Introduction

Mindfulness practice is a systematic training of attention and self-control that involves paying attention to the present moment without judgment and accepting whatever happens in our sensations, emotions, and thoughts1,2,3. Mindfulness-based interventions (MBIs) have been widely used in education, healthcare, and the workplace to improve attention and learning outcomes, boost performance, reduce stress, and manage symptoms2,4. MBIs have shown a range of benefits for physical and mental health, such as reducing stress, symptoms of anxiety, depression, inflammatory diseases, and chronic pain, while also improving immune function, cardiovascular health, and brain energy metabolism4,5,6,7,8,9. Moreover, MBIs have also demonstrated cognitive benefits in improving attention and memory, suggesting their potential in promoting healthy aging and reducing risks for cognitive impairment and Alzheimer’s disease (AD)4,6.

In our previous studies2,10,11,12,13,14,15,16, Integrative Body-Mind Training (IBMT) has shown that short-term practice (2–10 h) reduces stress hormones, improves immune function, attentional control, emotional regulation, cognitive performance, quality of life, and brain plasticity. Moreover, these positive changes are supported by the interaction and optimization of the central (e.g., increased brain functional, structural plasticity, and neurochemistry levels) and autonomic nervous systems (e.g., physiological changes – heart rate variability, skin conductance response, secretory immunoglobulin A, and reduced cortisol concentrations), which is different from relaxation training (RT)2,10,11,12,13,14,15,16. Similarly, the studies from other labs have also shown that mindfulness practice outperforms RT in stress management, inflammatory biomarkers, and related brain changes17,18. However, how brief mindfulness meditation may affect and regulate the metabolome, the index of the body’s systematic changes, remains largely unknown and underexplored.

Mass spectrometry (MS)-based metabolomics is a powerful approach that evaluates the physiological state of metabolic pathways through the measurement of thousands of low molecular weight metabolites in parallel19,20,21,22,23,24,25,26,27,28,29,30. It has shown significant potential for elucidating disease mechanisms and identifying therapeutic targets as well as diagnostic and prognostic markers31,32,33,34,35,36. Recently, many studies have used metabolomics to investigate the metabolome for identifying biomarkers of healthy lifestyles and whole-body health37,38,39. In addition, there has been a recent emphasis on gaining a better understanding of how the metabolome is affected during stressed states and how it may be linked to emotion and cognition40,41,42,43. Metabolomics has also shown the ability to accurately assess age-related conditions and potentially identify predictors of healthy aging43,44,45,46. Moreover, it has proven effective in uncovering metabolic alterations observed in individuals with cognitive decline and AD47,48,49,50.

A recent cross-sectional study combined an 8-day residential meditation/yoga retreat with a prepared 60-day vegan diet, which showed lipid profile changes that may relate to increased cellular anandamide levels, anti-inflammation, analgesia, and vascular relaxation51. The same research group also used the same design to examine the intervention effects on aqueous metabolites and the gut microbiome52. Increases in gut beta diversity and significant alterations to forty-two metabolites were noted, including elevated branched short-chain fatty acids, iso-butyrate and iso-valerate, which potentially signified an effect on the gut-brain axis52. Similarly, a 10-day training program consisting of meditation, breathing exercises, and cold exposure showed alterations to lactate and pyruvate, which correlated with the anti-inflammatory marker, interleukin-10, when compared to a waitlist control53. In addition, another previous study examined the effect of mindfulness meditation for 8 weeks on metabolic changes between healthy and depressive participants54. The intervention included intensive training with Saturday classes and daily home practice for two months. Results showed improved blood glucose and lipid metabolism and reduced depressive symptoms. Moreover, the meditation effect on metabolic profiles was greater in healthy participants54. While these previous studies are informative, the use of mixed interventions, such as dietary change, make it challenging to distinguish which metabolite alterations are related to meditation itself. Therefore, how mindfulness-based interventions regulate metabolic profiles to support physical and mental health is not well understood.

