The relationship between heartbeat counting and heartbeat discrimination: A meta-analysis
Introduction
In recent years the importance of interoception, the perception of the body’s internal state (Craig, 2002, 2003, 2009), for health and higher-order cognition has begun to be appreciated (Barrett & Simmons, 2015; Khalsa et al., 2018; Murphy, Brewer, Catmur, & Bird, 2017). Indeed, numerous theoretical models posit a fundamental role for interoception in various aspects of health and cognition (Barrett & Simmons, 2015; Brewer, Cook, & Bird, 2016; Murphy et al., 2017; Paulus & Stein, 2006; Quattrocki & Friston, 2014). These models are supported by a growing body of evidence demonstrating links between interoception and fundamental cognitive abilities including learning and decision making (Werner, Jung, Duschek, & Schandry, 2009), emotional processing (Füstös, Gramann, Herbert, & Pollatos, 2013; Herbert, Pollatos, Flor, Enck, & Schandry, 2010; Schandry, 1981), and social cognition (Quattrocki & Friston, 2014; Seth, 2013). Furthermore, atypical interoception has also been observed across several mental health conditions, including Autism Spectrum Disorder (ASD; Garfinkel, Tiley et al., 2016), alexithymia (Brewer et al., 2016), depression (Harshaw, 2015; Pollatos, Traut-Mattausch, & Schandry, 2009), anxiety (Domschke, Stevens, Pfleiderer, & Gerlach, 2010; Pollatos et al., 2009), and eating disorders (Herbert & Pollatos, 2014; Klabunde, Acheson, Boutelle, Matthews, & Kaye, 2013; Pollatos et al., 2008) as well as physical health conditions such as obesity (Herbert & Pollatos, 2014) and diabetes (Pauli, Hartl, Marquardt, Stalmann, & Strian, 1991). Such evidence has led to suggestions that atypical interoception may represent a common risk factor for poor mental and physical health (Barrett & Simmons, 2015; Brewer et al., 2016; Murphy et al., 2017).
Increasing recognition of the importance of interoception for our understanding of pathology and cognition has prompted much research (Khalsa & Lapidus, 2016); however, progress in the field has been hampered by difficulties with the measurement of interoception (Brener & Ring, 2016; Murphy, Brewer, Hobson, Catmur, & Bird, 2018; Zamariola, Maurage, Luminet, & Corneille, 2018). Indeed, whilst there are many aspects of interoception that may be quantified (e.g., the perception of respiratory, gastric or urinary signals; Khalsa et al., 2018), most studies of interoception have utilised one of two measures of cardiac interoceptive accuracy, the heartbeat counting task (HCT; Schandry, 1981) or the heartbeat discrimination1 task (HDT; Katkin, Reed, & Deroo, 1983; Whitehead, Drescher, Heiman, & Blackwell, 1977). In the HCT, participants are asked to count the number of heartbeats they can feel during a series of time intervals (typically 3–6 intervals). Their response is compared to an objective record to determine accuracy. In the HDT, participants are required to determine whether an auditory or visual signal is presented synchronously or asynchronously with their heartbeat (typically across 15–60 trials). For the purposes of the present study it is relevant to note that the HDT can be administered in several variant forms (Brener & Ring, 2016), including the two-alternative forced choice procedure (2AFC; e.g., Whitehead et al., 1977), six-alternative forced choice (Brener-Kluvitse) procedure (6AFC; e.g., Brener & Kluvitse, 1988) and the method of constant stimuli (MCS; e.g., Brener, Liu, & Ring, 1993; Yates, Jones, Marie, & Hogben, 1985). Whilst all of the above HDT variants require synchronicity judgements, they differ in terms of the delays at which the signal is presented with respect to the heartbeat and the analysis method used to determine accuracy (although notably moderate correlations have been observed between these HDT variants; > r = .50; Brener et al., 1993). Importantly, despite the existence of these variants and other tasks of cardiac interoceptive accuracy (e.g., heartbeat tapping, adjustment methods and perturbation methods; Carroll & Whellock, 1980; Gannon, 1980; Khalsa, Rudrauf, Sandesara, Olshansky, & Tranel, 2009; McFarland, 1975), it is the HCT and the 2AFC HDT that are used most frequently, and interchangeably, as measures of cardiac interoceptive accuracy.
It is evident from the above descriptions that the HCT and HDT likely make different demands on cognitive processes. Indeed, whilst both presumably involve the perception of cardiac signals, the HCT requires sustained attention to heartbeat sensations over time whereas the HDT requires participants to integrate the cardiac signal with an external stimulus (Garfinkel, Seth, Barrett, Suzuki, & Critchley, 2015). Given these differences, it is perhaps unsurprising that several factors are thought to influence performance on the HCT and HDT selectively; for example, good performance on the HCT can be achieved through the use of non-interoceptive strategies. Indeed, better performance on the HCT has been associated with participants’ beliefs regarding their resting heart rate (Brener & Ring, 2016; Ring & Brener, 1996; Ring, Brener, Knapp, & Mailloux, 2015; Windmann, Schonecke, Fröhlig, & Maldener, 1999) and their time estimation abilities (Murphy et al., 2018), factors that are unrelated to performance on the HDT (e.g., Knoll & Hodapp, 1992; Phillips, Jones, Rieger, & Snell, 2003). These dissociations suggest that different abilities may be quantified by the HCT and HDT and question the validity of interoceptive accuracy scores obtained from the HCT (Desmedt, Luminet, & Corneille, 2018; Murphy et al., 2018; Zamariola et al., 2018; but see Ainley, Tsakiris, Pollatos, Schulz, & Herbert, 2020).
