Mindfulness meditation increases interictal epileptiform discharges in meditation‐naïve surgical epilepsy patients Authors: Shweta Soni, Karine Brochu, Paolo Federico, Walter Hader, Julia W Y Kam, Fady Girgis Published: 2026 ABSTRACT AbstractObjectiveMeditation has widely recognized psychological and neuromodulatory benefits, yet its effects on epileptiform activity remain unclear. This study examined whether novice meditation modulates interictal epileptiform discharges (IEDs) in patients with drug‐resistant epilepsy using intracranial electroencephalography (iEEG).MethodsTwenty‐seven meditation‐naïve patients undergoing iEEG monitoring performed two meditation conditions, focused attention meditation (FAM) and open monitoring meditation (OMM), as well as two control conditions, story listening (STL) and classical music listening (CMM). IEDs were visually identified across the brain and manually labeled by an experienced neurologist. Linear mixed‐effects models were applied to assess the effect of meditation on IED density (IEDs/min). Subjective meditative depth was evaluated using the Meditation Depth Questionnaire (MEDEQ), and associations between IED density and subjective experiences were examined.ResultsBoth focused attention (p = .027) and open monitoring (p = .041) meditation increased IED density relative to the non‐meditative STL control condition, whereas no differences were observed relative to the CMM control condition. Although both meditation conditions produced numerically comparable IED increases relative to STL control at the individual level, there was notable inter‐individual variability depending on type of meditation. Subjects overall reported higher meditation depth scores during meditation than the STL control, and perceived depth increased with progressive meditative levels. Crucially, the relationship between IED density and perceived meditation depth changed across levels (p < .01), suggesting a dynamic interplay between neural excitability in the context of epilepsy and the subjective ease of meditative engagement.SignificanceThis exploratory study provides direct intracranial evidence that meditation modulates interictal epileptiform activity in meditation‐naïve epilepsy patients. Increased IED density during meditation suggests that these internally directed cognitive states can transiently influence epileptic network dynamics. These preliminary findings indicate a functional coupling between meditative states and cortical excitability in epilepsy, offering crucial insights into the clinical relevance of meditation in epilepsy.Keywords: epilepsy, interictal epileptiform dischargs, meditation, neural excitability, seizure risk Key points We investigated the effects of meditation on interictal epileptiform discharges (IEDs) in surgical epilepsy patients. We investigated the effects of meditation on interictal epileptiform discharges (IEDs) in surgical epilepsy patients. Both focused attention and OMM increased IED density in novice meditators relative to a control condition. Both focused attention and OMM increased IED density in novice meditators relative to a control condition. Individuals showed variability in IED density across meditation types, suggesting a heterogeneous relationship between meditative states and epileptiform processes. Individuals showed variability in IED density across meditation types, suggesting a heterogeneous relationship between meditative states and epileptiform processes. IED density was higher when participants self‐reported fewer hinderances and greater relaxation and concentration during meditation, suggesting a functional link between neural excitability and subjective experience of meditative states. IED density was higher when participants self‐reported fewer hinderances and greater relaxation and concentration during meditation, suggesting a functional link between neural excitability and subjective experience of meditative states. These findings provide insights into meditation's neurophysiological impact in epilepsy and inform future evaluation of its potential as a complementary intervention. These findings provide insights into meditation's neurophysiological impact in epilepsy and inform future evaluation of its potential as a complementary intervention. INTRODUCTION Epilepsy is a major neurological condition characterized by recurrent, unprovoked seizures, affecting over 50 million people globally 1 (World Health Organization, 2024). Despite advances in pharmacotherapy and surgical interventions, about one‐third of patients continue to experience drug‐resistant epilepsy, 2 contributing to substantial disability, psychosocial burden, and premature mortality, particularly in low‐resource settings. 3 Seizure occurrence can be influenced by psychological and environmental stressors, 4 , 5 highlighting the complex interplay between these stressors and neural excitability. These challenges underscore the need for complementary and more accessible non‐pharmacological interventions, such as meditation, which has been studied widely for its psychological and neuromodulatory benefits. 6 , 7 However, its effects on seizure outcomes and underlying neural activity in patients with epilepsy remain poorly understood. Meditation encompasses diverse mental techniques aimed at cultivating attention, mindfulness, and awareness, and is typically grouped into two broad categories. Focused attention meditation (FAM) involves voluntarily directing attention to a single chosen object such as breath, whereas open monitoring meditation (OMM) involves maintaining a receptive, non‐judgemental awareness of the field of experience without focusing on any specific object. 8 Evidence suggests that meditation can modulate both subjective experience and neural activity, with consistent changes in large‐scale brain networks involved in attention and emotion regulation. 9 , 10 , 11 In healthy populations, electroencephalography (EEG) studies have revealed reliable spectral changes during meditation, including alterations in theta, alpha, and gamma power, associated with relaxation and attentional engagement. 12 , 13 , 14 These findings support meditation's neuromodulatory potential, yet it remains unclear whether such effects are observed in patients with epilepsy. Early studies investigating the relationship between epilepsy and meditation yielded mixed outcomes. Some suggested that meditation might reduce seizure frequency and improve quality of life, 15 , 16 , 17 whereas others cautioned that it could increase susceptibility to seizures, potentially due to neural hypersynchrony and paradoxical activation. 18 , 19 , 20 Most of these investigations, conducted over two decades ago and restricted to a specific technique (i.e., Transcendental Meditation), relied on scalp EEG recordings. The observed inconsistencies, therefore, may partly reflect methodological limitations of scalp EEG, including low spatial resolution and limited sensitivity to focal epileptiform discharges. In addition, self‐reported measures of seizure frequency have inherent limitations as they are subject to reporting bias, recall errors, and both under‐ and over‐reporting, 21 , 22 making conclusions difficult to interpret and undermining their reliability as an efficacy measure in anti‐epileptic treatments. Therefore, approaches that more accurately quantify seizures or related clinical phenomena can provide a more objective, temporally precise and robust marker of cortical excitability, overcoming many of these methodological limitations. Intracranial EEG (iEEG), routinely employed in the presurgical evaluation of drug‐resistant epilepsy, 23 enables direct recording of local field potentials with high spatial and temporal precision. 