In this study, we aim to use a randomized controlled trial (RCT) design (integrative body-mind training, IBMT vs. relaxation training, RT) and an untargeted metabolomic approach to examine the brief intervention effects on metabolic profiles. Unlike RT, which only focuses on muscle relaxation, IBMT also emphasizes accepting all sensations, emotions, and thoughts while paying attention to the present moment without judgment. Given that brief IBMT works by improving both brain and physiological changes and optimizing the central and autonomic nervous system interaction10,11,12,13,14,15,16, we hypothesize that compared to the active control RT, IBMT improves metabolites associated with increased energy metabolism, reduced (oxidative) stress, and decreased systemic inflammation or neuroinflammation, which altogether support overall physical and mental health benefits.

Methods

Study overview

Similar to our previous studies2,10,11,12,13,14,15,16, we employed a RCT design in this metabolomic study (Fig. 1). Fasting serum samples were collected at baseline (T1) and completion (T2) of 10 one-hour online sessions from forty-two participants randomly assigned to an experimental group (IBMT) or an active control group (relaxation training, RT)2,11,12,45. The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of Arizona State University. Informed consent was obtained from all subjects involved in the study. The study was conducted as a preliminary feasibility trial before a full-scale clinical trial and was not registered with clinicaltrials.gov.

Fig. 1
figure 1

Schematic overview of the study protocol.

Full size image

Participants

Participants were recruited through online services and printed materials in this preliminary study. The study period was from October 2023 to August 2024. All participants were screened to protect against potential confounding variables that would impact short-term mindfulness effects on metabolic changes. Participants were excluded if they had (1) prior or current training experiences such as meditation, relaxation, yoga, or cognitive training; (2) a past or present history of psychiatric, immune, metabolic, or endocrine diseases, or use of medication with known immune or endocrine effects using the Short Form-36 Health Survey. Forty-two healthy and eligible participants (the scores of the Short Form-36 Health Survey > 60) completed the IBMT intervention (n = 23) or the active control RT (n = 19). The IBMT group had a mean age of 55 years (SD = 6y), and the RT group had an average age of 53 years (SD = 6y). Participants in the IBMT and RT groups were comprised of five males and one male, respectively. In our study sample, 88% were non-Hispanic or Latino, while 12% were Hispanic or Latino. The average years of education was 16.8, approximately equivalent to a bachelor’s degree. Participants between the two groups were not significantly different in age, sex, ethnicity, and education level (P > 0.05).

Interventions

Integrative body-mind training (IBMT)

Mindfulness meditation involves paying attention to the present moment with a nonjudgmental attitude and is commonly categorized into two forms: focused attention and open monitoring. According to the literature, IBMT is considered a form of open-monitoring mindfulness meditation2,4,56. As outlined in our previous studies4,10,11,12, IBMT differs from other mindfulness approaches that emphasize deliberate efforts to control or manipulate thoughts and feelings. Instead, IBMT encourages minimal effort, allowing thoughts and emotions to arise naturally while maintaining nonjudgmental awareness. This practice emphasizes awareness of the body’s and mind’s natural state, fostering acceptance of whatever enters consciousness in each moment.

To cultivate this mindful state, IBMT promotes body-mind integration by first engaging the body through gentle stretching postures, which help ground the individual and facilitate a shift into full awareness3,11. In this study, under the guidance of an experienced trainer, participants began with gradual adjustments of body posture performed with full awareness, aiming to achieve physical presence, balance, and integration. Once the body is naturally relaxed, the mind follows—becoming calm yet al.ert, without forceful effort. Our series of studies have demonstrated that this effortless body-mind practice is supported by interactions between the central and autonomic nervous systems, which collectively enhance mindful awareness3,10,11,12.

