Magnetoencephalography monitoring of hypnosis-induced modulation of event-related brain responses by hypnosis Authors: Hafid Sid-Ahmed, Julie Alayrangues, Lilia Langar, Nelly Richard, Véronique Albaladejo, Stéphane Pezzani, Virginie Brun, Daniel Anglade, Vincent Auboiroux Published: 2026 ABSTRACT AbstractHypnosis is increasingly used in a medical context. However, the hypnotic process is dynamic, and we currently lack specific and reliable biomarkers to diagnose and rigorously monitor the patient’s response to hypnotic suggestions. Here, we aimed to identify physiological biomarkers by investigating cerebral event-related responses in an oddball paradigm using a magnetoencephalography-based method. Twenty volunteers with different suggestibility levels were exposed to standard and deviant sounds as part of an oddball paradigm. Subjects were monitored in three distinct conditions: critical consciousness, while distracted by listening to a story, and during hypnotic trance. In each condition, event-related responses were analyzed in both a source space (auditory cortex and dorsal anterior cingulate cortex) and a sensor space in a minimalistic scenario (two sensors). The statistical analysis revealed significant reductions in amplitudes in the auditory cortex and in sensor measurements during hypnotic trance and distraction relative to critical consciousness. These reductions in amplitudes were observed for both standard and deviant stimuli, regardless of the level of hypnotic suggestibility. In particular, this study aimed to determine whether auditory responses are reduced even without explicit suggestions of induced deafness. A more subtle, but nevertheless statistically significant, difference was also found between the distraction and the hypnotic conditions in the source space. The modulation of the cortical response to auditory stimuli during hypnosis could be used as a physiological marker of the hypnotic process. Hypnotic trance and distraction do not appear to differ significantly, as both seem to induce a similar alteration of critical consciousness.Keywords: hypnosis, magnetoencephalography, MEG, oddball, auditory, biomarker Highlights Identification of electrophysiological biomarkers of neutral hypnosis using magnetoencephalography. Identification of electrophysiological biomarkers of neutral hypnosis using magnetoencephalography. Amplitude of the auditory evoked response (AER) in an oddball paradigm can differentiate hypnosis and distraction from critical consciousness. Amplitude of the auditory evoked response (AER) in an oddball paradigm can differentiate hypnosis and distraction from critical consciousness. The modulations of AER appear in late latencies regardless of the subject’s suggestibility to neutral hypnosis, and without need to suggestion deafness. The modulations of AER appear in late latencies regardless of the subject’s suggestibility to neutral hypnosis, and without need to suggestion deafness. Introduction In a medical context, hypnosis is defined as an altered state of consciousness induced by suggestion (Elkins et al. 2015). Hypnosis, also referred to as a hypnotic trance or state, is a dynamic process characterized by a detachment from the outside world, decreased sensory perception, and increased suggestibility. The first two aspects are related to the fact that hypnosis relies on the natural ability of individuals to enter a state of focused attention, which is associated with detachment from the external environment and, consequently, decreased sensory perception and altered internal consciousness (Mouheb et al. 2025). Lastly, suggestibility denotes the degree to which an individual responds to hypnotic suggestions, while hypnotizability is defined as an individual’s capacity to experience suggested alterations in perception, memory, or behavior during hypnosis (Montgomery et al. 2011). Therefore, the hypnotic trance is thought to involve a temporary reduction in critical consciousness, which is the brain’s typical resting state responsible for monitoring and analytically evaluating verbal communication (Halsband et al. 2009). In a medical context, the reduction of this evaluative process during hypnotic suggestion is of particular interest. It has been increasingly used in routine clinical practice in recent decades as part of pain management (Thompson et al. 2019; Rousseaux et al. 2025) and surgical procedures (Faymonville et al. 1998; Defechereux et al. 2000; Holler et al. 2021). However, the clinical application of hypnosis is constrained by its dynamic and, consequently, unstable nature. Indeed, subjects retain the ability to