Cortical correlates of face and scene inversion: A comparison
Introduction
Many kinds of objects are more difficult to recognize when they are upside-down than when they are in their canonical orientation. Behavioral studies have demonstrated that this inversion penalty is often stronger for faces than for other stimulus categories (Valentine, 1988, Yin, 1969). This phenomenon – known as the face inversion effect (FIE) – has been taken as support for the claim that faces and nonface objects recruit different processing streams (Farah, Wilson, Drain, & Tanaka, 1998). In particular, face recognition is believed to rely more than nonface recognition on configural or holistic information (Carey & Diamond, 1977; Farah, Tanaka, & Drain, 1995; Maurer, Grand, & Mondloch, 2002; Rhodes, Brake, & Atkinson, 1993). Recent studies have demonstrated that inversion impairs the ability to detect the configural aspects of faces (such as the distance between the two eyes) but not individual faces features (such as the color or shape of the eyes) (Bartlett & Searcy, 1993; Freire, Lee, & Symons, 2000; Searcy & Bartlett, 1996). There is evidence that similar configural or holistic processing mechanisms might be recruited by other object categories for which we have extensive expertise and are also more difficult to recognize after inversion (Diamond & Carey, 1986; Gauthier & Tarr, 2002; Rossion & Gauthier, 2002).
Claims for specialized processing streams have also been made for scenes (Aguirre, Zarahn, & D’Esposito, 1998; Epstein, 2005; Epstein & Kanwisher, 1998) and bodies (Downing, Jiang, Shuman, & Kanwisher, 2001). Faces, scenes, and bodies differentially activate specific regions of cortex that tend to be found in the same location across subjects (Kanwisher, 2004, but see Haxby et al., 2001) and both scene and face perception can be specifically impaired by brain damage (Aguirre & D’Esposito, 1999; Epstein, DeYoe, Press, Rosen, & Kanwisher, 2001; Mendez & Cherrier, 2003). Given these claims of specialization, one might expect to find scene and body inversion effects comparable to the FIE. Indeed, a “body-inversion effect” has recently been reported (Reed, Stone, Bozova, & Tanaka, 2003) in which subjects are impaired at discriminating between images of bodies exhibiting different postures when the images are inverted. The magnitude of the body inversion effect was similar to the magnitude of the face inversion effect. In contrast, the few studies that have examined scene inversion reported to have either no effect on recognition performance (Diamond & Carey, 1986; Wright & Roberts, 1996) or a small effect that is significantly less than the FIE (Scapinello & Yarmey, 1970; Yin, 1969).
In the current study, we measure the behavioral effects of face and scene inversion and also use functional magnetic resonance imaging (fMRI) to identify cortical regions that respond differentially to upright versus inverted faces and upright versus inverted scenes. This experiment follows on from a number of earlier studies that examined the effect of inversion on the neural response to faces (Aguirre, Singh, & D’Esposito, 1999; Gauthier, Tarr, Anderson, Skudlarski, & Gore, 1999; Haxby et al., 1999; Kanwisher, Tong, & Nakayama, 1998). Several of these studies focused on the fusiform face area (FFA), a region of cortex that responds much more strongly to faces than to nonface objects and is believed to be critical for face recognition (Kanwisher, McDermott, & Chun, 1997; McCarthy, Puce, Gore, & Allison, 1997; Puce, Allison, Asgari, Gore, & McCarthy, 1996; Rhodes, Byatt, Michie, & Puce, 2004). One might expect that inversion would strongly reduce FFA response to faces. Somewhat surprisingly, these earlier studies indicated that inversion reduces face response in the FFA only mildly (Gauthier et al., 1999, Haxby et al., 1999, Kanwisher et al., 1998; Yovel & Kanwisher, 2004) or not at all (Aguirre et al., 1999). In fact, the greatest effect of face inversion was found in putative “object” areas, which responded more strongly to inverted faces than to upright faces (Aguirre et al., 1999, Haxby et al., 1999). This neural effect was interpreted as reflecting greater engagement of generic object-processing mechanisms to inverted faces. However, a limitation of these earlier studies was the fact that occipitotemporal visual regions were classified as either “face” or “object” areas based on differential response to either faces versus nonface objects (Aguirre et al., 1999) or faces versus houses (Haxby et al., 1999). Thus, these studies did not distinguish between voxels involved in generic processing of objects from voxels that might be preferentially involved in house/scene processing. Furthermore, most of these earlier studies focused on particular regions of interest in occipitotemporal cortex and thus might have overlooked regions outside of these ROIs that responded differentially to upright versus inverted faces. An exception is a study by Leube et al. (2003), which examined effects of face inversion across the whole brain and found greater response to upright than to inverted faces in the right superior temporal sulcus and right insular cortex.
