Interoception-Based Yoga for Chronic Pain: A Pilot Feasibility Study

Abstract

Background

Chronic pain has been linked to increased awareness of but decreased accuracy in detecting internal sensations. Interoception, the sense of signals from inside the body, may be a therapeutic target to improve pain. No studies have examined the feasibility of targeting interoceptive skills within a yoga intervention for chronic pain – where direct engagement with challenging sensations may be difficult – or whether yoga improves any measure of interoception among persons with chronic pain.

Objective

The primary aim was to determine the feasibility and acceptability of a 6-week interoception-based yoga intervention in a population of mixed chronic pain. The second aim was to examine the preliminary effectiveness on 2 measures of interoception (interoceptive sensibility and interoceptive accuracy), pain and psychosocial outcomes.

Methods

This was a single-arm feasibility study. Primary outcomes were feasibility and acceptability. Secondary outcomes included self-report measures of interoceptive sensibility (Multidimensional Assessment of Interoceptive Awareness, Version 2), pain intensity, pain interference, mindfulness, spiritual well-being and physical and mental health-related quality of life. Interoceptive accuracy was assessed with heartbeat counting. All measures were collected at baseline and after 6 weeks.

Results

The 6-week interoception-based yoga intervention demonstrated feasibility and high acceptability in a sample of mixed chronic pain. Twenty-six participants consented, 24 started the intervention (mean age 42.5 years, 75% female) and 19 (79%) completed 6-week post-intervention testing. The average attendance rate was 69%. There were no serious adverse events and acceptability ratings were high (>5/7). While not powered to make conclusions, preliminary analyses revealed a statistically significant increase in interoceptive sensibility and interoceptive accuracy, and reduction in pain (p’s < .05). There were no correlations between the 2 measures of interoception.

Conclusion

The intervention appears feasible and acceptable and shows promise for improving pain and multiple dimensions of interoception, especially those involved in consciously attending to the body. Future studies examining the efficacy of interoception-oriented yoga interventions for chronic pain are warranted.

ClinicalTrials.gov registration

NCT06268197 Registered February 12, 2024.

Keywords

yoga interoception chronic pain

Introduction

Chronic pain impacts approximately 1 in 5 adults in the United States.¹ Characterized by pain persisting for longer than 3 months, chronic pain has been linked to increased awareness of, but decreased accuracy in detecting internal sensations.² Interoception, broadly defined as the sense of signals from inside the body,³ may thus be a novel therapeutic target that can address the ongoing and urgent need to improve pain outcomes and is emerging as a means by which mind-body practices (MBP), such as yoga, may improve pain.^3,4 While enhanced interoception is generally considered a positive outcome of MBP, it has also been cautioned that training greater awareness of internal sensations has the potential to increase distress, especially in clinical populations like chronic pain, where direct engagement with challenging sensations may exacerbate symptoms.^3,5 As such, examining the feasibility and acceptability of interoception-oriented interventions for chronic pain is warranted.

Interoception

Interoception is a multidimensional construct that includes the conscious and unconscious sensing, integration and interpretation of bodily signals across neural, behavioral and cognitive domains.^3,6 Conceptualizations and definitions of interoceptive dimensions continue to evolve,^3,6-8 with recent recommendations highlighting the importance of distinguishing between and assessing across interoceptive dimensions to improve research on interoceptive function.^6,8 The terminology used herein will follow the definitions from Garfinkel and Critchley (2013) 3-dimensional framework.⁷ This work will examine 2 dimensions of interoception: interoceptive sensibility, defined as dispositional tendencies to be internally focused and beliefs about internal sensations⁷ and interoceptive accuracy, defined as the accurate detection of internal signals.^6,7

Self-report measures of interoception vary widely and should be selected to measure the specific construct intended.⁹ The Multidimensional Assessment of Interoceptive Awareness (MAIA)¹⁰ is a commonly used self-report measure of interoceptive sensibility that evaluates 8 dimensions regarding attention to, regulation and appraisal of interoceptive cues. The MAIA is particularly useful for behavioral pain research by providing a comprehensive overview of how patients notice, interpret and react to internal sensations that extends beyond awareness of sensations to include the appraisal and behavioral response to those sensations.¹⁰ Interoceptive accuracy is often measured via cardiac perception, as in heartbeat counting (HCT) but includes other axes like gastric or respiratory sensitivity that do not generalize to cardiac perception.^11,12 HCT is commonly used in interoception research.⁶ However, the validity of the HCT has been questioned¹³ as task performance can be influenced by beliefs about one’s heartrate or time estimation.^13,14 Thus, studies using the HCT are recommended to control for these confounds.¹³

A recent meta-analysis reports that individuals with chronic pain demonstrate greater self-reported awareness of internal signals but lower interoceptive accuracy.² While the self-report measures varied, the MAIA was used most often, and higher scores were identified primarily in dimensions related to greater attention to bodily sensations. The relationship between HCT accuracy and pain has received more attention,⁶ but findings are mixed as to whether interoceptive accuracy corresponds to pain levels. For example, reduced interoceptive accuracy was associated with greater pain symptomology in fibromyalgia¹⁵ whereas others report that higher interoceptive accuracy predicted greater pain severity in mixed chronic pain.¹⁶ These inconsistencies could reflect a complex, non-linear relationship between interoceptive accuracy and pain, limitations of the HCT or differences between chronic pain conditions that require further study. However, if individuals with chronic pain do indeed attend more to bodily sensations, enhancing this awareness through an interoception-oriented yoga practice has the potential to increase distress and needs feasibility testing.

Yoga

The term yoga means ‘yoke’and encompasses a wide range of practices that are not restricted to physical postures.¹⁷ In modern and scientific usage, yoga typically refers to Hatha yoga, with an emphasis on postural practice, and is the style most commonly practiced¹⁸ and utilized in research settings.¹⁹ An umbrella term, Hatha yoga is broadly inclusive of any style that includes a combination of asana (postures), pranayama (breathwork), and dhyana (meditation), 3 of the 8 limbs of yoga described in Patanjali’s Yoga Sutras.²⁰ We hereby use the term yoga to refer to Hatha yoga, unless otherwise specified.

Yoga has become increasingly popular for pain management and is the MBP with the largest increase in utilization among U.S. adults for pain management since 2002.²¹ In 2022, 28.8% of adults who practiced yoga within the past year reported using yoga for managing or treating chronic pain.²² The evidence in support of yoga for pain management has also grown, especially in support of yoga for chronic low back pain (cLBP). Multiple randomized controlled trials (RCTs) have demonstrated the efficacy of yoga in improving pain in comparison to exercise,²³ stretching,²⁴ physical therapy,²⁵ delayed treatment²⁶ and in virtual formats.²⁷ Such tightly controlled research is essential for demonstrating the initial efficacy of yoga within specific pain populations, but the translation of these findings may be limited in clinical and community settings that serve varied pain populations. Few trials to date have examined whether yoga can feasibly be delivered to mixed chronic pain populations²⁸ that may improve the external validity of the findings.

While the evidence in support of yoga for chronic pain continues to grow, how yoga improves pain remains to be shown. Several hypotheses suggest a complex interplay between biological and psychological mechanisms including stress, immune, neurological and perceptual processes.^4,29 Increasingly, interoception is proposed to be a key component of mind-body practices for chronic pain^3,4 However, no studies have specifically examined whether it is feasible to use yoga to target interoceptive skills or whether yoga improves any measure of interoception among persons with chronic pain.^30,31 While many yoga practices inherently train interoceptive attention (eg, via intentional breath regulation and mindful movement), it is also possible that in the absence of specific cues 1 can move through yoga postures with little awareness of the breath or without consideration for using discomfort to inform the use of variations. These differences could influence how individuals experience yoga postures and may affect how people relate to their bodies and pain. For example, some may continue to push through pain in an attempt to achieve a particular shape, whereas others may learn to observe discomfort with curiosity to determine whether a prop or postural variation would be more appropriate.

Complementary findings suggest that examining the impact of yoga on interoceptive measures is worth investigating. Pain modulating benefits of MBP are associated with greater volume³² and activity^33,34 in the insular cortex, the primary interoceptive neural region,³⁵ in expert practitioners of yoga³² and meditation.^33,34 Improved self-reported interoceptive skills have been reported after 3-month of contemplative training³⁶ and yoga³⁷ in healthy individuals, as well as in response to self-compassion training,³⁸ qigong³⁹ and mindfulness-based approaches^40,41 in chronic pain. Interoception-oriented mindfulness treatments have also demonstrated improved health outcomes in populations with interoceptive dysfunction, including substance use disorder⁴² and opioid-treated chronic pain.⁴¹ The effects of MBP on interoceptive accuracy are more limited and mixed. Some report an increase in interoceptive accuracy in healthy individuals after contemplative training⁴³ and body scanning,⁴⁴ whereas others report no change after Mindfulness Based Stress Reduction (MBSR).⁴⁵ To our knowledge, no studies have examined the impact of yoga on multiple dimensions of interoception or the relationship between changes in interoceptive sensibility and interoceptive accuracy after yoga for chronic pain.