critically evaluate suggestions during hypnosis, and the effectiveness of these suggestions depends on active cognitive processes and the interactive, dynamic nature of hypnosis (Kirsch 2000; Lynn et al. 2020, 2015; Zahedi et al. 2024). In the context of using hypnosis in the operating theater, continuous monitoring of the hypnotic trance is required to detect and rapidly correct any issues. Unfortunately, there are currently no specific clinical signs that indicate hypnosis (Lynn et al. 2020). Consequently, intraoperative monitoring of the hypnotic process could benefit from the use of a brain physiological biomarker. Hypnotic suggestions are highly heterogeneous in their content and goals, ranging from neutral hypnosis—where the induction involves no specific suggestion beyond entering a hypnotic state—to targeted suggestions designed to modulate specific sensory, perceptual, or motor functions. Neuroimaging evidence demonstrated that distinct suggestions (e.g. analgesia, motor inhibition, or hallucination) elicit ‘type-specific modulation’ of brain activity in regions subserving their intended functions (Jensen et al. 2017). Among the various means of studying brain activity, electroencephalography (EEG) is a technique that measures brain electric activity. Many studies, described hereafter, have indicated that this technique is capable of detecting changes in cerebral function during a hypnotic trance. Most studies report changes in the power and coherence of brain oscillations during the hypnotic trance, with specific sensitivities for each frequency band. Alpha, beta, and theta oscillations were the subject of several studies, but these oscillations appear to be only slightly sensitive to differentiate hypnosis and critical consciousness. Some studies have demonstrated modulations of gamma power during hypnosis, although the direction of modification is unclear, with some authors reporting a decrease in gamma power (Kirenskaya et al. 2011), whereas others report an increase (De Pascalis 1999). Currently, while many publications seem to favor a change in theta oscillations associated with the hypnotic trance (Sabourin et al. 1990; Rainville et al. 1999; Jensen et al. 2015; De Pascalis 2024, 1999; Taghilou et al. 2025), controversy remains and their use as a marker of hypnotic trance is still debated (Hiltunen et al. 2021). The most significant EEG differences between highly hypnotizable (HH) subjects and less hypnotizable (LH) subjects primarily involve the theta band: HH subjects frequently exhibit higher baseline theta activity in both the waking state and during hypnosis (Jensen et al. 2017; De Pascalis 2024). Finally, a single magnetoencephalography (MEG) study found that an out-of-body experience induced by hypnosis led to a reduction in beta power in the right insula and a reduction in alpha power in the left inferior frontal gyrus (Zeev-Wolf et al. 2017). The heterogeneity of all these results from different studies is likely due to the use of distinct hypnosis techniques, experimental paradigms, and subjects. As an example, a study by Jensen et al. (2015) attempted to explain the discrepancies between results reported in the literature, and suggested that the neural changes produced by hypnosis depend on the type of hypnosis (Jensen et al. 2017; Landry et al. 2017). According to previous studies on somatosensory responses to painful electric stimulation, a reduction in pain perception during hypnoanalgesia is associated with a change in evoked responses for highly suggestible individuals (De Pascalis et al. 2008). Thus, analysis of event-related potentials (ERPs) during trance can provide information about how our brains process sensory information during hypnosis. Several methods exist to monitor event-related responses in real-time. The most common of these is EEG, allowing the measurement of ERPs. EEG studies investigating ERP and hypnosis used a protocol based on either external (e.g. sensory stimulation) or internal (e.g. voluntary focusing of attention, engaging in a motor act) stimulation. Their results revealed a modification of the late component of ERP in highly suggestible subjects (Franz et al. 2020; Munoz Musat et al. 2022). Among ERP measurements, auditory evoked potentials in general and mismatch negativity (MMN) in particular are of particular interest. Auditory evoked potentials are a measure of changes in the brain’s electrophysiological activity over time, triggered by a sound stimulus. The early components primarily reflect the initial processing of auditory information in the primary and secondary auditory cortices. In contrast, the