The stimuli drive the response: An fMRI study of youth processing adult or child emotional face stimuli

Highlights

• Behavior of youth differs in response to child versus adult emotion faces.

• Neural systems in youth process child and adult emotion faces differently.

• The immature amygdala responds to happy faces of peers, and negative faces of adults.

• Emotional signals from adults signify something different than peers.

Abstract

Effective navigation of the social world relies on the correct interpretation of facial emotions. This may be particularly important in formative years. Critically, literature examining the emergence of face processing in youth (children and adolescents) has focused on the neural and behavioral correlates of processing adult faces, which are relationally different from youth participants, and whose facial expressions may convey different meaning than faces of their peers. During a functional magnetic resonance imaging (fMRI) scan, we compared concurrent neural and behavioral responses as youth (N = 25) viewed validated, emotionally varied (i.e., anger, fear, happy, and neutral) adult and child face stimuli. We observed that participants made fewer errors when matching adult, compared to child, face stimuli, and that while similar brain regions were involved in processing both adult and child faces, activation in the face processing neural network was greater for adult than child faces. This was true across emotions, and also when comparing neutral adult versus neutral child faces. Additionally, a valence by stimuli-type effect was observed within the amygdala. That is, within adult face stimuli, negative and neutral face stimuli elicited the largest effects, whereas within child face stimuli, happy face stimuli elicited the largest amygdala effects. Thus, heightened engagement of the amygdala was observed for happy child and angry adult faces, which may reflect age-specific salience of select emotions in early life. This study provides evidence that the relational age of the perceived face influences neural processing in youth.

Introduction

Affective facial displays are central to the human experience. Seminal work by Paul Ekman showed that these displays are universally expressed, unbidden responses to internal feelings of emotion (Ekman, 1993), which largely extended Charles Darwin's early consideration of emotions and their expressions as physiological responses to environmental stimuli (Darwin and Ritter, 1916). Facial displays of others communicate key information, and set the emotional tone of interpersonal interactions. The predilection of infants to readily habituate to facial displays (Simion et al., 2007) suggests that in humans the ability to extract emotion from faces is present early, and perhaps foundational to later social and emotional development. The ability to process facial displays of emotion develops across the first two decades of life, commensurate with brain maturation and with the accumulation of experience (Hoehl and Peykarjou, 2012, Leppanen and Nelson, 2009). Overall, successful transition into adulthood is thought to involve growing capability to understand emotion and to regulate one's response to emotionally evocative stimuli. Because late childhood and adolescence are marked by the emergence of psychiatric illness (Dahl and Gunnar, 2009, Paus et al., 2008), and because aberrant processing of facial displays has been demonstrated in youth at risk for developmental psychopathology (Tottenham et al., 2009), there is reason to further examine the neural and behavioral bases of face processing in youth (children and adolescents). Strong characterization of these processes in typically developing youth will improve our understanding of the emotional armory of childhood, while also providing a means for making future comparisons in at-risk samples.

Studies of the neural and behavioral bases of face processing in adults are numerous. Literature review of the past seventeen years reveals more than 600 published functional magnetic resonance imaging (fMRI) studies of face processing in adults (Fig. 1A). Adult fMRI studies show that regions of the fusiform gyrus, inferior and middle occipital gyrus, lingual gyrus, middle/superior temporal gyrus, limbic, and prefrontal cortex (PFC) are involved in processing emotional facial displays (Fusar-Poli et al., 2009, Palermo and Rhodes, 2007). Adult fMRI and lesion studies corroborate early theories that select areas of the face processing network contribute to different aspects of face processing (Adolphs, 2002, Haxby et al., 2002). For example, occipital cortices process perceptual properties of faces (Adolphs, 2002), the amygdala responds to threat-related stimuli (e.g., fearful, angry faces; Breiter et al., 1996, Whalen et al., 2001), and PFC aspects of the response are attributed to emotion regulation processes (Hariri et al., 2000, Phillips et al., 2003). The PFC is also recognized as playing a central role in social cognitive functions such as evaluation of social reinforcement in faces (Haxby et al., 2002). Enhanced responses to emotionally expressive face stimuli have also been observed in the face-selective region of the fusiform gyrus, the fusiform face area (FFA), but this more likely reflects enhanced attention allocation, possibly via feedback from the amygdala to early visual regions (Kanwisher and Yovel, 2006, Vuilleumier and Pourtois, 2007). Taken together, prior evidence in adults indicates that affective facial displays are processed via a distributed neural network, and in a valence-specific manner.

In contrast to the more than 600 adult fMRI studies of face processing, fewer than 50 published fMRI studies have examined face processing in youth. Functional MRI studies of face processing in youth have demonstrated that the same core regions of the brain that process emotional faces in adults also underlie face processing in youth (Hoehl et al., 2010, Lobaugh et al., 2006). Differences between children and adult participants primarily reside in the magnitude and extent to which face regions are engaged during face processing fMRI tasks, and in how particular brain regions respond selectively to specific valences. Indeed, compared to adults, youth show attenuated activation in the PFC to emotional faces (Monk et al., 2003), which has been attributed to protracted maturation of prefrontal regions (Gogtay et al., 2004, Luna, 2009). Youth also show increased selectivity in the FFA (Golarai et al., 2010), and decreased amygdala activity (Guyer et al., 2008, Monk et al., 2003, Thomas et al., 2001) with age, fitting with knowledge that the FFA (Scherf et al., 2007) and amygdala (Giedd et al., 1996, Tottenham and Sheridan, 2009) continue to develop beyond adolescence.