Purpose

This study sought to examine whether interoception might be a key therapeutic ingredient around which to structure a Hatha yoga intervention for chronic pain. The protocol emphasized core practice themes that progressively directed attention inwards to train attentional and regulatory interoceptive skills. As no studies to date have examined whether yoga can feasibly train interoceptive skills among individuals with chronic pain,^30,31 the primary aim was to determine the feasibility and acceptability of an interoception-based yoga intervention in a mixed population of chronic pain. The second aim was to examine the preliminary efficacy of the intervention on 2 separate dimensions of interoception (interoceptive sensibility and interoceptive accuracy), pain and psychosocial outcomes.

Methods

Procedures

Study participants were recruited from the community around the University of Illinois Urbana-Champaign campus via community listservs and flyers. Inclusion criteria broadly targeted individuals ages 18-64 with any form of self-reported chronic pain (>3 months), self-reported pain interference in the past week (Yes/No), ability to get on and off the floor without assistance, no regular, ongoing MBP (>1x/weekly) in the past 6 months, and Physical Activity Readiness Questionnaire (PARQ) Screen Pass or physician consent. Exclusion criteria included use of an assistive device to move around the home or community, active or planned worker’s compensation or personal injury claim, and the following conditions: known pregnancy, recent surgery or acute bone, joint or nerve injury (<6 months), severe or progressive neurological conditions such as Parkinson’s disease or multiple sclerosis, as well as a few complex pain syndromes and pain-related comorbidities like cancer-related pain, complex regional pain syndrome, postural orthostatic tachycardia syndrome, and functional neurological/movement disorders. All participants provided written informed consent and completed a 1-hour testing session before and after intervention during which they completed survey measures and study procedures. Participants were compensated up to $30 ($15 at each time point) for their time. This study was approved by the University of Illinois Urbana-Champaign Institutional Review Board (IRB23-0329), conducted in compliance with the Helsinki accords and HIPAA law, and registered on ClinicalTrials.gov (NCT06268197).

Yoga Intervention

The yoga intervention is detailed following the CLARIFY (CheckList stAndardising the Reporting of Interventions For Yoga) guidelines⁴⁶ in Appendix 1. Briefly, the intervention consisted of twice weekly 75-minute, in-person yoga classes for 6 weeks delivered in 2 cohorts from April – July 2024 by a 500-hour certified yoga instructor with over 500 hours of yoga teaching experience and clinical experience as an occupational therapist for chronic pain. The Panchamayakosha model⁴⁷ (5 sheaths; hereafter abbreviated to Kosha model) from the yoga tradition was used as a framework to progressively and intentionally direct attention inwards. The 5 sheaths, bodies or layers of being from this framework include the annamayakosha (outer physical body), pranamayakosha (breath/energetic body), manomayakosha (mental/emotional body), vijnanamayakosha (wisdom/intellectual body), and the innermost anandamayakosha (bliss body) that overlap neatly with the biopsychosocial-spiritual frameworks from rehabilitation science.⁴⁸ Participants were also provided a cultivated list of pre-recorded breathing, meditation, and short movement-based yoga practices as an optional resource accessible throughout and after the intervention. During the intervention period, participants completed a weekly home practice journal to report on the type and frequency of pain management strategies utilized in the previous week.

The specific asana, pranayama and dhyana selected, how postures were taught (eg, static postural holds vs dynamic transitions) and cueing emphases (eg, on joint alignment vs breath awareness) varied based on the weekly theme and postural variations were graded to gradually increase in difficulty over 6 weeks. Corresponding interoceptive and pain management skills were emphasized in accordance with the weekly themes (Appendix Table A.2) The intervention emphasized using interoceptive cues to inform the choice of postural variations. The attentional focus during postures was emphasized via verbal cues and linked to the kosha and interoceptive theme for that week.

Measures

Participants provided demographic information, completed a general health history, objective anthropometric measures (height, weight, and body mass index (BMI)), and the measures described below. Demographics included age, sex, race/ethnicity, and highest level of education. Pain-related demographics included self-reported primary pain complaint, self-reported pain duration (years), a numeric rating scale of average pain intensity in the past week (0-10), and average pain interference in the past week (0-10). Participants also reported their previous experience with yoga and/or other MBP including frequency and type of practice.

Primary Outcomes

Feasibility

Feasibility targets for this study were defined as achieving the recruitment target (N = 25) in a 9-month period, 80% retention (N = 20), 70% overall attendance, and no serious adverse events.

Acceptability

Acceptability was evaluated via a post-intervention anonymous feedback form (Appendix 2). Participants rated 9 items on a 7-point Likert scale regarding program value (1 = lowest rating, 7 = highest rating). The acceptability target for this study was >5/7 on overall satisfaction and overall enjoyment questions. Additional program feedback questions were collected to improve future intervention design.

Fidelity

A fidelity checklist was created with the core practices for each class and included a rating for adhering to the class theme (Y/N), and instruction of the proposed postures, breathing and meditation practices (%). An independent rater (IP) listened to class recordings to provide a rating of instructor fidelity.

Interoception Outcomes

Multidimensional Assessment of Interoceptive Awareness, Version 2 (MAIA-2)

The MAIA-2 is a 37-item questionnaire evaluating 8 domains of interoceptive sensibility, including: Noticing, Not Distracting, Not Worrying, Attention Regulation, Emotional Awareness, Self-Regulation, Body Listening and Trusting.⁴⁹ Participants rate how much each statement applies to them generally in daily life from 0 (Never) to 5 (Always). Total and subscale scores are calculated as the average of item responses. Higher scores indicate greater interoceptive skill in that domain. All MAIA-2 subscales demonstrated acceptable internal consistency in healthy individuals (α′s = .64-.83)⁴⁹ and good reliability in this sample (α′s = .80-.91). While the MAIA-2 was designed for subscale interpretation,¹⁰ we also calculated a total MAIA-2 score to represent overall interoceptive sensibility, as in previous findings.^38,41,50 The full MAIA-2 demonstrated excellent internal consistency (α = .90).

Heartbeat Counting Task (HCT)

The HCT is a measure of interoceptive accuracy where participants are asked to feel and count their heartbeats. Participants were connected to a portable electrocardiogram (ECG; eMotion Faros 180, sampling rate 250 Hz) with non-polarizable Ag-AgCl electrodes in a 2-lead arrangement under the right mid-clavicle and lower left ribcage. Participants first completed a rest period during which they completed the survey measures. They were then instructed to silently count their own heartbeats when indicated, without taking their pulse at a pulse point (eg, radial or carotid artery), and to stop when instructed. Participants had a 2-minute practice trial and were told that they would perform the task 4 more times (for 25-, 35-, 45- and 60-second intervals) but were not told how long each trial would last or in which order they would occur. The trial order was generated randomly, and each trial was separated by a 30-second break. Following each trial, participants verbally reported the number of counted heartbeats. If they did not feel anything, they were asked not to guess and report 0 as their answer.

The actual number of recorded heartbeats was extracted using the Kubios HRV Scientific analysis software (Version 4.1.0, Kubios). Interoceptive accuracy was calculated as the mean accuracy across 4 trials, where trial accuracy was calculated as the difference between recorded and counted heartbeats. Scores range from 0-1, with higher scores indicating greater interoceptive accuracy, using the following formula:

Interoceptive accuracy = 1 / 4 \sum [1 - (| recorded heartbeats - counted heartbeats |) / recorded heartbeats]

Time Estimation (TE) Task

To control for the influence of time perception on HCT performance^13,51 participants also completed a time estimation (TE) task in which they counted seconds instead of heartbeats. Participants were told they would perform the task 4 times (for 23-, 38-, 42-, and 62-second intervals) but were not told how long each trial would last or in which order they would occur. The trial order was generated randomly, and each trial was separated by a 30-second break. Following each trial, participants verbally reported the number of seconds counted. Timing accuracy was calculated similarly to HCT accuracy with the following equation.

Time estimation = 1 / 4 \sum [1 - (| actual seconds - counted seconds |) / actual seconds]

Beliefs About Heartrate and Strategy Usage

To control for the influence of knowledge or beliefs about one’s own heartrate,^13,51 participants answered the question “How many heartbeats would you expect to have over a 60-second period when sitting down and feeling relaxed?” If participants provided a range (eg, 70-75bpm), then the midpoint was taken (eg, 72.5bpm). At baseline, HCT accuracy was not correlated with heartrate estimation accuracy (P > .27), therefore this variable was not considered further.

To control for task misunderstanding or alternative strategy usage (eg, estimating heartrate)¹³ participants answered the question “When you were monitoring your heartrate, did you use any particular strategy to count your heartbeats?” Participants reporting alternate strategy usage were excluded from analyses.

Pain and Psychosocial Outcomes

Patient Reported Outcomes Measurement Information System (PROMIS) – Pain Intensity and Pain Interference Scales

The Pain Intensity⁵² measure is a 3-item questionnaire, providing an assessment of pain intensity over the past 7 days at its worst, average, and “right now.” The Pain Interference Short Form 8a⁵³ is an 8-item questionnaire providing a global rating of how much pain interferes with daily life, including activities of daily living, social activities, and leisure. Individual items are rated on a 5-point scale from 1 (No pain/No interference at all, respectively) to 5 (Very severe pain/Pain interferes very much, respectively). Item responses are summed and correspond to T-Scores (mean = 50, sd = 10), with a T-score of 50 representing the average of the reference population. Higher scores indicate a greater presence of the measured concept (eg, more pain interference). The 3-item PROMIS Pain Intensity measure produces a T-score distinct from the standard 11-point numeric rating scale and should be interpreted separately. These measures demonstrate good-excellent reliability and validity^52-54 and demonstrated acceptable-good internal consistency in this sample (Intensity α = .74, Interference α = .89).