later components are associated with the detection of novelty, the orientation of attention, and the conscious cognitive evaluation of the stimulus (Polich 2007). MMN is measured by comparing the evoked potential associated with a repetitive, regular stimulus to that associated with a deviant stimulus. The latter shows a stronger response to the deviant stimulus. Current thinking suggests that this discrepancy in brain activity is indicative of a mechanism that compares sensory inputs with a mnemonic or predictive representation of auditory regularities (Näätänen et al. 2005; Auksztulewicz and Friston 2015). Contemporary models of hypnosis describe it as a process involving increased attentional focus and the reallocation of cognitive resources toward specific internal representations or suggestions (Oakley and Halligan 2009). This attentional focus may reduce the processing of competing environmental stimuli. From this perspective, hypnosis could modify auditory ERPs through at least two complementary mechanisms. Firstly, the attentional focus induced by hypnosis could reduce attention to irrelevant external auditory stimuli, thereby decreasing the amplitude of MMN and P3 responses. Secondly, hypnotic suggestions could modify the internal models used to predict auditory stimuli, leading to a change in the dynamics of auditory prediction errors. This hypothesis is consistent with the work of Terhune et al. (2017), which shows that hypnotic suggestions can profoundly alter top-down perceptual and attentional processing. While studies have reported minimal to no significant differences in MMN amplitude between the control and hypnosis conditions in EEG (Jutai et al. 1993; Hiltunen et al. 2019), the weight of evidence appears to support differences in MMN amplitude depending on whether subjects are in a state of critical consciousness or in a hypnotic trance (Kallio et al. 1999; Jamieson et al. 2005; Facco et al. 2014). These results suggest that analysis of ERPs resulting from a combination of standard and deviant stimuli may distinguish hypnosis from critical consciousness. However, a potential bias or source of influence must be taken into account: all these studies preferentially selected highly suggestible participants, and the hypnosis protocol included the use of direct hypnotic suggestions aiming to reduce sensitivity to visual or auditory stimuli. For the study presented here, we aimed to help hypnotherapists to determine the patient’s phase in the hypnotic process using MEG. MEG has the advantage over EEG of not requiring sensors to be directly applied to the subject’s scalp, meaning that the recorded signals are more stable over time (Munoz Musat et al. 2022). We characterized hypnosis by measuring cerebral responses to sensory stimulation. We used MEG to analyze event-related fields (ERFs) elicited by standard and deviant tones in an oddball paradigm, including both direct stimulus responses and the derived MMN component. The main objective was to identify potential biomarkers discriminating hypnosis from critical consciousness. Uniquely for a study of hypnosis, we tested this method in a group of healthy volunteers of varying degrees of hypnotic suggestibility, and our hypnosis-induction protocol was based on the recall of a pleasant moment. Contrary to the methods used in other studies involving auditory responses, our hypnosis-induction protocol did not include any suggestion of sound perception. These conditions ensured we were close to the medical use of hypnosis in a surgical context. The analysis was performed in two cases: after source reconstruction in the auditory and dorsal anterior cingulate cortices (dACC), and at sensor level, using a minimalistic approach (two sensors) to mimic operating theater-compatible measurement. Finally, given that hypnosis is closely associated with alterations in attentional processes, characterized by sustained internal focus and reduced responsiveness to competing external stimuli, we also tested whether distraction could be distinguished from a hypnotic phase or critical consciousness using the same paradigm. Material and methods Study population and ethical considerations The clinical trial (ID: NCT04020731) was conducted between December 2019 and March 2021. The study was approved by the institutional review board (Protocol: 38RC19.060, ID RCB: 2019-A00575-52) and was performed in accordance with the Declaration of Helsinki, with Good Clinical Practice guidelines, and with Good practice for conducting and reporting MEG research (Gross et al. 2013). All participants gave informed consent before enrollment. Twenty