Possible Association Between Interoception and Alexithymia in Patients With Addiction Authors: Yuichi Shimohara, Ariyuki Kagaya, Chika Shimohara, Wiktoria Chciuk, Maro G Machizawa, Shigeto Yamawaki Published: 2026 ABSTRACT ABSTRACTInteroception has been reported to be associated with several psychiatric disorders, including addiction. Alexithymia refers to difficulties in identifying and describing one's own emotions and has also been linked to addictive disorders. However, the relationship between self‐reported interoceptive sensitivity and alexithymia in patients with addiction remains insufficiently understood, particularly when current distress and functional impairment are considered. In this study, we examined the relationship between the Body Perception Questionnaire‐Body Awareness very short form Japanese version (BPQBAVSFJ), the Toronto Alexithymia Scale‐20 (TAS‐20), and the Distress and Impact Thermometer (DIT) in 94 patients with addiction. BPQBAVSFJ was significantly correlated with TAS‐20 (Pearson r = 0.405, p < 0.001; Spearman rho = 0.458, p < 0.001). This association remained significant after adjustment for age, sex, distress, and impact (B = 0.271, 95% CI = 0.105–0.438, p = 0.002). Distress showed a weak bivariate association with BPQBAVSFJ, whereas impact did not. In the multivariable model, neither distress nor impact independently predicted BPQBAVSFJ. No significant differences in BPQBAVSFJ were observed among the six addiction subtypes or between substance‐related and behavioral addictions after adjustment for age and sex. These findings suggest that self‐reported interoceptive sensitivity in patients with addiction may be more closely associated with alexithymia than with current distress or functional impairment.Keywords: addiction, alexithymia, distress, impact, interoception Introduction Interoception refers to the perception and processing of internal bodily signals. Altered interoception has been reported in several psychiatric disorders, including addiction, depression, eating disorders, obsessive‐compulsive disorder, and post‐traumatic stress disorder [1]. We have previously reported impaired interoceptive awareness in patients with alcohol use disorder and gambling disorder compared with healthy individuals, as well as longitudinal changes in interoception in patients with alcohol use disorder [2, 3]. Alexithymia is characterized by difficulties in identifying and describing one's own emotions and by externally oriented thinking [4, 5]. It has been associated with several psychiatric conditions, including psychosomatic disorders, depression, autism spectrum disorder, and addiction [4, 5, 6, 7]. Alexithymia has also been linked to interoception, because recognizing emotional states partly depends on the perception and interpretation of bodily signals [8, 9, 10, 11, 12, 13, 14]. However, previous findings on the relationship between alexithymia and interoception have been inconsistent, and this relationship remains insufficiently studied in patients with addiction [14, 15]. Several studies have reported alexithymia in addictive disorders, including gambling disorder, gaming disorder, and substance use disorder [16, 17, 18, 19, 20, 21]. However, few clinical studies have examined whether self‐reported interoceptive sensitivity is associated with alexithymia in patients with addiction while accounting for current distress and functional impairment. Therefore, we examined the relationships among BPQBAVSFJ, TAS‐20, and DIT in patients with addiction. We hypothesized that BPQBAVSFJ would be associated with TAS‐20 even after adjustment for age, sex, distress, and impact. Methods Participants Subjects were recruited from patients aged 15 to 75 years who were undergoing outpatient treatment for addiction at Senogawa Hospital and Yokogawa Ekimae Clinic in Hiroshima, Japan. Diagnoses were made according to the ICD‐10 Classification of Mental and Behavioral Disorders. Patients who were unable to complete the questionnaires