Relaxation training (RT)

Relaxation Training (RT) is a widely used form of progressive muscle relaxation (PMR) that is particularly popular in Western contexts. RT involves sequentially relaxing specific muscle groups throughout the body, including the face, head, shoulders, arms, legs, chest, back, and abdomen57. It has been commonly employed as an active control condition in RCTs, including our and others’ studies10,15,16,55,56,58. In this study, RT sessions were led by an experienced trainer. Participants were encouraged to stretch and adjust their body posture to facilitate physical comfort and optimal relaxation. With their eyes closed, they followed a guided progression through the muscle groups and were instructed to focus on subjective sensations of relaxation, such as warmth and heaviness. This process aims to induce both physical and mental calmness. While RT and IBMT share several components common to mind-body interventions (e.g., body awareness and breathing), the critical distinction lies in the presence or absence of the mindfulness component. RT does not incorporate mindfulness elements such as open monitoring or nonjudgmental awareness. As such, RT serves as a suitable and structurally equivalent control for studies examining the unique effects of IBMT2,11,13,47.

Metabolomics

Reagents

LC-MS grade acetonitrile (ACN), ammonium acetate, acetic acid, and methanol (MeOH) were obtained from Fisher Scientific (Pittsburgh, PA). Ammonium hydroxide was purchased from Sigma-Aldrich (Saint Louis, MO). Deionized water was provided by an onsite Water Purification System from EMD Millipore (Billerica, MA). Phosphate Buffered Saline was obtained from GE Healthcare Life Sciences (Logan, UT).

Serum preparation

Fasting serum samples were thawed overnight at 4oC, and 50 µL of each sample was transferred to a 2 mL Eppendorf vial. Protein precipitation and extraction of metabolites were achieved by adding 500 µL MeOH and 50 µL internal standard solution (containing 1810.5 µM 13C3-lactate and 142 µM 13C5-glutamic acid). The mixture was vortexed for 10 s and stored at –20 °C for 30 min, and then it was centrifugated at 14,000 RPM for 10 min at 4 °C. The supernatants (450 µL) were placed into a new Eppendorf vial and dried with a CentriVap Concentrator (Labconco, Fort Scott, KS). The samples were then reconstituted in 150 µL of 40% PBS/60% ACN. A mixture of all serum samples was pooled and used as the quality-control (QC) sample. In total, 84 serum samples were used in this study.

Untargeted LC-MS metabolomics

The untargeted LC-MS metabolomics method depicted here was formed from previous studies and published in several studies29,59,60,61. Every LC-MS experiment was performed on a Thermo Vanquish UPLC-Exploris 240 Orbitrap MS instrument (Waltham, MA). Every sample was injected twice, 4 µL for analysis using positive ionization mode and 10 µL for analysis using negative ionization mode. Both chromatography separations were achieved in hydrophilic interaction chromatography (HILIC) mode on a Waters XBridge BEH Amide column (150 × 2.1 mm, 2.5 μm particle size, Waters Corporation, Milford, MA). The column compartment was set at 40 ̊C, the flow rate used was 0.3 mL/min, and the auto-sampler temperature was held at 4 ̊C. The mobile phase was comprised of Solvents A (10 mM ammonium acetate, 10 mM ammonium hydroxide in 5% ACN/95% H2O) and B (10 mM ammonium acetate, 10 mM ammonium hydroxide in 5% H2O/95% ACN). The initial isocratic elution of 90% B was 1 min, then the percentage of Solvent B reduced to 40% at t = 11 min. The composition of Solvent B was held at 40% for 4 min (t = 15 min), and then the percentage of B slowly rose to 90%, to prepare for the following injection. The Orbitrap mass spectrometer with an electrospray ionization (ESI) source was used to collect untargeted data from 70 to 1050 m/z.

Thermo Compound Discoverer 3.3 software was used for the processing of aqueous metabolite data. The software processed untargeted data for peak picking, alignment, and annotation. The mass accuracy limit was set to 10 ppm and the absolute intensity threshold for the MS data extraction was 1,000. Available data for retention time, exact mass, MS/MS fragmentation pattern, and isotopic pattern were used for identification and annotation. These data were identified using in-house chemical standards (~ 600 aqueous metabolites), additionally with MS spectra being compared against the HMDB library, Lipidmap database, METLIN database, as well as commercial databases including mzCloud, Metabolika, and ChemSpider. To improve accuracy and reliability, only peaks with a CV < 20% across pooled quality control samples (once every 10 study samples) and signals showing up in greater than 80% of all the study samples were included for analysis.

