Elsevier

Behavioural Brain Research

Volume 251, 15 August 2013, Pages 155-162
Behavioural Brain Research

Research report
Fusiform Gyrus responses to neutral and emotional faces in children with Autism Spectrum Disorders: a High Density ERP study

https://doi.org/10.1016/j.bbr.2012.10.040Get rights and content

Abstract

Face processing is a neural mechanism that allows understanding social information and cues conveyed by faces, whose dysfunction has been postulated to underlie some of the behavioral impairments characterizing Autism Spectrum Disorders (ASD). A special region of the cortex, the Fusiform Gyrus (FG), is believed to be the specific area for processing face features and emotions. However, behavioral, fMRI and ERP studies addressed to investigate the role of FG dysfunction in ASD have led to conflicting results. Using a high-density EEG system, we recorded the face-sensitive ERP to neutral and emotional (happiness and fearful) faces, as a measure of early activity of the FG, in children with high functioning ASD. By controlling a number of experimental and clinical variables that could have biased previous research – such as gaze direction, attention to tasks, stimulus appearance and clinical profiles – we aimed to assess the effective role of the FG in the face emotion processing deficit hypothesized in ASD. No significant differences in early face-sensitive ERP components were found between ASD and neurotypical children. However, a systematic latency delay and amplitude reduction of all early potentials were observed in the ASD group, regardless of the stimulus, although more evident for emotions. Therefore, we can assume a diffuse dysfunction of neural mechanisms and networks in driving and integrating social information conveyed by faces, in particular when emotions are involved, rather than a specific impairment of the FG-related face processing circuit. Nevertheless, there is need of further investigation.

Highlights

► We recorded face-sensitive ERP to neutral and emotional faces in ASD. ► We found no specific delay of face processing over the FG region. ► Different face emotions do not modulate face-sensitive ERP components.

Introduction

The presence in humans of complex abilities specific for faces processing, strongly suggests that faces belong, in humans, to a very ‘special’ class of stimuli. Ontogenetically, faces represent the first and most fundamental channel of communication with caregivers. The ability to gather information from face features represents a pivotal skill in child development, and therefore a basic condition for social interactions.

Face processing and recognition have been deeply studied and discussed during the last decades. Neuro-scientific approach to this issue has allowed to recognizing specific brain areas and circuits involved in face processing. In the last years, two different neural systems deputed to faces recognition have been identified: a ‘core’ system which underlies the early visual processing of faces properties, such as shape and size, and an ‘extended’ system in which distinct brain areas perform a further processing of the information in a more sophisticated way [1], [2] (Fig. A.1).

The core system includes but is not limited to the right Fusiform Gyrus (FG). FG may be considered a sort of “front gate”, where the face is recognized as the “face” through the processing of its invariant aspects. It is well known that the FG is more active during face visualization compared to object visualization tasks. This activation engages a specific region of the FG, the lateral portion, referred to as Fusiform Face Area (FFA) [3], typically more strongly in the right hemisphere than in the left [4]. FFA seems to be involved in the face processing in a “high-level” manner, thus contributing to the processing of identity [5].

The temporal profile of the neural processing involved in face perception has been measured using event-related potentials (ERP) [6]. In neurotypical individuals, a negative deflection between 140 and 200 ms is found after the appearance of a face. This ERP component is significantly larger in amplitude for faces than for non-face objects [7], [8]. It has been proposed that this “face-selective N170” may be associated with the encoding of face structure [9], [10], [11] and/or detection of faces [12]. The N170 component is therefore considered as an established measure of early stages of face processing both in adults and children. By combining fMRI and ERP measures, a high correlation between N170 and activation in FFA and superior temporal sulcus (STS) has been found [13]. However, a previous study [14] had already detected a correlation between changes in the N170 and in the FG activation, when noise degraded facial images were manipulated.

On the other hand, it has been hypothesized that N170 could also be modulated by face emotional expressions rather than only by face structural features [15]; the debate is still open. Gunji et al. [16] underlined the specificity of N170 for face stimuli, without individuating any further specialization for specific emotions. Other studies noticed a shorter latency [15] and a wider amplitude [17] of N170 for fearful faces, rather than for other faces expressing neutral, anger or happiness emotions.