Five Facet Mindfulness Questionnaire (FFMQ)

The FFMQ is a 39-item questionnaire evaluating mindfulness across 5 domains including Observing, Describing, Acting with Awareness, Non-judging, and Non-reactivity.⁵⁵ Participants rate how much each item describes their own opinion of what is generally true for them on a 5-point scale from 1 (Never or very rarely true) to 5 (Very often or always true). Total and subscale scores taken as the mean of item responses. Higher scores indicate greater mindfulness. The full FFMQ and 4 of the 5 subscales demonstrated excellent internal consistency (FFMQ α = .95; subscale α′s = .91-.95). The FFMQ-Observing scale demonstrated questionable internal consistency (α = .67).

Functional Assessment of Chronic Illness Therapy – Spiritual Wellbeing Scale (FACIT-Sp)

The FACIT-Sp is a 12-item questionnaire evaluating spiritual wellbeing in individuals with chronic illness and is comprised of the Meaning, Peace and Faith subscales.⁵⁶ This study only administered the Meaning and Peace subscales as the components of spiritual wellbeing most relevant to yoga. Participants rate how applicable statements pertaining to chronic illness have been to them over the past 7 days on a 5-point scale from 0 (Not at all) to 4 (Very much). Items are summed to produce total and subscale scores. Higher scores represent greater spiritual wellbeing. The FACIT-Sp and each of the subscales demonstrated good internal consistency (α′s = .79-.84).

36-Item Short Form Health Survey (SF-36)

The SF-36 is a 36-item questionnaire of health-related quality of life (QOL).⁵⁷ This study focused on the Physical Component and Mental Component summary scales encompassing physical and mental health-related QOL (PCS and MCS, respectively). Following the procedures describes in the scoring manual,⁵⁸ the PCS and MCS utilize norm-based scoring to produce a summary T-score (mean = 50, sd = 10). Higher scores indicate better physical and mental health-related QOL. The PCS and MCS have demonstrated good-excellent internal consistency in multiple populations (PCS α′s = .89-.94; MCS α′s = .84-.91)⁵⁸ and were good in this sample (α′s = .88).

Analyses

All data was collated in REDCap and Excel and analyzed in SPSS. Missing data were handled following the scoring guidelines for each measure (MAIA: 2.1%, FFMQ: 1%, all others < 1%). Data was examined for outliers (Z-scores >± 3.29), extreme values (box plots > 2SD) and normality (Shapiro-Wilk) prior to conducting analyses. Statistical significance was set to P < .05. There were no outliers, but 1 participant demonstrated extreme values on both pain scales and multiple missing items on the MAIA-2 and was thus excluded from questionnaire analyses (N = 23). An additional participant was excluded from HCT analyses (N = 22) due to alternate strategy usage.

Descriptive statistics were calculated for participant characteristics, baseline measures, and resting heartrate, HCT accuracy and TE accuracy (Tables 1 and 2). Means and standard deviations were calculated for normal distributions, medians and ranges for skewed distributions. Demographic differences between dropouts and program completers were examined via Fisher’s exact and independent sample t-tests, for categorical and continuous variables, respectively. Differences in baseline measures were examined by program completion (completed vs dropout) and sex (Male vs Female) with independent sample t-tests and Mann-Whitney U tests for parametric and non-parametric variables, respectively. Associations between participant characteristics (Age, HR, BMI, pain duration) were examined using Pearson’s zero-order (r) and Spearman’s rank order (rho) correlations for parametric and non-parametric variables, respectively. Age had a normal distribution but ranged widely (22-64 years), therefore, associations between participant characteristics and all baseline outcome measures were examined with Pearson’s partial correlations controlled for age. Paired sample t-tests examined change in outcome measures after program completion. Due to the preliminary nature of these analyses, we did not correct for multiple comparisons and significance was set to P < .05.

Table 1.

Participant and Pain Characteristics

	All participants	Completers	Dropouts	Fisher’s exact test
N	24	19	5
Age (mean±sd)	42.5 ± 11.2	40.4 ± 10.0	50.2 ± 13.3	.08^#
Range	22-64	22-57	36-64
BMI (median)	28.6	28.2	28.9
Range	19-55.5	19-55.5	20.2-38.1
Female N (%)	18 (75%)	13 (68.4%)	5 (100%)	.28
Race/ethnicity N (%)				.64
Asian	3 (12.5%)	3 (15.8%)	0 (0%)
Black/African American	1 (4.2%)	1 (5.3%)	0 (0%)
Hispanic/Latin	1 (4.2%)	0 (0%)	1 (20%)
Middle Eastern	3 (12.5%)	3 (15.8%)	0 (0%)
More than 1 race	1 (4.2%)	1 (5.3%)	0 (0%)
White	15 (62.5%)	11 (57.9%)	4 (80%)
Education N (%)				.19
1-2 years college/2-year vocational school grad	1 (4.2%)	0 (0%)	1 (20%)
3 years of college	2 (8.3%)	1 (5.3%)	1 (20%)
College/University grad	6 (25%)	5 (26.3%)	1 (20%)
Master’s degree	10 (41.7%)	8 (42.1%)	2 (40%)
PhD or equivalent	5 (20.8%)	5 (26.3%)	0 (0%)
Yoga experience – Yes N (%)	15 (62.5%)	10 (52.6%)	5 (100%)	.12
Mind-body experience – Yes N (%)	13 (54.2%)	10 (52.6%)	3 (60%)	.59
Pain duration yrs (median)	9	8	15	.16⁺
Range	2-32	2-30	3-32
Pain intensity 0-10 (median)	5	5	5	.73⁺
Range	0-7	0-7	4-7
Pain interference 0-10 (mean ± sd)	3.9 ± 2.4	4.0 ± 2.4	3.4 ± 2.6	.63^#
Range	0-9	0-9	0-7
Primary pain condition N (%)				.88
Ehler’s danlos syndrome	1 (4.2%)	1 (5.3%)	0 (0%)
Fibromyalgia	2 (8.3%)	1 (5.3%)	1 (20%)
Migraine	1 (4.2%)	1 (5.3%)	0 (0%)
Musculoskeletal	15 (62.5%)	12 (63.2%)	3 (60%)
Osteoarthritis	3 (12.5%)	2 (10.5%)	1 (20%)
Peripheral neuropathy	2 (8.3%)	2 (10.5%)	0 (0%)
Primary pain region N (%)				.96
Cervical	3 (12.5%)	2 (10.5%)	1 (20%)
Head	1 (4.2%)	1 (5.3%)	0 (0%)
Lower Extremity	2 (8.3%)	2 (10.5%)	0 (0%)
Lumbar	10 (41.7%)	7 (36.8%)	1 (20%)
Thoracic	1 (4.2%)	1 (5.3%)	0 (0%)
Upper extremity	3 (12.5%)	3 (15.8%)	0 (0%)
Widespread	4 (16.7%)	3 (15.8%)	1 (20%)
Additional pain – Yes N (%)	20 (83.3%)	15 (78.9%)	5 (100%)	.54
Pain meds – Yes N (%)	9 (37.5%)	7 (36.8%)	2 (40%)	1.0
Comorbidities N (%)				.63
Anxiety	10 (41.7%)	8 (42.1%)	2 (40%)
Autoimmune disease	4 (16.7%)	3 (15.8%)	1 (20%)
Comorbid anxiety & depression	10 (41.7%)	8 (42.1%)	2 (40%)
Depression	15 (62.5%)	12 (63.2%)	3 (60%)
Diabetes	3 (12.5%)	2 (10.5%)	1 (20%)
HTN	4 (16.7%)	3 (15.8%)	1 (20%)
IBD	1 (4.2%)	1 (5.3%)	0 (0%)
Osteoporosis	1 (4.2%)	0 (0%)	1 (20%)
Panic disorder	2 (8.3%)	2 (10.5%)	0 (0%)
PTSD	4 (16.7%)	2 (10.5%)	2 (40%)
Pulmonary disease	1 (4.2%)	0 (0%)	1 (20%)

Note: # = independent sample t-test, + = Mann Whitney U-test. Abbreviations: BMI = body mass index; PTSD = post-traumatic stress disorder; HTN = hypertension; IBD = irritable bowel syndrome.

Table 2.

