Probing the Effect of Mindfulness Training on Glutamate/GABA Ratio and Its Association with Dopamine Availability: A Preliminary Single-arm Pilot Study

Abstract

Objective

Mindfulness is an effective non-pharmacological intervention for mental health. However, little is known about the underlying neurological mechanisms involved, especially glutamate, γ-aminobutyric acid (GABA), and the excitation-inhibition (E-I) balance. This study examined changes in these markers after mindfulness training and whether these changes are related to the baseline dopamine transporter (DAT) availability. We also probed the associations with outcome markers of mindfulness-based stress reduction (MBSR) training.

Methods

In this single-arm pilot study, 11 participants with Graves’ disease were enrolled and received MBSR training. Single-photon emission computed tomography study using [99mTc]TRODAT-1 was employed to assess DAT availability before the intervention, and ¹H magnetic resonance spectroscopy was used to assess the concentrations of glutamate, GABA, and the E-I balance. Attention and heart rate variability (HRV) were also measured before/after MBSR training.

Results

An elevated E-I balance and reduced GABA concentration after MBSR training were found. A lower baseline DAT availability was associated with a greater increase in glutamate and a greater increase in the E-I balance. A greater E-I balance after training was related to a greater frequency of HRV and better attention performance after the intervention.

Conclusion

The findings are novel. An elevated E-I balance could be the short-term effect of MBSR training, and this effect was greater among participants with a lower level of DAT availability at baseline. This finding may imply that excitation-related processes, such as neuroplasticity, could be the overlooked mechanism of mindfulness training.

INTRODUCTION

Mindfulness has been proven to be an effective non-pharmacological intervention for a wide range of mental illnesses [1,2]. A meta-analysis indicated a medium to large effect of mindfulness interventions on outcomes related to psychiatric disorders [3]. However, the neurobiological mechanism related to mindfulness could be complex. Although a previous meta-analysis indicated that the brain structure changed after meditation practice [4], a recent study with a large sample size reported no significant changes in structural parameters, including gray matter volume and cortical thickness [5]. A functional magnetic resonance imaging study demonstrated elevated activity in the anterior cingulate cortex (ACC), dorsolateral prefrontal cortex (DLPFC), and insula during an interoceptive and cognitively challenging task [6]. Other brain areas, such as the lateral orbitofrontal cortex, right middle frontal gyrus, and left caudate, were also noted [7]. Studies with functional connectivity approaches have indicated that the default mode network (DMN), salience network (SN), and frontoparietal network connectivity could be related to mindfulness interventions [8,9]. In particular, the cingulate cortex could play a pivotal role in the enhanced attention process [8].

The observed psychological (attention) and neurobiological (DMN and SN) mechanisms may imply that glutamate, γ-aminobutyric acid (GABA), and the excitation-inhibition (E-I) balance are potential underlying mechanisms [8,10]. Glutamate, a naturally occurring amino acid, serves as the most common and primary excitatory neurotransmitter in the brain. Approximately 80−90% of synapses in the brain are glutamatergic [11]. In contrast, GABA, derived from glutamate, functions as an inhibitory neurotransmitter. Higher glutamate concentrations are related to better performance on attention tasks [12], and higher levels of GABA are positively related to performance tasks [12]. However, the effects of mindfulness training on glutamate and GABA concentrations remain unclear. The balance between excitatory glutamate and inhibitory GABA, the E-I balance, is essential for optimal brain function, ranging from the regulation of neuronal firing rates at the cellular level to the stability of activity at the network level [13]. Altered E-I balance is an important phenomenon observed in patients with mood disorders [14,15]. Currently, little is known about whether mindfulness interventions alter the E-I balance. A very recent unpublished study involving transcranial magnetic stimulation (TMS) and electroencephalography (EEG) reported different E-I balance patterns between experienced mindfulness meditators and controls [16].