24 A few studies have investigated meditation using iEEG in patients with epilepsy, demonstrating changes in different aspects of the EEG signal during meditation. 25 , 26 , 27 , 28 Of note, iEEG also facilitates the detection of clinically relevant neurophysiological markers of epilepsy, such as interictal epileptiform discharges (IEDs), which are brief, pathological bursts of activity occurring between seizures, and generated by hyperexcitable neuronal populations. 29 , 30 To date, only one published case study 31 has examined the effects of meditation on IEDs and reported an increased rate of IEDs during meditation compared to a mind‐wandering control condition in a single patient with mesial temporal epilepsy. Whether such modulation of IEDs generalizes across individuals with different forms of epilepsy and whether it depends upon the type of meditation remains unknown. The present study addresses this gap by systematically examining whether meditation affects IEDs that are observed in iEEG recordings in patients with drug‐resistant epilepsy. Individuals with no prior meditation experience briefly engaged in two classical mindfulness practices, FAM and OMM, along with two control conditions, music and story listening. IEDs were quantified during each condition and statistically compared between the meditation and control conditions (Objective 1). Patients' self‐reported meditation depth experience was measured using a validated scale following each condition and its association with IED occurrence was evaluated (Objective 2). Given the mixed findings in the literature, we hypothesized that meditation would modulate IED occurrence relative to control conditions but did not predict a specific direction. Such modulations would reflect changes in clinically relevant neural activity, which may be associated with meditative depth. By integrating objective neural measures with measures of subjective experience, this exploratory work seeks to clarify the neurophysiological and clinical implications of meditation in epilepsy and inform the development of safe, accessible, and evidence‐based complementary interventions, while establishing a baseline framework for future prospective studies with larger, stratified cohorts. MATERIALS AND METHODS Participants Twenty‐seven patients (mean age = 35.70 years, range = 20–55 years, 15 female, 24 right‐handed) undergoing stereotactic EEG monitoring for drug‐resistant epilepsy in the seizure‐monitoring unit participated in the study. All participants provided written informed consent in accordance with the guidelines approved by the Research Ethics Board of the University of Calgary (REB‐20‐0598). Inclusion criteria required participants to have the ability to understand and follow instructions, perform the cognitive task, and have normal or corrected to normal vision. Although most patients (n = 15) had no diagnosed psychiatric and behavioral comorbidities, a significant subset (n = 10) had stress‐related conditions including anxiety, depression, hypertension, and migraines. Notably, no patients were experiencing active, symptomatic episodes such as a panic attack, depressive crisis, or an active migraine at the time of the experimental recordings. None of the patients exhibited cognitive deficits that impaired their ability to maintain a standard conversation while the consent form and task requirements were being explained to them, or to follow the instructions provided; nor did they report engaging in regular meditation practice of any kind prior to participation. Participant demographic details and clinical information are presented in Table 1. TABLE 1Patients’ demographic and clinical summary.Sub IDAge (in years)SexHandednessEducation (in years)Epilepsy duration (in years)Seizure‐onset zoneEtiology (MRI findings)Comorbidities diagnosed (psychiatric/behavioral)133MR1212L mTRNon‐lesionalAnxiety and depression227FR12+7L mTRPeriventricular Nodular Heterotopia (left temporal)None328MR126.5R HipNon‐lesionalNone455MR1714L HipNon‐lesionalInsomnia537FR171.3R ant. InsNon‐lesionalAnxiety, depression, and migraine627MR127.3Bi Temp (Hip)Non‐lesionalNone720FR1315.9L mTR/OFCNon‐lesionalMigraine848MR1622.9L TNCNon‐lesionalHypertension934MR127.5Bi mTRNon‐lesionalAnxiety and depression1024ML124.6R mTRNon‐lesionalNone1132FL12+15.6Bi mTRMultiple L Hip Cavernomas, DVA, and MTSLearning and memory issues1228FR16+27.8L InsNon‐lesionalNone1347MAD16+38R InsR MTSNone1421MR164.9R FTRNon‐lesionalAnxiety and depression1544FR16+30R mTRNon‐lesionalNone1639FR1429.9L HipL MTSNone1735MR1415.3L mTRL Parietal/Perisylvian PolymicrogyriaNone1832FR1310L AmyL Temp and OF Encephaloceles; localized gliosisNone1954MR1625Bi mTRBi Frontotemporal encephalomalaciaNone2042FR1324L Amy/HipNon‐lesionalAnxiety and depression2155FR16+39R FTRR AmygdalohippocampectomyMigraine and irritable bowel syndrome2228FR144R hand SMNon‐lesionalMigraine2343FR1725R InsNon‐lesionalNone2428FR1423R MotorUnknownDepression2522FR1615R FrNon‐lesionalNone2627MR126L ant. HipNon‐lesionalNone2754FR1649Bi multi‐focalR frontal and insular resectionNoneAbbreviations: AD, ambidextrous; Amy, amygdala; ant., anterior; Bi, bilateral; DVA, developmental venous anomaly; F, female; Fr, frontal region; FTR, frontotemporal region; Hip, hippocampus; Ins, insula; L, left; M, male; mTR, mesial temporal region; MTS, mesial temporal sclerosis; OF, orbitofrontal; OFC, mesial/orbitofrontal region; R, right; SM, sensory and motor area; temp, temporal; TNC, temporal neocortical region. Patients’ demographic and clinical summary. Abbreviations: AD, ambidextrous; Amy, amygdala; ant., anterior; Bi, bilateral; DVA, developmental venous anomaly; F, female; Fr, frontal region; FTR, frontotemporal region; Hip, hippocampus; Ins, insula; L, left; M, male; mTR, mesial temporal region; MTS, mesial temporal sclerosis; OF, orbitofrontal; OFC, mesial/orbitofrontal region; R, right; SM, sensory and motor area; temp, temporal; TNC, temporal neocortical region. Intracranial electrode implantation and localization As part of standard clinical care, each patient was implanted with 9 to 20 multi‐contact depth electrodes placed stereotactically to identify the seizure‐onset zone. Electrode implantation covering several cortical and sub‐cortical regions was determined solely by clinical needs. Each electrode (made of platinum plates with a diameter of 1–2 mm) consisted of 4–10 contacts of 2 mm length with a center‐to‐center gap of 5 mm (AdTech Medical instruments). Across all patients, a total of 353 electrodes comprising 2303 contacts were implanted, with 194 electrodes (1260 contacts) in the left hemisphere and 159 electrodes (1043 contacts) in the right (see Table S1 for detailed electrode information). For patients with bilateral implants, the number of contacts ranged from 10 to 83 (mean = 46.67 ± 26.21) and 24–80 (mean = 38.63 ± 28.24) in the left and right hemispheres, respectively. To determine contact locations, post‐implantation structural magnetic resonance imaging (MRI) scans were obtained from a 3 T scanner. For each patient, T1‐weighted scans were preprocessed in the Fieldtrip toolbox, aligned to the anterior and posterior commissures, and then saved in .nifti format. 32 Subsequent anatomic processing followed the iELVIS pipeline, 33 including segmentation, reconstruction of various brain structures (recon‐all), and co‐registration with the post‐implantation images using an online neuroimaging software suite (FreeSurfer, Charlestown, MA, USA; version 7.4.1) in MATLAB (Mathworks, Natick, MA, USA). Electrodes were displayed on each patient's reconstructed three‐dimensional (3D) cortical surface and then manually annotated with the Bioimage Suite software. 