Task-irrelevant fear enhances amygdala-FFG inhibition and decreases subsequent face processing

Social Cognitive and Affective Neuroscience, 2016, 1440–1448 doi: 10.1093/scan/nsw054 Advance Access Publication Date: 5 June 2016 Original article Task-irrelevant fear enhances amygdala-FFG inhibition and decreases subsequent face processing Barbara Schulte Holthausen1,2, Ute Habel1,2, Thilo Kellermann1,2, Patrick D. Schelenz1, Frank Schneider1,2, J. Christopher Edgar3, Bruce I. Turetsky4, and Christina Regenbogen1,2,5 1 Department of Psychiatry, Psychotherapy and Psychosomatics, Medical Faculty, RWTH Aachen, Aachen 52074, Germany, 2JARA - BRAIN Institute 1: Structure Function Relationship, Jülich 52428, Germany, 3 Department of Radiology, The Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA, 4Department of Psychiatry, University of Pennsylvania, Philadelphia, PA 19104, USA, and 5Department of Clinical Neuroscience, Karolinska Institutet, Stockholm 17177, Sweden, Correspondence should be addressed to Barbara Schulte Holthausen, Department of Psychiatry, Psychotherapy and Psychosomatics, Medical Faculty, RWTH Aachen, Pauwelsstraße 30, D-52074 Aachen, Germany. E-mail: . Abstract Facial threat is associated with changes in limbic activity as well as modifications in the cortical face-related N170. It remains unclear if task-irrelevant threat modulates the response to a subsequent facial stimulus, and whether the amygdala’s role in early threat perception is independent and direct, or modulatory. In 19 healthy participants, crowds of emotional faces were followed by target faces and a rating task while simultaneous EEG-fMRI data were recorded. In addition to conventional analyses, fMRI-informed EEG analyses and fMRI dynamic causal modeling (DCM) were performed. Fearful facial crowds reduced EEG N170 target face amplitudes and increased responses in a fMRI network comprising insula, amygdala and inferior frontal cortex. Multimodal analyses showed that amygdala response was present 60 ms before the right fusiform gyrus-derived N170. DCM indicated inhibitory connections from amygdala to fusiform gyrus, strengthened when fearful crowds preceded a target face. Results demonstrated the suppressing influence of task-irrelevant fearful crowds on subsequent face processing. The amygdala may be sensitive to task-irrelevant fearful crowds and subsequently strengthen its inhibitory influence on face-responsive fusiform N170 generators. This provides spatiotemporal evidence for a feedback mechanism of the amygdala by narrowing attention in order to focus on potential threats. Key words: amygdala, EEG-fMRI, faces, priming, threat processing Introduction Emotional face processing is an universal and adaptive process that facilitates verbal and non-verbal communication as well as the evaluation of potential threats (Dolan, 2002; Ekman and Friesen, 1971). Fearful and angry facial expressions signal potential danger, and rapid threat decoding is hypothesized to provide a survival benefit even when the nature of the threat is not fully evaluated (Adolphs, 2008). On a behavioral level, this ‘threat-priority’ processing results in faster reaction times and increased detection performance (Hansen and Hansen, 1988; € Eastwood et al., 2001; Ohman et al., 2001; but see Purcell and Stewart, 2010; Savage et al., 2013). Perceiving and detecting an emotional expression is associated with amygdala blood oxygenation level dependent (BOLD) changes, with strong evidence for the amygdala’s involvement in fearful stimulus processing (Davis, 1992; LeDoux, 2003). More precisely, the amygdala is hypothesized to modify visual encoding of salient social information (Morris et al., 1996; Furl et al., Received: 23 October 2015; Revised: 18 March 2016; Accepted: 18 April 2016 C The Author (2016). Published by Oxford University Press. For Permissions, please email: V 1440 B. Schulte Holthausen et al. 2013), even when outside of the attentional focus (Vuilleumier € et al., 2001; Ohman, 2005). The amygdala is further interconnected with visual areas, and disruption to these pathways results in impaired perception and recognition of emotional facial expressions (Adolphs et al., 1994). Its clinical relevance is evident in several psychiatric conditions; e.g. trait anxiety and depression are associated with hyperactivity of the amygdala (Sheline et al., 2001; Bishop et al., 2007). Intracranial findings indicate amygdala involvement soon after stimulus presentation, up to 100 ms before higher-order saliency evaluation and action preparation analyses in areas such as the insular cortex (Krolak-Salmon et al., 2004; Menon and Uddin, 2010). Contemporary models focus on the amygdala’s exact contribution in threat perception. Models in favor of a socalled ‘low road’ approach hypothesize that the amygdala is directly involved in the automatic processing of affective information via a rapid subcortical pathway that bypasses primary cortical areas (Garrido et al., 2012; Garvert et al., 2014). Other models assume that the amygdala acts modulatory via connections with primary visual and prefrontal areas, thus highlighting the amygdala’s function in coordinating cortical networks during danger and threat evaluation (Morris et al., 1998; Pessoa and Adolphs, 2010). Studies examining the above have focused on automatic threat processing and the results have increased our understanding of the basic mechanisms. (Whalen et al., 1998; Liddell et al., 2005; Williams et al., 2006). However, in real life, potential dangers are not isolated from their surroundings. Contextual factors such as scenes, voices, or even other faces influence identification accuracy and reaction times, as well as the perceived intensity levels of target faces (Masuda et al., 2008; Righart and de Gelder, 2008b; Aviezer et al., 2011). On a neural level, context information is processed at an early perceptual stage (N170), with joint activation in temporal and occipital brain regions (Righart and de Gelder, 2006, 2008a) as well as in prefrontal regions (Lee and Siegle, 2014). This suggests that context has an effect on early markers of visual perception via rapid recruitment of a wide cortical network. It remains unclear how fearful context influences the subsequent processing of a target face and how the amygdala contributes to early threat processing. The present study examined the impact of an emotional task-irrelevant facial crowd on behavioral and neural parameters of a subsequently-presented target face. Obtaining simultaneous EEG-fMRI, the present study sought to delineate when and where in the brain an emotional crowd influences target face processing, focusing on the N170 event-related potential (ERP). The N170 is associated with early face processing and assumed to be generated by the fusiform face gyrus (FFG, Herrmann et al., 2005; Deffke et al., 2007). Whereas some studies speak in favor of an emotion sensitivity of the N170 (Blau et al., 2007), others found the N170 to be emotion-unspecific (Eimer and Holmes, 2002; Herrmann e (...truncated)