Novel Assessment of the Impact of Irritability on Physiological and Psychological Frustration Responses in Adolescents

ABSTRACT

Objective

Irritability, typically defined as a proneness to anger, particularly in response to frustration, falls at the intersection of emotion and disruptive behavior. Despite well-defined translational models, there are few convergent findings regarding the pathophysiology of irritability. Most studies utilize computer-based tasks to examine neural responses to frustration, with little work examining stress-related responding to frustration in social contexts. The present study is the first to utilize the novel Frustration Social Stressor for Adolescents (FSS-A) to examine associations between adolescent irritability and psychological and physiological responses to frustration.

Method

The FSS-A was completed by a predominantly male, racially, ethnically, and socioeconomically diverse sample of 64 12- to 17-year-olds, who were originally recruited as children with varying levels of irritability. Current irritability was assessed using the Multidimensional Assessment Profiles-Temper Loss scale (MAP-TL-Youth). Adolescents rated state anger and anxiety before and after the FSS-A, and usable salivary cortisol data were collected from 43 participants.

Results

Higher MAP-TL-Youth scores were associated with greater increases in anger during the FSS-A, but not increases in anxiety, or alterations in cortisol. Pre-task state anger negatively predicted the slope of the rise in cortisol observed in anticipation of the FSS-A.

Conclusions

Results provide support for unique associations between adolescent irritability and anger during, and in anticipation of, frustrating social interactions. Such findings lay a foundation for future work aimed at informing physiological models and intervention targets.

Acknowledgments

We would like to thank Andres Salgado, Gabriela Marcelino, Chris Hall, David Rosen, Alexandra Prado, and Sydney Taylor for serving as confederates for the FSS-A task, as well as Robert Garvey, M.S., Kaley Davis, M.S., and Jamie Listokin, M.S. for their contributions as graduate assistants.

Disclosure statement

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

Supplementary data

Supplemental material for this article can be accessed online at https://doi.org/10.1080/15374416.2024.2301753.

Notes

¹ We did not adopt the popular model fit cutoff values for RMSEA, comparative fit index (CFI), and SRMR by Hu and Bentler (1999) because their guidelines may not be applicable to growth curve models which involve model constraints on the variables’ means. Further, SRMR and RMSEA demonstrate inconsistent performance in growth curve models (Wu & West, 2010) and Kenny et al. (2015) recommend not computing RMSEA when degrees of freedom of the model is small, as is the case for all of our models (Table 3). The SRMR reported by Mplus is different from that studied by Hu and Bentler (1999), as Mplus incorporates the model misfit of the variables’ means. To our knowledge, there are no validated cutoff values of SRMR in growth curve models (Pavlov et al., 2021). We did not report the CFI because the suggested baseline model for growth curve models (Widaman & Thompson, 2003) did not converge properly.

² We also explored a penalized spline growth model, which fits a smooth and nonlinear longitudinal trajectory while minimizing overparameterization (Suk et al., 2019). We tested 2 to 4 turning points, as well as linear and quadratic splines with the suggested knots by Suk et al. (2019). Results supported a quadratic and piecewise growth curve model at Sample 3 (T + 15) for the cortisol variable. Results are available upon request.

Additional information

Funding

This work was supported by the National Institutes of Health under Grant [R15MH115356-01] and Grant [R01MH091140-0181A].