Acute response and neurophysiological mechanisms of mindfulness meditation in emotional eating

Abstract

Introduction

Emotional eating often leads to adverse health outcomes. Understanding how mindfulness influences emotional eating is crucial for developing strategies to promote healthier eating behaviors and well-being. This study aims to investigate the mechanism and acute response to a mindfulness meditation practice on emotional eating by examining immediate post-intervention effects on brain activity and neurophysiological differences between individuals with and without emotional eating.

Methods

Forty-nine adults were recruited and allocated into two groups based on the emotional eating subscale score of the Dutch Eating Behavior Questionnaire: emotional eating (EE group) and non-emotional eating (comparison group). EEG was captured from three 14-min audio-guided mindfulness practices (MPs; each comprising three consecutive phases: mindful breathing [MPbreathing], emotion evocation [MPevocation], and meditation [MPmeditation]) and 5-min resting periods before and after each mindfulness practice. We compared between-group and between-condition differences in the EEG band power and examined potential correlations between EEG data and other participant characteristics.

Results

Following three MPs, the EE group exhibited a higher frontal midline theta power (p = 0.051) and a lower frontal midline alpha power (p = 0.059) compared with the comparison group. These effects primarily originated from the MPmeditation (theta band: corrected ps ≤ 0.03; alpha band: corrected ps ≤ 0.08). Two negative correlations were specifically found in the EE group: one between changes in frontal midline delta power and emotional eating behaviors, and another between beta-1 power and cognitive reappraisal that emerged in the MPmeditation.

Conclusions

Our findings suggest that the modulation of frontal midline theta and alpha activity may contribute to the immediate effects of mindfulness meditation in remediating emotional eating. The clinical significance of the correlation between delta activity and emotional eating severity warrants further investigation.

Plain language summary

We compared two groups of adults: those who often eat emotionally and those who do not. Both groups took part in three mindfulness practice sessions, and we measured their brain activity before, during, and after each session. We found that people who struggle with emotional eating showed different patterns of brain activity during meditation compared to those who do not eat emotionally. Specifically, certain brain signals linked to emotional processing were stronger in the emotional eating group during meditation. However, these brain changes did not directly relate to how much people reported emotional eating. Our findings suggest that mindfulness may affect the way the brain processes emotions in people who eat emotionally. This could help guide new ways to support people in managing emotional eating and improving their overall well-being.

Introduction

Emotional eating is a behavioral response to subjective emotional states, more frequently triggered by negative than positive emotions, and often presents as an increased food intake without physiological hunger [1]. This maladaptive behavioral response affects up to 44.9% of overweight and obese populations and 36.6% of people with normal weight, resulting in an increased risk of binge eating disorder, cardiovascular diseases, and cancer [2, 3]. The interaction between eating and emotion is complicated, has both physiological and psychological implications, and involves different mechanisms or pathways. Unveiling specific underlying mechanisms is crucial for developing targeted strategies to remediate emotional eating more effectively.

Mindfulness is defined as the awareness that arises when intentionally paying attention to the present moment non-judgmentally [4]. Mindfulness can be cultivated through the systematic practice of mindfulness meditation, which is one of several methods to develop mindfulness [5]. In recent years, mindfulness-based interventions (MBIs), which involve diverse mindfulness practices such as mindfulness meditation, have shown promise in reducing emotional eating [6]. Psychologically, MBIs can enhance emotional regulation skills by adaptively accepting uncomfortable emotions instead of turning to food for comfort [7, 8]. Physiologically, MBIs may help people focus on body sensations, restoring the body’s natural ability to respond to natural cues of hunger and satiety [9]. However, only a few studies have employed electroencephalography (EEG) to provide insight on how mindfulness meditation influence emotions [10, 11]. Among them, one study with a relatively long intervention time (8 weeks) found MBIs enhanced frontal EEG asymmetry during emotional challenge, underscoring the linkage between MBIs, the frontal area, and emotion regulation [11]. On the other hand, the importance of the frontal area in regulating emotional eating was supported by previous EEG studies. For example, individuals with high-level emotional eating behavior showed food-evoked event-related potentials (ERPs) over the right frontal regions that were modulated by emotional state [12]. Another ERP study found greater N2 activation of the mediofrontal area, indicating more impaired conflict processing, and elevated anxiety levels contribute to increased emotional eating behavior [13]. However, little is known about the underlying neurophysiological mechanisms that mindfulness meditation draws on to improve emotional eating in sub-clinical populations.