The neural consequences of scene inversion have been comparatively less studied, probably because the behavioral effects of scene inversion have not been reliably established. Haxby et al. (1999) used upright and inverted houses as the nonface comparison stimuli in an fMRI study of the face inversion effect. Although house inversion per se was not a focus of this study, the authors reported reduced response to inverted compared to upright houses in “house-selective” voxels in the medial fusiform part of ventral temporal cortex, but a nonsignficant trend towards greater response for inverted houses in more posterior occipital regions. Insofar as the medial temporal “house-selective” voxels abut the scene-selective parahippocampal place area (PPA; Epstein & Kanwisher, 1998), these results suggest that scene inversion may result in decreased engagement of scene-selective regions but increased engagement of generic object processing regions. In other words, scene inversion and face inversion may have qualitatively similar effects on their respective processing streams. However, in the absence of a nonhouse object comparison condition, it is unclear whether the effects of house inversion in this study are specific to houses, or whether they would apply to a wide variety of stimuli.
In the current study, subjects were scanned while they performed a continuous visual matching task on upright and inverted versions of faces, “scenes” consisting of images of houses set in their natural surroundings, and nonface objects. Our primary goal was to compare the effects of face inversion to the effects of scene inversion in order to gain new insight into the neural origins of the face inversion effect. We were particularly interested in four issues. First, would face inversion lead to greater engagement of regions involved in generic object processing, as suggested by earlier studies? Second, to what extent would scene inversion lead to a behavioral penalty, and would that penalty be reflected in a similar shift towards greater engagement of generic object-processing regions? Third, are changes in the fMRI response engendered by face and scene inversion closely tied to the corresponding behavioral effects? Fourth, would inversion of scenes and faces have effects on regions outside of the occipitotemporal visual areas that were the focus of earlier studies? To anticipate, we found that face inversion does lead to greater engagement of generic object processing regions, scene inversion has a qualitatively similar though less dramatic effect, the magnitude of these neural effects are not tightly coupled to the magnitude of the behavioral inversion effects, and both face and scene inversion have large effects on many cortical regions that may be involved in post-recognition processing.
Section snippets
Subjects
Twelve healthy right-handed volunteers (six males, median age 20.5 years) were recruited from the local community and gave informed consent according to procedures approved by the University of Pennsylvania institutional review board. All subjects had normal or corrected-to-normal vision.
fMRI parameters
Subjects were scanned at the Hospital of the University of Pennsylvania on a 3 T Siemens Trio scanner equipped with a Siemens body coil and a four-channel head coil. T2* weighted images sensitive to blood
Behavioral results
Reaction times to correct trials were examined (Table 1). An omnibus analysis of variance revealed significant main effects of stimulus type [F(2,22) = 12.4, p < 0.001] stimulus orientation [F(1,11) = 40.6, p < 0.001] and run [F(3,33) = 18.8, p < 0.001]. Critically, we also observed a significant stimulus type × stimulus orientation interaction [F(2,22) = 4.8, p < 0.02], reflecting the fact that increases in reaction time after both face and scene inversion were significantly greater than the increases in RT
Discussion
We examined the behavioral and neural consequences of face and scene inversion with the aim of identifying similarities and differences between the effects induced by these two manipulations. In particular, we addressed four questions. First, does face inversion lead to greater engagement of regions involved in generic object processing? Second, does scene inversion lead to a behavioral penalty and a similar shift towards greater engagement of generic object-processing regions? Third, can we
Acknowledgements
We thank Elaina Lord, Justin Lacasse, Nick Tang, Paul Hirshorn, and Mark Brack for their assistance with this experiment, and Paul Downing and Martha Farah for comments on the manuscript. This research was supported in part by NIMH/NINDS grant NS045839 to the Center for Functional Neuroimaging at the University of Pennsylvania and funds provided by the School of Arts and Sciences.