The suggestion that the tasks may index slightly different abilities is supported by the differential impact of pathology on task performance; it is not always the case that the HCT and HDT exhibit the same patterns across clinical groups. For example, Hina and Aspell (2019) reported that non-smokers performed better on the HCT compared to smokers, but this difference was not seen for the 2AFC auditory HDT. Similar dissociations have been observed with other populations such as individuals with ASD (Garfinkel, Tiley et al., 2016) and hypermobile individuals (Mallorquí-Bagué et al., 2014). Additionally, Rae, Larsson, Garfinkel, and Critchley (2019) reported a positive association between tic severity in Tourette syndrome and interoceptive accuracy as indexed by the 2AFC auditory HDT, but no such relationship was observed when interoceptive accuracy was indexed by the HCT. Such evidence again questions whether a common ability is quantified by these tasks of cardiac interoceptive accuracy and whether they can be used interchangeably, as one would expect a similar impact of pathology on task performance if the tasks index a common ability.
Despite indirect evidence suggestive of dissociations between performance on the HCT and HDT, studies directly comparing the two tasks are inconclusive regarding the presence or absence of a relationship; for example, early reports by Knoll and Hodapp (1992) suggest a moderate correlation (r = .59) between performance on the HCT and 2AFC auditory HDT. Similarly, other studies suggest a small but significant correlation between accuracy scores on the tasks (r = .36; Hart, McGowan, Minati, & Critchley, 2013). Such evidence of a small-to-moderate correlation between these measures is often used to justify the use of a single measure of cardiac interoceptive accuracy, as performance is presumed to generalise from one task to the other (e.g., Borhani, Ladavas, Fotopoulou, & Haggard, 2017; Herbert, Blechert, Hautzinger, Matthias, & Herbert, 2013; Pollatos, Traut-Mattausch, Schroeder, & Schandry, 2007; Scarpazza, Sellitto, & di Pellegrino, 2017; Werner et al., 2009). However, there are instances where performance on the HCT and HDT has not been found to correlate; for example, Forkmann et al. (2016) found no significant association between performance on the HCT and the 2AFC auditory HDT. This lack of an association was replicated by Schulz, Lass-Hennemann, Sutterlin, Schachinger, and Vogele (2013) who tested participants on the HCT and both the auditory and visual versions of the 2AFC HDT. Whilst a significant correlation was found between performance on the two versions of the HDT (r = .63; i.e. 39.7 % of variance in one task is explained by the other), no relationship was found between the HCT and either version of the HDT. Finally, a study by Ring and Brener (2018) which tested participants on the HCT and MCS auditory HDT also observed no significant association between performance on the two measures. It is clear that these inconsistent reports from single studies must be considered together before concluding whether there is a relationship between performance on the two tasks, and in turn whether they might index a common ability. Indeed, quantifying the relationship between these two tasks is important for determining whether the HCT and HDT can be used interchangeably as measures of cardiac interoceptive accuracy, and the generalisability of studies that have employed one task.
Thus far, we have focused on interoceptive accuracy, but there are other aspects of interoceptive ability that may be quantified using the HCT and HDT. In addition to accuracy, it is now common for studies to obtain confidence ratings during tasks of interoceptive accuracy in order to assess both one’s interoceptive sensibility (self-reported beliefs regarding interoceptive accuracy) and to calculate interoceptive awareness (a metacognitive measure reflecting the correspondence between interoceptive accuracy and interoceptive sensibility; Garfinkel et al., 2015; Murphy, Catmur, & Bird, 2019). For both tasks, interoceptive sensibility is calculated by averaging the confidence ratings obtained across trials. However, it is notable that there are differences in the assessment of interoceptive sensibility and awareness for the HDT and HCT; for example, 1) far fewer trials are used for the HCT (typically 3–6) compared to the HDT (typically 15–60) thus reducing the reliability of the HCT accuracy, sensibility and awareness indices, and 2) the analysis strategy for calculating interoceptive awareness differs for the HCT and HDT. Whilst HDT interoceptive awareness is usually calculated using Receiver Operating Characteristic (ROC) curves (but see Palser, Fotopoulou, Pellicano, and Kilner (2018) for an alternative method for calculating HDT interoceptive awareness), confidence-accuracy correlations are generally used to calculate interoceptive awareness for the HCT (but see Murphy et al. (2020) for an alternative scoring method for calculating HCT interoceptive awareness). In terms of the relationship between these aspects of interoception, confidence ratings for the HCT and HDT (indexing interoceptive sensibility) are often correlated with one another, but the strength of this association has been found to vary substantially across studies, with Forkmann et al. (2016) reporting a relatively low correlation (r = 0.348) and Garfinkel et al. (2015) reporting a much stronger correlation (r = 0.711). Conversely, awareness scores obtained using these two tasks have not been found to correlate (Forkmann et al., 2016; Garfinkel et al., 2015) and show different relationships across pathologies. For example, Ewing et al. (2017) found that interoceptive awareness on the HCT was predicted by an interaction between sleep effectiveness and mixed anxiety and depressive disorder, but no such relationship was observed for HDT interoceptive awareness. With increasing interest in these aspects of interoception (Forkmann et al., 2016; Garfinkel et al., 2015), understanding the generalisability of interoceptive sensibility and awareness scores calculated using the HCT and HDT is a priority.