Given that overlapping brain areas among mindfulness meditation, emotional eating, and emotional regulation were localized to the frontal midline area, we conducted a pilot study using EEG to record brainwave activity in this region in order to investigate the neurophysiological mechanisms of mindfulness meditation on emotional eating [14]. The study objectives were to (1) examine the acute response to a mindfulness meditation practice on emotional eating, (2) elucidate the neurophysiological effects of mindfulness meditation on emotional eating by comparing differences in EEG band power between individuals with emotional eating and those without emotional eating, and (3) explore potential correlations between frontal midline EEG band power and anthropometric measures, eating behavior, food craving, positive and negative affect, and emotional regulation.

Materials and methods

A quasi-experimental, non-randomized, parallel-group comparison design trial was conducted. The study was registered at the Hong Kong University Clinical Trial Registry (HKUCTR-3034) and obtained ethical approval from the Hong Kong Polytechnic University (HSEARS20240507003). Informed consent was obtained from all participants prior to conducting the study in accordance with the Declaration of Helsinki principles.

Participants

Forty-nine eligible volunteers (age: 24.80 5.90 years, 11 males and 38 females) were recruited from the Hong Kong Polytechnic University through email advertisements and posters distributed throughout the campus. Inclusion criteria: (1) over 18 years of age, (2) had completed the online emotional eating subscale of the Dutch Eating Behavior Questionnaire (DEBQ) in English, and (3) could read and understand English. Exclusion criteria: (1) a history of acquired brain injury or (2) bariatric surgery (Wong et al., 2020), (3) a BMI < 18.5 [15], (4) a history of an eating disorder (e.g., anorexia, bulimia, or binge eating disorder) or (5) a severe mental disorder (e.g., psychosis, major depressive disorder, schizophrenia) [16], (6) currently receiving pharmacological interventions, psychotherapy, or treatments for chronic illness (e.g., diabetes mellitus) [17].

Sample size estimation

The sample size was calculated using the G*Power 3.1 software. At least 20 participants for each group (40 participants in total) were needed when considering a small effect size of Cohen's d = 0.3, a power of 95%, and a significant level of α = 0.05. Given an attrition rate of 10%, 22 participants for each group (44 participants in total) were required.

Assignment method

Allocation of participants was based on a standardized cut-off score of 2.8 on the emotional eating subscale of the DEBQ (English version, digital administration) [18]. This 10-item subscale examines whether eating behavior is triggered by negative emotions instead of internal signals of hunger and satiety; higher scores indicate greater frequency of eating when experiencing negative emotions (1 = never to 5 = very often) [19]. Twenty-four participants with scores ≥ 2.8 on the emotional eating subscale of the DEBQ were classified as exhibiting emotional eating and allocated to the EE group, while the others (n = 25) were assigned to the comparison group (Fig. 1).

Fig. 1.