Children with Autism Spectrum Disorders (ASD) typically show early and severe impairments in understanding social and emotional information conveyed by faces, and this could be strictly connected to the disruption of the ability to activate specific brain circuits involved in face processing. Several authors argue that such difficulties represent a core deficit of autism [18]. In the last years, many authors have performed behavioral studies on face processing in ASD, but findings remain unclear: some studies have found intact face processing abilities (i.e.: emotion recognition) in ASD [19], [20], [21], while others found profound deficits [22], [23].

Functional neuroimaging and ERP studies investigating general and emotional face processing in ASD, have only partially contributed to resolve this issue, leading to mixed results. However, these instruments remain the most powerful tools for understanding connections between neurobiological and neuropsychological functioning and behavioral symptoms.

fMRI studies investigating the FG activity as a measure of face processing have reported a decreased activation in ASD [24], [25], [26]. On the contrary, other studies using eye-tracking, or manipulating gaze to faces during fMRI [27] indicated that individuals with ASD show a typical activation of FG in response to faces when looking at the eyes region. This last finding suggests that the lack of control of gaze direction toward the target during experiments may represent a crucial bias.

Also high-density EEG and ERP studies led to conflicting results. Analyzing ERP responses to both familiar and unfamiliar faces and objects, Dawson et al. [28] found that neurotypical children showed differential responses, while children with ASD did not. The same authors [15] also investigated the emotional modulation of face processing using both neutral and angry faces as stimuli. Authors found that their low-functioning ASD sample, differently from chronological age-matched control individuals, did not show a differential modulation in early and later ERP components. In contrast, Wong et al. [29] reported normal ERP in children with high-functioning ASD, although they found differences in dipole source localization. In particular ASD group showed weaker activation of ERP dipole sources in frontal, fusiform, and visual cortices, along with slower and larger responses in parietal somatosensory cortices. The latter finding may relate to a more effortful processing of facial emotions in autism and indicate the presence of more complex and wider circuits involved in neuro-behavioral impairments in individuals with ASD. Taken together, these two studies highlight the probable influence of cognitive functioning on processing. It is presumable that intellectual disability might per se affect ERP results.

To summarize, despite the huge bulk of studies, evidence suggests the need for additional experimental paradigms with a stronger control of the numerous bias in order to detect differences in face processing abilities of children with typical development and ASD. We have designed the paradigm of this study with the purpose of controlling at least some of the confounding variables, at different levels (e.g. stimulus, task, attention, basic gaze control, clinical heterogeneity, cognitive abilities), and processing data in different ways (e.g. automatic and manual tools, different references, different ROIs).

The aim of this study was to record face-sensitive ERP to neutral and emotional (happy and fearful) faces, and to a third control stimulus (trees) as a measure of early (first 200 ms) activity of the FG in children with ASD. This information could help to understand the effective role of the FG in the face emotion processing deficit encountered in ASD.

Section snippets

Subjects

A total of 15 children with a diagnosis of “Autistic Disorder” (AD) following DSM-IV-TR criteria, confirmed by ADOS-G and ADI-R, were recruited for this study. Despite the care given to the preparation, 2 children ended up refusing to wear the EEG cap. Moreover, 13 children completed the protocol, but 3 of them were rejected for the presence of too many artifacts in their EEG recordings.

The experimental sample consisted of 10 children with AD aged 6–13 years (mean 10.2 years). Other criteria of

Results

Recorded ERP latencies and amplitudes for each condition in both groups are reported in Table B.1. Waveforms are showed in Graph C.1, Graph C.2, Graph C.3, Graph C.4, Graph C.5. The two main early ERP components were clearly identifiable in the two groups of children. For all stimuli, in both hemispheres, the AD group showed higher latency and lower amplitude values.

Methodological matters

In Section 1, we have emphasized how evidences on the FG role in processing face and emotions in ASD are still contradictory [33], [34]. Conflicting results may be at least partially due to heterogeneity in methodology and experimental setups. In particular, most studies investigating the FG activity in ASD lack of explicit control of gaze direction, attention to tasks, stimulus characteristics and clinical profiles.

Since our experimental setup did not include a standardized system of

Conclusions

The results of the statistical analysis revealed no significant differences between groups with regard to latencies and amplitudes of the N170 (and ppN170), suggesting that the early FG face-specific responses are preserved in ASD. In this study, however, children with AD showed systematically delayed latencies and reduced amplitudes of all the early potentials, regardless of the stimulus “face” itself. The evaluation of peak-to-peak potentials has been necessary to distinguish the actual

Acknowledgements

This research was granted by the Italian Ministry of Health, Strategic Research Program “IDIA: Inquiry into Disruption of Intersubjective equipment in Autism spectrum disorders in childhood”.

