Changes in male brain responses to emotional faces from adolescence to middle age

Abstract

Facial emotion perception is fundamental to human social behaviour, and changes with age. Nevertheless, age-related differences in the relative activation of components of emotion processing networks are poorly understood. Thus we measured brain activity with event-related fMRI in 40 right handed healthy male controls, age range 8–50 years, during implicit processing of fearful, disgusted, and a control condition of neutral facial expressions. There was a significant negative correlation between increasing age and neural response to fearful and disgusted expressions in dorsomedial prefrontal cortex (BA 10), and middle frontal gyri (BA 6). Hence, in healthy subjects, the functional anatomy of facial emotion processing is not ‘hard-wired’, but undergoes progressive change into adulthood. Possible explanations for the age-related changes in dorsomedial and middle frontal cortical activity may include a reduction in the attentional demands of appraising facial expressions as perceptual skill increases, or changes in processing the self-relevance of facial expressions during social and cognitive development.

Introduction

Both developmental disorders of childhood (e.g. autism) and adult-onset psychiatric disorders (e.g. schizophrenia), are associated with impairments in facial expression recognition, and functional anatomical differences in facial emotion processing compared to healthy controls (Phillips et al., 2003a, Phillips et al., 2003b, Herba and Phillips, 2004). Also, the incidence of some psychiatric disorders changes with increasing age — for example, the onset of paranoid schizophrenia (in which people misconstrue the social behaviour of others, and/or differ in emotional reactivity) normally peaks at age 18 in men, and is rare before age 10 and past age 40 (Sham et al., 1994, Aylward et al., 2005). Hence, a necessary first step in determining the pathogenesis of disturbances of social cognition is to understand how the functional anatomy of facial emotion processing develops and changes in healthy individuals throughout life.

In healthy adults, facial emotion processing is associated with activation in core visual analysis areas (inferior occipital gyrus, fusiform gyrus, and superior temporal gyrus); subcortical and limbic regions (such as basal ganglia, insula, and amygdala); and prefrontal regions (such as ventromedial and ventrolateral regions, and anterior cingulate gyrus) (Haxby et al., 2002, Phillips et al., 2003a, Phillips et al., 2003b). Also, there is differential activation of limbic regions depending on emotion type (e.g. greater amygdala activation for fear, and insula activation for disgust) (Phillips et al., 2003a, Phillips et al., 2003b). Recent evidence also suggests that children may activate similar neural systems when processing facial emotions (Lobaugh et al., 2006) and have differential emotion-specific responses (e.g. greater amygdala and insula activation to fear and disgust than to sadness) (Lobaugh et al., 2006).

Nevertheless, the developmental trajectories of changes in the relative activation of components of emotion processing networks are poorly understood. Some prior studies have reported age-related reductions in limbic and/or visual cortical responses to emotional expressions (Iidaka et al., 2002, Gunning-Dixon et al., 2003, Mather et al., 2004). For example, one study employing a gender discrimination task reported that compared to younger subjects (mean age 25.1 ± 5.0 years), older subjects (mean age 65.2 ± 2.6 years) showed reduced activity in left amygdala to negative expressions, and right parahippocampal gyrus and parieto-occipital cortex to positive expressions (Iidaka et al., 2002). Another study employing an emotion discrimination task reported that compared to younger subjects (mean age 25.8 ± 3.1 years), older subjects (mean age 72.3 ± 7.6 years) showed reduced activation in visual cortices and amygdala (Gunning-Dixon et al., 2003). A further study reported reduced arousal ratings and amygdala activation in older (mean age 78.41 ± 4.86) compared to younger adults (mean age 23.41 ± 3.24) to negative but not positive facial expressions (Mather et al., 2004). Nevertheless, contrasts of neural responses across younger age ranges have produced conflicting findings. For example, one study showed reduced amygdala activation in adults (mean age 30.76 ± 3.07 years) relative to adolescents (mean age 13.2 ± 2.60 years) in a contrast of fearful to neutral expressions (Monk et al., 2003). By contrast, another reported that adults (mean age 24 ± 6.6 years) showed greater amygdala activation to fearful relative to neutral expressions compared to children (mean age 11 ± 2.4 years) (Thomas et al., 2001). Hence, there is accumulating evidence from contrasts of younger and older adults that older age is associated with reductions in limbic and/or visual cortical responses to facial expressions of emotion (with evidence that this effect may be restricted to negative expressions; Mather et al., 2004). However, it remains unclear as to whether the pattern of age-related reductions in limbic and/or visual cortical activity to expressions is present from adolescence through to middle age. This is an important period for investigations of the effects of aging on social cognition because it encompasses the transition from adolescence to adulthood, a critical period of social development associated with the peak incidence of adult-onset psychiatric disorders such as schizophrenia (Sham et al., 1994). It also includes key periods of brain maturation (e.g. myelination of the frontal cortex continues to the thirties) (Craik and Bialystok, 2006).

Hence, we investigated age-related changes in the functional anatomy of facial emotion processing spanning the whole period from adolescence to middle age while including a variety of facial emotions. We examined neural responses to fearful, disgusted, and a control condition of neutral facial expressions, using event-related fMRI, in 40 healthy right handed males of normal intelligence, age range 8–50 years. We examined neural responses to fearful expressions because of their proposed key role in socialization, whereby developing individuals learn to avoid behaviours that elicit the aversive arousal associated with fearful expressions (Blair, 2003). Also, prior studies had not investigated age-related changes in neural responses to disgusted expressions per se, although two studies had included disgust in mixed emotion blocks (Iidaka et al., 2002, Gunning-Dixon et al., 2003). We therefore included disgusted expressions as important social signals which convey information about avoidance of aversive or potentially noxious stimuli (e.g. rotten foods) in addition to some forms of socially unacceptable behaviour (Blair, 2003).

We postulated that the affective response elicited by facial expressions of emotion per se (the reinforcement value) would be maximal in childhood and adolescence to motivate attention to important social cues during the most formative period of socialization (Joffe, 1997, Blair, 2003, Monk et al., 2003), in contrast to adults who can enlist mature social perceptual skills and contextual knowledge to determine the meaning and emotional significance of facial expressions (Herba and Phillips, 2004). Emotionally salient expressions elicit increased activity in visual cortices and limbic regions in contrast to neutral expressions (Surguladze et al., 2003a, Surguladze et al., 2003b). Hence we predicted that age-related reductions in the affective response elicited by expressions of emotion would be associated with attenuation of limbic and visual cortical activity (Surguladze et al., 2003a, Surguladze et al., 2003b, Vuilleumier et al., 2004)—in other words, that the pattern of reduced limbic and visual cortical activity reported in contrasts of young and old adults would also be shown in the period from adolescence to middle age.