Moreover, mindfulness can restore the function of the reward system [17], which could be an important system related to mental health. However, currently, evidence from neuroimaging studies on the role of dopamine in the mechanism of mindfulness is scarce [18]. It is also unclear whether dopamine tone is related to the outcome of mindfulness. Previous studies have shown that specific groups identified as vulnerable due to lower levels of dopamine or decreased dopaminergic neuronal density may be more susceptible to experiencing increased dopamine levels [19,20]. This phenomenon, known as “dopamine compensation”, [21] can occur as a result of interventions or exposure to substances that enhance dopaminergic activity [19,20]. Our study also demonstrated that participants with lower dopaminergic activity benefitted more from exercise intervention [22]. These findings may imply a potential influence of dopamine on interventional outcomes.

Furthermore, the dopaminergic system mainly resides in the striatum and ventral tegmental area, and the glutamatergic systems of cortical regions, mainly in frontal regions, are interconnected and modulate each other’s effects [11,23]. Glutamatergic projections connecting these systems extend both to and from these two systems. Previous studies have reported that glutamatergic modulations in frontal regions modulate striatal dopaminergic outputs, and vice versa [11,23]. As a metabolite of glutamate, GABA also engages in crosstalk with the dopaminergic system, and there is a significant interplay between the GABAergic and dopaminergic nervous systems [24]. GABA also acts as a mediator between the dopaminergic and glutamatergic systems [25].

Another unclear marker associated with mindfulness is heart rate variability (HRV). A meta-analysis indicated that mindfulness tended to increase the root mean square of successive difference, a time domain HRV index [26]. Increased HRV has also been reported during mindfulness training [27], along with increased high frequency (HF) activity, a parameter associated with parasympathetic activity [28]. However, little is known about whether mindfulness is more positively related to sympathetic or parasympathetic activity.

Emerging studies indicate that hyperthyroidism may also be associated with depression and anxiety [29,30], which may impair the quality of life among these patients [31]. Our findings demonstrated that mindfulness-based stress reduction (MBSR) intervention could reduce anxiety and depression while improving attention among patients with Graves’ disease (GD), an autoimmune disorder that can result in hyperthyroidism [32]. The evidence based on the mechanism of mindfulness from studies with molecular imaging techniques, such as positron-emission tomography, single-photon emission computed tomography (SPECT), or magnetic resonance spectroscopy (MRS), is scarce [33]. Despite growing interest in the neurobiological mechanisms underlying mindfulness training, little is currently known about its effects on glutamate, GABA, and excitatory–inhibitory balance in the ACC. In addition, multimodal studies integrating neurochemical measures with physiological and dopaminergic markers remain scarce. Therefore, preliminary hypothesis-generating studies may help identify potential associations and inform the design of future larger-scale investigations. The primary aim of this preliminary hypothesis-generating study was to examine potential changes in glutamate and GABA concentrations and the E-I balance in the ACC, an important region of interest (ROI) related to attention, following mindfulness training. We also conducted two supplemental analyses. First, we examined whether changes in outcomes resulting from mindfulness training were associated with pre-training dopamine transporter (DAT) availability. Second, we explored the associations between the levels of glutamate and GABA (baseline, after intervention, and change) and outcome markers (depression, anxiety, attention, and HRV).

METHODS

Ethical Statement

The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2013. The Institutional Review Board at National Cheng Kung University Hospital approved the study (IRB protocol number = A-BR-105-096, date of approval = May 19, 2017) and all study participants provided written informed consent.

Design and Subjects

The study reported here is a single-arm pilot study, rather than a conventional clinical trial with placebo arm design. All subjects completed clinical assessments (Hamilton Depression Rating Scale [HDRS], Hamilton Anxiety Rating Scale [HARS], and attention test) [34,35], HRV, SPECT, and MRS before the 8-week MBSR intervention. One week after completion of the 8-week MBSR intervention, the clinical assessments, HRV, and MRS were repeated.