because of dementia, intellectual disability, higher brain dysfunction, or severe physical illness were excluded. A total of 97 patients were enrolled. Three were excluded because of insufficient responses, dementia, or higher brain dysfunction. Ultimately, 94 patients were included in the study. The sample consisted of 66 men, 27 women, and one participant not classified as male or female. The diagnostic breakdown was as follows: alcohol use disorder (AUD), n = 20; substance use disorder (SUD), n = 14; gambling disorder (GAMB), n = 23; kleptomania (KLEP), n = 13; gaming disorder (GAME), n = 10; and disorders of sexual preference (SXPD), n = 14. For analyses requiring binary sex adjustment, the one participant not classified as male or female was excluded, resulting in an analytic sample of 93 patients. For exploratory analyses, the diagnostic subtypes were also grouped into substance‐related addictions and behavioral addictions. AUD and SUD were classified as substance‐related addictions, whereas GAMB, KLEP, GAME, and SXPD were classified as behavioral addictions. Clinical status was classified as active addiction or remission. Active addiction was defined as substance use or addictive behavior within the past three months, or initiation of outpatient treatment within the past three months. Remission was defined as abstinence from substance use or addictive behavior for at least three months. Measures BPQBA consists of 12 items, each rated on a 5‐point scale. Total scores range from 12 to 60, with higher scores indicating greater self‐reported interoceptive sensitivity. The Japanese version of the BPQBA has been shown to have reliability and validity [22, 23]. TAS‐20 consists of 20 items, each rated on a 5‐point scale. Total scores range from 20 to 100, with higher scores indicating more severe alexithymia [24]. The DIT consists of two items assessing psychological distress and the impact of problems on daily life. Each item is scored from 0 to 10, with higher scores indicating more unfavorable conditions [25]. Statistical Analysis Variables are presented as means ± standard deviations. Comparisons between men and women were performed using the Mann–Whitney U test. Associations between BPQBAVSFJ, TAS‐20, distress, and impact were assessed using Pearson's correlation coefficient and Spearman's rank correlation coefficient. Partial Spearman correlations were calculated after adjustment for age and sex. Differences among the six addiction subtypes were examined using analysis of covariance (ANCOVA) adjusted for age and sex. In response to the small and uneven subgroup sizes, an additional exploratory analysis was conducted by grouping the diagnoses into substance‐related addictions and behavioral addictions. Multiple linear regression analysis was performed with BPQBAVSFJ as the dependent variable. TAS‐20, distress, impact, age, and sex were entered as independent variables. Multicollinearity was assessed using variance inflation factors. Studentized residuals were examined, and a sensitivity analysis was performed after excluding cases with studentized residuals greater than ±3. As an additional exploratory sensitivity analysis, addiction category and active addiction/remission status were added to the regression model. Statistical significance was set at p < 0.05. Analyses were performed using EZR [26]. Ethics This study was approved by the Senogawa Medical Corporation Ethical Committee, Approval Nos. R05‐04, R05‐04‐01 and R05‐04‐02. Written informed consent was obtained from all participants before the study procedure. All study activities were performed in accordance with the Declaration of Helsinki and relevant guidelines and regulations in Japan. Results The mean age of the 94 participants was 41.0 ± 14.1 years. Mean BPQBAVSFJ, TAS‐20, distress, and impact scores were 25.1 ± 8.4, 54.8 ± 11.3, 4.0 ± 2.9, and 3.0 ± 2.8, respectively. Among the 93 participants classified as male or female and included in sex‐adjusted analyses, there were 