Introduction

Interoception is understood as the sensing and representation of signals concerning the internal state of the body [1,2]. There is a growing and ongoing interest in this construct because of the evidence that different mental disorders show impaired interoceptive processes and that mental processes are embodied (e.g., [3–8]). Consequently, an interplay between the mind and physiological states can be assumed.

Moreover, different research findings showed that an increased interoception is associated with an enhanced emotion perception, processing and regulation (e.g., [9–16]). These results are also supported by different emotional theories, indicating that the perception of internal bodily signals is essential for the development of emotions [17–19]. In concordance with these results and theories, neurological studies found similar patterns of brain activity during interoception and emotion elicitation, especially in the anterior insula [20–24]. To conclude, the perception of internal bodily signals, namely interoception, is central to the processing of emotions.

Based on the results of a disturbed interoception in mental disorders and the involvement of the insular cortex in the pathophysiology of obsessive-compulsive disorders (OCD) it is necessary to investigate the relationship between interoception and OCD. Its cardinal features are obsessions, i.e. unwanted and distressing thoughts or urges, and compulsions, i.e. repetitive behaviors that the person feels compelled to perform, typically with a desire to resist [25]. OCD symptoms can include excessive washing, constant checking (e.g., of door locks or electrical appliances), and recurrent, distressing thoughts, for example, of harming loved ones.

Song et al. [26] examined whether the volume of insular cortices of patients with OCD differed in regards to several symptom dimensions. They reported enlarged anterior and reduced posterior insular cortices in patients with OCD. The insular volumetric alterations were more significant in patients with OCD with predominant checking rather than cleaning symptoms when compared to healthy controls. Importantly, the authors suggested that insular morphometric alterations may influence the modulation of interoceptive processing in patients with OCD. So far, there is only one study investigating the relationship between interoception and OCD [27]. Authors reported that patients with OCD were better in interoceptive accuracy (IAcc) as measured by the tapping accuracy task, while less confident and aware of their interoceptive abilities. The authors interpreted their results as evidence that overactive monitoring in OCD extends to the sensing of internal bodily signals.

In contrast, we hypothesized that patients with OCD show a decreased IAcc level in comparison to matched healthy controls. This assumption is based on the theoretical ideas of predictive coding in the context of interoceptive processing (see, e.g. [28–30]). The idea is that people make individual perceptions based on their own working models, so-called priors. In this context, interoceptive signal processing provides an important source of information regarding the evaluation of such priors [28]. The logic of these models is that incoming sensory data are compared with internal models, i.e., the brain’s or “top-down” “predictions” about the body and the environmental causes that affect the organism, and thus can lead to “bottom-up” prediction errors (PE) that finally have to be minimized. Using a predictive coding formulation, the ability of interoceptive directed predictions to suppress otherwise inappropriate interoceptive PE allows the individual to infer the autonomic and emotional state of another. Having in mind typical OCD symptoms, patients with OCD regularly realize upcoming thoughts or urges to conduct a specific behavior (e.g., washing hands). When these thoughts are related to compulsive behavior they usually also experience a desire to resist such internal thoughts and urges, associated with anxiety or concern what could happen if they do not engage in compulsive behavior. Interoceptive signals are therefore a source of mistrust and anxiety, resulting in priors or beliefs that determine information from the body as not reliable and thus inadequate for reducing the PE. One possible solution to reduce bodily PE is to rely on other modalities like the visual channel as a compensatory strategy. This might contribute to a greater dependency on input from the external world to understand the causes of sensations evoked by emotionally salient stimuli which contribute to deviations in affective experience.

Former research provides some evidence for the assumption above. For example, empirical data report that patients with OCD show tendencies to doubt their feelings and internal states (e.g., [31,32]). More specific, Shapiro [32] described that patients lost the ability to access internal states and instead use external indicators (e.g., biofeedback) to make conclusions about these states. For example, when patients with OCD had to answer the question whether they had fallen in love with another person, they did not trust their feelings. Instead, they checked external indicators (such as general rules or norms), which represent the feeling of falling in love [32]. Consequently, Shapiro [32] would argue, that patients with OCD fall in love when the partner shows all the right attributes but not because of a feeling.