Statistical analyses

Paired t-tests were performed to compare the pre- and post-differences within groups (T1 vs. T2). Analysis of covariance (ANCOVA) was used to compare group differences in metabolites at post-intervention, controlling for baseline group differences in metabolite levels (IBMT-T2 vs. RT-T2). Analysis of Variance (ANOVA) was also used to compare within and between group differences. All statistical analyses were performed using SPSS 29.0 (SPSS Inc., Chicago, IL, USA) and R version 4.4 (Vienna, Austria). Fold Change analysis, volcano plots, heat maps, partial least squares-discriminant analysis (PLS-DA), Variable Importance Projection (VIP) scores, and Pathway and Enzyme Enrichment analyses were performed using the MetaboAnalyst software62. A threshold α-level of 0.05 was used to define statistical significance and false discovery rate (FDR) was applied to all analyses unless otherwise stated.

Results

Pre- and post-IBMT intervention (IBMT-T1 vs. IBMT-T2)

To investigate the effects of IBMT on metabolic profiles, paired sample t-tests of baseline (T1) vs. post-intervention (T2) metabolites of the IBMT group were performed. A total of 923 metabolites were identified in serum at pre- and post-intervention timepoints. The IBMT group reported a total of 13 significant compounds related to central carbon metabolism and lipid metabolism, including several derivatives of glycine metabolism (N-nonanoyl glycine, capryloyl glycine, N-lauroyl glycine, and formyl glycine), as well as amino acid-related metabolites like N-acetyl-beta-lysine, 3-azetidinecarboxylic acid, and 2-methylhippuric acid (Table 1). Compounds associated with lipid metabolism, tetrahexosylceramide, 9-hydroxy-5Z-nonenoic acid, and 6-apo-y-caroten-6-al, and other metabolisms, 4-isothiocyanoatobutyl methyl sulfide, dulcin, and thiosulfate, were also significantly changed. Box plots of the significant metabolites have been included in Supplementary Fig. 1.

Table 1 Significant metabolites (FDR q < 0.05) comparing IBMT-T1 vs. IBMT-T2.
Full size table

A heatmap of significant metabolites showing normalized mean relative abundance within the IBMT group was generated to illustrate the directionality of each metabolite from Table 1 (Fig. 2A). After the IBMT training, amino acid metabolites (caprylol glycine, N-lauroyl glycine, 3-azetidinecarboxylic acid) and a lipid metabolite (tetrahexosylceramide) were all increased (red indicating a higher abundance), while all other compounds shown in Fig. 2A were decreased (blue representing a lower abundance). A volcano plot was created highlighting p < 0.05 and a significantly increased fold change of > 2 and a decrease of < 0.5 (Fig. 2B). Fold Change was calculated as IBMT T2/T1, and values larger than 1 demonstrated an increase after the IBMT intervention, and vice versa. For those metabolites with p < 0.05, Capryloyl glycine (Fold Change = 2.04) was the only metabolite to significantly increase after IBMT, while N-nonanoyl glycine (Fold Change = 0.48), formyl glycine (Fold Change = 0.34), N-acetyl-beta-lysine (Fold Change = 0.27), 9-Hydroxy-5Z-nonenoic acid (Fold Change = 0.26), 4-isothiocyanatobutyl methyl sulfide (Fold Change = 0.21), and dulcin (Fold Change = 0.27) were all significantly decreased.

Fig. 2
figure 2

A A heatmap of significant metabolites (FDR q < 0.05) comparing IBMT-T1 vs. IBMT-T2. B Volcano plot of IBMT-T1 vs. IBMT-T2, showing significant metabolites demonstrating p values < 0.05 and fold changes > 2 or < 0.5.