Baseline Outcome Measures by Program Completion

	All participants	Completers	Dropouts	Independent t-test (p)
Survey measures (N)	23	18	5
MAIA-2 (0-5)	2.43 ± .62	2.35 ± .61	2.70 ± .66	.29
Noticing	2.93 ± 1.15	2.85 ± 1.13	3.25 ± 1.32	.50
Not distracting	1.60 ± 1.02	1.76 ± .1.06	1.03 ± .71	.17
Not worrying	2.75 ± 1.09	2.49 ± 1.05	3.70 ± .61	.03*
Attention regulation	2.34 ± .98	2.19 ± .79	2.86 ± 1.47	.18
Emotional awareness	3.32 ± 1.07	3.32 ± 1.14	3.33 ± .87	.98
Self-regulation	2.23 ± 1.04	2.17 ± .99	2.45 ± .57	.60
Body listening	1.83 ± 1.23	1.76 ± 1.23	2.10 ± 1.27	.60
Trusting	2.43 ± 1.38	2.27 ± 1.35	3.17 ± 1.45	.25
Pain (T-score = 50, sd = 10)
PROMIS-intensity	59.75 ± 4.66	59.86 ± 4.93	59.38 ± 3.98	.85
PROMIS-interference	60.80 ± 6.14	60.59 ± 6.58	61.54 ± 4.75	.97^#
FFMQ (1-5)	3.17 ± .66	3.06 ± .65	3.55 ± .58	.14
Observing	3.53 ± .60	3.47 ± .61	3.75 ± .58	.37
Describing	3.58 ± 1.02	3.43 ± 1.05	4.13 ± .78	.18
Acting with awareness	2.93 ± .84	2.87 ± .82	3.18 ± .94	.48
Non-judging	3.10 ± 1.04	2.93 ± .90	3.70 ± 1.38	.15
Non-reacting	2.62 ± .88	2.55 ± .88	2.87 ± .94	.48
FACIT-Sp (0-32)	18.22 ± 5.95	17.44 ± 5.77	21.0 ± 6.40	.25
Meaning (0-16)	11.48 ± 3.29	11.11 ± 3.29	12.8 ± 3.27	.32
Peace (0-16)	6.74 ± 3.70	6.33 ± 3.73	8.20 ± 3.56	.33
SF-36 (T-score = 50, sd = 10)
PCS	39.06 ± 8.07	39.10 ± 8.89	38.90 ± 4.66	.96
MCS	40.51 ± 13.0	38.69 ± 12.37	47.05 ± 14.53	.21
Objective measures (N)	22	17	5
HR (mean ± sd)	76.2 ± 12.8	75.7 ± 13.2	77.6 ± 12.9	.78
Range	50.0 – 95.2	50.0 – 95.2	61.5 – 92.7	.78
HCT accuracy (mean ± sd)	.23 ± .18	.22 ± .15	.28 ± .25	.59
Range	0 – .53	0 – .52	0 – .53	.59
TE accuracy (mean ± sd)	.71 ± 11	.70 ± .11	.72 ± .12	.72
Range	.55 – .92	.55 – .92	.57 – .91	.72

Note: Means and standard deviations. # = Mann Whitney U-test. * = P < .05. Abbreviations: MAIA-2 = Multidimensional Assessment of Interoceptive Awareness, Version 2; PROMIS = Patient Reported Outcomes Measurement Information System; FFMQ = 5 Facet Mindfulness Questionnaire; FACIT-Sp = Functional Assessment of Chronic Illness Therapy – Spiritual Wellbeing Scale; SF-36 = 36-Item Short Form Health Survey; PCS = Physical Component Summary Scale; MCS = Mental Component Summary Scale. HR = heartrate, HCT = heartbeat counting task, TE = time estimation task.

Results

Participant Characteristics

A CONSORT is reported in Figure 1. A total of 163 respondents completed the initial prescreening eligibility questionnaire and were assessed for eligibility. Twenty-six participants completed baseline testing, of which N = 24 started the intervention and completed at least 1 yoga session. Two participants became ineligible prior to the start of the intervention due to new pregnancy (N = 1) and change in work schedule (N = 1). As these participants did not complete a single yoga session, they were removed from analyses for a total sample of N = 24.

Figure 1.

CONSORT

Participant and pain characteristics are reported in Table 1. Participants were predominantly middle-aged (mean = 42.5 years), female (75%), white (62.5%), highly educated (87.5% college graduate or higher) and overweight (median BMI = 28.6). There were no demographic differences between program completers and dropouts, but dropouts trended slightly older (P = .082). Most participants had previously tried yoga or another MBP at least once, of which 4 participants had engaged in sporadic practice (<1x/week) in the past 6 months. The most common forms of MBP previously trialed were mindfulness (29.2%) and meditation (29.2%).

Pain Characteristics

Chronic pain duration ranged widely from 2-32 years (median = 9 years), and participants reported moderate levels of pain interference (mean = 3.9) and intensity (median = 5). Most participants reported chronic musculoskeletal pain (62.5%), and the most common primary pain locations were lumbar pain (41.7%) and widespread pain (16.7%). Most participants reported more than 1 pain region (83.3%) of which 50% also reported migraine pain. Less than half (37.5%) reported using pain medication. Depression and anxiety were the most common comorbidities (62.5% and 41.7%, respectively). Given the high rate of comorbid mood disorders, we checked for demographic differences between those with and without anxiety and/or depression and individuals with comorbid anxiety were younger (P = .024).

Feasibility and Acceptability

Retention and Attendance

Nineteen participants completed the program for a retention rate of 79% (19/24). Reasons for dropout included family/personal conflict (N = 2), no show/unable to contact for follow-up (N = 2) and did not like yoga (N = 1). The average attendance rate was 69% with individual attendance ranging from 25% (3 classes attended) to 100% (12 classes attended). Fourteen participants (73.7%) completed at least 8 of the 12 classes. The most common reasons for absences were illness (32.9%), personal/family conflicts (18.6%), and travel/out of town (18.6%). There was an average rate of 23.7% of participants using the home practice recordings each week. Most participants (67.5%) reported using additional pain management strategies on non-intervention days, of which the most common types were stretching (30.7%), intentional rest (21.9%), walking (21.1%), and exercise/strength training (21.1%).

Adverse Events

One adverse event was reported to the IRB after a participant noted increased dizziness following a week 1 class. This participant subsequently reported a suspected diagnosis of benign paroxysmal positional vertigo and was provided additional modifications (eg, gaze stabilization and wall support). The symptoms were self-limiting and resolved after 1 session. No serious adverse events were reported.

Acceptability

All program completers submitted an anonymous acceptability survey. Ratings were high for overall satisfaction (6.6/7), overall enjoyment (6.6/7), appropriateness to daily life (6/7), how beneficial (6.2/7), usefulness of home practice (5.1/7) and appropriateness for chronic pain (6.5/7). Ratings were neutral for ability to manage pain in daily life (4.8/7). Participants were very likely to report recommending yoga to other individuals with chronic pain (6.5/7) and planning to continue practicing yoga (6.1/7). In addition to the acceptability ratings described above, participants provided open-ended feedback on the program (Appendix 2). The aspects participants liked most were breathing practices (N = 8, 42.1%), modifications/variations (N = 6, 31.6%), stretching (N = 6, 31.6%), and specific alignment cues (N = 5, 26.3%). Participants specifically noted the benefit of modifications/variations to “encourage me to listen to me and my body-not comparing me to others” and “that didn’t make me feel bad if I couldn’t do something.” The aspect of the program participants liked least was finding some postures to be uncomfortable (N = 7, 36.8%). No 2 participants identified the same type (eg, supine vs seated twist) or specific posture to be uncomfortable, likely reflecting the variation in pain symptoms in a mixed chronic pain population.

Fidelity

Overall fidelity was very high, with 100% adherence to the class theme and 98.3% adherence to teaching the proposed postures, breathing and meditation practices. As the instructor was responsible for developing the protocol, a high fidelity rating was expected.

Baseline Measures

Baseline values for all outcome measures are reported in Table 2. Participants reported above average pain intensity and interference levels (T-scores>50) and below average physical and mental health (T-scores<50). HCT accuracy was below chance performance (mean ± sd = .23 ± .18) with individual accuracies ranging from 0 to .53. TE accuracy was above chance performance (mean ± sd = .71 ± .11) with individual accuracy ranging from .55 to .92. There were no differences by program completion or sex except that dropouts reported lower Not Worrying score (P = .025), females were older (P = .035) and male participants reported lower Body Listening scores (P = .001).

There were no correlations between participant characteristics (age, BMI, pain duration and HR) and these variables were not associated with any of the total survey measures at baseline (p’s > .05). Older age was associated with lower TE accuracy (r = −.462, P = .03) and had a trend association with lower HCT accuracy (r = −.418, P = .053). A trend partial correlation between TE accuracy and HCT accuracy (r = .426, P = .054) remained when controlled for age.

Change Scores

A total of N = 19 completed the intervention; N = 18 and N = 17 were retained for analyses of questionnaire and HCT change scores, respectively. There was a statistically significant increase in mean MAIA-2 and HCT accuracy and reduction on both pain measures (Table 3; Figure 2). Individual pain ratings were variable across participants, with a small proportion reporting increased pain after intervention (Figure 2B). HCT accuracy post-intervention remained below chance performance (mean ± sd = .33 ± .22) and ranged from .04 to .87. Individual performance on the HCT was highly variable at both baseline and post-testing (Figure 2C). There were no changes in mindfulness, spiritual wellbeing, mental and physical health-related QOL or TE accuracy.

Table 3.