healthy volunteers (13 men and 7 women, min-median-max age: 18–32–69 years) participated in the study. The inclusion criteria were: right-handed as assessed by the Edinburgh Handedness Inventory (Oldfield 1971), non-smoking adults, fluently speaking French, with normal hearing. The exclusion criteria included any contraindication to MEG or MRI, any previous experience of hypnosis, and any history of neurological or psychiatric disorders. After recruitment and initial screening, none were excluded from the experiments. Hypnosis The hypnotic trance was induced using a non-directive approach. Participants were asked to relive a pleasant memory that had been determined in a preliminary interview. According to the first description of hypnoanalgesia used by the Faymonville team, we use ‘hypnosis as a state of mental focalization on a pleasant life experience’ (Faymonville et al. 1997). Participants were seated in the MEG chamber and instructed to focus on internal kinesthetic sensations of comfort and safety, followed by the recall of multisensory details from pre-identified pleasant memories—‘without any suggestions targeting the perception of ambient sounds’. The complete transcription of each individual hypnotic script is provided in the supplementary materials (see Appendices 2 and 3). Three experienced hypnotherapists took part in these experiments. They are anesthesiologists and nurse anesthetists who have undergone training in Ericksonian hypnosis and regularly perform hypnosedation in the operating room. For each session involving hypnosis, a video recording was reviewed blindly, without audio, by the two other hypnotherapists to evaluate the start and end of the deep phase of hypnosis. The evaluation was based on physiological indicators of hypnotic trance, including eyelid tremulations, the emergence of involuntary movements such as spontaneous swallowing, and unexpected changes in respiratory patterns. Due to the subjective nature of these assessments, achieving strict consensus among all three hypnotherapists was uncommon. A conservative approach was adopted, in which only the temporal intersection of the trance periods identified by all three evaluators was considered for subsequent analysis. Following the MEG recordings, participants were asked to complete a retrospective questionnaire for each hypnosis session. Given the current absence of robust physiological markers for hypnotic trance (Lynn et al. 2020), these questionnaires were used to assess their subjective perception of whether or not they had experienced a hypnotic trance. The questionnaire was based on three phenomenological dimensions (alterations in time perception, loss of agency, and unfamiliar bodily sensations) (Vanhaudenhuyse et al. 2019), with scores ranging from 2 to 14 (see Appendix 1 for a translated version of the questionnaire and Table 1 for the detail of individual scores). Table 1Cohort description, individual scores of subjective perception of hypnosis and Stanford scores.GenderAgeSubjective scoreStanford scoreSubject #1F69106Subject #2M29109Subject #3M51129Subject #4F641010Subject #5M3551Subject #6F4475Subject #7F40108Subject #8F2456Subject #9M2678Subject #10M3254Subject #11M3235Subject #12F2064Subject #13M2369Subject #14F3232Subject #15M3552Subject #16M1859Subject #17M4375Subject #18M2377Subject #19M4879Subject #20M2750 Cohort description, individual scores of subjective perception of hypnosis and Stanford scores. Finally, the hypnotic suggestibility of each participant was assessed at the end of the study using the French version of the Stanford Hypnotic Susceptibility Scale form C (Weitzenhoffer and Hilgard 1926; Gay 2007). The SHSS-C score is the sum of 12 dichotomous tasks (1 point for each successful task). Experimental design Over the course of half a day, each subject underwent the following steps: medical assessment to determine eligibility for inclusion, MEG recordings, MRI examination, and evaluation of hypnotic suggestibility. All sessions were scheduled in the morning to avoid confounding effects due to the subjects’ chronobiology (Green et al. 2015). The experimental set-up was divided into different steps including critical consciousness (CC), distraction (DIS), and hypnosis (HYP). These steps occurred in the following sequence: CC1, CC2, DIS, HYP1, HYP2, HYP3, and CC3 (see Fig. 1). In the absence of prior results for determining an appropriate duration for a washout period of hypnotic trance, the DIS and HYP conditions have not been randomized. Similarly, the absence of prior experience with hypnosis among the participants was intended to minimize the risk