That these external cues or interpretation are important for patients with OCD is supported by the research of Lazarov and colleagues [33,34]. Based on the theory of Liberman and Dar [35], they assume that patients with OCD display a decreased sense of subjective confidence and this leads to a higher focus on external “proxies” (i.e., opportunities to prove internal states) to compensate their deficits regarding inner subjective experience [34]. Results indicate that patients with OCD perform poorly on tasks with a focus on inner subjective feelings or states, in comparison to healthy controls. However, if they have the opportunity to use external proxies (e.g., biofeedback), they improve their task performance compared to healthy controls. Lazarov and co-workers [33] observed similar results with a non-clinical sample differentiating between low and high OCD tendencies. Furthermore, research showed that participants with high OCD tendencies relied on false feedback from biofeedback instead of their own internal state of relaxation [33]. This tendency could also be confirmed in the clinical sample [34]. We therefore follow that patients with OCD exhibit disturbances in the complex integration of information from interoceptive and exteroceptive channels resulting in severe deviations of affective experience of oneself and others. The perception of interoceptive signals is assumed to be reduced. One of the reason could be that typical OCD symptoms such as the occurrence of intrusive thoughts as well as the distrust of patient’s attention, perception and senses might lead to a distraction from heart rate counting or other mindfulness processes with a focus on internal bodily signals [36–40].

To sum up, interoceptive alternations might affect priors that determine information from the body as not reliable and thus are deemed inadequate for updating the prior and reducing the predictive error. As suggested by Lazarov and colleagues [33,34], one possible solution to reduce bodily predictive error from the view of a patient with OCD is to rely on other external modalities such as biofeedback as a compensatory strategy. Based on these findings, we will first investigate our assumption that IAcc, which is defined as the objective performance marker for interoception, is diminished in OCD. It should be noted that this hypothesis is not in line with the data of Yoris et al. [27]. Second and not investigated until now, we will examine the effect of cognitive-behavioral therapy for OCD on IAcc. Based on the assumed positive effects of cognitive-behavioral therapy, we hypothesized that patients with OCD would improve their IAcc during the time course of therapy compared to healthy controls. We also assessed OCD-, depression- and anxiety symptoms to identify possible working mechanism of the therapy, as well as potential factors influencing IAcc throughout the course of therapy.

Materials and methods

Results

All descriptive data of IAcc and OCD-, depression- and anxiety symptoms separated by patients and healthy controls are represented in Table 1. As shown in this table, OCD-, depression- and anxiety symptoms decrease during the time course of cognitive-behavioral therapy in the OCD sample. Furthermore, there is a descriptive increase in IAcc over time for patients with OCD. Moreover, patients have a decreased IAcc and more depressive- and anxiety symptoms in comparison to the healthy controls. These differences are significant for most group comparison at the different measurement points. Inferential statistic results will be described in more detail in the following sections.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive data for OCD-, depression- and anxiety symptoms and interoceptive accuracy contrasting patients with OCD and healthy controls.

https://doi.org/10.1371/journal.pone.0217237.t001

Interoceptive accuracy (heartbeat perception task)

The mean and standard deviation obtained heartbeat perception scores for both groups are summarized in Fig 1 and Table 1 depicting T1, T2 and T3. We observed a significant main effect of Group (F(1, 50) = 10.94; p = .002; η² = .17; ε = .87), indicating that patients with OCD had a lower IAcc (average score for all three measurements 0.63) at all three measurement points as compared to healthy controls (average score for all three measurements 0.74). No other main or interaction effect occurred (Time: F(2,100) = 1.64; p = .200; Time x Group: F(2,100) = 0.41; p = .668). It should be noted, that there was a descriptive higher improvement of IAcc for the OCD group (mean T1 = .60 (SD = .17), mean T2 = .62 (SD = .15), mean T3 = .66 (SD = .17)) compared to the healthy controls (mean T1 = .74 (SD = .16), mean T2 = .74 (SD = .19), mean T3 = .76 (SD = .16)). Similar to the results before, we did not find any influence of medication (F(1, 23) = .35; p = .560) or comorbidities (F(1, 23) = .07; p = .796) for the OCD group. Moreover, we integrated depression and anxiety symptoms as co-variates, which did not change the main effect of Time F(2,96) = .13; p = .882. Further, it could be shown that there is no influence of depression (F(1, 48) = 1.26; p = .268) nor anxiety (F(1, 48) = .51; p = .480) regarding the results in both groups.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Mean and standard deviation of interoceptive accuracy in patients with OCD compared to healthy controls in the time course of therapy (p < .01 = **).

https://doi.org/10.1371/journal.pone.0217237.g001

Discussion

To the best of our knowledge, these are the first empirical data demonstrating that IAcc is attenuated in OCD and that these interoceptive deficits do not change in the time course of cognitive-behavioral therapy. Nonetheless, there were definite improvements in OCD-, depression- and anxiety symptoms for patients with OCD. Additionally, we also showed negative associations between IAcc and OCD-, depression- and anxiety symptoms.