Change in Total Outcome Measures After Yoga Intervention

Measure	Pre (mean ± sd)	Post (mean ± sd)	Change (mean [95% CI])	p
Interoceptive sensibility (MAIA-2)	2.35 ± .61	2.71 ± .65	.35 [.10, .60]	.007
Noticing	2.85 ± 1.13	3.32 ± .85	.47 [.17, .77]	.004
Not distracting	1.76 ± 1.06	1.98 ± 1.12	.22 [−.34, .78]	.415
Not worrying	2.49 ± 1.05	2.41 ± 1.03	−.08 [−.38, .22]	.596
Attention regulation	2.19 ± .79	2.62 ± .96	.43 [.03, .82]	.036
Emotional awareness	3.32 ± 1.14	3.56 ± 1.01	.23 [−.29, .73]	.340
Self-regulation	2.17 ± .99	2.83 ± 1.08	.67 [.32, 1.02]	<.001
Body listening	1.76 ± 1.23	2.52 ± 1.23	.76 [.35, 1.17]	.001
Trusting	2.27 ± 1.35	2.67 ± .99	.40 [−.09, .89]	.104
PROMIS
Pain interference	60.59 ± 6.58	58.09 ± 6.70	−2.50[−4.94, −.06]	.045
Pain intensity	59.86 ± 4.93	56.47 ± 8.06	−3.38 [−6.67, −.10]	.044
Mindfulness (FFMQ-39)	3.06 ± .65	3.19 ± .57	.13 [−.01, .26]	.059
Observing	3.47 ± .61	3.47 ± .77	.01 [−.25, .26]	.955
Describing	3.43 ± 1.05	3.42 ± 1.05	−.01 [−.21, .20]	.944
Acting with awareness	2.87 ± .82	3.09 ± .80	.22 [.001, .44]	.049
Non-judging	2.93 ± .90	3.19 ± .77	.26 [−.09, .62]	.132
Non-reacting	2.55 ± .88	2.71 ± .78	.17, [−.03, .36]	.092
Spiritual wellbeing (FACIT-Sp)	17.44 ± 5.77	18.87 ± 6.03	1.42 [−.66, 3.50]	.168
Meaning	11.11 ± 3.29	10.72 ± 3.92	−.39 [−1.91, 1.13]	.596
Peace	6.33 ± 3.73	8.13 ± 2.96	1.80 [.61, 2.99]	.005
Health-related quality of life (SF-36)
Physical component Score	39.10 ± 8.89	41.65 ± 9.04	2.55 [−1.19. 6.29]	.168
Mental component Score	38.69 ± 12.37	39.13 ± 11.37	.44 [−3.51, 4.40]	.816
HCT accuracy	.22 ± .15	.33 ± .22	.11 [.02, .21]	.019
TE accuracy	.70 ± .11	.74 ± .10	.04 [−.03, .10]	.240

Note: Results of paired sample t-tests (N = 18) of all outcome measures and subscales. Bold text denotes significant changes. Abbreviations: MAIA-2 = Multidimensional Assessment of Interoceptive Awareness, Version 2; FFMQ = 5 Facet Mindfulness Questionnaire; FACIT-Sp = FACIT-Spiritual Wellbeing Scale; SF-36 = 36-Item Short Form Health Survey; HCT = heartbeat counting task, TE = time estimation task.

Figure 2.

Change in Pain, Self-Reported Interoceptive Tendencies and Interoceptive Accuracy after Yoga Intervention

Subscale analyses (Table 3; Figure 2D) revealed a statistically significant increase in 4 of the 8 MAIA-2 subscales: Noticing, Attention Regulation, Self-regulation, and Body Listening(p’s < .05). There was also an increase in the Acting with Awareness component of mindfulness and the Peace component of spiritual wellbeing (p’s < .05).

Exploratory Associations

Associations between baseline measures and change scores were examined on an exploratory basis to identify hypothesis-generating patterns with partial correlations controlled for age (Supplemental Tables 1-3). Briefly, at baseline higher overall MAIA-2 was associated with less pain interference (r = −.456, P = .033) but not intensity (P > .05) and there was no association between HCT accuracy and overall MAIA-2 or any of the subscales (P > .05). More pain interference was associated with lower Noticing (r = −.485, P = .018) and Not Worrying scores (r = −.423, P = .05). Similarly, greater pain intensity was associated with lower Not Worrying (r = −.510, P = .015) and Self-Regulation scores (r = −.45, P = .036). There were no associations between change in HCT accuracy and any of the survey measures (p’s > .05). However, increased HCT accuracy demonstrated a trend relationship with increased pain intensity (r = .472, P = .065; Supplemental Table 3; Supplemental Figure 1).

Discussion

Based on recruitment of the target population, retention, attendance, lack of serious adverse events, and participant feedback on acceptability surveys, a 6-week yoga intervention using the kosha framework to target interoceptive skills was shown to be feasible and highly acceptable in a sample of mixed chronic pain. While this study was not powered to make conclusions and data was not corrected for multiple comparisons, preliminary analyses yielded a statistically significant increase in interoceptive sensibility and interoceptive accuracy, and reduction in pain. There was no change in overall mindfulness, spiritual wellbeing, physical or mental health-related QOL and no associations between interoceptive accuracy and interoceptive sensibility. Thus, the intervention appears feasible and acceptable and shows promise for improving pain and multiple dimensions of interoception, especially those involved in consciously attending to the body.

Feasibility and Acceptability

Retention and overall attendance fell just short of the a priori goals. These targets were based on high rates achieved in the same community for a yoga intervention in healthy older adults,⁵⁹ which may be a population with greater scheduling flexibility. Participants in this study were predominantly middle-aged, engaged in full-time work or graduate education, and childcare was often cited as a scheduling conflict. Furthermore, illness was the most common reason for absence at a time with increased cold and flu prevalence in the local community. One participant dropped out of the program due to not liking yoga. The remaining dropouts were unrelated (personal/family conflicts) or possibly related to the intervention (no show/unable to contact).

The retention rates achieved (79%, N = 19/24) were greater than the only other yoga trial in heterogeneous chronic pain, reporting 63.6% (N = 28/44) after 8 weeks of twice weekly Hatha yoga.²⁸ In contrast, pilot studies in fibromyalgia have reported higher or comparable retention rates, including 86.4% (N = 19/22) after twice weekly Hatha yoga,⁶⁰ 88% (N = 22/25) after once weekly yoga of awareness,⁶¹ and 78.3% (N = 36/46) after 6 weeks of weekly Satyananda yoga.⁶² It may be more challenging to address the diverse pain needs of a mixed chronic pain population within a single intervention, potentially contributing to the lower retention rates seen in the current and previous trial in heterogeneous chronic pain.²⁸ However, yoga instructors in group-based community and clinical settings often encounter this challenge, improving the generalizability of these findings.

Mean attendance rates for this study (69%) were slightly lower than other small pilot trials of yoga for chronic pain. Two 8-week trials of twice weekly Hatha yoga achieved 75% (12/16 classes) with heterogeneous chronic pain²⁸ and fibromyalgia.⁶⁰ Another 8-week trial in fibromyalgia achieved an attendance rate of 87.5% (7/8 classes) and in a trial of 6 weeks of weekly yoga for fibromyalgia, 74% of participants attended at least 4/6 (66.7%) of classes.⁶² However, the current study achieved greater attendance across 12 total classes, condensed within 6 weeks, than 2 large RCTs for cLBP in which 12 classes were spread over 12 weeks.^24,25 These differences may reflect study design or population differences. Participant preferences and optimal dosage of yoga interventions both within and across pain populations would benefit from further study.

A key feasibility concern for this study was the tolerability of a yoga intervention that emphasized attending to pain and internal sensations to inform the intentional use of variations. However, acceptability ratings were high, and the intervention components participants liked most were the interoceptively intensive aspects of the program, including breathing and listening to their body/symptoms to choose modifications/variations. The aspect they liked least was finding some postures to be uncomfortable. The overall high acceptability ratings for the intervention suggest that this level of discomfort was deemed acceptable. We suggest this to be an important outcome as fear of movement (kinesiophobia) predicts low physical activity participation in chronic pain.⁶³ Learning to tolerate and adjust to challenging interoceptive input during yoga, without fear, may be a core regulatory component of the practice that could translate to sensory experiences ‘off-the-mat’.⁴

Preliminary Change in Interoceptive and Pain Outcomes

There was a statistically significant increase in overall MAIA-2, HCT accuracy and the Noticing, Attention Regulation, Self-Regulation and Body Listening subscales. To our knowledge, these are the first findings to examine and report a change in any measure of interoception after a yoga intervention for chronic pain. We present this data with caution as the study was not powered to make conclusions. However, the increase in overall interoceptive sensibility aligns with previous reports of improved body responsiveness in individuals with heterogeneous chronic pain after 8 weeks of Hatha yoga²⁸ and cross-sectional findings that regular yoga practice may be associated with the use of mental strategies involving interoceptive attention, such as breathing through, acceptance of, and observing pain, rather than ignoring or distracting, during acute pain tasks.³²

The reduction in both pain outcomes aligns with much of the previous literature reporting yoga to be beneficial for chronic pain in a variety of populations and formats.^{23,25-28,60-62,64} However, changes in pain intensity and interference were variable across participants, with a small number of participants reporting increased pain. Other findings have also highlighted the variability in pain scores after yoga, noting that such variability may reflect differences in pain experience and that some individuals may benefit more or less from yoga for managing pain.⁶² We further suggest that some individuals may benefit from a longer yoga intervention to notice an appreciable symptom improvement as some may experience an immediate benefit whereas others may experience an early increase in symptoms, potentially due to unfamiliar movement patterns, before later improvement. Additionally, training that enhances awareness of bodily sensations could result in an enhanced perception of pain, at least in the short-term, in those who may have previously relied on distraction or avoidance-based coping.