34 Raw electrode coordinates obtained through labeling were then imported into MATLAB for visualization onto the average 3D brain, with corrections for potential brain shift (Figure 1 shows implanted electrodes within mesial temporal structures where IEDs were observed in most participants). Finally, electrode locations for individual participants were verified by a neurosurgeon (author F.G.). Electrode visualization. (A) Location of electrodes and contacts within hippocampus and amygdala reconstructed in the standardized brain. (B) Exemplar contacts within amygdala (yellow), anterior hippocampus (orange), and posterior hippocampus (green) that showed interictal epileptiform discharges and are represented in the native brain space. Experimental procedure Participants completed a task where they engaged in two different meditation practices as separate conditions: FAM, and OMM. During FAM, participants were instructed to direct and maintain attention on their breathing. In OMM, they were asked not to anchor their attention on a specific mental object, but instead to remain aware of any thoughts, bodily sensations, or experiences arising in their awareness. Detailed and standard instructions were read aloud by the experimenter and simultaneously displayed on‐screen immediately before each meditation condition (see Supporting Information). The experimenter ensured that participants clearly understood all instructions and participants were offered opportunities to raise questions throughout the instruction period. There was no ongoing auditory guidance provided during the meditation conditions to allow participants to meditate without any external distraction. Two control conditions involving listening to a story (STL) and listening to a classical music piece (CML; Mozart's Piano K.467) were also included. In both conditions, participants were asked to relax, listen attentively, and refrain from thinking about anything in particular. The control conditions were chosen to match key aspects of each meditation condition: STL paralleled the sustained attention demands of FAM on an external object, whereas CML was intended to match the open and receptive awareness of OMM. This design allowed us to distinguish meditation‐specific effects from general attention or open awareness effects. Each condition lasted for 5, 8, or 10 min, considering participants' comfort and alertness. Each recording session, therefore, lasted for ~30–60 min. The four conditions were separated by brief periods of up to 3 min and presented in a pseudo‐randomized, alternating sequence. For the majority of the cohort (n = 24), meditation conditions (FAM/OMM) were interleaved with control conditions (STL/CML) to facilitate a functional cognitive reset and minimize order effects. In a small subset (n = 3), clinical constraints necessitated the consecutive presentation of meditation blocks. To maintain environmental consistency, recordings were performed in private monitoring rooms with restricted access, and an experimenter was present all the time for continuous monitoring of behavioral signs of exhaustion and drowsiness (i.e., head drooping or prolonged unresponsive eye closure). The clinical staff was provided with “notified that Research is in progress” beforehand to minimize interruptions, and non‐urgent clinical interventions were coordinated during inter‐condition intervals. Furthermore, the testing was scheduled during periods with no visitors and scheduled doctor rounds to ensure a quiet environment. Notably, recordings were conducted 3–4 days post‐implantation to ensure that patients reached a stable, reduced anti‐seizure medication regimen for clinical monitoring (see Table S1 for a detailed medication list). To avoid post‐ictal confounds, sessions were performed only on days when the patient had remained seizure‐free for at least 24 h. To avoid vigilance and sleep‐intrusion confounds, the recordings were strictly coordinated when patients were not sleep deprived the previous night and were performed only during waking hours. Furthermore, we collected self‐ratings on fatigue levels from all participants on a 5‐point scale (1: not tired at all, 5: completely exhausted) before the experiment began. The average fatigue score was 2.8 ± 0.85 (mean ± standard deviation) indicating their cognitive alertness and wakefulness. To validate task adherence, participants retrospectively evaluated the quality of their experience during that condition by completing a short, post‐condition questionnaire adapted from the Meditation Depth Questionnaire (MEDEQ). 35 The MEDEQ was designed to assess the depth of meditation across five progressive levels: L1 (hindrances), including difficulties such as drowsiness and distraction while practicing meditation; L2 (relaxation), capturing subjective ease and calmness; L3 (concentration), including experiences such as effortless attentive control over the mind, internal focus, or mindfulness; L4 (transpersonal qualities), including interpersonal and positive affect experiences often observed in advanced practitioners; and L5 (nonduality), characterizing states of diminished subject–object distinction and self‐dissolution, typically demanding extensive training. The meaning of each question was explained beforehand to ensure that the participants understood the question. Participants rated their experience on a 5‐point scale (1 = Not at all; 2 = Rarely; 3 = Sometimes; 4 = Often; 5 = Completely) for each level. These self‐ratings data also served as a behavioral proxy for task engagement and were used to ensure that participants were able to differentiate between the distinct cognitive demands of each condition. To ensure precise synchronization of behavioral and neural data, a photodiode attached to the presentation laptop was used to time‐lock the beginning and ending of all events within each condition, including instructions, the experimental block (meditation, CML, or STL), and the questionnaire, to the ongoing iEEG recordings. Most participants finished all conditions on the same day, whereas three participants performed the task across two consecutive days (two conditions per day, one meditation and one control). Data from four patients were incomplete due to fatigue or loss of motivation (two patients), or technical issues with the photodiode signals in between some conditions (two patients). iEEG data recording, pre‐processing, and IED annotation Electrophysiological data were recorded with a clinical video‐EEG monitoring system (Natus, NeuroWorks, Middleton, WI) at a sampling rate of 2048 Hz (12 participants) or 1024 Hz (15 participants). The raw signals were then imported into MATLAB 2024a (Mathworks, Inc.) for preprocessing and analyses using custom scripts and the Fieldtrip toolbox. 