The Neural Basis of the Behavioral Face-Inversion Effect

McGovern Institute for Brain Research, Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA.
Current Biology (Impact Factor: 9.57). 01/2006; 15(24):2256-62. DOI: 10.1016/j.cub.2005.10.072
Source: PubMed


Two of the most robust markers for "special" face processing are the behavioral face-inversion effect (FIE)-the disproportionate drop in recognition of upside-down (inverted) stimuli relative to upright faces-and the face-selective fMRI response in the fusiform face area (FFA). However, the relationship between these two face-selective markers is unknown. Here we report that the behavioral FIE is closely associated with the fMRI response in the FFA, but not in other face-selective or object-selective regions. The FFA and the face-selective region in the superior temporal sulcus (f_STS), but not the occipital face-selective region (OFA), showed a higher response to upright than inverted faces. However, only in the FFA was this fMRI-FIE positively correlated across subjects with the behavioral FIE. Second, the FFA, but not the f_STS, showed greater neural sensitivity to differences between faces when they were upright than inverted, suggesting a possible neural mechanism for the behavioral FIE. Although a similar trend was found in the occipital face area (OFA), it was less robust than the FFA. Taken together, our data suggest that among the face-selective and object-selective regions, the FFA is a primary neural source of the behavioral FIE.

  • Source
    • "Dipole source analyses of the N250r and its MEG counterpart (M250r component) suggest that these components are generated in posterior regions of the fusiform gyrus that are likely to correspond to the FFA (Bindemann et al., 2008; Schweinberger et al., 2002a, 2002b; Schweinberger et al., 2007). This is in line with fMRI and intracranial recording studies which found that FFA activity in response to faces is modulated by repetitions versus change of facial identity (Engell and McCarthy, 2014; Kanwisher et al., 1997; Rotshtein et al., 2005, Winston et al., 2004; Yovel and Kanwisher, 2005). The identity matching processes that give rise to the N250r component are thus likely to take place within the core visual-perceptual face processing network. "
    [Show abstract] [Hide abstract]
    ABSTRACT: It is frequently assumed that facial identity and facial expression are analysed in functionally and anatomically distinct streams within the core visual face processing system. To investigate whether expression and identity interact during the visual processing of faces, we employed a sequential matching procedure where participants compared either the identity or the expression of two successively presented faces, and ignored the other irrelevant dimension. Repetitions versus changes of facial identity and expression were varied independently across trials, and event-related potentials (ERPs) were recorded during task performance. Irrelevant facial identity and irrelevant expression both interfered with performance in the expression and identity matching tasks. These symmetrical interference effects show that neither identity nor expression can be selectively ignored during face matching, and suggest that they are not processed independently. N250r components to identity repetitions that reflect identity matching mechanisms in face-selective visual cortex were delayed and attenuated when there was an expression change, demonstrating that facial expression interferes with visual identity matching. These findings provide new evidence for interactions between facial identity and expression within the core visual processing system, and question the hypothesis that these two attributes are processed independently.
    Full-text · Article · Nov 2015 · Neuropsychologia
  • Source
    • "This pattern of response is consistent with the better discrimination of upright than inverted faces – the behavioral face inversion effect. Furthermore, correlational analyses of the behavioral and fMRI inversion effects across individuals revealed a strong correlation (r ¼ 0.5) between the two measures in the FFA (Yovel and Kanwisher, 2005). Interestingly, the pSTS-FA showed higher response to upright than inverted faces but did not show correlation with the face inversion effect that was measured on a face identity task, suggesting that its neural response to upright and inverted faces does not underlie the behavioral effect. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Faces elicit robust and selective neural responses in the primate brain. These neural responses have been investigated with functional MRI and EEG in numerous studies, which have reported face-selective activations in the occipital-temporal cortex and an electrophysiological face-selective response that peaks 170ms after stimulus onset at occipital-temporal sites. Evidence for face-selective processes has also been consistently reported in cognitive studies, which investigated the face inversion effect, the composite face effect and the left visual field (LVF) superiority. These cognitive effects indicate that the perceptual representation that we generate for faces differs from the representation that is generated for inverted faces or non-face objects. In this review, I will show that the fMRI and ERP face-selective responses are strongly associated with these three well-established behavioral face-selective measures. I will further review studies that examined the relationship between fMRI and EEG face-selective measures suggesting that they are strongly linked. Taken together these studies imply that a holistic representation of a face is generated at 170ms after stimulus onset over the right hemisphere. These findings, which reveal a strong link between the various and complementary cognitive and neural measures of face processing, allow to characterize where, when and how faces are represented during the first 200ms of face processing.
    Preview · Article · Sep 2015 · Neuropsychologia
  • Source
    • "Although it has been argued that the FFA might not be specialized for faces per se, but for individuation within any object class of high expertise (Gauthier et al., 2000, 1999), there is evidence that the FFA is genuinely face-sensitive (Rhodes et al., 2004; see also McKone, Kanwisher, and Duchaine, 2007, for a review arguing that the FFA is specialized for faces and not for any object class of expertise). Individual faces are assumed to be represented in a holistic manner in the FFA (e.g., Schiltz and Rossion, 2006; Yovel and Kanwisher, 2005). According to the neurocognitive model proposed by Haxby et al. (2000), face perception and recognition are based on a distributed neural network. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Misidentifications are a common phenomenon in unfamiliar face processing, but little is known about the underlying cognitive and neural mechanisms. We used the face identity-sensitive N250r component of the event-related brain potential as a measure of identity-sensitive face matching process in visual working memory. Two face images were presented in rapid succession, and participants had to judge whether they showed the same or two different individuals. Identity match and mismatch trials were presented in random sequence. On similar mismatch trials, perceptually similar faces of two different individuals were shown, while two physically distinct faces were presented on dissimilar mismatch trials. Misidentification errors occurred on 40% of all similar mismatch trials. N250r components were elicited not only in response to an identity match, but also on trials with misidentification errors. This misidentification N250r was smaller and emerged later than the N250r to correctly detected identity repetitions. Importantly, N250r components were entirely eliminated on similar mismatch trials where participants correctly reported two different facial identities. Results show that misidentification errors are not primarily a post-perceptual decision-related phenomenon, but are generated during early visual stages of identity-related face processing. Misidentification errors occur when stored representations of a particular individual face in visual working memory are incorrectly activated by a perceptual match with a different face.
    Full-text · Article · Sep 2015 · Neuropsychologia
Show more