Full size image

In Fig. 3A, a partial least squares – discriminant analysis (PLS-DA) was conducted within the IBMT group (IBMT-T1 vs. IBMT-T2), and variable importance projection (VIP) scores denoted significant drivers of the models (VIP > 1.0) (Fig. 3B). The models highlighted the within-condition similarities and the significant differences between the IBMT baseline (T1) and post-intervention (T2). As shown in Fig. 3A, clear separation was observed between the two time points, and N-nonanoyl glycine, capryloyl glycine, and N-lauroyl glycine had the most significant impacts on the score plot (VIP > 4.0). Additional compounds identified by the paired t-test, such as 6-Apo-y-caroten-6-al, 9-Hydroxy-5Z-nonenoic acid, 4-isothiocyanoatobutyl methyl sulfide, dulcin, thiosulfate, and 2-methylhippuric acid all had greater VIP scores than 2.4. Cross validation tests can be seen in Supplementary Fig. 2A, and the PLS-DA modeling is relatively reliable.

Fig. 3
figure 3

PLS-DA score plot (A) and variable importance projection scores (B) comparing IBMT-T1 vs. IBMT-T2 groups. Each dot in the PLS-DA score plot (A) represents a sample from baseline (T1) or post-intervention (T2). The directionality and influence of metabolites are depicted as a VIP score (B).

Full size image

Pathway and Enzyme Enrichment analyses of the IBMT-T1 vs. IBMT-T2 groups were conducted to identify significantly impacted metabolic pathways from the intervention and their potentially affected and enriched enzymes. The Pathway analysis found significant differences in nitrogen and sulfur metabolism with a greater impact on sulfur metabolites (Fig. 4A). In particular, we observed post-intervention reductions in sulfate and thiosulfate as well as an increase in glutamate. The increased levels of glutamate as well as glycine also had a near significant impact on Glutathione metabolism. The Enzyme Enrichment analysis also identified several significantly enriched enzymes related to sulfur metabolism, cyanide transport, and NAD synthesis (Fig. 4B). Additional enzymes related to glycolysis (pyruvate kinase, carboxylase, and enolase) and the citric acid (TCA) cycle (citrate synthase) were identified, but not significant.

Fig. 4
figure 4

Pathway (A) and Enzyme Enrichment (B) analyses within the IBMT group (IBMT-T1 vs. IBMT-T2). Data were plotted as -log10(p) versus pathway impact (A) and enrichment ratio with P-value versus enrichment (B).

Full size image

Pre- and post-RT intervention (RT-T1 vs. RT-T2)

To investigate the effects of RT on metabolic profiles, paired sample t-tests of baseline and post-condition metabolites of the RT group were conducted. The RT group did not yield any significant metabolites after the intervention (all FDR q > 0.05, Supplementary Tables 1 and Supplementary Fig. 3). Although no significant results were noted, PLS-DA demonstrated some time point differences in the score plot (Supplementary Fig. 4A). Notably, the significant drivers of the score plot, as depicted by VIP scores (Supplementary Fig. 4B), are largely shown to be metabolically unrelated to central carbon metabolism and lipid metabolism, while cross validation testing (Supplementary Fig. 2B) demonstrates an overfitting of the PLS-DA data.

Evidence-based mindfulness intervention vs. active control (IBMT-T2 vs. RT-T2)

Considering pre-existing baseline differences (T1) in metabolites between the two groups, we performed analysis of covariance (ANCOVA) by adjusting for baseline levels of metabolites to compare group differences in metabolites following 10 sessions of the interventions (IBMT-T2 vs. RT-T2)63,64. There were significant differences in 106 metabolites between IBMT vs. RT at post-intervention (FDR-corrected p < 0.05) (Supplementary Table 2). While IBMT demonstrated profound changes throughout the diverse metabolome, we focused on metabolites linked to central carbon metabolism, amino acid metabolism, and lipid metabolism. The IBMT group showed higher levels in 4 metabolically relevant metabolites than the RT group at post-intervention, including capryloyl glycine, PC (17:1/0), PE (13:0/22:6), and PS (16:0/20:5) (Table 2). In contrast, 12 metabolites were higher in the RT group than the IBMT group at post-intervention from both amino acid and lipid metabolism. Box plots of the significant metabolites have been included in Supplementary Fig. 5. Within and between-group differences as determined by ANOVA are reported in Supplementary Table 3, which are consistent with ANCOVA results.

