Gender differences in hemispheric asymmetry for face processing

Using functional magnetic resonance imaging (fMRI), Kanwisher and coworkers [1] found an area in the fusiform gyrus that was significantly more active when the subjects viewed faces than when they viewed assorted common objects. The authors concluded that this area, hereafter called the fusiform face area (FFA), might be specifically involved in the perception of faces, not ruling out that other structures might be play a role in this process.

Indeed, Haxby et al. (2000) provided evidence that face perception involves a distributed and hierarchically organized network of the occipito-temporal regions. In this model, the core system consists of the extrastriate visual cortex (FFA), which mediates the analysis of face structure, while the superior temporal sulcus (STS) mediates the analysis of changeable aspects of the face, such as eye gaze, facial expression, and lip movements.

Interestingly, the Kanwisher et al. study [1] showed an activation of FFA only in the right hemisphere in about half the subjects (both men and women), whereas the other subjects showed bilateral activation. These results raised the possibility of functional hemispheric asymmetry in the FFA. Studies addressing this possibility have provided conflicting evidence: several human [2–6] and animal studies [7] found stronger activity in the right hemisphere, while other studies failed to support the notion of a strict right-lateralization (e.g. [8] performed in 5 men and 7 women).

Closer examination of several studies offers more details, but no consensus, on hemispheric asymmetry in areas devoted to face processing. Yovel and Kanwisher [9] found significantly higher fMRI responses to faces than to objects in both the left and right mid-fusiform gyrus regions, although this effect was slightly greater in the right than the left FFA. In another fMRI study [10], a region that responded more strongly to faces than to objects was found within the right fusiform gyrus in 8 subjects (both women and men); however, in 6 of these subjects the same significant pattern was also found in the left fusiform gyrus. Recently, Pourtois and coworkers [11] performed an fMRI study on face identity processing on a group of 8 men and 6 women. Results revealed a reduced response in the lateral occipital and fusiform cortex with face repetition. Specifically, view-sensitive repetition effects were found in both the left and right fusiform cortices, while the left (but not right) fusiform cortex showed viewpoint-independent repetition effects. These findings were interpreted as a sign of left hemisphere dominance in terms of the ability to link visual facial appearance with specific identity knowledge. In line with this, a case has been reported of hyperfamiliarity for unknown faces after left lateral temporo-occipital damage in a female patient [12], suggesting a possible role of the left hemisphere in identity processing. Again, a recent fMRI study [13] carried out on 8 women and 7 men provided evidence of a significant activation of right fusiform and occipital gyrus (2260 voxels), left fusiform gyrus, left inferior, and middle temporal gyrus (3022 voxels) for a face familiarity effect during gender classification, thus providing a complex lateralization pattern for processing face structures and properties.

It is of great interest to note that face-specific N170 responses were found to be bilateral or even left-sided in studies involving a sample in which women were the majority [20–24]. Equally interesting, in a recent paper on prosopagnosia in which both male and female patients were considered [20], 2 out of the 3 male patients showed an M170 response which was not sensitive to faces (as opposed to houses) while the third patient showed a right-sided sensitivity to faces. As for the two female patients, one of them showed a lack of sensitivity to faces at M170 level, while the second one showed a left-sided sensitivity.

Many face processing studies using MEG, ERP, neuroimaging, or neuropsychological data do not take subject gender into account as a variable that might affect asymmetry in brain activation. The specific goal of this study was to investigate the timing and topography of brain activity in men and women during processing of neutral and affective faces in order to detect whether there are gender differences in lateralization. To address this question, early face-sensitive P1 and N1 responses over the occipital/temporal cortices to neutral and affective expressions were measured in strictly right-handed men and women.

The P1 response was larger and earlier in women than in men, probably suggesting a female preference for the visual signal (infants' faces). This hypothesis is supported by a recent fMRI study showing a stronger activation of the fusiform gyrus in women (compared to men) in response to children's faces [25]. In our study, both P1 and N1 were affected by the emotional content of faces, being earlier (P110) and larger (N160) in response to distressed faces as opposed to neutral faces. These data fit with the available literature which supports the notion of early effects of emotional [14–17] and attentional factors [26–28] in the first stages of visual cortical processing. Overall, the P1 component was larger over the right occipital area in all individuals, and to all stimuli, as clearly visible from the topographic maps in Fig. 5.

This effect might be due either to sensory or cognitive factors. Since all stimuli were faces, in this experiment the asymmetry cannot be ascribed to a generic effect of face processing. Indeed, the literature on face recognition does not support the evidence of a right lateralization for the P1 response, but, rather, a bilateral distribution is often reported (when P1 is considered, see Table 1). Furthermore, in studies involving visual-spatial or selective attention tasks, the P1 component is often described as larger at the right than the left occipital lateral sites both for space orienting (e.g., [29]) and processing of global configurations (e.g., [30]). In addition, P1 is always right-lateralized in response to low spatial frequency patterns even in passive viewing conditions [31, 32]. For these reasons, we cannot discuss the P1 right lateralization as an index of a hemispheric dominance for face processing.

On the other hand, the face-specific N160 component was clearly lateralized differently in the men and women in our study. Indeed, a strong gender effect in the hemispheric lateralization of N1 component was observed, both in the latency and amplitude of cerebral response. This hemispheric asymmetry in men was not restricted to the processing of affective faces, and was significant in response to both neutral and distressed faces (see topographic maps in Fig. 6, displaying N1 scalp voltage distribution). Thus, a right hemispheric dominance is suggested for face processing in men but not in women. This may explain the many inconsistencies present in the relevant ERP and neuroimaging literature, which sometimes predicts a bilateral effect and other times a strong right-sided activity in regions devoted to face processing. These conclusions often rely on a mixed gender population, in which men and women are not necessarily equally represented (see Table 1).

Our results are also in line with many studies that show gender differences in the degree of lateralization of cognitive and affective processes. Considerable data support greater hemispheric lateralization in men than women for linguistic tasks [33] and for spatial tasks [34]. Gender differences have also been found in the lateralization of visual-spatial processes such as mental rotation [35] and object construction tasks [36], in which males are typically right hemisphere (RH) dominant and females bilaterally distributed. More relevant to the present experiment are the data provided by Bourne [37], who examined the lateralization of processing positive facial emotion in a group of 276 right-handed individuals (138 males, 138 females). Subjects were asked to observe a series of chimeric faces formed with one half showing a neutral expression and the other half showing a positive expression in the left or right visual field, and to decide which face they thought looked happier. The results showed that males were more strongly lateralized than women in the perception of facial expressions, showing a stronger perceptual asymmetry in favour of the left visual field. There are also a number of studies that have found different degrees of lateralization in the cerebral response of men and women to emotional stimuli [38–41]: men tend to demonstrate an asymmetric functioning, and women a bilateral functioning [42].

Our study found a lesser degree of lateralization of brain functions related to face and expression processing in women than men. Furthermore, these results emphasize the importance of considering gender as a factor in the study of brain lateralization during face processing. In this light, our data may also provide an explanation of the inconsistencies in the available literature concerning the asymmetric activity of left and right occipito-temporal cortices devoted to face perception during processing of face identity, structure, familiarity or affective content.