A total of 11 GD patients with mood symptoms (age = 43.09 ± 10.22 years; 8 females and 3 males) who signed a written informed consent form were enrolled, as illustrated in Supplementary Figure 1 (available online). A senior endocrinologist confirmed the diagnosis of GD in all participants. Prior to study enrollment, serum levels of triiodothyronine, tetraiodothyronine, and thyroid-stimulating hormone were measured, with mean ± standard deviation values of 110.50 ± 55.49 ng/dl, 7.82 ± 2.32 μg/dl, and 0.71 ± 1.02 μU/ml, respectively. Psychiatric diagnoses, including major depressive disorder and dysthymia, were assessed using the Mini International Neuropsychiatric Interview. The mean duration of GD, defined as the interval between initial diagnosis and study recruitment, was 9.12 ± 9.69 years. Among the participants, 36.4% had undergone surgery for GD, with a mean interval of 8.13 ± 10.36 years between disease onset and surgery. In addition, 81.81% of participants had received pharmacological treatment during the course of illness, and the mean duration between disease onset and initiation of medication was 3.71 ± 10.23 years. No significant differences were found in the distributions of sex, age, DAT availability, glutamate concentration, GABA concentration, E-I balance or clinical outcomes between completers and non-completers, as shown in the Supplementary Table 1 (available online). The inclusion and exclusion criteria were identical to those of our previous study [32]. The sample in the present study was a subset of that study. The clinical characteristics are shown in Table 1.

Table 1.

Clinical characteristics of the subjects

Category	Value
Sex
Male	3
Female	8
Age (yr)	43.09 ± 10.22
HDRS before mindfulness intervention	5.09 ± 3.39
HARS before mindfulness intervention	5.82 ± 6.75
HDRS after mindfulness intervention	2.64 ± 2.25
HARS after mindfulness intervention	2.73 ± 1.79
Duration of illness (yr)	9.12 ± 9.69
T3 (ng/dl)	110.50 ± 55.49
T4 (mg/dl)	7.82 ± 2.32
TSH (mU/ml)	0.71 ± 1.02

Values are presented as number only or mean ± standard deviation.

HDRS, Hamilton Depression Rating Scale; HARS, Hamilton Anxiety Rating Scale; T3, triiodothyronine; T4, thyroxine; TSH, thyroid-stimulating hormone.

A structured group program, MBSR [36], was provided for all participants. Our program was developed based on key elements of the Stress Reduction Clinic at the University of Massachusetts Medical Center [37]. The structure of this program included eight weekly classes of 2.5−3.5 hours in duration. The methods used in this program included meditation, yoga, and the enhancement of awareness in daily life. The program was directed only by a senior clinical psychologist with a PhD to prevent therapist variation. This senior psychologist has been well-experienced in conducting mindfulness programs in Taiwan for many years.

Instrument

SPECT

The participants underwent a SPECT examination with [99mTc]TRODAT-1 (a radio-labeled form of tropane derivative for the selective labeling of DAT). The imaging procedure (note 1 of Figure 1) and ROI (Figure 1A) were identical to that used in our previous studies [38,39].

Fig. 1.

(A) Typical single photon-emission computed tomography (SPECT) image. Dopamine transporter (DAT) was concentrated in the striatum (St). (B) The anterior cingulate cortex (ACC) as a region of interest (ROI) in the magnetic resonance spectroscopy (MRS). (C) Typical spectra for point-resolved spectroscopy (PRESS) (top) and Meshcher-Garwood (MEGA)-PRESS (bottom). Note 1. SPECT images were acquired and then coregistered with a T2-weighted magnetic resonance imaging scan (Signa CV-I, 1.5 Tesla; GE Medical Systems) (slice thickness: 3.3 mm) using the commercial software PMOD (PMOD Technologies). The ROIs included the St and occipital cortex (Oc). The striatal DAT availability, evaluated according to the ratio of [99mTc]TRODAT-1 binding in the striatal area ([(St–Oc) / Oc ratio]), was then derived by dividing the difference between the average activity in the St and the average activity in the Oc by the average activity in the Oc. Typical SPECT images and ROIs are shown in Figure 1A. Note 2. PRESS: echo time (TE) = 35 ms, repetition time (TR) = 2,000 ms, spectral width = 2,500 Hz, 2,048 data points, 128 water-suppressed, 16 water-unsuppressed averages, and eight excitation numbers. MEGA-PRESS: TR = 2,000 ms, TE = 68 ms, number of excitations = 8, number of scans = 128, no water suppression. Note 3. The voxels were placed in the bilateral ACC (voxel size = 30 × 25 × 25). Spectra were preprocessed using free induction decay (FID)-Appliance (https://github.com/CIC-methods/FID-A), and concentrations of neurometabolites were measured using linear combination (LC) models. Measurements with signal-to-noise ratios ≤ 10, full width at half-maximum ≥ 10 Hz, or Cramer‒Rao lower bound values ≥ 20% were considered of poor quality and were excluded from the analyses. MRS data were fitted and analyzed using LC model (version 6.3-1K; LCModel Inc.). The ROI and a typical MRS spectrum are shown in Figure 1B and 1C.