of self-hypnosis during the initial DIS stage. To assess the stability and replicability of potential findings, the three HYP conditions were sequentially repeated under identical parameters. Experimental design. CC: critical consciousness, DIS: distraction, HYP: hypnosis, HT: hypnotherapist. The auditory oddball paradigm consists of 75% of standard tones (1 kHz) and 15% of deviant ones (2 kHz). The mean number of standard and deviant tone for each subject and each block is indicated above the blocks. The tones were delivered every 800 ms. MEG was first recorded while the subject was sitting alone in a passive state in the magnetically-shielded room during an auditory oddball paradigm (Squires et al. 1975) for about 5 minutes (CC1). The subject was assumed to be in critical consciousness during this step. Following this recording, the hypnotherapist entered the room. The subject and the hypnotherapist remained silent for about 5 minutes while the MEG recorded the subject’s brain activity during the same oddball paradigm (CC2). During the DIS stage, the hypnotherapist then read a story to the subject (The script is available in the supplementary materials), during which the same oddball paradigm was repeated for about 5 minutes. Following this recording, the participant completed a Stroop task (Stroop 1935) to ensure that they had regained critical awareness. The Stroop task is a test of attention that consists of displaying color names written in a non-matching color typeface, for example ‘green’ written in blue (Stroop 1935). The patient was asked to press a button with their right index finger if they saw a congruent display (match between color of text and color name), otherwise they were to press with their left index finger. This task activates the neurocognitive networks responsible for executive control and conflict monitoring (MacLeod 1991). This activation is closely associated with high levels of cognitive alertness (Cavanagh et al. 2009; Cavanagh and Frank 2014). Therefore, a strong Stroop effect can be interpreted as an operational indicator of critical alertness, which is characterized by an enhanced capacity to identify and address cognitive conflicts. During the Stroop tasks, no sound was emitted. Then, the MEG recording was started, and the subject was guided through the hypnotic process. To make sure the patient was in a hypnotic trance for 10 minutes, the hypnotherapist pressed a button at the end of the induction phase. The recording continued for about 10 minutes after that, during which time the oddball paradigm described above was repeated. A slight change in the lighting in the experimental room signaled to the hypnotherapist when the 10 minutes were up, so that he/she could bring the subject back to critical consciousness. The subject once again completed a Stroop task to ensure that they had regained critical awareness. These steps were repeated three times (HYP1, HYP2, and HYP3). The experiment ended with the subject alone in the closed room for about 5 minutes (CC3). Each subject met only one hypnotherapist and performed the three HYP phases experiment with him/her. Subjects were instructed to keep their eyes closed throughout the experimental sessions, except during the Stroop tasks. This instruction was linked to the fact that people in a hypnotic trance tend to close their eyes, and to harmonize all conditions for the corresponding generation of alpha waves in the brain (Berger 1929). With the eyes closed throughout the experiment, the recorded signal was not impacted by variations in alpha waves. To accurately select the MEG signal corresponding to the deep phase of hypnosis, the subjective estimation of trance onset and duration was made both by the hypnotherapist in charge (by pressing a button during the experiment) and later confirmed by the two other hypnotherapists reviewing the recorded video. Due to variability in achieving trance onset, each HYP block recording can last longer than 10 minutes. In total, each MEG session lasted about 70 minutes. Auditory oddball paradigm Sound stimuli were 100-ms tone bursts with an inter-stimulus delay of 800 ms, generated using Psychtoolbox v3.015 on MATLAB R2019b (http://psychtoolbox.org) (Brainard 1997). Standard stimuli (1 000 Hz) were presented to the subject, 85% of the time; the remaining 15% were deviant stimuli (2000 Hz) (see Fig. 1). The stimuli were presented in a pseudo-randomized order with each deviant stimulus separated by at least two standard stimuli. Stimuli were presented binaurally via headphones (Nicolet—Biomedical Madison TIP-300) with an intensity of 75 dB. Data acquisition