Anorexia nervosa as a model for increased volitional control: temporal dynamics, intrinsic costs and alternative explanations

A major goal of the CRC-940 is to elucidate the neural mechanisms underlying dysfunctional volitional control as expressed in mental and neurodegenerative disorders. Suggesting that anorexia nervosa (AN) might serve as a model for “too much” cognitive control, patients typically ) i) have an elevated willingness to forgo primary rewards (e.g. food), ii) exert an abnormal amount of effort to avoid weight gain (e.g. calorie counting, excessive exercise) and iii) are typically characterized by performance-driven, perfectionistic personality traits. Results from our experiments in the first CRC-940 funding phase largely support this hypothesis. First, using fMRI we found tonically-elevated activation of a region of dorsolateral prefrontal cortex associated with cognitive control in a sample of former patients recovered from AN (recAN) during a task designed to investigate reward-related processing (Ehrlich et al., 2015). Second, indicative of increased neural efficiency, fMRI data obtained during an intertemporal choice paradigm showed decreased response in the frontoparietal control network (FPN; and notably ACC) for difficult decisions in acutely ill patients (acAN) relative to healthy controls (HC) (King et al., submitted), despite generally normal (but faster) reward-based decision making (Ritschel et al., 2015). Third, in an associated project, we found increased functional connectivity in resting state fMRI data in AN participants in the FPN which correlated with the personality dimension “persistence” (Boehm et al., 2014). Last but not least, results of our ongoing ecological momentary assessment (EMA) study illustrate the everyday “costs” of increased cognitive control: acAN patients invest three times as much time as HC ruminating – particularly about body shape and food intake (Seidel et al., submitted) which was correlated with decreased striatal responses following volitional emotion regulation. These results are encouraging, but we have yet to develop a fundamental understanding of the specific mechanisms that may underlie increased cognitive control AN and why some sub-processes seem to be more affected than others.
To gain new insight, one aim for the second funding phase is to quantify the temporal dynamics of cognitive control and reward-based decision making with higher ecological validity. Specifically, we intend to use delay discounting tasks (Scherbaum, Dshemuchadse, Leiberg, & Goschke, 2013), that track responding (mouse movements) continuously and deliver rewards in real-time in a virtual world - video game like environment. Importantly, this procedure allows for quantification of “conflicts” in decision making, detection of habitual response tendencies and computational modeling of the temporal evolution of decisions (jointly with project A8). A second goal is to test the hypothesis of decreased effort discounting in AN, that is – an abnormally increased willingness to engage in effortful tasks. To this end, we will modify an influential cognitive effort discounting paradigm (Westbrook, Kester, & Braver, 2013) for event-related fMRI which will enable us to disentangle neural systems representing reward anticipation and cognitive effort, i.e. intrinsic costs. Analyses will focus on group differences in cognitive effort discounting parameters, brain reward as well as cognitive control systems associated with valuation, and relationships with self-reported “need for cognition”. A third objective is to test an alternative explanation for the excessive cognitive control in AN – that AN is characterized by extraordinarily rapid formation and/or persistence of habits. To this end, we will develop two versions of an event-related fMRI instruction-based stimulus-response-outcome learning paradigm (in collaboration with A2) in which goal-directed behaviors stand in conflict with competing habitual actions (that are induced by reward and punishment) and must be resolved on a trial-by-trial basis. Furthermore, we plan to extend our investigation of an alternative account of elevated self-control by using EMA to measure habit strength, habit frequency and self-control conflicts in everyday life and evaluate associations not only with context variables using hierarchical linear modelling, but also with the meta-control parameters and neural correlates of dynamic decision making, effort discounting and habit formation as measured in the above described experiments. Finally, we plan to extend our investigation of the potential costs of increased cognitive control by assessing clinical outcome (after 3 month and one year). As in the first funding phase, all assessments in acAN (including the acquisition and analyses of resting state functional connectivity data) will be followed-up after treatment and compared to recovered patients and HC to dissociate state from trait factors. Together, the proposed projects promise to add important insight into the mechanisms underlying excessive cognitive control in AN by complementing our previous work with a range of novel paradigms that target key clinical features of disorder and allow for exploration of alternative hypotheses.

Project Members

Principal Investigators


News Project C3

no news in this list.