36 Following visual inspection, electrode contacts exhibiting poor signal quality due to amplifier saturation, excessive environmental noise within the hospital room, or poor contact with brain tissue were excluded. Signals were down‐sampled to 1000 Hz, high‐pass filtered at 1 Hz, low‐pass filtered at 200 Hz using Butterworth filters, and notch filtered to remove line noise (59–61 Hz) and its harmonics (119–121 Hz, 179–181 Hz). Electrode contacts contaminated by movement artifacts or large transients were manually removed from further analysis. The continuous iEEG recordings were not divided into separate trials, but instead, four separate blocks corresponding to the experimental conditions, with onset defined at the beginning of each experimental block and offset just prior to the questionnaire period. Across participants the dataset comprised 27 FAM blocks (2 × 5 min; 17 × 8 min; 8 × 10 min), 25 OMM blocks (3 × 5 min; 15 × 8 min; 7 × 10 min), 27 CML blocks (2 × 5 min; 18 × 8 min; 7 × 10 min), and 25 STL blocks (2 × 5 min; 16 × 8 min; 7 × 10 min) for subsequent analysis. IEDs, including sharps, spikes, and spike–wave discharges, were annotated by a neurologist and certified EEG reader (author K.B.) in accordance with the criteria recommended by the International Federation of Clinical Neurophysiology (e.g., di‐ or tri‐phasic waveform, sharp peak and duration <70 ms, and the presence of logical physiological field spread across adjacent contacts). 30 To prevent over‐counting of networked activity, synchronous discharges occurring across multiple contacts or distant regions within a 50 ms window were treated as a single event. Consecutive spikes were distinguished as separate events only if a clear return to the background rhythm was observed. Complex waveforms, such as spike‐and‐slow‐wave complex or polyspikes or bursts were marked as a single event representing a discrete period of increased cortical excitability. All annotations relied on expert clinical judgment without additional assumptions. Considering the standard clinical practice in epilepsy research, we chose IEDs as the primary outcome measure, as participants were more likely to exhibit IEDs than seizures during the brief experimental blocks. The IEDs were identified through visual inspection of condition‐specific signal epochs using the ft_databrowser function in the FieldTrip toolbox (Figures S1 and S2). To preserve the integrity of the spatial distribution of neural signal and minimize signal cancellation or subtraction artifacts, we utilized a monopolar montage referenced to a patient‐specific inactive lead (typically located far from the suspected seizure zone or in non‐cortical tissue). Reference and ground contacts were strictly excluded from quantitative analysis. All identified IEDs across all artifact‐free stereoelectroencephalography (sEEG) contacts, regardless of their anatomic location, were included to enable us to capture the effect of meditation on the whole brain. To examine the impact of meditation on neurophysiological markers of epilepsy, we quantified IED density, defined as the number of IED per minute for all conditions. This measure, instead of raw IED counts, was used as the dependent variable to account for variability in block duration across conditions and participants. We additionally implemented a control analysis to examine whether IED density was associated with the number of minutes of meditation completed. The Pearson correlation was non‐significant (r = .25, p = .09), suggesting that the duration of meditation condition did not systematically influence our measure of IED density. Statistical analyses Statistical procedures were performed using custom scripts in MATLAB (2024a) using the Statistics and Machine Learning Toolbox. Descriptive data are reported as mean ± SD unless otherwise specified. For group‐level analyses, linear mixed‐effects models (LMMs) were fit using the fitlme function with maximum likelihood estimation, including a random intercept for participants to account for repeated measurements. Model coefficients are reported as estimates (β), 95% confidence intervals (CIs), standard errors (SEs), t‐statistics, and p‐values (alpha = .05). Planned hypothesis‐driven comparisons were conducted, testing whether meditation conditions differed from the story listening (STL) control and the classical music listening (CML) control conditions. As these were planned a priori orthogonal contrasts embedded within the experimental design, no corrections were made for multiple comparisons. Pairwise contrasts between conditions were computed using model‐based linear combinations of fixed‐effect coefficients. For models that were used to test interactions between two fixed‐effects predictors, a likelihood ratio test was first performed using compare function to evaluate whether including an interaction term improved model fit. In the absence of a significant interaction, the final model included only main effects. The significance of fixed‐effects predictors was assessed using the F‐tests obtained from anova(lme). In LMMs, STL served as the reference for the condition predictor and L1 (hinderances) as the reference for the level predictor, unless stated otherwise. Individual‐level comparisons were conducted to evaluate the consistency and directionality of changes in IED density across participants and were quantified as symmetrized percentage change. This approach reduces the asymmetry inherent in traditional percent change scores, accounts for non‐linear scaling of increases vs decreases, and provides a more balanced measure of relative change across participants. 37 RESULTS Meditation increased IED density relative to control To test our main hypothesis of whether meditation modulates interictal epileptiform activity relative to control, we examined IED density across all conditions: STL, CML, FAM, and OMM. Data from one participant were excluded prior to this stage because no IEDs were observed in any condition, resulting in a final sample of n = 26. To evaluate the effects of meditation on IED density, we implemented an LMM in which condition predicted IED density. Although the omnibus main effect of condition on IED density was not significant (F(3, 91) = 1.95, p = .106), we proceeded with a priori planned contrasts, which revealed higher IED density during both meditation conditions relative to STL (Figure 2A): FAM vs STL (t(91) = 2.25, p = .027) and OMM vs STL (t(91) = 2.07, p = .041). Crucially, the 95% CIs for these effects ([.79, 12.90] and [.26, 12.51], respectively) did not cross zero, further supporting the directionality of the meditation‐induced modulation. In contrast, the CML condition did not differ significantly from STL (t(91) = 1.64, p = .104). Other planned comparisons between conditions showed no significant differences in IED density between the meditation conditions and CML, as well as between the two meditation conditions (see Table 2, Panel A for all parameter estimates). 38 , 39 Variance component estimates indicated high between‐subject variability in baseline IED density (random intercept SD = 21.46 IEDs/min), which likely reduced sensitivity to detect an overall effect. Estimated marginal means of IED density were highest for FAM (M = 22.19, 95% CI 6.76–37.63) and OMM (M = 21.73, 95% CI 6.22–37.24), followed by CML (M = 20.27, 95% CI 4.93–35.62), and lowest for STL (M = 15.35, 95% CI 5.96–24.73). Comparison of interictal epileptiform discharge (IED) density across experimental conditions: Story listening (STL), classical music listening (CML), focused attention meditation (FAM), and open‐monitoring meditation (OMM). (A) Group‐level data showing model predicted means (colored bars), raw data (gray boxplots underneath), 95% confidence intervals (transparent colored bands), and within‐subject repeated‐measures trends (gray lines). Black horizontal lines indicate significant effects (p < .05), showing higher IED density during meditation compared to STL (control). (B) Individual‐level symmetrized % changes in meditation conditions (in teal) and music condition (in gray) relative to STL. Positive and negative values indicate higher and lower IED density compared