MEG data were recorded using a 306-channel Elekta Neuromag MEG (MEGIN, Helsinki, Finland) with a sampling frequency of 1 000 Hz and a 0–330 Hz band-pass filter. The participant’s head position was tracked using five head position indicator (HPI) coils (two on each mastoid and one each on the right, middle, and left of the forehead). The reference points of each participant’s head and face, as well as the five HPIs used for co-registration, were recorded at the start of the experiment using a hand-held 3D digitizer (Polhemus FASTRACK). In addition, the MEG system recorded the subject’s electrocardiogram (ECG) and electrooculogram (EOG) with surface electrodes. Respiration was monitored using a respiratory belt (MP-160, Biopac Systems Inc., USA). At the end of the MEG recording, a 3D anatomical T1 sequence was acquired for each participant using a 1.5 T Espree MRI (Global Siemens Healthcare Headquarters, Erlangen, Germany) with a resolution of 1 × 1×1 mm. MEG preprocessing Signal was analyzed using the MNE-Python 1.2 toolbox (Gramfort et al. 2013). First, noisy channels and time periods were removed from raw datasets following manual inspection. This de-noising ensured the removal of artifacts related to strong muscle activity or sensory anomalies (signal jump). The MEG recording was then treated with band-pass filters (two Butterworth, fourth order, cut-off frequencies: 0.1 Hz and 60 Hz) and notch filters (50 Hz power line and harmonics), before finally removing cHPI signals by iterative filtering (Gramfort et al. 2013). Subsequently, the signal space was separated to filter out artifactual signal components coming from outside the brain (Taulu et al. 2003; Taulu and Hari 2009). Finally, head movements were corrected: for each subject, the head position was aligned with the average head position during all experiments. The resulting signals included brain signals, sensor noise, and some other electrophysiological signals, such as signals from the cardiac and facial muscles. Heart and eye muscle signals were removed by applying independent component analysis (ICA): Independent components were computed from the data; those containing cardiac or eye signals were flagged by Kuiper statistics based on the ECG and EOG measurements before removal from the signals. The pre-processed MEG signals were then segmented into epochs starting 200 ms before the tone stimulus and ending 600 ms after the tone stimulus (time segments: −200 ms to +600 ms). Because of the paradigm design, large differences in numbers of epochs are expected at each step in our dataset. To avoid bias when comparing the evoked responses in the three conditions, the same number of standard and deviant stimuli were selected uniformly for the analysis of each experiment: after removal of noisy epochs, the number of epochs was standardized for each condition (CC, DIS, and HYP) by uniformly choosing 300 standard epochs and 51 deviant epochs. For the HYP condition, the epochs were chosen from the time period during which participants were confirmed to be in the deep phase of hypnosis by the hypnotherapists. ERF analysis and MMN ERFs for each stimulation were obtained by averaging epochs. Thus, an ERF measured on 306 sensors was obtained for each participant, each condition (CC, DIS, and HYP), and each type of auditory stimulation (standard and deviant). With these 306 sensors, two analyses were performed in this study: a sensor space analysis, based on one sensor pair in each hemisphere, and a source space analysis, that uses all sensors in each hemisphere to estimate source brain signals in the cortex. MMN is defined as the difference between standard and deviant ERFs. Since standard epochs are about five times more numerous than deviant ones, and because the MMN depends on the relation between a deviant stimulus and its preceding stimulus (Garrido et al. 2009), the MMN was calculated using 51 deviant epochs and the 51 standard epochs preceding them. Sensor space analysis Spatial and temporal modulations of auditory evoked responses linked to hypnosis were sought in the sensor space (gradiometers and magnetometers). To handle inter-subject variability of source orientations, the norm for each gradiometer pair was calculated, as follows: Where and are the measurements taken by gradiometers x and y at time step t, and g is the module of the measurement of the two gradiometers. Then, the magnetometer activity was defined as the absolute value of the amplitude measured for that magnetometer. By taking absolute values, the result reflects brain activity related to the stimuli and can be compared between subjects. The