Processing of Image Categories Prior to Awareness in Children with Autism Spectrum Disorder

Background: Amygdala activation is observed even when face stimuli are rendered invisible via suppression techniques (Whalen et al. 1998,2004;Morris et al. 2001;Pasley et al. 2004).  This amygdala activation is thought to guide attention towards salient locations, such as the eye region of faces (Adolphs, 2008).  We have recently identified an amygdala-driven pre-attentive response to fearful faces in the absence of awareness in typical adults (Troiani et al, submitted).  Because attention to faces is atypical in autism, understanding the response to face and non-face stimuli in the absence of awareness is critical to capturing the full integrity of face processing systems in autism.

Objectives: We compared neural activation in adolescents with an autism spectrum disorder (ASD) compared to typically developing children (TDC) to better understand if images are processed differently in these populations prior to awareness.

Methods: Continuous flash suppression (CFS) is a method whereby an image presented to one eye is suppressed from conscious awareness by a stream of rapidly flashed colorful patterns presented to the opposite eye.  A stimulus of interest (i.e. target) is projected to one eye, but due to suppression, participants only perceive the flashing colorful pattern. We used continuous flash suppression (CFS) in conjunction with fMRI while participants were presented with target images including fearful faces, houses, or a no stimulus control outside of conscious awareness. Target images were rendered invisible via flashing colorful images presented to the opposite eye while participants performed an orthogonal task. This task was to detect whether centrally presented letters that appeared overtop the flashing colorful background were vowels or consonants. Whole brain fMRI images were collected from 28 participants (18 ASD;10 age-/IQ-matched TDCs;mean age:14;mean IQ:101).  Five ASD participants were excluded from analysis because they perceived the stimuli (experienced breakthrough from suppression).

Results: First, we did a focused analysis based on our previous findings of amygdala and parietal activation in the absence of awareness in adults. We found increased amygdala activation in TDCs relative to ASDs for subliminally presented fearful faces compared to subliminal houses (p<.001 uncorr., apriori).  Consistent with our previous findings, we also found increased parietal activation for face stimuli compared to house stimuli in TDCs (p<.001 uncorr., apriori), but not in ASD.  Unexpectedly, a whole brain analysis of TDCs compared to ASDs showed lingual gyrus activation for houses compared to control condition (random effects; p<.05 FWE corrected). The lingual landmark area is thought to respond to visual configurations that may have wayfinding value (Aguirre et al., 1998).

Conclusions: While results are preliminary in light of ongoing data collection, we find that TDCs demonstrate associated neural signatures for social and non-social objects (amygdala & parietal activation in response to faces and lingual landmark area in response to houses). We interpret these findings as the intact differentiation of object categories in typical children, even when these images are presented without awareness. The lack of distinct neural signatures to faces and houses in ASD prior to awareness suggests disturbance of adaptive processing of visual categories even prior to awareness.