The financial support of Telethon - Italy (Grant no. GGP11188) is gratefully acknowledged.

Also, the authors thank Dr. Natasha Chericoni and dr. Sara Baldini for assistance.

References (40)

  • N. Hadjikhani et al.

    Activation of the Fusiform Gyrus when individuals with autism spectrum disorder view faces

    Neuroimage

    (2004)
  • N. Tottenham et al.

    The NimStim set of facial expressions: judgments from untrained research participants

    Psychiatry Research

    (2009)
  • D. Tucker

    Spatial sampling of head electrical welds: the geodesic sensor net

    Electroencephalography and Clinical Neurophysiology

    (1993)
  • B.A. Corbett et al.

    A functional and structural study of emotion and face processing in children with autism

    Psychiatry Research

    (2009)
  • A. Senju et al.

    Atypical eye contact in autism: Models, mechanisms and development

    Neuroscience and Biobehavioral Reviews

    (2009)
  • S.M. Letourneau et al.

    Behavioral and ERP measures of holistic face processing in a composite task

    Brain and Cognition

    (2008)
  • K. Howner et al.

    Brain processing of fearful facial expression in mentally disordered offenders

    Journal of Behavioral and Brain Science

    (2011)
  • N. Kanwisher et al.

    The fusiform face area: a module in human extrastriate cortex specialized for face perception

    The Journal of Neuroscience

    (1997)
  • K. Nagy et al.

    The lateral occipital cortex in the face perception network: an effective connectivity study

    Frontiers in Psychology

    (2012)
  • S. Bentin et al.

    Electrophysiological studies of face perception in humans

    Journal of Cognitive Neuroscience

    (1997)
  • Cited by (46)

    • Facial emotion processing and language during early-to-middle childhood development: An event related potential study

      2022, Developmental Cognitive Neuroscience
      Citation Excerpt :

      With respect to FEP, the N170 is typically associated with the initial processing of face characteristics (Eimer and Holmes, 2007). Previous studies have primarily found no emotion effect on the N170 in children (Apicella et al., 2013; Usler et al., 2020), however some studies have reported emotion effects (Battaglia et al., 2017; Batty and Taylor, 2006). Previous FEP studies support a non-linear relationship between age and N170 amplitude in children, however, overall, N170 amplitude tended to become stronger with increasing age (though for contrasting results see (Battaglia et al., 2007).

    • The development of neural responses to emotional faces: A review of evidence from event-related potentials during early and middle childhood

      2021, Developmental Cognitive Neuroscience
      Citation Excerpt :

      There were differences in the average P100 amplitude between studies, ranging from 3.12 μV (Miki et al., 2011) to 27.5 μV (Batty et al., 2011). Sample size across studies ranged from 12 (Apicella et al., 2013) to 69 (Batty and Taylor, 2006), with a median of 26 participants. Studies included a variety of stimulus sets, however the most commonly used was the NimStim set (Tottenham et al., 2009).

    • Linear and nonlinear analysis of intrinsic mode function after facial stimuli presentation in children with autism spectrum disorder

      2021, Computers in Biology and Medicine
      Citation Excerpt :

      The whole analysis is carried out in the MATLAB software package R2018a. The dataset was taken from a previous EEG-based EFE ERP studies [7,8]. The dataset contains EEG signals from 24 subjects −12 ASD and 12 TD; age group 6–13 years (mean age 10.2 and 9.7 years for ASD and TD, respectively).

    • Classification of evoked responses to inverted faces reveals both spatial and temporal cortical response abnormalities in Autism spectrum disorder

      2021, NeuroImage: Clinical
      Citation Excerpt :

      We found that while it was possible to accurately classify evoked responses to upright faces relative to houses for both groups, differences between the TD and ASD groups emerged only for classification accuracies in response to inverted faces relative to houses. The results for upright faces align with prior studies from our group and others showing no group differences in evoked responses to upright faces in ASD (Apicella et al., 2013; Khan et al., 2013; Mamashli et al., 2018), and confirms that a multivariate approach to this question yields results that are in line with previously used univariate approaches, as expected. The results for inverted neutral faces, in contrast, were novel, and did show a group difference that did not manifest for upright neutral faces.

    View all citing articles on Scopus
    View full text