to STL, respectively. Overall, meditation consistently increased IED density relative to STL across participants. TABLE 2Model‐derived estimates for pairwise comparisons of conditions and meditation depth levels for IED density and self‐reported scores.Pairwise contrastβ [95% CI]SE t p Panel A: IED densityCML vs STL4.93 [−1.04, 10.89]3.001.64.104FAM vs STL6.84 [.79, 12.90]3.052.25.027 OMM vs STL6.38 [.26, 12.51]3.092.07.041 FAM vs CML1.92 [−3.96, 7.80]2.96.65.519OMM vs CML1.46 [−4.61, 7.53]3.06.48.635OMM vs FAM−.46 [−6.57, 5.65]3.07−.15.881Panel B: Self‐reported ratings for the MEDEQ at meditation depth levelsCML vs STL.24 [.07, .41].092.78.006 FAM vs STL.18 [.01, .35].092.09.037 OMM vs STL.27 [.10, .44].093.07.002 FAM vs CML−.06 [−.23, .11].09−.70.487OMM vs CML.03 [−.15, .21].09.29.770OMM vs FAM.09 [−.09, .26].09.99.323L2 vs L11.29 [1.02, 1.56].149.43 <.001 L3 vs L11.07 [.88, 1.26].1011.04 <.001 L4 vs L1.38 [.21, .55].094.34 <.001 L5 vs L1.27 [.08, .46].092.78 .005 Note: Bold values indicate p 〈 0.05. Model‐derived estimates for pairwise comparisons of conditions and meditation depth levels for IED density and self‐reported scores. Note: Bold values indicate p 〈 0.05. At the individual‐level, condition‐specific symmetrized percentage changes relative to STL confirmed the directionality of meditation‐related effects based on numerical comparisons (Figure 2B). After excluding four participants with incomplete data (i.e., no data in the STL condition), subsequent analyses included 22 participants. Consistent with group‐level trends, most participants exhibited higher IED density during meditation compared to the STL condition (FAM > STL = 17/22; OMM > STL = 16/22), whereas IED density was comparable between the meditation conditions and CML (see Figure S3). To explore the potential differences between the two meditation conditions, we numerically compared IED density between FAM and OMM within individuals (Figure 3). This revealed a heterogeneous pattern: approximately half of the participants exhibited higher IED density during FAM (12/22), whereas the others showed higher density during OMM (10/22). Individual‐level comparison of IED density between focused attention meditation (FAM) and open‐monitoring meditation (OMM). Positive and negative values indicate higher IED density during FAM and OMM, respectively. No consistent pattern emerged across participants, suggesting inter‐individual variability in response to distinct meditation types. Meditation increased subjective perception of meditation depth To evaluate how subjective experiences varied across different levels of meditation depth and conditions, we first fitted an LMM with self‐reported ratings for the MEDEQ as the outcome measure, with condition, level of meditation depth, and their interaction as fixed effects, and a random intercept for participants. The condition × level interaction did not improve model fit (χ 2 (12) = 15.14, p = .23), indicating that the effect of condition did not vary systematically across levels (see Table S2). The final model revealed a significant main effect of condition (F(3, 1252) = 3.84, p = .009). Specifically, self‐report ratings of perceived meditation depth increased significantly during meditation conditions relative to the STL control condition (Figure 4A): FAM vs STL (t(1252) = 2.09, p = .037) and OMM vs STL (t(1252) = 3.07, p = .002). In addition, CML showed higher self‐reported ratings than STL (t(1252) = 2.78, p = .006). No other pairwise differences between conditions were statistically significant. Furthermore, there was a significant main effect of level (F(4, 1252) = 45.83, p < .001), such that self‐report ratings for the MEDEQ systematically increased with meditation depth. The largest increases were observed at L2 (relaxation) and L3 (concentration) relative to L1 (p < .001), with smaller but significant increases at L4 and L5 (transpersonal qualities and non‐duality (p < .005), respectively). Model coefficients and pairwise differences are reported in Table 2, Panel B and estimated means with 95% CI for all conditions and levels are summarized in Table S3. (A) Self‐reported rating for the MEDEQ at the first level of meditation depth (L1; hindrances) across all experimental conditions. Raw data are shown as colored box plots (story listening (blue), classical music listening (orange), focused attention meditation (yellow), and open monitoring meditation (magenta)), with black squares indicating the mean and error bars representing the standard error of the mean. Participants reported significantly more hinderances during meditation compared to the STL control condition, as well as during music listening compared to the STL control condition (dashed black horizontal line, p < .05). (B) Slopes of IED density predicting self‐reported rating across five levels of meditation depth. Lines represent slopes estimated from the linear mixed‐effects model, and shaded areas indicate 95% confidence intervals. At L1 (blue), IED density negatively predicted self‐reported experience, whereas the association became positive at the intermediate deeper levels (L2, orange; L3, yellow) and was non‐significant at L4 (magenta) and L5 (green). Participants rated their experience on a 5‐point scale (1 = Not at all; 2 = Rarely; 3 = Sometimes; 4 = Often; 5 = Completely) for each level. Together, these results suggest that meditation heightened participants' perception of meditation depth compared to the STL control condition, and that this conditional effect was consistent across levels, with overall self‐reported ratings systematically changing by meditation depth. The association between IED density and self‐reported experiences varied with levels of meditation depth To investigate the relationship between IED density and subjective experience of meditation depth, we first tested a full LMM with a three‐way interaction between IED density, condition and level as fixed effects, self‐reported ratings for the MEDEQ as the outcome measure, and participants as a random intercept. The three‐way interaction did not improve model fit over a model including all possible two‐way interactions (IED density × condition, IED density × level, condition × level) additively (χ2(12) = 6.72, p = .87). Next, all two‐way interactions were examined. The IED density × level interaction was significant (χ2(4) = 12.65, p = .013), indicating that the association between IED density and self‐reported ratings varied across levels of meditation depth. In contrast, the IED density × condition interaction (χ2(3) = 1.86, p = .601) and the condition × level interaction were not significant (χ2(15) = 15.33, p = .428), suggesting that condition did not moderate the relationship of subjective experience of meditation depth with IED density or with meditation depth level. (These results are reported in Tables S4 and S5). Follow‐up analyses on the significant IED density × level interaction revealed that IED density negatively predicted self‐reported ratings at L1 (β = −.011, 95% CI −.018 to −.004, SE = .004, t(407) = −2.96, p = .003), whereas this association was reversed at L2 (β = .014, 95% CI .005 to .023, SE = .004, t(407) = 3.18, p = .002) and L3 (β = .011, 95% CI .002 to .019, SE = .004, t(407) = 2.47, p = .014) and was non‐significant at higher levels (L4 and L5, p > .07). These results indicate that IED density is differentially associated with the subjective experience of meditation depth at different levels of meditation depth. These results are visualized in Figure 4B, and detailed model‐derived parameters are reported in Table