Attentional selection may be directed based on

Attentional selection may be directed based on the desires or goals of the individual (i.e. top-down control) or can be driven by salient information within the environment (i.e. bottom-up modulation). Adaptive allocation of attention rarely consists of exclusively top-down or bottom-up mechanisms; rather, successful attentional selection requires the integration of these two processes. For example, a child\'s attention may be directed at a person entering a room on the basis of bottom-up modulation of attention; however, the child\'s attention may be more likely to be captured by the new person if they are expecting guests or it is the time of day when their parents arrive to pick them up from school (i.e., based on the top-down expectations). Contingent attentional capture, as when a stimulus-driven shift of attention is contingent upon on a pre-existing top-down attentional setting, is one such example of a combination of top-down and bottom-up systems (Folk et al., 1992). Thus, this form of attentional orienting may provide insight into top-down and bottom-up processes, as well as the interaction between these neurocognitive networks. Corbetta and colleagues (2008; 2002) have proposed that two distinct networks underlie top-down and bottom-up modulation of attention; a bilateral dorsal frontal-parietal network, which includes frontal eye fields (FEF) and intraparietal sulci (IPS) and is responsible for voluntary control of attention, and a right-lateralized ventral frontal-parietal network, which includes ventral frontal retinoic acid receptor (VFC)/anterior insula (AI), middle frontal gyrus (MFG), and temporal parietal junction (TPJ) and is responsible for behavioral re-orienting of attention on the basis of bottom-up information. fMRI studies investigating the developmental differences in the activation (Konrad et al., 2005) and connectivity (Farrant and Uddin, 2015a) of these networks in typically developing (TD) children (7–12 years old) and adults have shown that these regions undergo continued maturation into adolescence and adulthood, with patterns of hyper- and hypo-connectivity varying according to age in children and adults with ASD (Farrant and Uddin, 2015b). Studies examining these networks in adults have demonstrated that behaviorally-relevant stimuli (e.g., a target/deviant in oddball paradigm) result in activation of ventral, and to a lesser extent dorsal, attentional networks (Kim, 2014), reflecting detection of salient, environmental changes. Previous functional magnetic resonance imaging (fMRI) studies have also demonstrated that contingent attentional capture by task-irrelevant stimuli results in increased activation of both networks and, further, results in enhanced activation of visual cortex for irrelevant information that shares task-relevant features (Serences et al., 2005). In their task, Serences et al. (2005) had participants view three streams of continuously changing letters of varied colors and instructed them to attend only to the central stream (see Fig. 1). The task was to respond only to red letters (which appeared infrequently) within the central stream. Letters in the to-be-ignored peripheral distractor streams most frequently appeared in gray; however, occasionally, peripheral distractors were either target-colored (red) or a unique, non-target color (green; a color not included in the central stream). Behavioral and neuroimaging results of their study showed evidence of attentional capture by the target-colored, but not the non-target-colored distractors (i.e., contingent attentional capture), and engagement of both dorsal and ventral attentional networks. Importantly, the task design permitted analysis of peripheral distractors in the absence of any target-related processing (isolated activation due solely to appearance of to-be-ignored information). In the present study, we employed a modified version of the rapid serial visual presentation (RSVP) paradigm by Serences et al. (2005) to investigate activation of dorsal and ventral attentional networks and capture-related activity in visual cortex in a group of children and adolescents with ASD. The goal of the current study was three-fold: first, we sought to investigate target-related processing (in the absence of colored distractors) in ASD. Based on previous fMRI and event-related potential (ERP) studies we hypothesized that individuals with ASD would show atypically reduced activation in regions of both dorsal and ventral attentional networks. Our second goal was to investigate attention capture by irrelevant information as a function of whether that information shares characteristics with attentional set. Given evidence of either over-focused or distractible states reviewed above, three potential outcomes were possible: (1) Over-focused attention would result in limited attentional capture by irrelevant distractors regardless of task-relevance (e.g., Liss et al., 2006); (2) alternatively, over-focused attention might engender enhanced capture of distractors that share a target-defining feature (i.e., enhanced top-down processing); (3) conversely, inability to filter irrelevant information could result in capture by irrelevant distractors regardless of task relevance (e.g., Murphy et al., 2014). Lastly, because links between non-social attentional impairments and social deficits in ASD have been reported (Fan et al., 2012; Keehn and Joseph, 2008; Keehn et al., 2010), we investigated the association between ASD symptomatology and behavioral and neural indices of target-related processing and attentional capture to further determine how attentional dysfunction may be linked to sociocommunicative impairments in children and adolescents with ASD.