MRS image acquisition and processing

We used a 3.0 Tesla MRI scanner (GE Discovery MR750; GE Medical Systems). A three-dimensional T1-weighted inversion recovery MRI scan was performed as described in our previous study [40]. ¹H-MRS was performed using point-resolved spectroscopy (PRESS, parameters in the note 2 of Figure 1). A Mescher-Garwood PRESS pulse sequence was used for GABA analysis. Spectral processing and concentrations of neurometabolites were measured using a linear combination model. Measurements with signal-to-noise ratios ≤ 10, full width at half-maximum ≥ 10 Hz, or Cramer‒Rao lower bound values ≥ 20% were predefined as poor-quality data and would have been excluded from the analyses. If any of these criteria were not met for glutamate or GABA, the entire measurement would have been excluded from further analysis. However, in the present study, all MRS data met the predefined quality control criteria, and no scans/spectra were excluded on this basis. ROIs and typical spectra were shown in Figure 1B (ROI) and 1C (typical spectra).

Attention test

This test includes a total of 780 signals. Subjects were asked to identify 107 target signals out of the 780 trials using paper and pencil [34]. A higher d’ score indicates better processing capability. In addition, the δ’/T score included the total time consumed (in seconds) for signal discrimination. Our previous study has confirmed the validity of this test [41].

Measurement of cardiac autonomic function

Cardiac autonomic function was determined using an ANSWatch^Ⓡ wrist monitor, which calculates HRV (Taiwan Scientific Corp.), as identical to our previous study [42]. The intervals between adjacent normal R waves (NN intervals) and the quantified cyclic fluctuations in R-R were measured over the recording period. HRV was analyzed through frequency (low frequency/HF ratio) and time domain methods [43].

Statistical Analysis

Due to the limited sample size, non-parametric analyses were employed. For the main aim of the present study, the Wilcoxon signed-rank test was used to test changes in the E-I ratio, glutamate and GABA concentrations, and HRV parameters after the mindfulness intervention. Spearman’s rho correlation analysis was used to test the associations between DAT availability and the E-I ratio or between DAT availability and glutamate and GABA concentrations. The magnitude of the correlation coefficient was interpreted as an index of effect size (small: 0.10−0.29; moderate: 0.30−0.49; large: ≥ 0.50). Given the small sample size, 95% confidence intervals (CI) were estimated using bootstrap resampling. The associations between neuroimaging markers (DAT availability, the E-I ratio, and glutamate and GABA concentrations after intervention, which changed from baseline to intervention) and clinical outcomes (HDRS, HARS, attention, and changed HRV parameter) were also probed with Spearman’s rho. Considering that DAT availability could decrease with age, a supplemental partial correlation test was conducted on the correlations with DAT. The threshold for statistical significance in all analyses was set at p < 0.05. All the statistical analyses were conducted using IBM SPSS Statistics, version 22 (IBM Corp.), and the figures were plotted using GraphPad Prism 10.5.0 (GraphPad Software).