Neural studies of face processing in youth have relied almost exclusively on adult emotion face stimuli (Fig. 1B). The reliance on adult face stimuli for understanding the emergence of neural substrates for face processing in children is a chief limitation recognized by the field (Somerville et al., 2011). This concern arises from observations that age of the face stimuli (Hoehl et al., 2010), familiarity with the actor (Nielson et al., 2010), and relationship of the face to the viewer (Leibenluft et al., 2004, Taylor et al., 2009) alter neural responses in fMRI studies. In a developmental context, the amygdala, in particular, may respond differently to the emotional faces of similarly aged children, so-called “peers”, than to the emotional faces of adults that are inherently authority figures. After all, activation in the amygdala is driven by the emotional significance and behavioral relevance of faces (Anderson and Phelps, 2001, Davis and Whalen, 2001), and adult faces convey specific behavioral relevance to children. Thus, faces of emotional peers may be processed differently than emotion faces of adults, in children.

Indeed, Hoehl and colleagues report evidence for altered processing of adult versus child angry and happy emotion faces in 5–6 year old children (Hoehl et al., 2010). Specifically, activation was observed in the amygdala for angry adult and happy child faces (Hoehl et al., 2010), which was attributed to the regular occurrence of angry expressions of adults to children, and the relevance of peer smiles, respectively. Other recent youth fMRI studies have demonstrated that amygdala response is influenced by learned biases to select races (Telzer et al., 2012). The latter reinforces that experience can alter amygdala response in youth. Together, these findings underscore that prominent features of a face, such as age or emotion, may alter the ways in which that face is processed. Data such as these motivate discussion about the role of experiential learning, or socially relevant cues, in shaping amygdala response (Hooker et al., 2006, Todd and Anderson, 2009).

Behavioral studies of face processing provide further support that characteristics of the stimuli affect speed and accuracy. Recognition memory is best for in-group stimuli (e.g., species, race; Meissner and Brigham, 2001, Scott et al., 2005), and for similarly aged faces (review by Rhodes and Anastasi, 2012). In contrast to a theoretical framework that suggests a processing advantage for own-age face stimuli (see Rhodes and Anastasi, 2012), children appear to be more accurate at recognizing faces of adults (Macchi Cassia, 2011, Macchi Cassia et al., 2012), which may reflect the frequent interaction with adult faces from a young age. In addition, those who interact with children more frequently, such as school teachers, show improved capacity to recognize child faces (Harrison and Hole, 2009), suggesting that these biases are experientially-driven rather than inherent. Hoehl et al. did not find accuracy differences for happy adult compared to happy child faces in a post hoc emotion recognition task, but for angry faces, children were less accurate at recognizing anger in their peers relative to that in adults (Hoehl et al., 2010). Overall, it is apparent that experience or familiarity with select facial stimuli influence processing. This work has motivated interest in how salient aspects of a face affect neural and behavioral processing and raise important questions regarding the exclusive use of adult face stimuli in facial emotion processing studies of youth.

As the transition into adolescence brings about changes in social behavior and the primacy of peer relationships is established, there is need to extend work by Hoehl and colleagues beyond young childhood, using validated child face stimuli. The aim of the present study was to examine differences in brain engagement when late childhood/adolescent participants viewed emotional faces of peers compared to faces of adults. Using a well-vetted experimental paradigm that consists of matching the identity of the face, we compared within-subjects neural and behavioral responses to two variations of a task; one with the most widely used adult emotional face stimuli (Ekman faces; see Figs. 1B and 2A) and the other with newly validated child emotional faces (Fig. 2B; Egger et al., 2011), in which actors used in stimuli closely match the ages of our participants. We examined the following hypotheses: (1) that the same basic neural systems underlie processing of adult and peer faces, (2) that magnitude and extent of response in face processing brain regions will vary between child and adult face stimuli, and (3) that the profile of amygdala response will differ across emotional valences of peer and adult faces. Specifically, we predicted that youth will show more neural activation for adult compared to child emotion faces. This prediction is based on evidence that children's face representation is tuned to adult faces (Macchi Cassia et al., 2012) and also the idea that adult faces should better signal behaviorally relevant outcomes, and thus be associated with enhanced neural processing. Given that Hoehl et al. (2010) observed greater amygdala response for angry adult faces compared to angry child faces, we expected to observe greater amygdala response to negative emotion faces (e.g., fearful, angry) of adults compared to child same-emotion faces. The perception of peer faces, in contrast, may increase amygdala activation for happy faces compared to other peer emotions, as observed in younger children (Hoehl et al., 2010).