baseline was adjusted by subtracting the mean amplitude of the pre-stimulus period (−200 to 0 ms) for each trial from all data points in the segment, make it possible to compare activity between conditions. The evoked activity was first spatially analyzed after a standard and a deviant stimulus by performing statistical analysis on topomaps. The topomap for each subject is defined as the evoked activity measured on each sensor over a given time period. Then, the mean of the topomaps was compared across all subjects to produce the mean topomap. Finally, a temporal analysis of the evoked activity was performed by taking signal from two sensors close to the auditory area, one for each side. As a result, the signals compared in the results section corresponding to those measured by the MEG023 (left) and MEG134 (right) gradiometers. Source space analysis For each participant, their 3D MRI acquisition was processed as follows: gray and white matter were segmented, scalp and cortical surfaces were reconstructed and the parcellation of their folding pattern was calculated using the ‘recon-all’ pipeline available in the FreeSurfer 6.0 software package [http://surfer.nmr.mgh.harvard.edu, (Fischl 2012)]. MRI data, meshes for scalp and cortex surfaces, and cortical surface parcellation were then imported into MNE to estimate the sources of the recorded signals. Fiducials and digitized head points were used to align the anatomical MRI scans with the MEG recording, using the co-registration tool included in the MNE software. Following these reconstructions, a head model was computed using the default linear collocation method implemented in MNE, based on the overlapping sphere measured on the cortical surface (Mosher et al. 1999). In addition, noise covariance was computed on a signal with signal-space separation (Taulu et al. 2003), recorded in the empty magnetically shielded room on the day of each participant’s experiment. Finally, source estimates were computed from all sensors in both hemispheres for an equal number of epochs for each participant, each condition, and each stimulus type (standard/deviant). Minimum norm estimation (MNE) of the sources, normalizing noise covariance based on the recording for the empty room (empirical estimation). This process allowed the estimation of signal sources over approximately 20 000 points on the cortical surface for each subject and each ERF. The main parameters for source calculation were loose=0, , snr=3, and depth=0.8. Analysis was restricted to four regions of interest (ROI), based on indications from the literature (Franz et al. 2020; Munoz Musat et al. 2022): the left and right auditory cortices and the dACC. The auditory cortex was defined as the temporal transverse sulcus based on the APARC 2009 parcellation (Destrieux et al. 2010), and the dACC as the combination of the rostral anterior cingulate cortex and the caudal anterior cingulate cortex according to the DKT parcellation (Klein and Tourville 2012). For each ERF, one source was extracted per hemisphere by applying the ‘mean_flip’ method (Gramfort et al. 2013) to the cortical points located within the spatial ROI mentioned above. Finally, for each subject, the absolute value of the evoked response was used to analyze cortex activity. Stability of the evoked responses Before comparing evoked responses between conditions, it is important to check their stability. Therefore, for conditions CC and HYP, the sub-conditions CC1, CC2, and CC3; HYP1, HYP2, and HYP3 were first compared in the sensor and source space to assess a possible overall habituation effect. Those comparisons were performed with 300 standard epochs and 51 deviant ones for each sub-conditions CC1, CC2, CC3, and HYP1, HYP2, and HYP3. The comparison was based on the statistical analysis described below. Statistical analysis To assess the effect of subject’s consciousness state on brain activity, repeated measures permutation analysis of variance (rm-permANOVA) was used. This test is defined as a permutation test where the statistic is the F value computed for a rm-ANOVA: When the rm-permANOVA returned a p-value of less than 0.05, post hoc permutation tests were applied to identify which specific conditions showed significant variations. For these two statistical tests, 10 000 permutations were used. To control for multiple comparisons and to maintain a 5% family-wise error rate, the FDR correction was applied to the post hoc analysis. The amplitude of a component was defined as the average of the brain activity within a given time period. Then, the statistical tests were applied to those averages. The ERF