Individual differences in affective flexibility predict future anxiety and worry

ABSTRACT

Deficits in cognitive flexibility have been associated with anxiety and worry, however few studies have assessed cognitive flexibility in the context of emotional stimuli (i.e. affective flexibility). The present study (n = 79) investigated whether individual differences in affective flexibility predict levels of trait anxiety and worry over a period of seven weeks. Affective flexibility was measured using a task-switching paradigm. Results showed that less efficient shifting of attention towards affective aspects of positive stimuli predicted higher anxiety over time. Additionally, more efficient shifting of attention away from affective towards non-affective aspects of negative stimuli predicted higher anxiety and worry over time. This latter finding may be understood by considering theoretical models and empirical evidence associating avoidance of negative information with increased anxiety. The effects were small and require replication in larger, representative samples, but they are an initial indication that anxiety may not be associated with general impairments in cognitive flexibility. Instead, our study emphasises the importance of breaking down cognitive flexibility into different components to investigate more nuanced relationships.

KEYWORDS:

• Cognitive flexibility
• affective flexibility
• anxiety
• worry

Authorship

All authors contributed to the study concept and design. Data collection, analysis and interpretation was performed by E. Twivy, under the supervision of M. Grol and E. Fox. E. Twivy and M. Grol drafted the paper, and E. Fox. provided critical reviews. All authors edited and approved the final version of the paper for submission.

Open practices statement

The data for this study are available on the Open Science Framework: https://osf.io/2vb8x/?view_only=88f9ea5a2cb741af9d1fdf5a65d177c1

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 An a priori power analysis for a hierarchical linear multiple regression, based on the R² increase when testing four predictors and total number of predictors is five (i.e. including baseline anxiety/worry), assuming a medium effect size f²= 0.15 and power of 0.8, shows the necessary sample size is 85. When testing two predictors (i.e. switch costs of interest) and the total number of predictors is three, the necessary sample size is 68.

2 We also administered the Connor-Davidson Resilience Scale and the Cognitive Emotion Regulation Questionnaire. The Hassles and Uplifts Scale was administered once a week in between the two test sessions. However, these will not be used to answer the current research questions.

3 Exploratory hierarchical regression analyses with accuracy-based switch costs showed that none of these switch costs significantly predicted anxiety or worry.

Additional information

Funding

This research was supported by a grant awarded to E. Fox from the FP7 Ideas: European Research Council under the European Union's Seventh Framework Programme (FP7/2007-2013) / ERC grant agreement no. [324176] for the CogBIAS project, and a studentship awarded to E. Twivy, funded by the Medical Research Council and Department of Experimental Psychology, University of Oxford. M. Grol is currently supported by a postdoctoral fellowship of the Research Foundation Flanders (FWO2017/PDO/201).