Task-switching inefficiencies in currently ill, but not remitted anorexia nervosa

Abstract

Objective: Models of anorexia nervosa (AN) posit that set-shifting deficits may contribute to behavioral inflexibility and extreme dietary restriction. Findings from neurocognitive studies of set-shifting in AN have been somewhat mixed, perhaps due to the use of tasks that cannot distinguish shifting from other processes (i.e., learning). To more precisely characterize cognitive flexibility and selectively assess this process independent of rule learning and feedback sensitivity, we examined task-switching ability in AN.

Method: Women ill with AN, subthreshold AN or atypical AN (IAN; n = 40), women remitted from AN (RAN; n = 24), and age-matched healthy control women (n = 42) completed a computerized cued color-shape task-switching paradigm. Groups were compared on mix costs (reflecting global cognitive control) and switch costs (reflecting transient cognitive control).

Results: Although mix costs were equivalent across groups, switch costs were more pronounced in the IAN group, as indicated by a group-by-trial type interaction for reaction times on stay and switch trials.

Discussion: Findings indicate that IAN, but not RAN, have difficulty flexibly switching between cognitive task sets, and suggest that prior findings of set-shifting deficits in AN may reflect difficulty with cognitive flexibility independent of learning deficits. As such, task-switching may represent a promising adjunctive treatment target.

Keywords: anorexia nervosa; cognitive flexibility; executive function; task switching.

Figure 1.

A) Cued Color-Shape Switching Task. Participants completed a cued color-shape switching task (Weissberger et al., 2012) comprised of two single-task blocks, in which the task remained the same across all trials (e.g., indicate color of the stimuli, with order of color and shape tasks counterbalanced across subjects), followed by four mixed-task blocks, in which half of the trials were sequential trials of the same type and did not require task-switching (“stay trials”), and half were sequential trials of different types (e.g., color followed by shape; “switch trials”), followed by two more single-task blocks (with color and shape presented in the opposite order of the first two blocks). In cued task-switching paradigms, switching costs capture the change in reaction time (RT) and accuracy from task-repetition (‘stay’) trials to task-switch (‘switch’) trials in the context of task blocks that involve switching between tasks. These costs are thought to reflect the influence of the increased cognitive demand of immediate, “transient,” or “local” task changes from not switching to switching. Mixing costs capture the change in RT and accuracy when comparing two types of task repetition trials--specifically, when comparing task repetition trials from task blocks that do not involve task switching (‘single’ trials) to task repetition trials from task blocks that involve task switching (‘stay’ trials). In mixing costs, the only difference between the two trial types is contextual. As a result, these costs are thought to reflect more “global” aspects of control, and have been shown specifically to represent cognitive conflict-related processing related to which task rule is required. B) Color-Shape Mix and Switch Costs across Groups. A group x trial-type interaction was detected only for switch costs analyses (p=0.022), and indicated that the IAN group showed the most exaggerated time (RT) difference between stay and switch trials.