The increase in the Attention Regulation, Self-Regulation and Body Listening subscales align with complementary reports that MBPs often yield an increase in the MAIA subscales that reflect tendencies to attend to, regulate responses to and trust bodily sensations, encompassed by the Attention Regulation, Self-Regulation, Body Listening and Trusting subscales. These findings include 12 weeks of qigong for breast cancer survivors with persistent post-surgical pain,³⁹ brief self-compassion training for cLBP³⁸ as well as 3-month of contemplative training in healthy individuals.³⁶ Significant increases in the Noticing scale, as in this report, are less often reported. However, of the 3 MBP noted above, the only other movement-based practice (qigong) also reported a trend increase in Noticing (P = .09)³⁹ and the only other yoga trial to examine changes on the MAIA (in healthy individuals),³⁷ similarly reported the greatest change in Noticing, Attention Regulation and Self-Regulation. It is possible that short-term instruction in movement-based mind-body practices may train interoceptive attention to the body but may be less effective at improving the cognitive appraisal of bodily sensations, as in the Not Distracting and Not Worrying scales. Training in these dimensions could require more time or benefit from formal psychotherapeutic interventions, like cognitive behavioral therapy, to address thought patterns and cognitive distortions about pain. Alternatively, this data may also indicate that the yoga intervention may benefit from refinement to better target the cognitive appraisal of sensations, particularly in individuals with comorbid pain and mood disorders. This could be achieved with greater cuing towards mindful appraisal of sensations during postures or integration of group discussions.

Interestingly, male participants reported lower baseline Body Listening scores and there was no significant change in overall mindfulness. As females consistently demonstrate higher rates of chronic pain⁶⁵ and tend to be overrepresented in yoga studies of chronic pain,^24,28 future studies should seek to examine the role of sex-differences in interoceptive processing on pain outcomes. The lack of change in mindfulness may reflect the interoception-oriented focus of this work and supports previous findings that mindfulness and interoception may be overlapping, yet distinct constructs.⁶⁶

We interpret the increase in HCT accuracy with caution due to very low baseline accuracy in a small sample. However, this result aligns with reports that HCT accuracy improved after 3 months of contemplative training⁴³ and 8 weeks of body scanning,⁴⁴ but is in contrast to reports of no change after MBSR,⁴⁵ all in healthy individuals. Exploratory examinations also identified a moderate, albeit trend, association between an increase in HCT accuracy and increased pain intensity, providing initial indications that training greater sensitivity to internal sensations has the potential to increase pain, at least in the short term. These data, if confirmed in larger trials, would be in line with precautions that “training and increasing [interoception] is not a panacea” (Mehling, 2016, p.7) and that attention to internal sensations should be distinguished from the behavioral and regulatory response to those sensations.⁵ The combined self-report and HCT data suggest that yoga may simultaneously increase attention and sensitivity to the body and tendencies to regulate responses to bodily sensation. Future large trials would benefit from examining how multidimensional interoceptive processes across bodily axes and attentional and regulatory domains relate to pain outcomes.

Of note, this study identified a much lower mean HCT accuracy (.23) than previous reports using the same task in chronic pain populations. These range from .3-.4 in mixed chronic pain,¹⁶ .43-.59 in fibromyalgia,^15,67-69 .52 in CRPS⁷⁰ to .65 in cLBP.⁷¹ HCT accuracy was highly variable in this study, which may reflect a small sample with mixed chronic pain. Some participants (N = 2) were unable to perform the task at baseline, which would have a greater effect on the sample mean than in larger studies. These participants were not excluded from analyses as inability to perform the task could reflect a floor effect of the task or characteristic of the population. The lower HCT accuracy may also reflect the use of strict instructions. As the task instructions explicitly limited guessing, participants who were less confident that felt sensations were a heartbeat could have been less likely to count those sensations. Previous findings that similarly used strict task instructions^16,69,70 also report lower mean accuracies than those that do not.^15,71 Our results combined with patterns in the literature suggest a likely impact of task instructions on HCT accuracy, aligning with recommendations that the interoceptive accuracy literature would benefit from replication with stricter instructions and control procedures.¹³

We further note that the moderate, albeit trend association between HCT and TE accuracies may indicate that HCT accuracy could be influenced by timing ability or attentional capacity. Participants were excluded if they relied on time estimation, but an unconscious awareness of the passage of time could still impact task performance. Furthermore, reduced attentional ability is part of the normal cognitive aging process⁷² and older participants performed more poorly on both tasks. Previous findings have also reported decreased interoceptive accuracy with older age.⁵¹ It is thus possible that the increase in HCT accuracy observed in this study could reflect improved attentional capacity, previously demonstrated to improve after yoga.⁷³ Future studies examining the impact of yoga on HCT accuracy would benefit from controlling for the influence of attentional capacity in addition to timing ability.

Strengths and Limitations

There are several limitations to this study. The small sample and single group design limits interpretation beyond feasibility and acceptability. This study was not powered to make conclusions, thus changes in pain and interoception outcomes are presented on a preliminary basis. While the inclusion of a mixed chronic pain population was intentional to be more representative of community settings to address the primary feasibility aim, the variability in pain conditions and pain scores may mask the identification of outcomes unique to specific pain populations. It is worth noting that in community-based settings, individuals seeking to use yoga for therapeutic purposes may benefit more from yoga classes targeted to meet their specific pain needs (eg, fibromyalgia, low back pain). Furthermore, correlation analyses were all conducted on an exploratory basis and were not adjusted for multiple comparisons, thus they are insufficient to determine strength of relationships beyond the identification of hypothesis-generating patterns. These associations were all controlled for age, another source of variability within the population.

The small scale and minimal funding also introduced potential sources of bias. The primary author was the yoga instructor and administered the testing sessions. It is thus possible that strong participant-instructor relationships increased the likelihood of participants completing follow-up testing, which could positively skew acceptability outcomes. However, this may also be a relative strength as participant-instructor dynamics could be an important component of the intervention and future trials would benefit from qualitatively examining these dynamics. Significant efforts were made to limit potential bias with data interpretation by including the use of intervention and testing protocols, standardized assessments, anonymous acceptability feedback and providing transparent statistical analyses. Furthermore, an independent rater evaluated intervention fidelity (IP), a second rater assisted in consolidating data for analyses (IP and CS), and the senior author (NG) provided supervision regarding data interpretation and analyses. There is also the potential of selection bias as recruitment materials specifically stated the intervention was yoga for chronic pain. Participants with previous experience with or positive opinions of yoga or mind-body practices may have been more motivated to try and stay with the intervention, which could influence feasibility metrics. However, while more than half the participants had previously tried yoga at least once, none of the dropouts were new to yoga and only 1 participant cited not liking yoga as the reason for dropout.

The 2 interoceptive measures also have their limitations. The HCT was selected as a measure of interoceptive accuracy for its relative ease of administration within the time and funding constraints of this study, and to allow for greater comparisons to the literature examining interoceptive accuracy in chronic pain.^15,16,67-71 While the validity of this task has been questioned,^13,14 we sought to mitigate these limitations by controlling for beliefs about heartrate and timing estimation. Furthermore, whether heartbeat counting is a relevant measure of interoceptive skills after yoga, which places a greater emphasis on breath than cardiac perception, must also be questioned as heartbeat counting does not generalize to other bodily axes of interoceptive accuracy.^11,12 A relative strength is that we included the recommended control procedures^13,14 to improve the validity of our results.

While both versions of the MAIA have previously demonstrated responsivity to change across mind-body interventions in both healthy and pain populations,^36-40 it is possible that the change in MAIA-2 scores could be the result of developing a greater understanding of the individual scale items and may not accurately reflect a change in self-reported interoceptive tendencies and beliefs, especially in the absence of a control group. For example, the participant excluded from analyses due to extreme values and multiple missing items reported difficulty understanding and applying statements from the MAIA-2 at baseline but was able to complete all items at follow-up. Lastly, 2 studies examining the factor structure of both the original MAIA and updated MAIA-2 have questioned the 8-factor structure of the scale and reported that the Not Distracting and Not Worrying subscales are weakly associated with an underlying general self-reported interoception construct captured in the remaining 6 subscales,^9,74 highlighting the importance of examining the subscales separately.

An additional strength of this study is that it is the first yoga intervention for chronic pain to specifically target interoceptive skills as a core component of the intervention and to include 2 separate measures of interoception. Non-correspondence between these 2 measures highlights that changes in 1 domain of interoception may not generalize to others. Furthermore, while including a mixed chronic pain population and wide age range introduces variability to the data, the findings are more generalizable to clinical and community-based populations that serve varied pain populations. The same instructor across cohorts is another strength by providing consistency of instruction. Lastly, this study used a novel approach of integrating weekly themes based on the kosha framework to target interoceptive and pain management skills while allowing for flexibility to adjust postural variations to the individual needs within a mixed group.

Conclusion

The results demonstrate that a 6-week yoga intervention using the kosha framework to train interoceptive skills was feasible and acceptable for individuals with chronic pain and shows promise for improving pain and multiple dimensions of interoception, especially those involved in consciously attending to the body. Participants identified breathwork and learning to listen to their body/symptoms to select postural variations to be the most beneficial components of the program. Interoceptive accuracy assessed via heartbeat counting with recommended control procedures was lower than previous reports and potentially influenced by timing ability and/or attentional capacity. Future studies examining the efficacy of interoception-oriented yoga interventions against other comparators and control interventions for chronic pain across multiple dimensions of interoception are warranted.