References (73)
- et al.
Empirical analyses of BOLD fMRI statistics. II. Spatially smoothed data collected under null-hypothesis and experimental conditions
Neuroimage
(1997) - et al.
An area within human ventral cortex sensitive to “building” stimuli: Evidence and implications
Neuron
(1998) - et al.
Inversion and configuration of faces
Cognitive Psychology
(1993) - et al.
Response and habituation of the human amygdala during visual processing of facial expression
Neuron
(1996) - et al.
A temporoparietal and prefrontal network for retrieving the spatial context of lifelike events
Neuroimage
(2001) - et al.
Neural basis of spontaneous thought processes
Cortex
(2004) - et al.
Bodies capture attention when nothing is expected
Cognition
(2004) - et al.
Other minds in the brain: A functional imaging study of “theory of mind” in story comprehension
Cognition
(1995) - et al.
Reading the mind in cartoons and stories: An fMRI study of ‘theory of mind’ in verbal and nonverbal tasks
Neuropsychologia
(2000) The neural basis of object perception
Current Opinion in Neurobiology
(2003)
The lateral occipital complex and its role in object recognition
Vision Research
The well-worn route and the path less traveled: Distinct neural bases of route following and wayfinding in humans
Neuron
Large-scale mirror-symmetry organization of human occipito-temporal object areas
Neuron
The effect of face inversion on activity in human neural systems for face and object perception
Neuron
Mental navigation in humans is processed in the anterior bank of the parieto-occipital sulcus
Neuroscience Letters
The effect of face inversion on the human fusiform face area
Cognition
Increased activity in human visual cortex during directed attention in the absence of visual stimulation
Neuron
Brain regions sensitive to the face inversion effect: A functional magnetic resonance imaging study in humans
Neuroscience Letters
The topography of high-order human object areas
Trends in Cognitive Sciences
The many faces of configural processing
Trends in Cognitive Sciences
Agnosia for scenes in topographagnosia
Neuropsychologia
Current directions in social cognitive neuroscience
Current Opinion in Neurobiology
Beauty in a smile: The role of medial orbitofrontal cortex in facial attractiveness
Neuropsychologia
What's lost in inverted faces?
Cognition
Face perception: Domain specific, not process specific
Neuron
Empirical analyses of BOLD fMRI statistics. I. Spatially unsmoothed data collected under null-hypothesis conditions
Neuroimage
Topographical disorientation: A synthesis and taxonomy
Brain
Stimulus inversion and the responses of face and object-sensitive cortical areas
Neuroreport
Human cortical activity correlates with stereoscopic depth perception
Journal of Neurophysiology
The size of the connected components of excursion sets of chi(2), t and F fields
Advances in Applied Probability
From piecemeal to configurational representation of faces
Science
Contextually evoked object-specific responses in human visual cortex
Science
Neural components of social evaluation.
Journal of Personality & Social Psychology
Why faces are and are not special: An effect of expertise
Journal of Experimental Psychology. General
A cortical area selective for visual processing of the human body
Science
The cortical basis of visual scene processing
Visual Cognition
Cited by (98)
The face inversion effect or the face upright effect?
2023, CognitionThe effects of stimulus inversion on the neural representations of Chinese character and face recognition
2022, NeuropsychologiaCitation Excerpt :In fact, category-selective regions may respond differently to upright and inverted faces. For example, early fMRI studies showed that inverted faces elicited stronger neural responses in the right middle FG and bilateral LOC (Epstein et al., 2006) or the same activity magnitude in the FFA as upright faces (Aguirre et al., 1999; Haxby et al., 1999). These findings suggested that inverted faces activated post-recognition processing and caused a shift from specialised processing towards generic object processing (Haxby et al., 2000).