Table 2 Significant (FDR-corrected p < 0.05) metabolites in the central carbon metabolism, amino acid metabolism, and lipid metabolism when comparing IBMT-T2 vs. RT-T2.
Full size table

Pathway and Enzyme Enrichment analyses of the IBMT and RT groups at post-intervention uncovered significant differences in sphingolipid, linoleic acid, histidine, alpha-linolenic acid, and pyruvate metabolism (Fig. 5). In particular, the IBMT group changed metabolite abundances that are associated with lipid metabolism (increases in N-acylsphingosine and decreases in linoleate and octadecatrienoic acid), histidine metabolism (increases in histidine and glutamate), and pyruvate metabolism (decreased pyruvate and lactate with increased fumarate) (Fig. 5A). As shown in Fig. 5B, several enriched enzymes in the IBMT group were related to fatty acid beta-oxidation (linoleic acid transport, fatty acyl-CoA desaturase and ligase, carnitine-O-palmitoyltransferase, linolenic acid exchange, and carnitine transferase), and methenyltetrahydrofolate cyclohydrolase was also significantly enriched.

Fig. 5
figure 5

Pathway (A) and Enzyme Enrichment (B) analyses of IBMT-T2 vs. RT-T2. Data are plotted as -log10(p) versus pathway impact (A) and enrichment ratio with P-value versus enrichment (B).

Full size image

In Fig. 6A, a PLS-DA was conducted between IBMT-T2 and RT-T2 with VIP scores denoting significant drivers of the models (VIP > 1.0) (Fig. 6B). The score plot (Fig. 6A) demonstrates tight clustering of the individual groups as well as significant separation and group differences. Consistent with the ANCOVA results (Table 2) as well as the IBMT-T1 vs. IBMT-T2 PLS-DA (Fig. 3), capryloyl glycine (> 3.1) had a significant impact on the model, while N-nonanoyl glycine was also a significant driver (> 3.2), respectively. Other bacterial and fungal-related compounds and metabolites not associated with central carbon metabolism, such as paecilopeptin, picaridin, fasciculic acid, and elansolid B2, were also among those with greater VIP scores. Cross validation tests can be seen in Supplementary Fig. 2C, and the PLS-DA modeling is reliable.

Fig. 6
figure 6

PLS-DA score plot (A) and variable importance projection scores (B) of IBMT-T2 vs. RT-T2. Each dot in the PLS-DA score plot (A) represents a sample from the post-interventions. The directionality and influence of metabolites are depicted as a VIP score (B).

Full size image

Discussion

Relaxation Training (RT) involves sequentially relaxing separate muscle groups57, but it lacks the mindfulness in an open and non-judgmental attitude that integrative body-mind training (IBMT) emphasizes. Studies using mindfulness and RT have shown the distinct brain mechanisms10,11,12,13,14,15,16,17,18,65. Recent evidence has shown the mechanisms involved with mindfulness-based interventions and how they elucidate both physical and mental health benefits4,5,6,7,8,9. We employed an untargeted metabolomic approach to investigate the mindfulness effects of IBMT on the metabolome in comparison with RT among healthy participants. After the IBMT intervention (Fig. 7), there were significant within-group changes in amino acid metabolism, particularly alterations to four glycine intermediates (increased N-lauroyl glycine and capryloyl glycine as well as decreased N-nonanoyl glycine and formyl glycine), which had been observed in a prior trial51. Notably, all significant metabolites were very strong drivers of the distinct and specific clustering of the PLS-DA models (VIP > 2.4), with the glycine intermediaries appearing to be the largest difference. Reduced glycine levels have been linked to several metabolic disorders like type 2 diabetes (T2D) and non-alcoholic fatty liver disease66. Additionally, the neuromodulatory capability of glycine to alleviate stress and depression has recently been shown67. Glycine binds to an orphan G protein-coupled receptor, GPR158, which is a metabotropic glycine receptor that regulates neuronal excitability in cortical neurons67. In addition, increased levels of glycine in the brain have been shown to improve cognitive performance, such as memory68. Its excitatory action on neurotransmission mediated by N-methyl-D-aspartate (NMDA) receptors, which also require glutamate binding, is essential for learning, memory, and cognition69. The pathway analysis also identified increased glutamate levels through nitrogen metabolism. Glutamate is essential for similar cerebral functions as glycine and its elevation following the intervention mirror previous IBMT findings after using 3T proton magnetic resonance spectroscopy to measure levels in the anterior cingulate cortex13. These parallels may reflect an indirect relationship after IBMT treatment. Similarly, tetrahexosylceramide was increased after the intervention. As a glycosphingolipid, it is a component of plasma membranes critical in cellular signal transductions70. Thus, the improved glycine metabolites and glutamate may suggest a molecular mechanism through which IBMT may promote physical and cognitive health.