RESULTS

For the primary aim of the present study, a series of Wilcoxon signed-rank tests were conducted. Improved performance on the attention test (baseline: 0.33 ± 0.06; after mindfulness intervention: 0.37 ± 0.05, Z = −2.67, p < 0.01) and a decrease in HF% were observed (baseline: 56.00 ± 21.41; after mindfulness intervention: 38.45 ± 13.82, Z = 2.05, p < 0.05, as shown in Table 2). An elevated E-I balance (glutamate/GABA, baseline: 4.67 ± 1.30; after mindfulness intervention: 5.50 ± 0.88, Z = −2.05, p < 0.05) and a reduced GABA concentration (baseline: 2.55 ± 0.42; after mindfulness intervention: 2.25 ± 0.38, Z = 2.67, p < 0.01) are shown in Table 2 and Figure 2.

Table 2.

HRV, E-I ratio and glutamate and GABA concentrations before and after mindfulness training

	Baseline	After mindfulness training	Z	p
Attention: d’/T	0.33 ± 0.06	0.37 ± 0.05	−2.67	0.008
HRV
HRV	35.73 ± 6.93	37.36 ± 16.77	−0.04	0.965
LF%	44.00 ± 21.41	61.55 ± 13.82	−2.05	0.041
HF%	56.00 ± 21.41	38.45 ± 13.82	2.05	0.041
LF/HF	1.22 ± 1.39	2.00 ± 1.28	−1.42	0.155
LF	202.73 ± 104.66	458.18 ± 506.79	−0.85	0.398
HF	297.00 ± 236.97	348.45 ± 530.00	−0.89	0.374
RMSSD	37.82 ± 19.58	32.55 ± 26.40	1.20	0.230
PNN50	18.64 ± 20.08	14.18 ± 21.47	1.16	0.250
MRS
E-I balance	4.67 ± 1.30	5.50 ± 0.88	−2.05	0.041
GABA concentration	2.55 ± 0.42	2.25 ± 0.38	2.67	0.008
Glutamate concentration	11.65 ± 2.48	12.14 ± 1.26	−0.36	0.722

Values are presented as mean ± standard deviation.

HRV, heart rate variation; E-I, excitation-inhibition balance; GABA, g-aminobutyric acid; LF, low frequency; HF, high frequency; RMSSD, root mean square of the successive differences; PNN50, the proportion of NN50 divided by the total number of NN (R-R) intervals; MRS, magnetic resonance spectroscopy.

Fig. 2.

Changes in the E-I balance (Wilcoxon Z = −2.05, p < 0.05), glutamate (Wilcoxon Z = −0.36, p = 0.72), and GABA (Wilcoxon Z = 2.67, p < 0.01) before (visit 1) and after the mindfulness intervention (visit 2) are illustrated. The median, upper, and lower quartiles are shown in the right panel. Each line is a subject.

E-I, excitation-inhibition; GLU, glutamate; GABA, γ-aminobutyric acid.

*p < 0.05, **p < 0.01.

No significant associations were found between demographic variables (age or sex) and neuroimaging variables (DAT, glutamate, GABA, or the E-I ratio), except for a marginal correlation between an older age and a lower DAT (rho = −0.55, p ≤ 0.08). A series of correlation analyses were conducted, as shown in Table 3. Baseline DAT availability was significantly associated with baseline glutamate concentration (rho = 0.68, p < 0.05, 95% CI: 0.21, 0.92) and with changes in glutamate concentration after mindfulness training (rho = −0.81, p < 0.01, 95% CI: −0.94, −0.46). Baseline DAT availability was significantly associated with baseline E-I ratio (rho = 0.62, p < 0.05, 95% CI: 0.06, 0.91) and with changes in E-I ratio after mindfulness training (rho = −0.74, p < 0.01, 95% CI: −0.94, −0.25), as illustrated in Figure 3. The trends of these findings were similar after controlling for the effect of age.

Table 3.

