Background: Historically, alexithymia has been characterized as a trait deficit specific to emotional processing. However, despite the fact that poor emotional self-insight is a central feature of alexithymia, extant work primarily focuses on decision accuracy (e.g., percent of emotional images correctly categorized) over metacognitive components of decision-making. Some research suggests alexithymia is associated with reduced confidence in emotional decision-making, but it is not clear to what extent this may reflect an accurate perception of skill versus a negative bias in self reflection. Moreover, it is not clear to what extent the relationships between alexithymia and any of these decision features are specific to emotional content or generalized across task domains (e.g., visual perception, social perception). To address these questions, we probed the relationship between alexithymia and decision accuracy, confidence, and metacognitive sensitivity (i.e., the alignment of confidence and accuracy) across visual, emotional, and social decision-making tasks in three samples.
Methods: We collected data from three samples: an in-person community sample (N = 123), an in-person student sample (N = 124) and an online student sample (N = 317). All participants completed the TAS-20 and three decision-making tasks indexing different task domains: coherent visual motion detection (Random Dot Kinematogram; RDK), emotional mentalizing (Revised Reading the Mind in the Eyes Test; RMET), and social decision-making (the Trust Game). After each trial of each task, participants were asked to rate how certain they felt that they answered correctly. We calculated metacognitive sensitivity scores using the area under the type-2 ROC curve for the RDK and RMET, estimating the extent to which an individuals’ confidence ratings aligned with their task accuracy. To examine the relationship between alexithymia and decision features, we conducted Spearman correlations between total TAS-20 scores and percent-correct and median confidence in all three decision-tasks.
Results: Alexithymia was associated with poorer accuracy in the RMET in-person students (ρ = -.26, p < .05) and poorer confidence in the RMET in the community sample (ρ = -.26, p < .05). However, these results did not replicate across different samples. Notably, our third and largest sample found no relationships between alexithymia and any of the three decision features. Preliminary analyses examining the relationship between alexithymia relative entropy (i.e., Kullback-Leibler divergence) in confidence ratings suggest convergent results. Exploratory regression models suggest that confidence ratings do not significantly change over the course of trials and alexithymia does not appear to alter this relationship.
Conclusions: Contrary to conventional reports on the topic, our results do not support the assumption that alexithymia is consistently associated with deficits in emotional mentalizing ability. Moreover, despite the explicit inclusion of poor emotional self-knowledge in the definition of alexithymia, alexithymia was not associated with either an emotion-specific or domain-general metacognitive deficit in our data. Future work seeking to elucidate the alexithymia or emotional awareness more broadly may benefit from examining more precise dimensions of the constructs and considering a wider variety of behavioral indices of metacognitive functioning.