S6. Overall, these findings suggest that IED density appears to be functionally relevant, such that it corresponds to subjective experience of meditation depth, with the association being more pronounced at lower levels and weakens at deeper levels of meditation. DISCUSSION Using iEEG, the present study investigated whether meditation modulates IEDs in meditation‐naïve patients with drug‐resistant epilepsy. Both FAM and OMM practices showed modest increases in IED density (IEDs per minute) relative to a non‐meditative control condition (STL), with no such effects when comparing meditation practices to another control condition (CML). Individual‐level patterns also converged on higher IED density during meditation than the STL control, with notable inter‐individual variability in responses to distinct meditation types. Of note, participants reporting fewer hinderances and greater relaxation and concentration exhibited higher IED density, suggesting a functional link between iEEG‐measured clinically‐relevant neural dynamics and the subjective perception of meditation depth. Meditation‐related modulation of interictal activity Our findings suggest that meditation can influence interictal epileptiform activity in meditation‐naïve individuals with epilepsy. The observed increase in IED density during meditation compared to a non‐meditative STL condition aligns with a previous case report that also reported elevated IED rates during meditation compared with a mind‐wandering control condition. 31 One possible explanation is that meditative states modulate neural excitability in regions overlapping with epileptogenic zones. Supporting this view, recent iEEG studies have demonstrated that meditation reorganizes local field activity across both epileptic and non‐epileptic areas, including the hippocampus and amygdala, 25 , 27 , 28 shifting oscillatory dynamics and enhancing transient synchrony. In our cohort, the majority of IEDs were detected in the mesial temporal structures. The similarity in the timing of spike activity observed between adjacent structures, such as the hippocampus and amygdala, likely represent the temporally synchronized activity of well‐connected epileptogenic network in these patients. Accordingly, the increased IED density observed here may reflect transient changes in neural synchrony associated with meditation engagement within circuits prone to epileptiform activity, highlighting a functional coupling between these types of internally directed cognitive states and epileptic network dynamics. Although the presence of IEDs in other regions, including frontal and parietal regions, suggests that meditation may impact a whole‐brain epileptic network, the lower IED density in these extra‐temporal areas precluded a secondary analysis of these subregions. The absence of differences in IED density between meditation and music listening likely reflects that music listening, characterized by attention to sounds, silences, and musical elements, could act as a form of contemplative focus, 38 , 39 particularly for novice meditators, producing similar effects. However, the discrepancy between our finding and well‐established reports of anti‐epileptogenic effects during music may be attributed to the state‐dependent nature of IED modulation. Unlike passive listening protocols, where music serves as a background stimulus, our participants were instructed to listen in a relaxed manner but attentively, fostering a state of active sensory monitoring. Such heightened attentional arousal can lead to transient increases in synchronization, manifesting as higher IED density. Notably, although both conditions showed a similar directional trend compared with the STL control, meditation significantly modulated the epileptic network, whereas the effect of music listening was quantitatively lower and failed to reach statistical significance. This suggests that compared to music, meditation provides a more effective and reliable way for probing and engaging deep functional epileptogenic networks. From a clinical perspective, an intervention with higher effect size offers a more robust framework for therapeutic interventions. An alternative explanation for our results is that the observed differences across conditions may partly reflect variations in cognitive load. Listening to a story or music likely imposes comparatively lesser demand, involving externally directed attention, whereas meditation (FAM, OMM) demands sustained internal attention and continuous monitoring of thoughts, making it potentially the more cognitively demanding condition, especially for novice practitioners. At first glance, it seems possible that the observed conditional differences in IED density may be attributed to task difficulty. However, given that increased IED density during the meditation conditions was only observed in comparison to the STL condition and that cognitive load likely would not differ between the two control conditions, it is unlikely that IED density solely reflected cognitive load differences across conditions. Future work could systematically manipulate or measure cognitive load to clarify its role relative to meditative state in IED occurrences. Notably, differences in the pattern of IED density between FAM and OMM conditions may reflect their distinct attentional demands. 40 FAM requires deliberate redirection of attention to a chosen object, engaging top‐down attentional control via dorsolateral prefrontal and parietal regions, whereas OMM emphasizes non‐reactive awareness of ongoing experience supported by monitoring‐related activity within insular and anterior cingulate cortices. 10 , 41 The observed heterogeneity across meditation types may therefore reflect individual differences in how these attentional demands interact with the underlying epileptiform network. (See Supporting Information for more details regarding the clinical differences in patients showing more IEDs in FAM than OMM and vice versa). Functional relevance of meditation‐related changes in IEDs We also found that the meditation conditions were associated with significantly higher self‐reported ratings of meditation depth compared with the STL control condition estimated across all five levels. The greater perceived depth during meditation is consistent with prior work 42 and confirms that engagement in structured mindfulness practices yields measurable experiential benefits compared to a control state such as listening to a story. Notably, self‐reported ratings enhanced at deeper levels of meditation, particularly for relaxation and concentration. This suggests that brief mindfulness practices can elicit systematic progression toward deeper meditative states, even in meditation‐naïve individuals. 