Study flowchart

Intervention

Mindfulness meditation practice

A previous study suggested 10- to 20-min mindfulness meditation practices were effective for people with emotional eating [20]. Therefore, 14-min mindfulness meditation practice (MP) was applied using an audio track that was prepared in English by a certified mindfulness instructor using a trauma sensitive mindfulness approach [21]. Given that emotion evocation used in the current MP poses potential risks, an orientation in a trauma sensitive approach was given before the practice to ensure participant’s safety, to reduce the risk of being emotionally overwhelmed, and to make them feel in control of their own MP [19]. Participants were instructed to evoke a different emotion or stop the experiment immediately when they felt uncomfortable or aware that their emotions were overwhelming. The audio-guided MP consisted of three consecutive phases: (1) a short mindful breathing phase (MP_breathing, ≈ 2 min) for relaxing and preparing participants for subsequent phases; (2) an emotion evocation phase (MP_evocation, ≈ 4 min) for participants to recall a past situation or event in which an emotion cause eating in the absence of physical hunger or overeating; (3) a meditation phase (MP_meditation, ≈ 8 min) for processing emotional eating. MP was performed three times with 5-min intervals using the same audio track. Prior to the intervention, participants were briefed on the meditation duration, break schedules, and procedures for communicating with the researchers (e.g., to report physical or mental discomfort or adjust seating). They were instructed to remain still and keep their eyes closed during EEG-recorded meditation sessions. Those reporting consistent daily practice (three times or more per week) were classified as experienced, while others were considered beginners, in which we also inquire about their prior experience (frequency and duration). Participants were also asked about prior mindfulness practice. Those reporting consistent daily practice (three times or more per week) were classified as experienced, while others were considered beginners in which we also inquire about their prior experience (frequency and duration).

Experimental procedure

The three rounds of MP were performed while recording EEG. A hunger visual analogue scale (100 mm) was used to identify participants’ hunger level before the resting-state EEG of the first session, those feeling hungry (scoring ≥ 40 mm) were rescheduled for data acquisition on an alternative day. This step was necessary to ensure the quality of the emotion evoked during the MP_evocation [22]. Participants who scored < 40 mm would proceed to complete all measures. For engaging in a variety of emotions, participants were informed to recall different emotions associated with eating in the absence of physical hunger or overeating in each round of MP. Participants were asked to identify the specific emotion that they evoked after each round of MP.

During the process, participants were seated on a comfortable chair in a quiet, light-attenuated, electromagnetically shielded room; they were instructed to keep their eyes closed and muscles relaxed, minimize body movements, and stay awake during EEG recording. The EEG signal was recorded using a 64-channel Quik-Cap with the SynAmps DC amplifier and Curry 8 (Compumedics Neuroscan, USA). The sampling rate was set at 1 kHz, and electrode impedance was maintained below 10 kΩ. Three experimental sessions were conducted, and participants were allowed to stand up and stretch between them. In each session, the 5-min eyes-closed resting-state EEG was recorded pre-MP additional to the EEG data of the 14-min eyes-closed MP to investigate condition-induced changes in neural activity. In the last experimental session, one more 5-min eyes-closed resting-state EEG data was recorded post the MP to examine the after-effects of three consecutive MPs (Fig. 2).

Fig. 2.

Experimental procedure

Questionnaires and anthropometric measures

All questionnaires (English version) were delivered via emails prior to the first experimental session and managed using REDCap. Four online questionnaires were administered after group assignment, including: the Salzburg Emotional Eating Scale (SEES), the Salzburg Stress Eating Scale (SSES), the Food Craving Questionnaire-Trait-reduced (FCQ-T-r), and the Emotional Regulation Questionnaire (ERQ), whereas anthropometric data (body weight, height, and BMI () and the Positive and Negative Affect Schedule (PANAS) scores were obtained on the experimental day before the first EEG recording. The body weight and height were the mean of three trials; the mean values were used for calculating the BMI. All these measures were conducted before proceeding to the first session only.

Salzburg emotional eating scale (SEES)

The SEES is a 20-item self-reported questionnaire consisting of four subscales (happiness, sadness, anger, and anxiety) that shows how eating behavior is affected by different types of emotions [23]. The mean scores of 20 items were calculated; people with a higher mean score on the SEES have a stronger tendency to engage in emotional eating.

Salzburg stress eating scale (SSES)

The SSES is a 10-item self-report questionnaire that depicts different stressful situations to show how eating behavior changes in response to stress. The mean scores of 10 items were calculated; people with a higher mean score on the SSES may eat more than usual when they feel stressed [24].