Fig. 7
figure 7

Summary of the metabolome that is regulated by the IBMT Intervention. Metabolites and derivatives that were increased after the IBMT intervention are shown in red, while decreases are denoted as blue. Compounds in black are pathway intermediaries. Created with BioRender.com.

Full size image

Aside from these compounds, the IBMT intervention had a largely lowering metabolite effect on amino acid derivatives (N-acetyl-beta-lysine, 3-azetidinecarboxylic acid, and 2-methylhippuric acid), as well as on a hydroxy fatty acid and lipid-like molecule (9-hydroxy-5Z-nonenoic acid and 6-apo-y-caroten-6-al, respectively), and a sugar substitute (dulcin). Moreover, the reduction in sulfur containing metabolites like thiosulfate and 4-isothiocyanatobutyl methyl sulfide is noteworthy. The Pathway and Enzyme Enrichment analyses also uncovered similar findings with sulfate, thiosulfate, and sulfur-related enzymes being reduced after IBMT. Many previous studies have demonstrated the positive association between high sulfide concentrations and cognitive dysfunction, atrophy, dementia, and Alzheimer’s disease71,72. As a result, the IBMT-induced reduction in sulfur metabolites may suggest protective effects on cognitive function.

As an active control for IBMT, the RT group demonstrated no significant within-group metabolite changes after 10 sessions of the active intervention. Although RT shares several techniques, such as body relaxation and mind calmness, IBMT highlights and targets a mindfulness component and strengthens body-mind interaction that may be more effective in holistic health, which may explain the lack of RT metabolic effect we observed in this study2,3,4,11,73.

Comparing the groups at post-intervention (IBMT-T2 vs. RT-T2), we found significant differences in metabolite levels; for example, there was a similarly greater impact of IBMT than RT on glycine derivative concentrations, including a glycine glycosphingolipid compound (N-glycine glucosylsphingosine). This result is consistent with our previous studies that IBMT outperformed RT in the brain physiological and behavioral improvement2,10,11,12,13,14,15,16, suggesting the superior effect of IBMT on the host metabolism. In addition to the cognitive benefits of glycine, it has been shown to possess other whole-body protective effects through anti-inflammation, immunomodulation, and the stabilization or hyperpolarization of the plasma membrane through the activation of chloride channels69. Additionally, IBMT also appeared to have a significant impact on methionine through its reduction compared to RT. Increased levels of the amino acid methionine result in elevated homocysteine concentrations and cardiovascular disease (CVD) risk74. The Pathway analysis found between-group increases in glutamate and histidine post-IBMT intervention. Previous studies have shown histidine treatment to be neuroprotective via astrocyte migration and also reduce insulin resistance through inflammation suppression75,76. In addition, prominent decreases in pyruvate and lactate with an increase in fumarate production may signify enhanced energy production via the TCA cycle, which could have anti-cancer effects as cancer cells prefer to exchange glucose-derived pyruvate into lactate (anaerobic glycolysis)77,78,79,80.

