Elsevier

NeuroImage

Volume 21, Issue 1, January 2004, Pages 75-83
NeuroImage

Dissociation between overt and unconscious face processing in fusiform face area

https://doi.org/10.1016/j.neuroimage.2003.08.038Get rights and content

Abstract

The precise role of the fusiform face area (FFA) in face processing remains controversial. In this study, we investigated to what degree FFA activation reflects additional functions beyond face perception. Seven volunteers underwent rapid event-related functional magnetic resonance imaging while they performed a face-encoding and a face-recognition task. During face encoding, activity in the FFA for individual faces predicted whether the individual face was subsequently remembered or forgotten. However, during face recognition, no difference in FFA activity between consciously remembered and forgotten faces was observed, but the activity of FFA differentiated if a face had been seen previously or not. This demonstrated a dissociation between overt recognition and unconscious discrimination of stimuli, suggesting that physiological processes of face recognition can take place, even if not all of its operations are made available to consciousness.

Introduction

Face processing is one of the most important functions in social interactions, because the information given in a face let us infer the identity of a person, the mood, the sex, the race, etc. Although its relevance, the neuronal correlates of face processing remain controversial (Haxby et al., 2001). Neuropsychological findings due to cerebral lesions, such as the syndrome of prosopagnosia Damasio et al., 1982, Meadows, 1974, McNeil and Warrington, 1993, face-selective neurons in monkey single-cell recordings (Perrett et al., 1984), face-selective responses during subdural electrode recordings in patients with epilepsy (Allison et al., 1994), and functional brain imaging studies by means of positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) that demonstrate face-selective activation patterns in the posterior, lateral fusiform gyrus Gorno-Tempini et al., 1998, Kanwisher et al., 1997, Leveroni et al., 2000, Nakamura et al., 2000, Sergent et al., 1992, even in 2-month-old infants (Tzourio-Mazoyer et al., 2002), led to the assumption that the posterior, lateral fusiform gyrus is face-selectively activated. Kanwisher et al. (1997) functionally defined the individually localized face-selective part of this region as the fusiform face area (FFA).

The exact role of the FFA in face processing remains a matter of debate. While it has been proposed that the primary function of the FFA may simply be to detect faces (“it is a face”) (Kanwisher et al., 1997), it has been shown that this region is also sensitive to experience Gauthier et al., 1999, Gauthier et al., 2000a, level of categorization Gauthier and Tarr, 1997, Gauthier et al., 1999, Gauthier et al., 2000a, priming (George et al., 1999), attention (Wojciulik et al., 1998), and is involved in individual face discrimination Gauthier et al., 2000b, Hoffman and Haxby, 2000. However, many of the effects deemed “face-specific” can be demonstrated for nonface objects of one's expertise as well. For instance, in car experts, the FFA demonstrated higher activation for cars than for birds, whereas the reverse was true for bird experts (Gauthier et al., 2000a). A recent study found that holistic processing of novel objects is also related to activity in the right FFA during the acquisition of expertise (Gauthier and Tarr, 2002).

Memory deficits associated with prosopagnosia are described as an inability to visually recognize faces that were well known before the cerebral damage and furthermore as impairment in recognizing faces that are frequently encountered after the brain injury (Bruyer, 1994). It seems that the FFA might be involved not only during the retrieval of faces from long-term memory but also during the encoding of new faces into memory, suggesting that the role of the FFA goes beyond the functions described above. Might activity in FFA even predict the fate of a seen face, whether it will be remembered or forgotten in future and does activity in FFA differentiate between faces stored in memory and faces never seen before?

Until now there has been little evidence suggesting that the FFA might be involved in encoding. The few studies Bernstein et al., 2002, Golby et al., 2001, Kuskowski and Pardo, 1999 correlating activity in fusiform gyrus across subjects with recognition performance reported contradictory results. This could be due to the following issues: First, these studies did not localize the individual FFA. However, given the variability of the location of the FFA across normalized subjects, it has been argued that studies without individual functional definitions of the FFA are inadequate (Tarr and Gauthier, 2000). Therefore, it is hard to conclude from these data alone whether the FFA has a significant role in successful face encoding (Phelps, 2001). Second, due to the restricted time resolution, they were not able to measure activations separately for faces later remembered or later forgotten, a critical requirement for valid results (Brewer et al., 1998).

A few studies have reported on the role of the fusiform gyrus in the retrieval of faces from long-term memory. Leveroni et al. (2000) did not detect any differential activation in the fusiform gyrus for familiar and unfamiliar faces. Katanoda et al. (2000) reported greater activation for blocks containing only previously seen faces than for blocks with both novel and previously seen faces. But again, both studies did not perform a functional localization of the FFA in each individual, and the latter did not measure activity for each single face separately. Furthermore, to our knowledge, until now no study has yet analyzed activity in FFA for false memories, that is, forgotten faces (misses) and never seen faces falsely labelled to be a known face (false alarms). The question remains, whether the activity in FFA reflects the history of faces presented (i.e. a previously seen or a new face) or whether a face is consciously recognized or not.