Food craving questionnaire-trait-reduced (FCQ-T-r)

The FCQ-T-r is a short version of the food craving questionnaire-trait, which measures the frequency and intensity of food craving [25]. It contains 15 items with excellent internal consistency (α = 0.94) in the English version [26]. Total scores of the FCQ-T-r were calculated; people with a higher total score have higher probability of experiencing food cravings.

Emotion regulation questionnaire (ERQ)

The ERQ is a 10-item questionnaire designed to measure a respondents' tendency to use two emotion regulation strategies: cognitive reappraisal (an adaptive manner involves reinterpreting an emotional situation to alter its impact) and expressive suppression (a maladaptive manner involves inhibiting ongoing emotion-expressive behavior) [27]. The mean scores of each factor were calculated; higher scores indicate more frequent use of either strategy.

Positive and negative affect schedule (PANAS)

The PANAS is a 20-item self-report questionnaire that tests two primary dimensions of mood and emotion (i.e., positive or negative affect) [28]. The sum scores for positive and negative affect were calculated separately.

EEG analysis

EEGLAB v2024.2 and MATLAB_R2024b were used to analyze EEG data. Initially, the continuous EEG data of the 14-min MP were divided into three individual datasets for condition-specific analysis according to the duration of each audio track: MP_breathing (0–110 s), MP_evocation (110–317 s), and MP_meditation (317–780 s), using pop_select function of EEGLAB. Because the audio instructed participants to stretch and leave the meditative state, the last seconds (780–840 s) of MP were excluded from analysis. Therefore, EEG data for all four conditions were obtained for subsequent preprocessing of EEG signals. The raw EEG signal was downsampled to 250 Hz, followed by visual inspection to remove data that contained considerable movement artifacts. After that, band-pass filters within the frequency range of 1–40 Hz were applied to the data. Other artifacts were detected and removed based on the results of an independent component analysis (ICA, EEGLAB: runica algorithm). The cleaned EEG data were subsequently referenced to a common average. Rejected channels were interpolated back using spherical interpolation. EEG spectral power values within bins of 0.5 Hz were obtained using Fourier decomposition of data epochs with the mtmfft method [29]. The frontal midline (F3-Fz-F4) power values within the delta (1–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta-1 (12–16 Hz), and beta-2 (16–30 Hz) bands were normalized to the relative percentage of the average power over the five bands [30, 31]. Data were extracted from the frontal midline area because it is a region where cortical changes induced by mindfulness meditation, emotional eating, and emotional regulation converge. A custom-made MATLAB script modified based on a tutorial of the Fieldtrip toolbox was used to calculate the relative power values. Finally, we averaged three EEG data of the same condition (i.e., resting before each MP round, MP_breathing, MP_evocation, and MP_meditation) at each frequency band to improve the signal-to-noise ratio. Therefore, the entire 14-min MP was analyzed as three phases/conditions to identify phase-/condition-specific effects.

Statistical analysis

The Statistical Package for Social Sciences, version 22, Chicago, IL, was used to analyze the data. The Shapiro–Wilk test was used to assess the normality of all variables. Between-group differences were analyzed using the independent t-test, Mann–Whitney test, or Fisher’s exact test. Baseline EEG differences were determined by the first resting-state EEG data of the first session. The immediate after-effects of MP were tested by comparing the last resting-state EEG data of the last session between two groups. In contrast, two-way repeated measures analysis of variance (rmANOVA) was conducted to test the effects of condition, group, and condition-by-group interaction using the mean frontal midline EEG band power calculated from the three sessions. If any significant results were found in rmANOVA, post hoc pairwise comparisons using paired t-tests were performed to examine EEG band power differences across conditions. Spearman’s correlation analyses were performed to test the potential association between mean frontal midline EEG band power in specific frequency bands and other participant characteristics (anthropometric measures, eating behavior, food craving, positive and negative affect, and emotional regulation). The statistical significance level was set as p < 0.05 and corrected as p < 0.0083 (0.05/6) for the paired t-test.