It is clear from the above review that questions exist as to the relationship between the HCT and HDT, which has implications for whether they can be considered to be testing the same ability (or set of abilities). Lack of clarity regarding the relationship between these measures is potentially problematic for cases where only one task is utilised as a measure of cardiac interoception, or where both tasks are employed but show differential relationships with a third variable. As such, in this study we investigate the relationship between the HCT and HDT in order to clarify the extent to which using these measures interchangeably should be a concern. Specifically, evidence from studies that utilised both the HCT and HDT was collated to determine the relationships between accuracy, confidence and awareness scores obtained using the two different tasks. This was achieved by employing a meta-analytical strategy to obtain the pooled effect sizes of the reported correlations.
Section snippets
Search strategy
A systematic literature search was conducted in PubMed, Web of Science, PsycINFO and Medline. All searches were restricted to the year 1976 onwards, 2 years prior to the first description of the HCT (Dale & Anderson, 1978). All searches were conducted on the 28th October 2019. The following search was employed across the 4 search engines:
(“interoceptive sensitivity” OR “interoceptive accuracy” OR “heartbeat perception” OR “heartbeat interoception” OR “cardiac perception” OR “cardiac
Primary meta-analysis: accuracy
Using all data obtained from the 22 selected studies (23 correlation coefficients), we employed the above analysis to uncover the pooled effect size of the relationship between accuracy as measured by the HCT and HDT. A significant Q statistic (Q = 41.47, p = .007) and an I2 value of 47.0 % supported the use of a random-effects model meta-analysis. The meta-analysis identified a pooled effect size of 0.21 (p < .001). Thus, with an R2 value of 0.044, 4.4 % of the variance in accuracy on one
Discussion
This study aimed to investigate the relationship between the HCT and HDT with respect to interoceptive accuracy, confidence and awareness. Meta-analyses conducted for each of these dimensions of interoception revealed a small but significant correlation between HCT accuracy and HDT accuracy (4.4 % variance shared), a moderate significant correlation between HCT confidence and HDT confidence (36.0 % variance shared), and no significant correlation between HCT awareness and HDT awareness (0.8 %
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
LH was supported by a BBSRC PhD studentship provided by the Midlands Integrative Biosciences Training Partnership [grant reference: BB/M01116X/1]. JC was supported by the European Union’s Horizon 2020 Research and Innovation Programme under ERC-2017-STG Grant Agreement No 757583. GB was supported by the Baily Thomas Trust.
References (88)
- et al.
Comment on “Zamariola et al. (2018), interoceptive accuracy scores are problematic: Evidence from simple bivariate correlations”- the empirical data base, the conceptual reasoning and the analysis behind this statement are misconceived and do not support the authors’ conclusions
Biological Psychology
(2020) - et al.
"Lacking warmth": Alexithymia trait is related to warm-specific thermal somatosensory processing
Biological Psychology
(2017) - et al.
Heart rate perception and the voluntary control of heart rate
Biological Psychology
(1980) - et al.
Contribution of time estimation and knowledge to heartbeat counting task performance under original and adapted instructions
Biological Psychology
(2020) - et al.
The heartbeat counting task largely involves non-interoceptive processes: Evidence from both the original and an adapted counting task
Biological Psychology
(2018) - et al.
Interoceptive sensitivity in anxiety and anxiety disorders: An overview and integration of neurobiological findings
Clinical Psychology Review
(2010) - et al.
Sleep and the heart: Interoceptive differences linked to poor experiential sleep quality in anxiety and depression
Biological Psychology
(2017) - et al.
Interoceptive accuracy and body awareness - Temporal and longitudinal associations in a non-clinical sample
Physiology & Behavior
(2018) - et al.
Making sense of what you sense: Disentangling interoceptive awareness, sensibility and accuracy
International Journal of Psychophysiology
(2016) - et al.
Knowing your own heart: distinguishing interoceptive accuracy from interoceptive awareness
Biological Psychology
(2015)