The correlations between DAT availability and MRS measurements

		(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
DAT	(1)	1.00
Glutamate baseline	(2)	0.68*	1.00
Glutamate after mindfulness training	(3)	−0.34	0.19	1.00
Glutamate change	(4)	−0.81**	−0.76**	0.45	1.00
GABA baseline	(5)	−0.24	−0.02	0.08	−0.03	1.00
GABA after mindfulness training	(6)	−0.06	0.26	0.11	−0.27	0.63*	1.00
GABA change	(7)	0.30	0.39	0.07	−0.24	−0.51	0.27	1.00
E-I baseline	(8)	0.62*	0.69*	0.21	−0.48	−0.62*	−0.15	0.60	1.00
E-I after mindfulness training	(9)	−0.14	−0.10	0.62*	0.52	−0.50	−0.67*	−0.11	0.32	1.00
E-I change	(10)	−0.74**	−0.86**	0.10	0.83**	0.20	−0.33	−0.61*	−0.67*	0.28

DAT, dopamine transporter availability; MRS, magnetic resonance spectroscopy; GABA, g-aminobutyric acid; E-I, excitation-inhibition balance.

*p < 0.05, **p < 0.01.

Fig. 3.

Baseline DAT availability (x-axis) related to baseline glutamate (y-axis), changes in glutamate concentration (dGLU, after subtraction) and the E-I balance and changes in glutamate concentration (dE-I, post-training minus pre-training).

DAT, dopamine transporter; GABA, γ-aminobutyric acid; E-I, excitation- inhibition.

*p < 0.05, **p < 0.01.

HF% before mindfulness training was related to the change in GABA levels after mindfulness training (rho = 0.83, p < 0.05, 95% CI: 0.16, 1.00). HF% after mindfulness training was related to the E-I balance after mindfulness training (rho = 0.71, p < 0.05, 95% CI: −0.11, 1.00). A greater change in HF% was related to lower GABA levels before (rho = −0.74, p < 0.05, 95% CI: −1.00, −0.13) and after (rho = −0.74, p < 0.05, 95% CI: −0.95, 0.14) mindfulness training. A better attention performance after mindfulness training was related to a lower GABA concentration after mindfulness training (rho = −0.61, p < 0.05, 95% CI: −0.94, −0.33) and was marginally related to a greater increase in the E-I ratio (rho = 0.55, p = 0.08, 95% CI: 0.20, 0.88). The anxiety level assessed by the HARS after mindfulness training was related to the increase in the E-I balance after training (rho = 0.63, p < 0.04, 95% CI: 0.04, 0.87). No significant association between glutamate concentration and HDRS was found. A heatmap summarizing the abovementioned findings is presented in Supplementary Figure 2 (available online).

DISCUSSION

Our previous findings confirmed that mindfulness training reduced depression and anxiety while improving attention performance among patients with GD [32]. The present preliminary study found an elevated E-I balance after mindfulness training, and this effect was associated with baseline DAT availability. An elevated E-I balance was also related to better attention performance, and a tendency toward anxiety after mindfulness training (Supplementary Figure 2). Although exploratory, these findings may provide preliminary insights into the neurobiological mechanisms underlying mindfulness training and help generate hypotheses for future studies.

An elevated E-I balance, characterized by a reduced GABA concentration after mindfulness training, was found in the present study. This finding is not entirely consistent with the previous evidence supporting the beneficial effects of increased GABA on improving cognitive function and mental health [44]. For example, reduced GABA could be related to increased activity in the DMN, which could be associated with depression [45]. Lower levels of GABA were also found in individuals with other mental disorders, such as schizophrenia [46] and attention deficit/hyperactivity disorder [47]. However, little is known about the positive effects related to lower levels of GABA. Lim and Aquili [48] reported that GABA supplementation may reduce cognitive flexibility. Moreover, reduced GABA could be a result of mental effort. A study with MRS indicated that the concentration of GABA decreased during a 50-minute working memory task [49]. As working memory performance involves components demanding mental efforts, the reduced GABA concentration could be a result of 8 weeks of high-intensity training. It is important to note that our follow-up visit was one week after MBSR training, and we speculated that the findings in the present study could be a trait marker of the short-term effect of MBSR training. The E-I balance and concentrations of glutamate and GABA within long-term, experienced mindfulness practice are unclear.