Results

This study aimed to elucidate the acute neurophysiological response to mindfulness meditation practice (MP) in individuals with emotional eating (EE). The principal finding was a distinct, though marginally significant, pattern of post-intervention resting-state activity in the EE group, characterized by increased frontal midline theta (p = 0.051) and decreased alpha power (p = 0.059), suggesting a unique neurophysiological signature. This effect was specific to the meditation component of the practice (theta band: corrected ps ≤ 0.03; alpha band: corrected ps ≤ 0.08). Furthermore, during active meditation, the EE group demonstrated stronger correlations between beta-band activity and emotion regulation strategies. Notably, however, no significant Condition × Group interaction was found in the rmANOVA (ps > 0.220), indicating the overall pattern of EEG change during the practice was not statistically distinct between groups. Also, no significant interaction was found in rmANCOVA that included positive affect as a covariate (ps ≥ 0.050). Correlation analyses also revealed that the EE group lacked the significant positive associations between low-frequency band power and positive affect that were present in the comparison group. Across the full sample, meditation-specific correlations were identified, including a negative association between alpha power and food craving. Furthermore, based on self-reported prior experience, the majority of participants had either no prior mindfulness practice or had only practiced on one or two occasions. No participants met the criteria for an experienced practitioner (daily practice), with most participants reporting not having any experience practicing mindfulness or having only a trial session. Consequently, all participants in this study were classified as beginners.

There was no significant difference between two groups at baseline, but participants in the comparison group had stronger positive affect than those in the EE group (p < 0.006) (Tables 1 and 2). Seventy-three percent of participants reported experiencing different emotions in each experimental session, none reported experiencing the same emotion in all experimental sessions.

Table 1.

Summary of participant characteristics

Variables	Comparison group (n = 25) mean SD	EE group (n = 24) mean SD	p
Age (years)	23.96 4.62	25.67 6.99	0.567a
Sex (F/M)	20/4	18/7	0.496b
Height (cm)	167.44 8.36	165.46 7.41	0.387c
Weight (kg)	63.95 10.93	66.82 14.95	0.741a
BMI (kg/m2)	22.70 2.73	24.37 4.84	0.337a
Emotional eating	2.78 0.40	3.25 0.54	0.002a
Stress eating	2.66 0.64	3.33 0.90	0.009a
Food craving	37.20 11.00	55.54 12.75	< 0.001a
Cognitive reappraisal	4.79 0.81	4.73 0.78	0.952a
Expressive suppression	4.43 0.85	4.15 0.95	0.415a
Positive affect	28.16 7.98	21.88 6.80	0.006a
Negative affect	12.80 2.61	14.42 3.36	0.070a

Significant results are shown in bold

^a Mann–Whitney test

^b Fisher’s exact test

^c Independent t-test

Table 2.

Resting-state relative EEG band power before and after three rounds of mindfulness meditation

Frequency bands	Time	Comparison group (n = 25) Median [IQR]	EE group (n = 23–24) Median [IQR]	p
Delta (1–4 Hz)	Pre	0.036 [0.021–0.050]	0.034 [0.024–0.050]	0.078
	Post	0.036 [0.024–0.054]	0.036 [0.025–0.059]	0.657
Theta (4–8 Hz)	Pre	0.015 [0.009–0.017]	0.015 [0.009–0.020]	0.503
	Post	0.015 [0.011–0.017]	0.018 [0.015–0.023]	0.051
Alpha (8-12 Hz)	Pre	0.054 [0.031–0.076]	0.046 [0.021–0.064]	0.322
	Post	0.045 [0.028–0.067]	0.026 [0.013–0.059]	0.059
Beta-1 (12–16 Hz)	Pre	0.009 [0.005–0.015]	0.009 [0.005–0.017]	0.865
	Post	0.009 [0.007–0.016]	0.010 [0.007–0.013]	0.975
Beta-2 (16–30 Hz)	Pre	0.003 [0.002–0.005]	0.004 [0.002–0.007]	0.516
	Post	0.004 [0.003–0.006]	0.005 [0.002–0.008]	0.451