Supplemental Material

Supplemental Material - Interoception-Based Yoga for Chronic Pain: A Pilot Feasibility Study

Supplemental Material for Interoception-Based Yoga for Chronic Pain: A Pilot Feasibility Study by Stephanie Voss, Ishaan Patel, Claire Skowron, Steven Petruzzello, and Neha P. Gothe in Global Advances in Integrative Medicine and Health

Footnotes

Acknowledgements

The authors would like to thank Linda Lehovec for her feedback on the intervention protocol.

ORCID iDs

Stephanie Voss

Neha P. Gothe

Ethical Considerations

This study was approved by the Institutional Review Board at University of Illinois Urbana-Champaign (IRB23-0329).

Consent to Participate

All participants completed written informed consent to participate.

Author Contributions

SV and NG contributed to the conceptualization and design of this study and analyzed the data. SV, IP and CS collected and collated the data. IP completed the fidelity ratings. SP and NG provided resources and study supervision. All authors approved the submitted version.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was partially supported by the National Institutes of Health award AG066630.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Study data are available from the corresponding author upon reasonable request.*

Supplemental Material

Supplemental material for this article is available online.

References

Rikard

Strahan

Schmit

Guy

. GP. Chronic pain among adults — united States, 2019–2021. MMWR Morb Mortal Wkly Rep. 2023;72:379-385. doi:10.15585/mmwr.mm7215a1

Horsburgh

Summers

Lewis

Keegan

Flood

. The relationship between pain and interoception: a systematic review and meta-analysis. J Pain. 2024;25(7):104476. doi:10.1016/j.jpain.2024.01.341

Farb

Daubenmier

Price

, et al. Interoception, contemplative practice, and health. Front Psychol. 2015;6(JUN):1-26. doi:10.3389/fpsyg.2015.00763

Gard

Noggle

Park

Vago

Wilson

. Potential self-regulatory mechanisms of yoga for psychological health. Front Hum Neurosci. 2014;8:1-20. doi:10.3389/fnhum.2014.00770

Mehling

. Differentiating attention styles and regulatory aspects of self-reported interoceptive sensibility. Philos Trans R Soc B Biol Sci. 2016;371:20160013. doi:10.1098/rstb.2016.0013

Suksasilp

Garfinkel

. Towards a comprehensive assessment of interoception in a multi-dimensional framework. Biol Psychol. 2022;168:108262. doi:10.1016/j.biopsycho.2022.108262

Garfinkel

Critchley

. Interoception, emotion and brain: new insights link internal physiology to social behavior. Commentary on: “anterior insular cortex mediates bodily sensibility and social anxiety” by Terasawa et al. (2012). Scan. 2013;8:231-234. doi:10.1093/scan/nss140

Khalsa

Adolphs

Cameron

, et al. Interoception and mental health: a roadmap. Biol Psychiatry Cogn Neurosci Neuroimaging. 2018;3(6):501-513. doi:10.1016/j.bpsc.2017.12.004

Desmedt

Heeren

Corneille

Luminet

. What do measures of self-report interoception measure? Insights from a systematic review, latent factor analysis, and network approach. Biol Psychol. 2022;169:108289. doi:10.1016/j.biopsycho.2022.108289

10.

Mehling

Price

Daubenmier

Acree

Bartmess

Stewart

. The multidimensional assessment of interoceptive awareness (MAIA). PLoS One. 2012;7(11):e48230. doi:10.1371/journal.pone.0048230

11.

Ferentzi

Bogdány

Szabolcs

Csala

Horváth

Köteles

. Multichannel investigation of interoception: sensitivity is not a generalizable feature. Front Hum Neurosci. 2018;12:223. doi:10.3389/fnhum.2018.00223

12.

Garfinkel

Manassei

Hamilton-Fletcher

den Bosch

Critchley

Engels

. Interoceptive dimensions across cardiac and respiratory axes. Philos Trans R Soc B Biol Sci. 2016;371(1708):20160014. doi:10.1098/rstb.2016.0014

13.

Desmedt

Corneille

Luminet

Murphy

Bird

Maurage

. Contribution of time estimation and knowledge to heartbeat counting task performance under original and adapted instructions. Biol Psychol. 2020;154:107904. doi:10.1016/j.biopsycho.2020.107904

14.

Desmedt

Luminet

Corneille

. The heartbeat counting task largely involves non-interoceptive processes: evidence from both the original and an adapted counting task. Biol Psychol. 2018;138:185-188. doi:10.1016/j.biopsycho.2018.09.004

15.

Duschek

Montoro

Reyes del Paso

. Diminished interoceptive awareness in fibromyalgia syndrome. Behav Med. 2017;43(2):100-107. doi:10.1080/08964289.2015.1094442

16.

Di Lernia

Lacerenza

Ainley

Riva

. Altered interoceptive perception and the effects of interoceptive analgesia in musculoskeletal, primary, and neuropathic chronic pain conditions. J Pers Med. 2020;10(4):1-24. doi:10.3390/jpm10040201

17.

Miller

. Yoga Discipline of Freedom: The Yoga sutra Attributed to Patanjali. Bantam Books; 1998.

18.

Mandlik

Nguyen

Ding

Edwards

. Not all yoga styles are the same: an international survey on characteristics of yoga classes. J Integr Complement Med. 2023;29(5):321-326. doi:10.1089/jicm.2022.0712

19.

Cramer

Lauche

Langhorst

Dobos

. Is one yoga style better than another? A systematic review of associations of yoga style and conclusions in randomized yoga trials. Complement Ther Med. 2016;25:178-187. doi:10.1016/j.ctim.2016.02.015

20.

Bryant

. The Yoga Sutras of Patanjali: A New Edition, Translation, and Commentary. North Point Press; 2009.

21.

Nahin

Rhee

Stussman

. Use of complementary health approaches overall and for pain management by US adults. JAMA. 2024;331(7):613-615. doi:10.1001/jama.2023.26775

22.

Elgaddal

Weeks

. Yoga Among Adults Age 18 and Older: United States, 2022. National Center for Health Statistics; 2024. doi:10.15620/cdc/156474

23.

Sherman

Cherkin

Erro

Miglioretti

Deyo

. Comparing yoga, exercise and a self-care book for chronic low back pain: a randomized, controlled trial. Ann Intern Med. 2005;143(12):849-856. doi:10.7326/0003-4819-143-12-200512200-00003

24.

Sherman

Cherkin

Wellman

, et al. A randomized trial comparing yoga, stretching, and a self-care book for chronic low back pain. Arch Intern Med. 2011;171(22):2019-2026. doi:10.1001/archinternmed.2011.524

25.

Saper

Lemaster

Delitto

, et al. Yoga, physical therapy, or education for chronic low back pain: a randomized noninferiority trial. Ann Intern Med. 2017;167(2):85-94. doi:10.7326/M16-2579

26.

Groessl

Liu

Chang

, et al. Yoga for military veterans with chronic low back pain: a randomized clinical trial. Am J Prev Med. 2017;53(5):599-608. doi:10.1016/j.amepre.2017.05.019

27.

Tankha

Gaskins

Shallcross

, et al. Effectiveness of virtual yoga for chronic low back pain: a randomized clinical trial. JAMA Netw Open. 2024;7(11):e2442339. doi:10.1001/jamanetworkopen.2024.42339

28.

Schmid

Fruhauf

Sharp

Puymbroeck

Bair

Portz

. Yoga for people with chronic pain in a community-based setting: a feasibility and pilot RCT. J Evid Based Integr Med. 2019;24:1-11. doi:10.1177/2515690X19863763

29.

Gupta

Gautam

Kumar

Arora

Dada

. Potential role of yoga intervention in the management of chronic non-malignant pain. Evid Based Complement Alternat Med. 2022;2022(1):5448671. doi:10.1155/2022/5448671

30.

Voss

Boachie

Nieves

Gothe

. Mind-body practices, interoception and pain: a scoping review of behavioral and neural correlates. Ann Med. 2023;55(2):2275661. doi:10.1080/07853890.2023.2275661

31.

Gnall

Sinnott

Laumann

Park

David

Emrich

. Changes in interoception in mind-body therapies for chronic pain: a systematic review and meta-analysis. Int J Behav Med. 2024;31:833-847. doi:10.1007/s12529-023-10249-z. Published online.

32.

Villemure

Čeko

Cotton

Bushnell

. Insular cortex mediates increased pain tolerance in yoga practitioners. Cereb Cortex. 2014;24(10):2732-2740. doi:10.1093/cercor/bht124

33.

Lutz

McFarlin

Perlman

Salomons

Davidson

. Altered anterior insula activation during anticipation and experience of painful stimuli in expert meditators. Neuroimage. 2013;64(1):538-546. doi:10.1016/j.neuroimage.2012.09.030

34.

Gard

Hölzel

Sack

, et al. Pain attenuation through mindfulness is associated with decreased cognitive control and increased sensory processing in the brain. Cereb Cortex. 2012;22(11):2692-2702. doi:10.1093/cercor/bhr352

35.

Critchley

Wiens

Rotshtein

Öhman

Dolan

. Neural systems supporting interoceptive awareness. Nat Neurosci. 2004;7(2):189-195. doi:10.1038/nn1176

36.