The associations between lower baseline DAT availability and mindfulness-related changes in glutamate concentrations and the E-I index are novel. Previous animal studies have shown that populations with a low dopamine tone, such as those with various diseases such as Parkinson’s disease, exhibit more pronounced therapeutic outcomes in vulnerable groups than in control groups [20]. These therapeutic outcomes included various indicators of fitness, such as motor coordination, balance, physical activity, and limb stride length [20]. Additionally, studies conducted on animals have indicated that exercise, as a therapeutic intervention, leads to greater increases in dopamine levels and other indicators of dopaminergic neurons in animal models of Parkinson’s disease than in healthy counterparts. These include increased dendritic spine growth and neuronal density within the basal ganglia [19,20]. Our results may be broadly consistent with previous animal studies suggesting that animals with low baseline dopamine have better outcomes related to motor performance and neuron function after intervention.

Similarly, previous human research has indicated that individuals with lower cognitive performance, possibly linked to lower dopaminergic activity, exhibit greater susceptibility to benefits from exercise in terms of cognitive benefits [50,51]. Researchers have suggested that a physiological ceiling may explain better outcomes for vulnerable (i.e., psychologically or physically less fit) people. Low dopamine and glutamate levels in the brain have been associated with cognitive decline [52]. Low baseline DAT availability was associated with change in glutamate and the E-I index in the present study. These findings may provide preliminary insights into previous studies suggesting that individuals with low cognition may benefit more. It is possible that people with lower brain DAT or D2/D3 receptor availability, which previous studies have shown to be a condition associated with low cognition [53,54], may be susceptible to greater improvement because of the greater room for improvement. These preliminary findings may help generate hypotheses for future studies investigating the role of dopamine tone in mindfulness-related neurobiological changes.

The findings of a reduced HF, elevated E-I balance, and the association between anxiety and E-I balance elevation may suggest that mindfulness training involves increased neuronal excitation or arousal. Although somewhat unexpected, a recent study using HRV and EEG, similarly reported increased parasympathetic activity and elevated cortical excitability during arousal-based meditation [55]. Although MBSR training was not the intervention used in that study, enhanced attention and effortful awareness may represent shared underlying processes.

Our findings regarding the associations of glutamate, GABA, and the E-I ratio with HF%, attention, and anxiety suggest that the underlying neurobiological mechanisms could be complex and require further investigation. A human development study indicated that elevated excitability could be related to neuroplasticity and new cognitive skills [56]. Based on these preliminary findings, we hypothesize that neuroplasticity could represent a potential therapeutic mechanism of mindfulness training [57,58]. A systematic review indicated that reduced GABA and elevated glutamate after transcranial direct current stimulation, a non-invasive brain stimulation method, may enhance neuroplasticity [59]. A similar effect of repetitive TMS was also reported in a recent animal study [60]. However, other plausible explanations, such as the stress related effects, cannot yet be ruled out. It should also be noted that GABA may exert excitatory effects under certain physiological conditions [61].

Several limitations of the present study should be noted. First, and most importantly, substantial attrition occurred between initial enrollment (n = 42) and completion of all study procedures included in the final analyses. This attrition may have introduced potential selection bias, as participants who completed the full protocol could differ systematically from those who did not. Therefore, the generalizability of the present findings may be limited, and replication in larger cohorts is warranted. Second, statistical power was limited by the small sample size, which may also have increased the risk of unstable effect estimates. Third, the single-arm design without a control group limits causal inference, as the observed effects may have been influenced by natural fluctuations, repeated testing effects, or regression to the mean. Fourth, the correction for multiple comparisons was not performed because of limited statistical power. Fifth, clinical and cognitive assessments were not conducted blind to assessment timepoint, which may have introduced potential observer or expectation bias. Additionally, the HRV measurement used in this study was a compact one, not the golden standard electrocardiogram. Finally, the study sample consisted exclusively of patients with GD; therefore, the generalizability of the findings to other populations remains uncertain. In addition, variations in medication history/status and thyroid function may have confounded mood symptoms, HRV, and neuroimaging outcomes.

To conclude, elevated E-I balance could be one consequence of mindfulness training. Whether this phenomenon contributes to the therapeutic mechanisms underlying mindfulness remains to be elucidated.