The idea that interoceptive processes might be affected in OCD was proposed mainly based on research concerning brain activities in the insula as well as the dysbalance of internal and external channels in patients with OCD. Song and co-workers [26] demonstrated altered insula activation patterns in OCD. Supporting the authors’ hypotheses that insular morphometric alterations may influence the modulation of interoceptive processing in patients with OCD, we could demonstrate that IAcc is reduced in patients with OCD. Furthermore, prior studies observed that patients with OCD have tendencies to not trusting in their internal states and feelings (e.g., [32,31]). Instead, patients seem to compare their assumptions of internal states with external indicators like general rules or norms. This postulation could be supported by a study of Lazarov and colleagues [33,34], showing that patients with OCD improved their performance when they had the opportunity to use biofeedback to prove their internal states. Conversely, Yoris and co-workers [27] demonstrated that patients with OCD show an increased IAcc in comparison healthy controls and patients with panic disorders. One reason for different results could be that they used the tapping accuracy task of Canales-Johnson et al. [55] instead of the heartbeat perception task [45]. For the tapping task, participants were instructed to firstly tap with their dominant-hand in synchrony with an audio-recorded heartbeat and secondly to their own heartbeat. For the IAcc score, time differences between the R-Peaks and the motor-taps incorrect answers were calculated. A smaller difference between both indicates a better IAcc. In their discussion, Yoris et al. [27] suggest replicating their results with non-motor interoceptive tasks such as the heartbeat perception task. We used the heartbeat perception task in our study. In comparison to tapping with the hand to different heartbeats in the tapping accuracy task, participants were instructed to focus on their heartbeat and to silently count it over three different time intervals. A smaller difference between recorded heartbeats via ECG and counted heartbeats indicate a better IAcc. It should be noted that there are different focuses in the tapping and heartbeat perception task. Whereas the tapping accuracy task includes the focus on interoceptive and exteroceptive signals (heartbeat and tapping), the heartbeat perception task only concentrates on interoceptive signals. To summarize, we could not show similar results in the heartbeat perception task, but in our opinion, these are two different measurements of IAcc, which make the comparison of the findings of the two studies difficult.

The observed interoceptive alternations did not change in the time course of cognitive-behavioral therapy in patients with OCD. So far, there are no other studies that also examine changes in IAcc during the time course of therapy. The present findings are of high importance for the pathophysiology of OCD. A reduced IAcc might constitute a risk factor for a more stressful everyday life as abilities to regulate one’s emotions are needed and are associated with more positive well-being [56]. Another point is that this form of cognitive-behavioral therapy does not focus on the perception of internal states. In a study of patients with anorexia nervosa, also only a trend regarding the improvement in IAcc was observed in the same clinic with the same therapy [57]. Consequently, it could be useful to develop an intervention focusing only on internal bodily signals and the self (e.g., looking in the mirror). There is evidence for positive effects on interoceptive ability when applying this kind of methods (e.g., [58,59]). Based on the findings of Lazarov and colleagues [33,34], another interesting point for future studies would be the inclusion of a biofeedback intervention with a reduction of the biofeedback over time to reduce the reliance on external stimuli’s and to increase the focus on internal signals step by step.

Different strengths and limitations of this study should be noted. An advantage of the study is that participants were matched according to age, sex, and educational background. Additionally, this is the first study investigating IAcc during the time course of cognitive-behavioral therapy. Potential limitations of this study refer to the fact that we used an OCD sample that was heterogeneous in terms of symptoms and comorbidities, as well as medication use. For example, comorbidities like depression, which highly occurs in the present sample as comorbidity, could be an influencing factor on observed results. This assumption is based on different studies, indicating an impaired IAcc in depressive participants (e.g., [3,60]). Though we addressed this issue and found no influence, the effects of medication on autonomic measures cannot be ruled out when using only small numbers of patients in each subgroup. Additional analyses with the inclusion of medication and comorbidities as co-variates showed no influencing effect regarding the reported results with OCD-, anxiety- or depressive symptoms nor IAcc in the OCD sample. Nevertheless, these results should be interpreted with caution due to the small sample size of OCD patients. Furthermore, Fischer et al. [58] found that a body scan training improved IAcc. Consequently, mindfulness should also be included in future studies as another confounding variable.

Conclusions

To conclude, our study highlighted a decreased IAcc in patients with OCD and the observation that these deficits do not seem to improve during the time course of cognitive-behavioral therapy. Future studies should use longer follow-ups and additionally examine more levels of interoceptive signal processing such as metacognitive confidence (i.e., interoceptive awareness) or subjective awareness (i.e., interoceptive sensibility). Besides, it could be interesting to differentiate between obsessive-compulsive thoughts and behaviors in future studies. It could also be useful to integrate another psychiatric control group (e.g., panic disorder) as in the setting of Yoris et al. [27]. Lastly, using methods in therapy that can improve IAcc (e.g., body scan; [58]) should be considered to be of high relevance.

Acknowledgments

The authors would like to thank all staff members of the hospital Windach am Ammersee as well as the secretaries, the nursing and medical doctors who supported the project. We also want to thank Dr. Sarah Wankner for her help in data assessment and data management, as well as all students and student assistants who were involved in assessing and post-processing of data.