Recent studies applying rapid event-related fMRI have shown correlations between neural activity and behavioral performance (Konishi et al., 2000) that had not previously been shown by slow event-related fMRI (here the unusually long interstimulus interval may interfere with task processing) or other imaging techniques. In the present study, we employed an event-related fMRI design with rapid trial presentation Clark et al., 1998, Dale and Buckner, 1997. This method significantly increases statistical power and allows an intertrial interval of less than the 12–16 s usually employed in slow event-related fMRI (Dale, 1999). By isolating the FFA in each subject (a condition for the definition of the FFA), we made sure to include only highly face-selective voxels for each subject and therefore to measure the optimal FFA responses.

The previous reports led us to formulate three hypotheses: First, in a face-encoding task, faces that will be remembered later will show different activity in the FFA than faces that will be forgotten. Second, in a face-recognition task, the activity for previously seen faces that are correctly recognized as known faces is different than the activity for faces correctly recognized as newly presented, and third, activity in FFA for false memories (misses and false alarms) does not differ from the activity for faces correctly recognized as newly presented, since for none of them a match between the cue and a trace in memory can take place.

Section snippets

Subjects

Seven volunteers (two females, five males, age range 25–37 years with a mean of 31 years) were recruited from an academic environment. All subjects had a normal or corrected-to-normal vision. Handedness (six right-handed, one left-handed) was assessed by the Edinburgh handedness inventory. None of the subjects reported a history of major medical, neurological, or psychiatric disorders or of psychotropic medication. Normal structural MR data sets were reported for all subjects. The study was

Behavioral results

Subjects correctly identified 64% of the learned faces (Fig. 1, Table 1). The correct rejection of newly presented faces was 67%. A χ2 test indicated that subjects discriminated between previously studied and newly presented faces (χ(1)2=55.22; P = 0.001). Repeated-measure ANOVA did not reveal a significant difference in accuracy across the two conditions (F(1,6) = 0.099; P < 0.764).

Reaction times did not differ between correct and incorrect responses (Table 1) as tested by two-factorial

Discussion

Here we demonstrated in a single-trial rapid event-related fMRI study that the function of the FFA is not restricted to perceptual face processing but goes beyond that. Whereas during encoding activity in FFA differentiated between later remembered or forgotten faces, this was not the case during recognition. On the other hand, during recognition, the activity of FFA differentiated if a face was seen previously or not independent of the conscious memory. This demonstrates a dissociation between

Acknowledgements

We are grateful to Rik Henson for advice on statistical analysis. This research was partly supported by an SNSF grant (3200-059077.99).

References (51)

  • P.A. Bandettini et al.

    Time course EPI of human brain function during task activation

    Magn. Reson. Med.

    (1992)
  • J.B. Brewer et al.

    Making memories: brain activity that predicts how well visual experience will be remembered

    Science

    (1998)
  • V. Bruce et al.

    Understanding face recognition

    Br. J. Psychol.

    (1986)
  • V. Bruce et al.

    What is distinctive about a distinctive face?

    Q. J. Exp. Psychol.

    (1994)
  • R. Bruyer

    Face recognition

  • R. Buckner

    The cognitive neuroscience of remembering

    Nat. Rev. Neurosci.

    (2001)
  • V.P. Clark et al.

    fMRI study of face perception and memory using random stimulus sequences

    J. Neurophysiol.

    (1998)
  • A.M. Dale

    Optimal experimental design for event-related fMRI

    Hum. Brain Mapp.

    (1999)
  • A.M. Dale et al.

    Selective averaging of rapidly presented individual trials using fMRI

    Hum. Brain Mapp.

    (1997)
  • A.R. Damasio et al.

    Prosopagnosia: anatomic basis and behavioral mechanisms

    Neurology

    (1982)
  • J. Fell et al.

    Human memory formation is accompanied by rhinal–hippocampal coupling and decoupling

    Nat. Neurosci.

    (2001)
  • I. Gauthier et al.

    Orientation priming of novel shapes in the context of viewpoint-dependent recognition

    Perception

    (1997)
  • I. Gauthier et al.

    Unraveling mechanisms for expert object recognition: bridging brain activity and behavior

    J. Exp. Psychol. Hum. Percept. Perform.

    (2002)
  • I. Gauthier et al.

    Activation of the middle fusiform ‘face area’ increases with expertise in recognizing novel objects

    Nat. Neurosci.

    (1999)
  • I. Gauthier et al.

    Expertise for cars and birds recruits brain areas involved in face recognition

    Nat. Neurosci.

    (2000)
  • Cited by (0)

    View full text