IBMT, in comparison with RT, also had a significant influence on phospholipid and sphingolipid metabolism while producing lower levels of circulating fatty acids, which may contribute to its overall reported health benefits. As a substrate of phospholipid metabolism, a diacylglycerol (DAG) compound was significantly reduced after IBMT, as depicted in Fig. 7. DAGs play a large role in phospholipid production, activation of protein kinase C (PKC), and cell signal transduction across the plasma membrane81. PKC has been largely implicated in cancer, CVD, T2D, and neurodegeneration due to its integral role in cell survival and death82. IBMT also resulted in alterations of phosphatidylcholine (PC) and phosphatidylethanolamine (PE) derivatives as well as reductions in phosphoglycerol (PG) and α-linolenic acid. PC and PE comprise most of the outer and inner cell membrane phospholipids, respectively, and are critical in membrane integrity and function, and thus have a vital role in metabolic health83. In addition, increased phosphatidylserine (PS) and reduced sphingomyelin (SM) concentrations in the IBMT group are notable due to their presence in myelin84, consistent with previous findings which showed that IBMT significantly improved brain white matter plasticity through myelination2,11,15,16. A review and meta-analysis indicated that increased PS levels have demonstrated improvements in memory and support cognitive function85. Moreover, elevated SM levels have been widely associated with atherosclerosis, CVD, insulin resistance, and T2D86,87. Pathway and Enzyme Enrichment analyses also identified the effect of IBMT on another sphingolipid with an increase in N-acylsphingosine. These bioactive compounds act on specific protein targets and facilitate neuronal plasticity88, but imbalanced compositions can lead to disrupted functioning and neurodegenerative disorders89. The analyses also revealed lower levels of linoleate after IBMT, the ester of inflammatory linoleic acid, and significant reductions in estimated enzymatic activity associated with linoleic metabolism at post-IBMT intervention90. In summary, with baseline as the covariate, a notable disparity in the efficacy of the two interventions was observed, with IBMT showing positive impacts on numerous metabolites associated with whole-body health, potentially signifying its holistic physiological effects4,12,55.

IBMT shares key components of mindfulness-based interventions with other mindfulness techniques and produces beneficial effects on physical and mental health2,3,11,56. However, a recent review suggests that different static meditation techniques, including mindfulness, have different effects on blood lipid levels91. To our knowledge, this is the first RCT to demonstrate the effects of a brief mindfulness intervention on the metabolome in a normal life setting. Previous studies explored metabolic changes following combined interventions such as a long-term vegan diet and intensive mindfulness retreat in an isolated environment, or a mixed intervention with several other non-mindfulness techniques51,52,53. Further investigations into whether and how various mindfulness-based interventions or certain intervention components produce common and/or unique metabolic changes are warranted53,91,92.

Finally, some limitations of this study must be acknowledged. As a pilot trial, a small sample size was used, and additional sampling would provide better estimates of metabolite trends. Future studies with larger sample sizes and metabolomic data will elucidate other potential benefits and validate the observed findings. Prior studies suggested that more females participated in mindfulness-based interventions6,93, which may reflect individual preferences in response to mindfulness interventions57. Our sample had a similar distribution, thus it’s impossible to study the potential gender differences following interventions. Future trials are needed to address these questions. Similar to other studies51,52,53, this pilot study did not measure the exact amount of daily food intake and calories during the interventions. However, the participants were focusing on mindfulness experiences and were not asked to alter dietary habits, and in addition, we used a repeated-measure experimental design (T1 vs. T2 from each participant) and ANCOVA to adjust for the baseline51,52,94. Therefore, our findings can be attributed to the mindfulness intervention. Taken together, future research should explore and elucidate specific mindfulness techniques, practice durations, and the demographic factors that may influence metabolic profile benefits following interventions.

Conclusions and future directions

In this study, we used an untargeted metabolomic approach to demonstrate the effects and pathways of a brief mindfulness intervention on metabolic profiles. Our results indicate that, compared to RT, brief mindfulness IBMT may improve physical and mental health through reprogramming amino acid and lipid metabolisms, particularly through many glycine derivatives, methionine, ceramides, and phospholipids. Further analyses identified histidine, sphingolipid, essential fatty acid, and energy-producing pathways as well as estimates of altered enzymatic activity that could support these positive health outcomes. Our findings suggest the promise of brief intervention in regulating metabolic factors contributing to health and well-being. This novel work may inform future mindfulness-based intervention recommendations on acute health-related problems and chronic diseases using larger-scale trials to further elucidate the underlying mechanisms and treatment effects.