Intra-individual variability identifies individuals vulnerable to contextually induced executive lapses

Abstract

Objective:

Contextual stressors, such as engagement in burdensome emotion regulation known as expressive suppression (ES), can result in transient but clinically meaningful decrement in performance on measures of executive functioning (EF). The goal of the present investigation was to examine whether intra-individual variability (IIV-I), which has been identified as an indicator of cognitive weakness, could serve as a marker of vulnerability to EF decrements due to both naturally-occurring and experimentally-manipulated ES.

Methods:

In Study 1, 180 cognitively healthy older adults completed the Push-Turn-Taptap (PTT) task to assess IIV-I, four Delis-Kaplan Executive Function System (D-KEFS) subtests to assess EF, and the Burden of State Emotion Regulation Questionnaire (B-SERQ) to assess naturally-occurring ES. In Study 2, a subset (n = 81) of participants underwent experimental manipulation to induce ES, followed by second administration of the D-KEFS to examine ES-induced decrements in EF.

Results:

In Study 1, hierarchical linear regression yielded a significant interaction between ES and IIV-I as predictors of EF performance, demonstrating that high ES was associated with low EF only among individuals with high IIV-I. In Study 2, repeated measures ANOVA demonstrated an interaction between time (pre- vs. post- manipulation), group (ES vs. control), and IIV-I (high vs. low), such that only individuals who exhibited high IIV-I were negatively impacted by the ES manipulation.

Conclusions:

IIV-I moderates the association between ES and EF, such that only individuals with high IIV-I exhibit vulnerability to the impact of ES. Thus, IIV-I may act as a marker of vulnerability to temporary EF depletion.

Keywords:

• Executive functioning
• intra-individual variability
• inconsistency
• expressive suppression
• emotion regulation
• contextual factors
• depletion
• cognitive aging

Acknowledgements

This research was supported by internal funding from the University of Utah College of Social and Behavioral Sciences and received no specific grant from any other funding agency, commercial, or not-for-profit sectors. Portions of this study were presented as an abstract at the meeting of the International Neuropsychological Society (2020).

Disclosure statement

This article was co-authored by Yana Suchy, who is the Editor-in-Chief of The Clinical Neuropsychologist (TCN). As a Taylor & Francis policy, and in compliance with publishing ethics, no TCN editors (EIC or Associate Editors) have access to any aspects of the peer review process pertaining to articles on which they are an author/co-author. For the purpose of this submission, Dr. Steven Paul Woods, a TCN Associate Editor, assumed 100% responsibility for the processing and editorial decisions regarding this manuscript.

Funding

The author(s) reported there is no funding associated with the work featured in this article.

Notes

1 Because IIV-I in this study was assessed using an experimental measure for which national norms are not available and because IIV-I is significantly correlated with age in this sample, it was crucial to avoid age-correcting the EF composite scores. Therefore, the D-KEFS raw scores in this study were converted to scaled scores using the normative reference group for adults aged 60-69 years. By doing so, D-KEFS scores could be standardized and combined into a single composite without correcting for age. The 60-69-year-old age band was selected because the scores within this age band encompassed the widest range of raw scores (as compared to other age bands) and would therefore have the highest probability of avoiding floor or ceiling effects (Delis et al., 2001). We used a similar procedure in other prior studies (Suchy, Brothers, et al., 2020; Suchy, Mullen, et al., 2020).

2 In our past research, we have consistently used the median rather than the mean as a measure of overall performance for the action planning latency as the mean variable is quite skewed, as would be expected with a response time variable (Kraybill et al., 2013; Kraybill & Suchy, 2011; Niermeyer et al., 2017; Suchy & Kraybill, 2007).

3 To aid in interpretation, recent ES and inconsistency were zero-centered (i.e., converted to z-scores).

4 In the later portion of the study, approximately one-third of the sample was pseudo-randomized to ensure matching of the groups on relevant characteristics (e.g., age, sex, education).