical activations during judgements about the self and another person

    Neuropsychologia

    (2004)
  • SemendeferiK. et al.

    The evolution of the frontal lobes: a volumetric analysis based on three-dimensional reconstructions of magnetic resonance scans of human and ape brains

    J. Hum. Evol.

    (1997)
  • SemendeferiK. et al.

    The evolution of the frontal lobes: a volumetric analysis based on three-dimensional reconstructions of magnetic resonance scans of human and ape brains

    J. Hum. Evol.

    (1997)
  • SperryR. et al.

    Self recognition and social awareness in the deconnected minor hemisphere

    Neuropsychologia

    (1979)
  • SuarezS.D. et al.

    Self-recognition in chimpanzees and orangutans, but not gorillas

    J. Hum. Evol.

    (1981)
  • SugiuraM. et al.

    Passive and active recognition of one's own face

    NeuroImage

    (2000)
  • SugiuraM. et al.

    Cortical mechanisms of visual self-recognition

    NeuroImage

    (2005)
  • SugiuraM. et al.

    Multiple brain networks for visual self-recognition with different sensitivity for motion and body part

    NeuroImage

    (2006)
  • AmsterdamB.

    Mirror self-image reactions before the age of two

    Dev. Psychobiol.

    (1972)
  • Baron-CohenS.

    Hey! It was just a joke! Understanding propositions and propositional attitudes by normally developing children and children with autism

    Isr. J. Psychiatry

    (1997)
  • Baron-CohenS. et al.

    Reading the eyes: evidence for the role of perception in the development of a theory of mind

  • Baron-CohenS. et al.

    Mechanical, behavioural, and intentional understanding of picture stories in autistic children

    Br. J. Dev. Psychol.

    (1986)
  • Baron-CohenS. et al.

    Another advanced test of theory of mind: evidence from very high functioning adults with autism or Asperger's syndrome

    J. Child Psychol. Psychiatry

    (1997)
  • Baron-CohenS. et al.

    Is there a language of the eyes? Evidence from normal adults and adults with autism or Asperger's syndrome

    Vis. Cogn.

    (1997)
  • Baron-CohenS. et al.

    A new test of social sensitivity: detection of faux pas in normal children and children with Aspergers syndrome

    J. Autism Dev. Disord.

    (1999)
  • Baron-CohenS. et al.

    The ‘reading the mind in the eyes’ test revised version: a study with normal adults, and adults with Asperger Syndrome or High-Functioning autism

    J. Child Psychol. Psychiatry

    (2001)
  • BescheC. et al.

    Syntactic and semantic processing in schizophrenic patients evaluated by lexical decision tasks

    Neuropsychology

    (1997)
  • Cited by (139)

    • Network of autoscopic hallucinations elicited by intracerebral stimulations of periventricular nodular heterotopia: An SEEG study

      2021, Cortex
      Citation Excerpt :

      According to Heydrich and Blanke (2013) most of the patients with AH exhibited lesions in the right superior occipital gyrus and cuneus (extrastriate visual cortex). The connectivity data of Blondiaux et al. (2021) places emphasis on the importance of visual mechanisms underlying AH by revealing the involvement of the extrastriate visual cortex (visual processing of bodily and facial signals (Allison et al., 1994; Downing et al., 2001)), ventral stream regions (pITG, proximal to the FFA, and also involved in face recognition (Allison et al., 1994; Ungerleider & Haxby, 1994)) and dorsal stream areas (bilateral precuneus and superior parietal cortex -(Ungerleider & Haxby, 1994)) involved in own face and body processing, self-location and self-projection (Cavanna & Trimble, 2006; Northoff et al., 2006; Platek, Wathne, Tierney, & Thomson, 2008). Blondiaux et al. (2021) revealed in their study regarding common and distinct brain networks underlying autoscopic phenomena lesions causing AH involved manily bilateral precuneus and adjacent regions in superior occipital cortex and superior parietal cortex, as well as bilateral occipito-temporal cortex (posterior inferior temporal cortex) and bilateral cerebellum.

    View all citing articles on Scopus
    View full text