66 men and 27 women; 56 were classified as active addiction and 37 as remission. Women had higher TAS‐20 scores than men (60.5 ± 10.6 vs. 52.6 ± 10.9, p = 0.001), whereas age, BPQBAVSFJ, distress, and impact did not differ significantly by sex (Table 1). TABLE 1Differences between men and women.Men (n = 66)Women (n = 27) p Age, years39.3 ± 13.444.7 ± 15.20.133BPQBAVSFJ24.8 ± 9.125.6 ± 6.80.332TAS‐2052.6 ± 10.960.5 ± 10.60.001Distress3.8 ± 2.84.6 ± 2.90.19Impact2.8 ± 2.63.3 ± 3.30.833 Note: Values are presented as means ± standard deviations. The one participant not classified as male or female was excluded from this analysis. p values were calculated using the Mann–Whitney U test.Abbreviations: BPQBAVSFJ, Body Perception Questionnaire‐Body Awareness very short form Japanese version; TAS‐20, Toronto Alexithymia Scale‐20. Differences between men and women. Note: Values are presented as means ± standard deviations. The one participant not classified as male or female was excluded from this analysis. p values were calculated using the Mann–Whitney U test. Abbreviations: BPQBAVSFJ, Body Perception Questionnaire‐Body Awareness very short form Japanese version; TAS‐20, Toronto Alexithymia Scale‐20. After adjustment for age and sex, there were no significant differences among the six addiction subtypes in BPQBAVSFJ, F (5,85) = 0.389, p = 0.855, or TAS‐20, F (5,85) = 1.710, p = 0.141 (Table 2). In the exploratory comparison between substance‐related and behavioral addictions, BPQBAVSFJ did not differ significantly between groups, F (1,89) = 1.067, p = 0.304. TAS‐20 tended to be higher in substance‐related addictions, but the difference was not statistically significant, F (1,89) = 2.978, p = 0.088. TABLE 2Characteristics according to addiction subtype.AUD (n = 20)SUD (n = 14)GAMB (n = 23)KLEP (n = 13)GAME (n = 10)SXPD (n = 14)Female/Male/Other5/15/07/6/12/21/011/2/00/10/02/12/0Age, years50.5 ± 11.441.1 ± 13.134.3 ± 9.752.9 ± 12.623.6 ± 9.734.9 ± 10.7BPQBAVSFJ25.3 ± 11.126.4 ± 7.323.7 ± 7.523.0 ± 8.827.6 ± 8.726.0 ± 6.5TAS‐2052.3 ± 10.863.1 ± 9.153.5 ± 13.056.8 ± 11.352.2 ± 11.252.4 ± 7.8Distress3.0 ± 2.85.0 ± 2.83.9 ± 2.94.0 ± 2.94.3 ± 3.34.5 ± 2.6Impact2.2 ± 2.74.7 ± 2.82.7 ± 2.72.5 ± 2.93.2 ± 2.93.1 ± 2.3 Note: Values are presented as means ± standard deviations. ANCOVA adjusted for age and sex showed no significant differences among the six addiction subtypes in BPQBAVSFJ, F (5,85) = 0.389, p = 0.855, partial η 2 = 0.022, or TAS‐20, F (5,85) = 1.710, p = 0.141, partial η 2 = 0.091.Abbreviations: AUD, alcohol use disorder; BPQBAVSFJ, Body Perception Questionnaire‐Body Awareness very short form Japanese version; GAMB, gambling disorder; GAME, gaming disorder; KLEP, kleptomania; SUD, substance use disorder; SXPD, disorders of sexual preference; TAS‐20, Toronto Alexithymia Scale‐20. Characteristics according to addiction subtype. Note: Values are presented as means ± standard deviations. ANCOVA adjusted for age and sex showed no significant differences among the six addiction subtypes in BPQBAVSFJ, F (5,85) = 0.389, p = 0.855, partial η 2 = 0.022, or TAS‐20, F (5,85) = 1.710, p = 0.141, partial η 2 = 0.091. Abbreviations: AUD, alcohol use disorder; BPQBAVSFJ, Body Perception Questionnaire‐Body Awareness very short form Japanese version; GAMB, gambling disorder; GAME, gaming disorder; KLEP, kleptomania; SUD, substance use disorder; SXPD, disorders of sexual preference; TAS‐20, Toronto Alexithymia Scale‐20. BPQBAVSFJ was significantly correlated with TAS‐20 (Pearson r = 0.405, 95% CI = 0.221–0.562, p < 0.001; Spearman rho = 0.458, p < 0.001). This association remained significant after adjustment for age and sex (partial Spearman r = 0.427, 95% CI = 0.245–0.581, p < 0.001). A scatterplot of TAS‐20 and BPQBAVSFJ scores is shown in Figure 1. BPQBAVSFJ showed a weak bivariate association with distress (Pearson r = 0.263, 95% CI = 0.064–0.442, p = 0.010), whereas its association with impact was not statistically significant (Pearson r = 0.140, 95% CI = −0.065–0.333, p = 0.179) (Table 3). Scatterplot of TAS‐20 and BPQBAVSFJ scores. Pearson r = 0.405, R2 = 0.164, p < 0.001. TABLE 3Correlations between BPQBAVSFJ, TAS‐20, distress, and impact.Pearson r95% CI p Partial spearman r adjusted for age and sex p TAS‐200.4050.221 to 0.562< 0.0010.427< 0.001Distress0.2630.064 to 0.4420.010.2370.022Impact0.14−0.065 to 0.3330.1790.1670.11 Note: Pearson correlations were calculated in the full sample (n = 94). Partial Spearman correlations adjusted for age and sex were calculated among participants classified as male or female (n = 93).Abbreviations: BPQBAVSFJ, Body Perception Questionnaire‐Body Awareness very short form Japanese version; CI, confidence interval; TAS‐20, Toronto Alexithymia Scale‐20. Correlations between BPQBAVSFJ, TAS‐20, distress, and impact. Note: Pearson correlations were calculated in the full sample (n = 94). Partial Spearman correlations adjusted for age and sex were calculated among participants classified as male or female (n = 93). Abbreviations: BPQBAVSFJ, Body Perception Questionnaire‐Body Awareness very short form Japanese version; CI, confidence interval; TAS‐20, Toronto Alexithymia Scale‐20. In the multiple linear regression model adjusted for age, sex, distress, and impact, TAS‐20 was independently associated with BPQBAVSFJ (B = 0.271, 95% CI = 0.105–0.438, p = 0.002). Distress, impact, age, and sex were not significant predictors. The overall model was significant, F (5,87) = 4.491, p = 0.001, with R2 = 0.205 and adjusted R2 = 0.160. All VIFs were below 3 (Table 4). TAS‐20 remained significant after additional adjustment for addiction category and clinical status (B = 0.260, 95% CI = 0.088–0.432, p = 0.004) and after excluding two outliers with studentized residuals greater than ±3 (B = 0.234, 95% CI = 0.086–0.382, p = 0.002). TABLE 4Multiple regression analysis with BPQBAVSFJ as the dependent variable.BSE95% CI t value p VIFIntercept12.1185.2921.600 to 22.6362.290.024—TAS‐200.2710.0840.105 to 0.4383.2340.0021.381Distress0.5610.408−0.249 to 1.3721.3770.1722.113Impact−0.3170.403−1.119 to 0.484−0.7870.4331.925Age−0.0710.06−0.191 to 0.048−1.1880.2381.098Female−1.2811.926−5.108 to 2.546−0.6650.5081.185 Note: Model: F (5,87) = 4.491, p = 0.001; R2 = 0.205; adjusted R2 = 0.160. The one participant not classified as male or female was excluded from this analysis. Sex was coded as male = 0 and female = 1. In sensitivity analyses, the association between TAS‐20 and BPQBAVSFJ remained significant after additional adjustment for addiction category and clinical status and after excluding two outliers.Abbreviations: BPQBAVSFJ, Body Perception Questionnaire‐Body Awareness very short form Japanese version; CI, confidence interval; SE, standard error; TAS‐20, Toronto Alexithymia Scale‐20; VIF, variance inflation factor. Multiple regression analysis with BPQBAVSFJ as the dependent variable. Note: Model: F (5,87) = 4.491, p = 0.001; R2 = 0.205; adjusted R2 = 0.160. The one participant not classified as male or female was excluded from this analysis. Sex was coded as male = 0 and female = 1. In sensitivity analyses, the association between TAS‐20 and BPQBAVSFJ remained significant after additional adjustment for addiction category and clinical status and after excluding two outliers. Abbreviations: BPQBAVSFJ, Body Perception Questionnaire‐Body Awareness very short form Japanese version; CI, confidence interval; SE, standard error; TAS‐20, Toronto Alexithymia Scale‐20; VIF, variance inflation factor. Discussion The main finding of this study was that BPQBAVSFJ was significantly associated with TAS‐20 in patients with addiction. This association remained significant after adjustment for age, sex, distress, and impact, whereas distress and impact did not independently predict BPQBAVSFJ. These findings suggest that self‐reported interoceptive sensitivity in patients with addiction may be more closely associated with alexithymia than with current psychological distress or functional impairment [14, 15]. The differences in TAS‐20 among addiction subtypes were attenuated after adjustment for age and sex. Therefore, subgroup findings should be interpreted as exploratory, particularly because the diagnostic groups were small and uneven. Although TAS‐20 tended to be higher in substance‐related addictions than in behavioral addictions, this difference was not statistically significant. Women had higher TAS‐20 scores than men, but this result should also be interpreted cautiously because comorbid depression, anxiety, and post‐traumatic stress symptoms were not systematically assessed. The association between BPQBAVSFJ and DIT was weaker than that between BPQBAVSFJ and TAS‐20. Distress showed only a weak bivariate association with BPQBAVSFJ and was not significant in the multivariable model, whereas impact was not significantly associated with BPQBAVSFJ. Thus, self‐reported interoceptive sensitivity may not simply reflect current distress or functional impairment. However, the DIT is a brief two‐item screening tool and may not fully capture the complex psychosocial state of patients with addiction [25]. These findings may have clinical implications. Difficulties in identifying and verbalizing emotions may be relevant when clinicians interpret bodily sensations or subjective bodily awareness in patients with addiction. Future studies should examine whether interventions targeting emotional awareness, emotion labeling, or body awareness can improve clinical outcomes in this population. This study has several limitations. The sample size was relatively small, especially for subgroup comparisons, and the sample included heterogeneous addictive disorders. We did not include a healthy control group. Potential confounders such as duration of illness, educational level, comorbid psychiatric symptoms, withdrawal symptoms, craving, and addiction severity were not systematically assessed. In addition, BPQBAVSFJ and TAS‐20 are self‐report measures and may be influenced by underreporting or limited awareness of internal states. The study was conducted at two clinics in a single region of Japan, which may limit generalizability. Finally, because of the cross‐sectional design, causal relationships cannot be inferred. In conclusion, BPQBAVSFJ was significantly associated with TAS‐20 in patients with addiction, and this association remained significant after adjustment for age, sex, distress, and impact. Self‐reported interoceptive sensitivity in addiction may be more closely associated with alexithymia than with current distress or functional impairment. Author Contributions A.K. and C.S. contributed to the conception and design of the study and data acquisition. W.C. and M.G.M. performed the statistical analyses. A.K., C.S., W.C., M.G.M., S.Y. and Y.S. interpreted the data. A.K., C.S. and Y.S. drafted the manuscript. All authors critically revised the manuscript and approved the final version. Funding The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest except M.G.M. is a founder of a profitable company, Xiberlinc Inc. Supporting information Data S1: This file contains the raw data for the subjects regarding age, gender, type of addiction (AUD, SUD, GAMB, KLEP, GAME, SXPD), BPQBAVSFJ, TAS‐20, and Distress and Impact, as evaluated in this study. Data Availability Statement All relevant data are included in this paper. Raw data from this study are available in Data S1 of the Supporting Information. No information that identifies the individual participants is included. References Associated Data Supplementary Materials Data S1: This file contains the raw data for the subjects regarding age, gender, type of addiction (AUD, SUD, GAMB, KLEP, GAME, SXPD), BPQBAVSFJ, TAS‐20, and Distress and Impact, as evaluated in this study. Data Availability Statement All relevant data are included in this paper. Raw data from this study are available in Data S1 of the Supporting Information. No information that identifies the individual participants is included.