CHARACTERISTICS OF COGNITIVE CONTROL SPECIALIZATION IN HEALTHY AND PATIENT POPULATIONS

Abstract Background Cognitive control mechanisms enable an individual to regulate, coordinate, and sequence thoughts and actions to obtain desired outcomes. A theory of control specialization posits that proactive control is necessary for anticipatory planning and goal maintenance and recruits sustained lateral prefrontal activity, whereas reactive control, essential for adapting to transient changes, marshals a more extensive brain network (Braver, 2012). Increased task errors and reduced frontoparietal activity in proactive contexts is observed in severe psychopathology, including schizophrenia (Poppe et al., 2016), leading to the prediction that patients rely on reactive control more when performing such tasks. However, evidence of primate prefrontal ‘switch’ neurons, active during both proactive and reactive contexts, challenges the notion that cognitive control relies on discrete processing networks (Blackman et al., 2016). To examine this contradiction, we sought to characterize the distinctiveness between proactive and reactive control in healthy and patient populations using the Dot Pattern Expectancy Task (DPX). We also examined if a bias toward proactive or reactive control predicted behavioral metrics. Methods 44 individuals with schizophrenia (SZ) and 50 matched healthy controls (HC) completed 4 blocks of the DPX during a 3-Tesla fMRI scan (Poppe et al., 2016). Participants followed the ‘A-then-X’ rule, in which they pressed one button whenever an A cue followed an X probe, and pressed a different button for any other non-target stimulus sequence. We examined bilateral frontoparietal ROIs from the literature for evidence of cognitive control specialization as well as whole-brain analyses. Subsequent nonparametric tests and measures of neural response variation strengthened our interpretations. Participant d’-context (dependent on task accuracy) measured their tendency to engage in proactive control. Results Behavioral data revealed that HC participants showed a greater proclivity for proactive control than did their SZ counterparts. HC reaction time outpaced SZ reaction time in trials requiring successful marshalling of proactive control. Preliminary neuroimaging analyses suggest marginal between-group differences in control specialization. HC specialization appeared to be most apparent in diffuse frontal lateral regions, and bilateral posterior parietal cortex. Within the SZ group, specialization was most evident in bilateral posterior parietal cortex. Between-group control specialization differences were most apparent in right hemisphere frontal regions. Superior frontal gyrus and medial temporal lobe activity during proactive processes accounted for modest variance in d’-context. Discussion There were significant between-group differences in goal maintenance behavioral metrics such as reaction time and a tendency to engage in proactive control. Control specialization occurred more diffusely in controls compared to patient counterparts. However, activity in these regions had minimal ability to predict behavioral metrics. Overall, the relatively small size of control-specific areas compared to regions involved in dual processing offers support for the malleable nature of regions implicated in human cognitive control.