One EE group participant did not complete the resting−state EEG recording following three rounds of mindfulness meditation practice

EEG power spectrum analysis

Three consecutive rounds of MP caused marginally significant changes in resting-state band-specific power, with increased theta (p = 0.051) and decreased alpha band power (p = 0.059) in the EE group, suggesting the neurophysiological after-effects of the MP (Fig. 3; Table 2). No significant differences were found for delta and beta band power between two groups following the three rounds of MP (ps > 0.450).

Fig. 3.

Boxplots summarize resting-state relative EEG power before and after three rounds of mindfulness meditation

Two-way rmANOVA with factors CONDITION (resting, MP_breathing, MP_evocation, and MP_meditation) and GROUP (EE group vs. comparison group) revealed significant effects of CONDITION in delta, theta, and alpha bands (ps < 0.001; Fig. 4a) and a marginally significant effect of GROUP in theta band (F_1,46 = 2.93, p = 0.094; Fig. 4b). Neither a significant CONDITION GROUP interaction (ps > 0.220) nor a significant change of the power in beta bands (ps > 0.170) was found. Specific condition effects were revealed by paired-t-tests that found MP_meditation caused a significant enhancement and attenuation in theta and alpha band power, respectively, compared with resting, MP_breathing, and MP_evocation conditions (ps ≤ 0.001; Fig. 4b). In addition, increased delta band power was observed in each phase of the MP compared to resting (ps < 0.001).

Fig. 4.

Condition comparisons of mean relative EEG band power across three rounds of mindfulness meditation. a Within-group comparison. b Between-group comparison. EEG power of the last resting stage was excluded here as it cannot be seen as baseline data according to the mindfulness meditation practice. Data was shown as mean standard error. *** Bonferroni-corrected p < 0.001, ** Bonferroni-corrected p < 0.01

Correlation analysis

Correlation results between participant characteristics and EEG activity during MP were summarized in Fig. 5. In addition, we found the change of delta band power (the last resting-state EEG data minus the first resting-state EEG data) was negatively correlated with baseline emotional eating scores (ρ = -0.481, p = 0.020). The changes of other frequency band power were not significantly correlated with the eating scale scores.

Fig. 5.

Spearman’s correlation between average EEG band power and interest behavioral and anthropometric measures. ** p < 0.01, * p < 0.05

Full sample

Solely in MP_meditation, alpha band power correlated with food craving (ρ = -0.285, p = 0.047) and beta band power correlated with emotion regulation (beta-1 expressive suppression: ρ = -0.290, p = 0.043; beta-2 cognitive reappraisal: ρ = -0.317, p = 0.026). In contrast, alpha band power was significantly and positively correlated with cognitive reappraisal in three phases of the MP (rs > 0.28, ps < 0.05). Theta, alpha, and beta-2 band power were consistently correlated with positive affect across conditions (ps < 0.05), while delta band power and positive affect were correlated in MP_evocation and MP_meditation (ps < 0.01).

EE group

The correlations between beta-1 band power and emotion regulation were established in MP_meditation (cognitive reappraisal: ρ = -456, p = 0.25; expressive suppression: ρ = -0.481, p = 0.017). Similar to the result of the full sample, alpha band power was positively correlated with cognitive reappraisal (ps < 0.05) in MP_breathing and MP_evocation. However, the significant correlations between low-frequency band power and positive affect did not exist in the EE group (ps > 0.1). In contrast, the beta-2 band power was moderately correlated with positive affect in MP_evocation (ρ = -0.448, p = 0.028) and MP_meditation (ρ = -0.420, p = 0. 041).