42 Of note, the association between IED density and participants' subjective perception of meditation depth varied across five depth levels, as indicated by changes in the slope of this association. At the initial level, participants who reported fewer hinderances showed higher IED density. One possible interpretation is that transient increases in cortical excitability or neural synchrony, indexed by increased IED activity, may co‐occur with early internal engagement, thereby reducing the subjective experience of distraction or challenges. Conversely, reduced IED density during experiences of greater hinderance may reflect states of lower cortical excitability or less coordinated neural activity, making it harder to maintain focus and leading to more frequent distractions. This relationship, however, shifted at intermediate levels, where higher IED density was associated with higher self‐reported relaxation and concentration, indicating that heightened neural excitability may support arousal regulation and attentive control over the mind, facilitating the transition into deeper meditative states. At the deepest levels, the non‐significant association suggests that changes in neural excitability may have minimal influence on meditation‐naïve participants' deeper internal engagement. Although speculative, these findings suggest that iEEG‐measured IEDs may be behaviorally meaningful rather than purely pathological events, potentially reflecting transient reorganization of neural synchrony within epileptogenic networks 43 , 44 that coincide with moments of heightened engagement in meditative states. Clinical implications of meditation The clinical implications of these findings depend upon the role IEDs play in the broader context of seizure dynamics and disease severity. The relationship between IEDs and seizure occurrence remains divided: some studies report no association between IEDs and seizure frequency, 45 whereas others have proposed that IEDs may inhibit ictal discharges by engaging adaptive inhibitory mechanisms within the epileptogenic networks. 46 Experimental evidence suggests that IEDs are often followed by a period of network inhibition or refractoriness, transiently reducing excitability and stabilizing local circuits. 46 Consequently, increases in IEDs, in certain contexts, could signal functional engagement of epileptogenic networks rather than necessarily indicating pathological consequences. Nevertheless, elevated IED occurrence has conventionally been viewed as a marker of increased cortical excitability and potential pathological risk, and recent evidence also shows that the presence of IEDs may lead to seizure recurrence. 47 However, it is important to note that no clinical seizures were observed during any meditation or control sessions in our study. In this context, meditation‐related increases in IED density observed in our study should be interpreted cautiously and should not discourage the use of meditation as a supportive practice for managing the comorbidities (i.e., anxiety and depression) of drug‐resistant epilepsy. Longitudinal clinical assessments tracking meditation‐related IED changes with seizure outcomes will be crucial for determining the therapeutic relevance of these findings. Furthermore, although the current study did not focus on a formal stratification by “epileptic terrain,” such as distinguishing between specific lesion types and seizure‐onset zone locations, and sought to investigate the overall impact of meditation on IEDs, these factors could influence spike dynamics and responses to interventions such as meditation. However, the observation of a consistent IED modulation across a heterogeneous cohort, including a majority of non‐lesional cases, suggests that meditation‐induced shifts in network excitability may be a generalized physiological response. Future research with larger, multi‐center cohorts is warranted to determine systematically if specific pathologies exhibit distinct sensitivities to the cognitive demands of meditation or if the magnitude of this effect is influenced by the underlying structural differences. Limitations and future directions The findings of this study should be interpreted with consideration of some limitations. First, all participants were novice meditators with no prior training, which limits the generalizability of our findings to long‐term practitioners who typically exhibit more stable attentional control and distinct neural signatures of meditation. 42 , 48 Future studies including experienced meditators could help determine whether sustained meditation practice leads to more robust or qualitatively different effects on interictal activity. Furthermore, the brief, single‐session meditation design likely constrained the depth of experiential engagement achievable during the experiment. The absence of meditation‐related effects at deeper levels suggests that participants may not have reached advanced meditative states, potentially limiting the expression of stronger neural–behavioral coupling. Longitudinal designs with repeated or extended meditation sessions would better capture meditation‐related variations in IEDs. A significant limitation is the lack of formal, standardized neuropsychological assessing baseline attention and executive functions for patients in the present study. Although our within‐subject design partially controls for individual variance, it is difficult to determine how specific cognitive traits influenced the observed spike modulation. Future research building on this work should aim for better‐stratified patient cohorts based on age, sex, epilepsy localization, etiology, and comprehensive neuropsychiatric profiles. This would allow for a more robust investigation into the mechanisms and predictors of responsiveness to cognitive interventions. Furthermore, although the current study focuses on transient interictal events (IEDs), the absence of spectral composition analysis (e.g., theta/alpha power) prevents a granular verification of the objective brain states achieved during sessions. Future studies integrating EEG spectral measures with IED detection will be a logical and valuable extension of our findings to determine how specific frequency bands relate with the observed changes in network excitability. We also observed variability in IED density between the two meditation types at the individual level, indicating that different meditation styles may interact differentially with the epileptogenic network across individuals. Given the diversity of existing techniques, each emphasizing different attentional and affective processes, 40 these findings should be interpreted within the specific context of the two practices tested. Furthermore, although IEDs provided a practical proxy for neural excitability in epilepsy, seizure counts would remain the clinically preferred outcome measure. Regarding procedural concerns, carry‐over effects stemming from the sequence of conditions could potentially confound the modulation of IED density. We mitigated this by interleaving meditation blocks with control blocks, combined with brief inter‐block gaps, to facilitate a cognitive reset. Notably, participants who performed meditation blocks consecutively also exhibited IED density modulation similar to those who followed the alternating sequence. This suggests that observed changes in IED density are driven by the cognitive state rather than carry‐over effects from the preceding block. Similarly, changes in anti‐seizure medication levels could potentially modulate IED density and therefore influence our results; however, almost all participants completed all conditions on the same day with the condition order pseudo‐randomized, with up to 3 min gaps between them. Therefore, substantial variation in pharmacological effects is unlikely across conditions, suggesting that the observed changes in IED density cannot be attributed directly to medication effects. Finally, there are anatomic and statistical constraints to consider. Our study did not distinguish whether meditation primarily modulates the primary generators of the seizure‐onset zones or the secondary propagation pathways. Individual differences in network architecture likely influence the spatial and temporal distribution of the observed spike changes. Moreover, although our findings were robust across a diverse cohort, we cannot rule out the possibility that specific clinical syndromes (e.g., temporal