Robustness of the relationship between the heartbeat counting task and measures of anxiety, depression, and alexithymia: A behavioural study and meta-analysis

Interoception - the process of receiving information from within the body - has been proposed as a factor in different psychological phenomena, including anxiety, alexithymia, and depression. A popular approach for investigating this potential relationship has been to combine the heartbeat counting task (HCT) with questionnaires that aim to measure these phenomena. In particular, the State-Trait Anxiety Inventory (STAI-T), the Toronto Alexithymia Scale (TAS-20), and the Beck Depression Inventory-II (BDI-II) have been widely used. More recent work has, however, suggested that the initial findings of correlations between the HCT and these questionnaires are not robust. Potential reasons for such a lack of robustness can include reporting bias in the literature, variation arising from methodological factors, or the use of under-powered sample in studies. To investigate whether these factors may be present in the literature looking at the HCT and the questionnaires of interest, we combined a well-powered behavioural study with a meta-analysis. Together, these showed no association between the HCT and STAI-T or TAS-20 responses. A weak correlation between the HCT and BDI-II responses was observed in the meta-analysis ($r = -0.08$). A moderator effect of the type of instruction was seen on the relationship between HCT scores and STAI-T responses, along with a moderating effect of population type on the correlation between the HCT and TAS-20 responses. No evidence of publication bias was observed but there was a notable issue of low-power across the studies included in the meta-analysis. In summary, we emphasise the absence of a meaningful relationship between the HCT and commonly used measures of psychological phenomena of interest. Although the relevant literature may not suffer from publication bias, we identify methodological issues that may contribute to a lack of robustness in this area of study.