Discussion

This EEG study investigated the acute response to mindfulness meditation and its neurophysiological mechanisms on emotional eating. The resting-state results revealed that the EE group exhibited increased theta-band activity, while showing decreased alpha-band activity, in the frontal midline area post the three rounds of MP. These effects primarily resulted from MP_meditation, rather than from MP_breathing and MP_evocation. The associations between beta-band activity and emotion regulation, which emerged during MP_meditation, were stronger in the EE group compared to comparison group. However, such statistically significant correlations between positive affect and the activity of low-frequency bands disappeared in the EE group. While the observed trends in theta and alpha band activity align with our hypothesis, the marginal statistical significance of these group differences, particularly the lack of a significant interaction effect in the rmANOVA, warrants careful interpretation. Despite this, the study has the potential to make an important contribution by providing preliminary evidence of the acute neurophysiological impact of mindfulness meditation in a population with emotional eating behaviors, an area previously underexplored.

EEG power spectrum analysis

Theta rhythm (4–8 Hz) is dominant in the frontal cortex, serving as a key role of mindfulness meditation known to improve cognitive processing, attention, and emotional regulation [32]. In line with previous studies, we found a stronger frontal midline theta power in MP_meditation than in other conditions [33, 34]. The increase in theta activity suggests the engagement of the limbic system, anterior cingulate cortex, and prefrontal cortex in emotional processes [35]. Literature review suggested that an increase in theta band activity was associated with emotional stimulus and information processing, which were affected by arousal levels [35]. Greater theta synchronization has been found in participants with sensitive or emotional experiences [36]. However, we did not observe significantly increased theta power during MP_evocation.

Alpha rhythm (8–12 Hz) correlates with relaxed states and passive attention [32]. Delta rhythm (0.5–4 Hz) relates to the orientation of attention to internal processes during wakefulness [37]. The changes of alpha and delta activity in response to mindfulness meditation are not consistent [34]. Previous studies supposed that alpha and delta activity vary depending on different meditation components (e.g., focused attention, open monitoring, transcendental meditation, and loving-kindness meditation) and signal processing approaches used [34, 37, 38]. One study proposed categorizing meditation as relaxation (associated with parasympathetic activation) or arousal (linked to sympathetic activation) [39]. Relaxation meditation appears to be related to lower alpha power [39]. Thus, the decreased theta power may be partially due to relaxation properties of MP_meditation.

Murphy and colleagues (2020) discovered that alpha suppression was associated with proactive control in mitigating emotional distraction [40]. Harmony [41] proposed the hypothesis that increased delta activity can help inhibit interference relevant to the increase of internal concentration [41]. Together, these findings support a potential mechanism for our study–that decreased alpha activity and increased delta activity may involve detachment from emotions elicited through visualization. A potential explanation for the increase in delta-band power during MP_breathing is that beginner practitioners are unfamiliar with mindfulness meditation and may lapse into ‘mind-wandering’. However, we cannot distinguish whether higher delta power may be attributed to different components of the MP or ‘mind-wandering’ during mindfulness meditation [36]. On the other hand, one study discovered that emotional stimulation can induce delta and theta synchronization, particularly in more sensitive participants, which may partially explain the increased delta power in MP_evocation [42].

Beta rhythm (12–35 Hz) is crucial in cognitive and affective processing. However, most participants in this study are beginners in MP, whose brain rhythmic activity may be less responsive to the MP, especially for beta band [43, 44]. In addition, the effects of mindfulness meditation on beta activity are inconsistent across previous studies [45], highlighting a need for further investigation.

Correlations between behavioral assessments and EEG powers

Our result showed the changes of delta power were significantly correlated with baseline emotional eating scores. This may indicate that individuals with more severe emotional eating require greater inhibitory control over emotion-related areas for emotional processing [46]. Further studies should investigate the origin of this correlation as it could be due to stronger evoked emotions or a stronger response to mindfulness meditation.