vs frontal lobe epilepsy) could exhibit distinct sensitivities to meditation. Although we attempted to account for individual differences using mixed‐effects models, future studies with larger, balanced sub‐groups are recommended to determine if meditation‐induced changes are syndrome‐specific and localized to the onset zone or represent the propagation to other regions. Finally, the modest sample size, further reduced by the exclusion of five participants with incomplete or non‐detectable IED data, may have limited statistical power of the omnibus test and sensitivity to conservative post hoc corrections. Consequently, our iEEG findings are reported as planned comparisons highlighting the modulatory potential of meditation. Larger, multi‐center iEEG studies will be essential to validate these findings and establish their reproducibility across heterogeneous patient populations. CONCLUSIONS In conclusion, this study provides direct intracranial evidence that meditation can modulate interictal epileptiform activity in patients with drug‐resistant epilepsy. Both focused attention and OMM increased IED density relative to a control condition (STL). Inter‐individual variability in IED density in response to distinct meditation practices further indicate that epileptiform networks are differentially sensitive to internal cognitive states. Notably, higher IED density was associated with fewer perceived hinderances, but showed a positive association with greater relaxation and concentration, suggesting a functional link between neural excitability and the subjective ease of meditation. Future research should determine whether meditation‐related modulation of IEDs simply represents transient increases in neural excitability or more permanent adaptive neural mechanisms and assess whether meditation can be used as a clinical tool in the treatment of epilepsy. AUTHOR CONTRIBUTIONS S.S., F.G., and J.W.Y.K. conceived of and designed the study; S.S. performed the experiments; W.H., P.F., and S.S. contributed to data collection; S.S. analyzed the data under the supervision of F.G. and J.W.Y.K.; S.S. wrote the original draft; and F.G. and J.W.Y.K. provided critical revisions. All authors reviewed and approved the final version of the manuscript for submission. CONFLICT OF INTEREST STATEMENT The authors declare no conflicts of interest. ETHICAL PUBLICATION STATEMENT We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. Supporting information Figure S1: Example of interictal epileptiform discharge (IED) annotation in control conditions in a representative participant. A certified electroencephalography (EEG) reader (neurologist) visually inspected intracranial EEG (iEEG) recordings across all experimental conditions and manually marked IEDs using the FieldTrip software. The annotated IEDs are highlighted as pink vertical sections over the clean, white background of the EEG trace. The Y‐axis represents contacts (amygdala, anterior and posterior hippocampus). STL, story listening; CML, classical music listening. Figure S2: Example of interictal epileptiform discharge (IED) annotation in meditation conditions in a representative participant. A certified electroencephalography (EEG) reader (neurologist) visually inspected intracranial EEG (iEEG) recordings across all experimental conditions and manually marked IEDs using the FieldTrip software. The annotated IEDs are highlighted as pink vertical sections over the clean, white background of the EEG trace. The Y‐axis represents contacts (amygdala, anterior and posterior hippocampus). FAM, focused attention meditation; OMM, open monitoring meditation. Table S1. Summary of patients' anti‐seizure medication status and electrophysiological recording. Table S2. Fixed‐effect estimates from the full linear mixed‐effects model predicting self‐reported ratings based on condition and five levels of meditation depth. Table S3. Estimated mean (with 95% confidence interval [CI]) of self‐reported scores corresponding to all meditation depth levels and across all conditions. Table S4. Fixed‐effect estimates from the full linear mixed‐effects model predicting self‐reported ratings based on IED density, condition, and five levels of meditation depth. Table S5. Stepwise comparison of linear mixed‐effects models to identify the optimal parsimonious model predicting self‐reported ratings for meditation depth. Table S6. Fixed‐effect estimates from the final linear mixed‐effects model with interictal epileptiform discharge (IED) density predicting self‐reported ratings at five levels of meditation depth. ACKNOWLEDGMENTS We are grateful to the patients for their valuable participation and to E. Mercer and the team for their assistance with data collection. This research was supported by the Harley Hotchkiss – Samuel Weiss Postdoctoral Fellowship (S.S.), Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant (J.W.Y.K.), and funding from the Hotchkiss Brain Institute (HBI) (F.G.). Soni S, Brochu K, Federico P, Hader W, Kam JWY, Girgis F. Mindfulness meditation increases interictal epileptiform discharges in meditation‐naïve surgical epilepsy patients. Epilepsia. 2026;67:4758–4773. 10.1002/epi.70332 Julia W.Y. Kam and Fady Girgis contributed as co‐senior authors. DATA AVAILABILITY STATEMENT The data that support the findings of this study are available from the corresponding author upon reasonable request. REFERENCES Associated Data Supplementary Materials Figure S1: Example of interictal epileptiform discharge (IED) annotation in control conditions in a representative participant. A certified electroencephalography (EEG) reader (neurologist) visually inspected intracranial EEG (iEEG) recordings across all experimental conditions and manually marked IEDs using the FieldTrip software. The annotated IEDs are highlighted as pink vertical sections over the clean, white background of the EEG trace. The Y‐axis represents contacts (amygdala, anterior and posterior hippocampus). STL, story listening; CML, classical music listening. Figure S2: Example of interictal epileptiform discharge (IED) annotation in meditation conditions in a representative participant. A certified electroencephalography (EEG) reader (neurologist) visually inspected intracranial EEG (iEEG) recordings across all experimental conditions and manually marked IEDs using the FieldTrip software. The annotated IEDs are highlighted as pink vertical sections over the clean, white background of the EEG trace. The Y‐axis represents contacts (amygdala, anterior and posterior hippocampus). FAM, focused attention meditation; OMM, open monitoring meditation. Table S1. Summary of patients' anti‐seizure medication status and electrophysiological recording. Table S2. Fixed‐effect estimates from the full linear mixed‐effects model predicting self‐reported ratings based on condition and five levels of meditation depth. Table S3. Estimated mean (with 95% confidence interval [CI]) of self‐reported scores corresponding to all meditation depth levels and across all conditions. Table S4. Fixed‐effect estimates from the full linear mixed‐effects model predicting self‐reported ratings based on IED density, condition, and five levels of meditation depth. Table S5. Stepwise comparison of linear mixed‐effects models to identify the optimal parsimonious model predicting self‐reported ratings for meditation depth. Table S6. Fixed‐effect estimates from the final linear mixed‐effects model with interictal epileptiform discharge (IED) density predicting self‐reported ratings at five levels of meditation depth. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.