No consistent correlation between baseline pupil diameter and cognitive abilities after controlling for confounds—A comment on Tsukahara and Engle (2021)

No consistent correlation between baseline pupil diameter and cognitive abilities after controlling for confounds—A comment on Tsukahara and Engle (2021). Nash Unsworth, Ashley L. Miller, Matthew K. Robison. Cognition, Volume 215, October 2021, 104825. https://doi.org/10.1016/j.cognition.2021.104825

There has been a recent surge of studies examining whether variation  in baseline pupil diameter is related to various cognitive abilities such as working memory capacity (WMC), fluid intelligence (Gf), and attention control (AC) to name a few (e.g., Aminihajibashi, Hagen, Foldal, Laeng, & Espeseth, 2019; Bornemann et al., 2010; Heitz, Schrock, Payne, & Engle, 2008; Ralph, Gibson, & Gondoli, 2020; Robison & Brewer, 2020, 2021, in press; Sibley, Foroughi, Brown, & Coyne, 2018; Tsukahara, Harrison, & Engle, 2016; Tsukahara & Engle, 2021; Unsworth & Robison, 2015, 2017a; Unsworth, Miller, & Robison, 2021; Unsworth, Robison, & Miller, 2019; van der Meer et al., 2010). These studies are based, in part, on the idea that baseline pupil diameter is related to functioning of the locus coeruleus-norepinephrine system, which is thought to be important for regulating arousal and alertness (AstonJones & Cohen, 2005; Gilzenrat, Nieuwenhuis, Jepma, & Cohen, 2010; Joshi, Li, Kalwani, & Gold, 2016; Szabadi, 2013; Unsworth & Robison, 2017b). Despite a number of studies examining potential relations, it is still unclear whether baseline pupil diameter is related to cognitive abilities. That is, many prior studies find near zero correlations between baseline pupil diameter and cognitive abilities (e.g., Aminihajibashi et al., 2019; Ralph et al., 2020; Robison & Brewer, 2020, 2021, in press; Sibley et al., 2018; Unsworth & Robison, 2015, 2017a; Unsworth et al., 2019; see Unsworth et al., 2021 for a meta-analysis of the relation between WMC and baseline pupil diameter), whereas some studies do find a relation (e.g., Bornemann et al., 2010; Heitz et al., 2008; Tsukahara et al., 2016; Tsukahara & Engle, 2021; van der Meer et al., 2010). In a recent attempt to examine discrepancies across studies, Tsukahara and Engle (2021) suggested that differences in luminance (both overall room lighting and screen brightness) influence the correlations between baseline pupil diameter and cognitive abilities such that the correlations seem to arise under dark conditions allowing for more variability between participants. Thus, Tsukahara and Engle (2021) concluded that baseline pupil diameter is related to cognitive abilities under proper lighting conditions. 

While the Tsukahara and Engle (2021) results are interesting and important in terms of providing information on how luminance can potentially impact correlations with pupil diameter, it is not clear how robust these results are and whether confounding variables account for the relations. In particular, in a prior study Tsukahara et al. (2016) noted that it was important to account for possible confounding variables such as age and race/ethnicity when examining correlations between baseline pupil dimeter and cognitive abilities. Tsukahara et al. found that there were race/ethnicity and age differences in baseline pupil diameter. 

Importantly, Gf still correlated with baseline pupil diameter after controlling for these confounding variables (along with others such as nicotine use, medication use, and whether the participant was a college student). Tsukahara et al. (2016) concluded that these analyses provided strong evidence suggesting that the relation between Gf and baseline pupil diameter (note that they did not report the corresponding analyses for WMC) was not due to confounding variables.