Bornemann

Herbert

Mehling

Singer

. Differential changes in self-reported aspects of interoceptive awareness through 3 months of contemplative training. Front Psychol. 2014;5(OCT):1-13. doi:10.3389/fpsyg.2014.01504

37.

Koncz

Nagy

Csala

, et al. The effects of a complex yoga-based intervention on healthy psychological functioning. Front Psychol. 2023;14:1-11. doi:10.3389/fpsyg.2023.1120992

38.

Berry

Lutz

Schuman-Olivier

, et al. Brief self-compassion training alters neural responses to evoked pain for chronic low back pain: a pilot study. Pain Med. 2020;21(10):2172-2185. doi:10.1093/pm/pnaa178

39.

Osypiuk

Ligibel

Giobbie-Hurder

, et al. Qigong mind-body exercise as a biopsychosocial therapy for persistent post-surgical pain in breast cancer: a pilot study. Integr Cancer Ther. 2020;19:1-12. doi:10.1177/1534735419893766

40.

Jong

Lazar

Hug

, et al. Effects of mindfulness-based cognitive therapy on body awareness in patients with chronic pain and comorbid depression. Front Psychol. 2016;7:13. doi:10.3389/fpsyg.2016.00967

41.

Roberts

Ledermann

Garland

. Mindfulness-oriented recovery enhancement improves negative emotion regulation among opioid-treated chronic pain patients by increasing interoceptive awareness. J Psychosom Res. 2022;152:110677. doi:10.1016/j.jpsychores.2021.110677

42.

Price

Thompson

Crowell

, et al. Immediate effects of interoceptive awareness training through mindful awareness in body-oriented therapy (MABT) for women in substance use disorder treatment. Subst Abus. 2019;40(1):102-115. doi:10.1080/08897077.2018.1488335

43.

Bornemann

Singer

. Taking time to feel our body: steady increases in heartbeat perception accuracy and decreases in alexithymia over 9 months of contemplative mental training. Psychophysiology. 2017;54(3):469-482. doi:10.1111/psyp.12790

44.

Fischer

Messner

Pollatos

. Improvement of interoceptive processes after an 8-week body scan intervention. Front Hum Neurosci. 2017;11:452. doi:10.3389/fnhum.2017.00452

45.

Parkin

Morgan

Rosselli

, et al. Exploring the relationship between mindfulness and cardiac perception. Mindfulness. 2014;5(3):298-313. doi:10.1007/s12671-012-0181-7

46.

Moonaz

Nault

Cramer

Ward

. CLARIFY 2021: explanation and elaboration of the Delphi-based guidelines for the reporting of yoga research. BMJ Open. 2021;11(8):e045812. doi:10.1136/bmjopen-2020-045812

47.

Le Page

. Creating yoga therapy classes and individual sessions that work. Int J Yoga Ther. 2002;12(1):13-23. doi:10.17761/ijyt.12.1.u14r026806426533

48.

Pearson

Prosko

Sullivan

. Yoga and yoga therapy. In: Yoga and Science in Pain Care: Treating the Person in Pain. Jessica Kingsley Publisers; 2019.

49.

Mehling

Acree

Stewart

Silas

Jones

. The multidimensional assessment of interoceptive awareness, version 2 (MAIA-2). PLoS One. 2018;13(12):e0208034. doi:10.1371/journal.pone.0208034

50.

Ahmadi

Adib

Selk-Ghaffari

, et al. Comparison of the effects of the feldenkrais method versus core stability exercise in the management of chronic low back pain: a randomised control trial. Clin Rehabil. 2020;34(12):1449-1457. doi:10.1177/0269215520947069

51.

Murphy

Geary

Millgate

Catmur

Bird

. Direct and indirect effects of age on interoceptive accuracy and awareness across the adult lifespan. Psychon Bull Rev. 2018;25(3):1193-1202. doi:10.3758/s13423-017-1339-z

52.

Cella

Riley

Stone

, et al. The patient-reported outcomes measurement information system (PROMIS) developed and tested its first wave of adult self-reported health outcome item banks: 2005–2008. J Clin Epidemiol. 2010;63(11):1179-1194. doi:10.1016/j.jclinepi.2010.04.011

53.

Amtmann

Cook

Jensen

, et al. Development of a PROMIS item bank to measure pain interference. Pain. 2010;150(1):173-182. doi:10.1016/j.pain.2010.04.025

54.

Purvis

Neuman

Riley

LHI

Skolasky

. Discriminant ability, concurrent validity, and responsiveness of PROMIS health domains among patients with lumbar degenerative disease undergoing decompression with or without arthrodesis. Spine. 2018;43(21):1512-1520. doi:10.1097/BRS.0000000000002661

55.

Baer

Smith

Hopkins

Krietemeyer

Toney

. Using self-report assessment methods to explore facets of mindfulness. Assessment. 2006;13(1):27-45. doi:10.1177/1073191105283504

56.

Peterman

Fitchett

Brady

Hernandez

Cella

. Measuring spiritual well-being in people with cancer: the functional assessment of chronic illness therapy—spiritual well-being scale (FACIT-Sp). Ann Behav Med. 2002;24(1):49-58. doi:10.1207/S15324796ABM2401_06

57.

Ware

Sherbourne

. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;30(6):473-483.

58.

Ware

Kosinski

Keller

. SF-36 physical and mental health summary scales: a user’s manual. Health Assessment Lab, New England Medical Center; 1994.

59.

Gothe

Erlenbach

. Feasibility of an 8-week yoga intervention among middle-aged and older adults. OBM Geriatr. 2020;4(4):1-15. doi:10.21926/obm.geriatr.2004137

60.

Curtis

Osadchuk

Katz

. An eight-week yoga intervention is associated with improvements in pain, psychological functioning and mindfulness, and changes in cortisol levels in women with fibromyalgia. J Pain Res. 2011;4:189-201. doi:10.2147/JPR.S22761

61.

Carson

Jones

Bennett

Wright

Mist

. A pilot randomized controlled trial of the yoga of awareness program in the management of fibromyalgia. Pain. 2010;151(2):530-539. doi:10.1016/j.pain.2010.08.020

62.

Lazaridou

Koulouris

Devine

, et al. Impact of daily yoga-based exercise on pain, catastrophizing, and sleep amongst individuals with fibromyalgia. J Pain Res. 2019;12:2915-2923. doi:10.2147/JPR.S210653

63.

Larsson

Ekvall Hansson

Sundquist

Jakobsson

. Impact of pain characteristics and fear-avoidance beliefs on physical activity levels among older adults with chronic pain: a population-based, longitudinal study. BMC Geriatr. 2016;16(1):50. doi:10.1186/s12877-016-0224-3

64.

Anheyer

Haller

Lauche

Dobos

Cramer

. Yoga for treating low back pain: a systematic review and meta-analysis. Pain. 2022;163(4):e504-e517. doi:10.1097/j.pain.0000000000002416

65.

Fillingim

King

Ribeiro-Dasilva

Rahim-Williams

Riley

. Sex, gender, and pain: a review of recent clinical and experimental findings. J Pain. 2009;10(5):447-485. doi:10.1016/j.jpain.2008.12.001

66.

Hanley

Mehling

Garland

. Holding the body in mind: interoceptive awareness, dispositional mindfulness and psychological well-being. J Psychosom Res. 2017;99:13-20. doi:10.1016/j.jpsychores.2017.05.014

67.

Valenzuela-Moguillansky

Reyes-Reyes

Gaete

. Exteroceptive and interoceptive body-self awareness in fibromyalgia patients. Front Hum Neurosci. 2017;11:1-14. doi:10.3389/fnhum.2017.00117

68.

Rost

Van Ryckeghem

DML

Schulz

Crombez

Vögele

. Generalized hypervigilance in fibromyalgia: normal interoceptive accuracy, but reduced self-regulatory capacity. J Psychosom Res. 2017;93:48-54. doi:10.1016/j.jpsychores.2016.12.003

69.

Borg

Chouchou

Dayot-Gorlero

, et al. Pain and emotion as predictive factors of interoception in fibromyalgia. J Pain Res. 2018;11:823-835. doi:10.2147/JPR.S152012

70.

Solcà

Park

Bernasconi

Blanke

. Behavioral and neurophysiological evidence for altered interoceptive bodily processing in chronic pain. Neuroimage. 2020;217:116902. doi:10.1016/j.neuroimage.2020.116902

71.

El Grabli

Quesque

Borg

, et al. Interoception and social cognition in chronic low back pain: a common inference disturbance? An exploratory study. Pain Manag. 2022;12(4):471-485. doi:10.2217/pmt-2021-0090

72.

Harada

Natelson Love

Triebel

. Normal cognitive aging. Clin Geriatr Med. 2013;29(4):737-752. doi:10.1016/j.cger.2013.07.002

73.

Gothe

McAuley

. Yoga and cognition: a meta-analysis of chronic and acute effects. Psychosom Med. 2015;77(7):784-797. doi:10.1097/PSY.0000000000000218

74.

Ferentzi

Olaru

Geiger

Vig

Köteles

Wilhelm

. Examining the factor structure and validity of the multidimensional assessment of interoceptive awareness. J Pers Assess. 2021;103(5):675-684. doi:10.1080/00223891.2020.1813147

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.40 MB