Positive emotions were correlated with low-frequency EEG activities in most conditions, mainly driven by the comparison group. Research has shown that people who have a higher level of positive emotions are less easily distracted [47]. Furthermore, increased delta and decreased alpha activities might be related to distraction inhibition. This finding may explain why individuals in the comparison group who exhibit higher level of positive emotions show lower delta activity and higher alpha activity during mindfulness meditation. This finding also suggests that by modulating delta and alpha activity in individuals with emotional eating and a lower level of positive emotions, we could potentially improve their engagement in mindfulness meditation.

The correlations between emotion regulation strategies (cognitive reappraisal and expressive suppression) and beta-1 activity were observed during MP_meditation in the EE group. This finding implies mindfulness meditation affects both emotion regulation strategies in a non-specific manner by modulating beta activity. However, multiple sessions specifically increase cognitive reappraisal while decrease expressive suppression [48, 49]. Taken together, with repeated practice, other mechanisms may become integrated, refining the regulation strategies, or the beta activity may become more specialized to modulate a specific strategy. However, previous findings regarding the correlation between beta power and emotion regulation strategies or positive affect are not consistent [50–53]. This inconsistency may be due to the absence of subdivision within the beta band.

Limitations

The study has several limitations. Firstly, the sex imbalance in the sample (i.e., more females [n = 38] than males [n = 11]), the quasi-experimental design, and the convenience sampling (recruitment via advertisements) limit the generalizability of the findings. Secondly, relying solely on EEG signals without incorporating other physiological data (e.g., fMRI or galvanic skin response) may weaken the comprehensive interpretation of emotion regulation mechanisms. Thirdly, novice participants are more prone to mind-wandering. Considering that repeated MP has been found to reduce mind-wandering and lead to greater EEG changes, it is recommended that future research include multiple MP sessions and questionnaires to track responses during mindfulness meditation practice and enhance the understanding of the underlying mechanisms [36, 43, 44]. Additionally, it is suggested to integrate EEG with other physiological measures to provide a more comprehensive assessment of the neurophysiological mechanisms of mindfulness meditation, thereby delineating the clinical utility of mindfulness meditation in managing emotional eating. Although prior mindfulness practice reported by participants before the intervention was explored, the classification relied on self-reported mindfulness practice. This is a limitation because the effects of mindfulness would arguably differ between individuals who are less versus more experienced. A deeper exploration of the level of mindfulness practice at the recruitment phase is advised. Finally, the interpretation of group differences in theta and alpha activity is limited by their marginal statistical significance, highlighting the need for future research with greater power.

Conclusion

The sub-clinical cohort involved in this study exhibited decreased alpha-wave activity and increased theta-wave activity in the frontal midline region following mindfulness meditation. These changes, which primarily originate from the meditation component of the practice, may contribute to the immediate effects of mindfulness meditation. The correlation between delta-wave activity and the severity of emotional eating needs further investigation. The correlation between mindfulness and beta-band activity may be changed after multiple sessions. The observed effects in the present study were transient and primarily attributable to a single session of mindfulness meditation practice. Therefore, the conclusions of this study are based on acute effects.

Author contributions

Z.R.R., H.R., L.H.Y., H.H.Y., L.K.S., H.K.Y., and K.R. collected data and wrote the main manuscript text. Z.R.R. conducted data analysis and prepared tables and figures. S.V.D.I. and Z.J.J. provided conceptualization, methodology, and guidance. All authors reviewed the manuscript.

Funding

The study was supported by the NSFC Young Scientists Fund (82402987) and the Start-up Fund for New Recruits, PolyU (P0059567) to JZ.

Availability of data and materials

Data available upon request.

Declarations

Ethics approval and consent to participate

The study was registered at the Hong Kong University Clinical Trial Registry (HKUCTR-3034) and obtained ethical approval from the Hong Kong Polytechnic University (HSEARS20240507003). Informed consent was obtained from all participants prior to conducting the study in accordance with the Declaration of Helsinki principles.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interest.