While the conceptual roots of interoception may be traced back to William James over a century ago (James, 1994), there has been a renaissance of interest in this topic within cognitive neuroscience in recent decades (Chen et al., 2021; Khalsa and Lapidus, 2016). Interoception refers to the perception of the body's internal milieu (mechanical condition, cellular integrity, and inflammatory status, etc.) and visceral organs (gastric, respiratory, and cardiac signals, etc.) (Craig, 2002, 2003; Critchley and Harrison, 2013). These internal signals are generally transmitted to the central nervous system via neural and humoral pathways, with the insula cortex being the brain region most extensively studied in relation to interoception (Critchley and Harrison, 2013). Such complex processes underpin homeostasis, ensuring that physiological states remain within reasonable bounds to sustain life (Critchley and Harrison, 2013; Wilfrid, 2006). Garfinkel and colleagues conceptualized interoception as consisting of at least three interrelated but distinct components: interoceptive accuracy (IAcc), interoceptive sensitivity, and interoceptive awareness (Garfinkel and Critchley, 2013; Garfinkel et al., 2015). IAcc represents an individual's ability to accurately detect internal body signals, typically assessed using behavioral tasks, such as the Heartbeat Counting Task (HCT) and the Heartbeat Detection Task (HDT). Interoceptive sensitivity reflects an individual's self-perceived acuity regarding internal physiological states and is usually measured by self-report questionnaires. interoceptive awareness is characterized by an individual's metacognitive insight into their internal feelings, representing the metacognitive level of interoception. IAcc is a critical dimension that, along with interoceptive sensitivity and awareness, constitutes the multidimensional framework of interoception (Garfinkel et al., 2015).

Interest in interoception has been driven by a variety of empirical and theoretical work suggesting that it plays an important role in both physical health conditions and the etiology of mental illness (Critchley and Garfinkel, 2017; Khalsa et al., 2018). Specifically, research has shown that IAcc is closely related to emotion regulation. Higher IAcc is generally associated with enhanced negative emotion regulation, such as the ability to manage emotions more effectively during stress (Fustos et al., 2013; Pollatos et al., 2015). Furthermore, IAcc is associated with vagus nerve activity, which plays an important role in emotion regulation (Pinna and Edwards, 2020). Pathologically, the relationship between mental illness and interoception has been widely investigated, with depression frequently cited as a key psychiatric condition linked to interoceptive processes (Khalsa and Lapidus, 2016). Clinical observations provide compelling support for this link. Individuals with depression often experience somatic symptoms such as pain, nausea, constipation, palpitations, shortness of breath, and dizziness (Eggart et al., 2019). Consistent with this, depression is notably prevalent in chronic medical and functional somatic conditions, and this clinical overlap is substantial (Zhu et al., 2024), as evidenced by a meta-analysis reporting a pooled prevalence of depression of approximately 60.8% in patients with chronic pain (Rayner et al., 2016). Together, these clinical findings suggest a robust association between depression and interoceptive processes. Theoretical frameworks offer a mechanistic explanation for this association. The interoceptive predictive coding model posits that the brain continuously generates top-down predictions about expected bodily states (e.g., heartbeat rhythm) based on prior experience (Paulus and Stein, 2010). Interoception relies on minimizing the prediction error—the discrepancy between these top-down predictions and actual bottom-up bodily signals (Paulus and Stein, 2010). For individuals with depression, this predictive mechanism may be impaired: their brains may fail to update predictions in response to real bodily inputs, or may overprioritize negative expectations of bodily states. This persistent mismatch between predicted and actual interoceptive signals is hypothesized to contribute to the development and maintenance of depressive symptoms, including somatic complaints and emotional dysregulation.

Although a solid theoretical basis exists to expect that depression is related to IAcc, current empirical evidence is equivocal. Some findings indicate that IAcc is decreased in patients with depression (Furman et al., 2013); however, Dunn et al. (2007) reported that cardiac IAcc decreased in patients with moderate depression but increased in patients with severe depression. Meanwhile, other studies have found no significant differences (Blickle et al., 2024; Schmitz et al., 2023). Given the inconsistencies and the significant global burden of depression, it is crucial to determine whether there is a correlation between depressive symptoms and IAcc and, if so, whether there are moderators that could explain the current mixed results. Such an understanding is needed to establish a clinical correlation between depressive symptoms and IAcc, which in turn may provide new opportunities for clinical intervention in depressive disorders from an interoceptive perspective.

Previous research has suggested several factors that may influence the association between depressive symptoms and IAcc. In the HCT, participants are asked to count their heartbeats over a series of intervals, which is then compared to an objective estimate of the actual heartbeats to determine IAcc (Schandry, 1981). Evidence indicates that HCT performance is better in individuals without “no guessing” instructions than in those with such instructions (Smith et al., 2021), implying that a stronger relationship might be observed in participants who did not receive “no guessing” instructions. Similarly, whether studies utilize a clinical diagnosis of depression and the measurement of depressive symptoms may also influence the relationship between depression and interoception accuracy (Dunn et al., 2007; Ricciardi et al., 2016b). In addition, since anxiety or other clinical conditions are related to both depression and interoception (Dunn et al., 2010; Pollatos et al., 2009), whether other clinical diseases are included may also affect the relationship.

Demographic characteristics of participants may also serve as potential confounders, given that both IAcc and depression present different features across various age groups (Pfeifer and Cawkwell, 2025; Yang et al., 2024). Research has shown that adolescence is a period of high incidence of depression (Yang et al., 2024), while interoceptive sensations capacity, which refers to the ability to detect and process internal bodily signals such as cardiac and respiratory cues, tends to decline beginning as early as during this stage (Pfeifer and Cawkwell, 2025). Consequently, a stronger relationship between interoception and depressive symptoms may be observed in adults. Additionally, considering sex differences in interoception, with men demonstrating greater accuracy in cardiac tasks (Prentice and Murphy, 2022), the gender ratio among participants may also serve as a moderating factor. Finally, concerns regarding the validity of the HCT have prompted a growing focus on alternative measures of interoception (Murphy et al., 2018). A common alternative is the HDT, in which participants are asked to judge whether auditory or visual stimuli are synchronized with their heartbeats (Garfinkel et al., 2015). Some studies have shown that the association between depression and cardiac IAcc varies between the HCT and the HDT (Michal et al., 2014). Considering these inconsistencies and the absence of evidence with correspondence between the two tasks, the outstanding question of whether the relationship between depressive symptoms and interoception differs depending on the task employed remains unresolved.

Given the ambiguous evidence on the correlation between depressive symptoms and IAcc, coupled with a lack of quantitative reviews, our study aims to quantify the association between depressive symptoms and cardiac IAcc, and to assess whether potential moderators influence this relationship. In terms of moderators, we focused on assessing the influence of the HCT instructions used, depression assessment methods, study design, demographic characteristics (age and sex), whether participants with depression were included, whether participants with anxiety were included, whether participants with any other diagnoses were included, and whether participants with physical/somatic diagnoses versus those with only psychiatric diagnoses were included on the correlation between depressive symptoms and cardiac IAcc.