was divided into six components: M100 (75–125 ms), M150 (125–175 ms), M200 (175–225 ms), M250 (225–275 ms), M300 (275–325 ms), and M3b (325–575 ms). Time analysis was performed at two levels: on source estimations (auditory cortex, dACC) and in the sensor space (MEG023, MEG134). Spatial analysis was performed on the sensor space (mean topomap). Results Hypnosis evaluation Participants’ subjective evaluations of hypnotic depth ranged from 3 to 14, with ‘all subjects reporting at least a perceived shift from critical consciousness’. Participants’ hypnotic suggestibility, as measured by their Stanford scores (Weitzenhoffer and Hilgard 1926), ranged from 1/12 to 10/12, correlating broadly with subjective evaluations. Six had low scores, six had moderate scores, and eight had high scores (see Table 1). Notably, we retained participants across the full suggestibility spectrum, including low scorers, to mitigate selection bias and preserve ecological validity. In all cases, there was consensus among the three hypnotherapists that the subject had entered a deep hypnotic phase. The mean durations (±SD) of the hypnosis induction phase for the three sessions were 267 ± 121 s, 173 ± 56 s and 155 ± 50 s, respectively. The corresponding maintenance phase was determined as the intersection of all hypnotherapists’ individual estimations. The corresponding durations were 421 ± 132 s, 473 ± 157 s, and 520 ± 140 s. Stability of responses Brain activity was monitored throughout the experiment. At the brain sources level, during the various stages of critical consciousness (CC1, CC2, and CC3), the amplitude of the ERF signal in both the auditory cortex and the dACC decreased slightly over time in response to both standard and deviant stimuli. This decrease was observed for both the left and right hemispheres (see Table 2). In contrast, for sub-conditions HYP1, HYP2, and HYP3 greater amplitude was recorded during the third repetition (see Table 3). Table 2Stability of the evoked responses in source and sensor spaces during the critical consciousness (CC) conditions.Left hemisphereRight hemisphereAmplitudes (mean [std]) P-value (post-hoc with FDR)Amplitudes (mean [std]) p-value (post-hoc with FDR)[nA.m for sources, fT/cm for sensors] [nA.m for sources, fT/cm for sensors] AreaROIMean CC1Mean CC2Mean CC3CC1-CC2CC1-CC3CC2-CC3Mean CC1Mean CC2Mean CC3CC1-CC2CC1-CC3CC2-CC3 Response to standard stimuli Auditory cortexM10030,7 [24,8]29,9 [23,8]24,1 [18,7]1,01,01,034,2 [29,2]34,1 [26,9]28,8 [22,8]1,01,01,0M15049,3 [44,5]40,6 [39,1]35,9 [28,6] 0,05 0,05 0,2560,9 [51,1]52,9 [49,5]46,7 [32,0]1,01,01,0M20038,7 [23,4]35,1 [21,4]35,3 [23,1]1,01,01,036,6 [24,6]37,9 [24,2]40,6 [25,1]1,01,01,0M25040,3 [22,8]41,6 [26,1]31,3 [19,5]1,01,01,041,2 [25,5]41,2 [22,3]38,7 [19,6]1,01,01,0M30048,8 [32,5]47,1 [35,3]35,7 [24,8]0,780,060,0648,4 [22,4]44,4 [25,0]40,6 [22,7]1,01,01,0M3B28,5 [18,6]25,2 [14,5]23,3 [11,9]0,130,070,436,0 [19,5]31,9 [19,3]28,3 [13,9]0,1 0,05 0,3 ACCM1006,4 [3,8]6,7 [3,6]5,5 [3,1]1,01,01,07,5 [5,1]6,4 [3,9]6,4 [3,7]1,01,01,0M15011,1 [9,5]11,0 [6,1]8,8 [5,5]1,01,01,011,9 [12,6]9,4 [7,0]9,4 [7,6]1,01,01,0M2009,8 [4,9]8,7 [4,8]8,7 [5,2]1,01,01,09,5 [6,4]8,0 [4,3]9,4 [5,8]1,01,01,0M25010,8 [7,0]10,4 [7,1]8,6 [5,2]1,01,01,09,0 [5,8]10,8 [7,4]9,4 [7,6]1,01,01,0M3007,3 [4,8]9,2 [5,4]7,5 [2,6]1,01,01,06,6 [4,8]10,7 [6,8]8,2 [3,3]0,060,30,2M3B6,6 [2,6]5,7 [2,6]5,8 [2,1]1,01,01,06,1 [3,2]6,8 [3,0]7,2 [2,3]1,01,01,0 GRADM10011,7 [8,7]11,2 [7,4]9,9 [5,7]1,01,01,012,9 [9,5]11,1 [6,9]10,4 [6,9]1,01,01,0M15017,8 [18,8]15,2 [15,1]14,3 [12,7]1,01,01,022,0 [19,5]18,3 [15,2]17,7 [11,2] 0,05 0,30,2M20013,5 [7,0]12,4 [7,1]14,9 [8,4]1,01,01,013,6 [6,1]12,5 [5,6]14,9 [9,2]1,01,01,0M25014,5 [7,8]13,5 [8,4]13,0 [10,8]1,01,01,013,8 [6,6]14,3 [6,0]14,8 [7,5]1,01,01,0M30014,2 [6,8]14,5 [9,2]12,1 [7,0]1,01,01,013,9 [7,6]14,1 [7,1]13,2 [6,3]1,01,01,0M3B10,7 [5,5]8,9 [4,3]8,8 [3,3] 0,05 0,060,89,9 [3,4]8,9 [2,5]8,7 [2,6]1,01,01,0 Response to deviant stimuli Auditory cortexM10043,7 [34,3]46,9 [35,1]37,2 [19,8]1,01,01,051,3 [46,4]57,0 [54,5]48,4 [42,0]1,01,01,0M150123,8 [125,8]125,1 [131,1]92,9 [76,7]0,80,060,06109,8 [92,0]112,7 [103,7]91,7 [75,2]1,01,01,0M20054,8 [36,0]65,8 [66,6]56,1 [31,2]1,01,01,057,6 [56,6]73,3 [64,1]52,1 [38,6] 0,05 0,6 0,05 M25065,3 [53,4]67,2 [42,4]70,7 [67,7]1,01,01,094,7 [64,2]78,5 [53,9]74,6 [52,7]0,060,060,7M30071,9 [61,5]77,8 [60,5]63,8 [54,3]1,01,01,087,9 [67,6]72,5 [55,3]68,6 [54,6]0,06 0,05 0,7M3B52,2 [32,9]45,3 [35,7]46,7 [23,0]1,01,01,062,0 [40,0]54,5 [32,0]58,8 [29,7]1,01,01,0 ACCM10011,7 [4,9]11,6 [7,9]11,9 [5,6]1,01,01,010,9 [6,9]13,7 [6,5]12,9 [6,8]1,01,01,0M15021,2 [17,2]17,7 [12,2]15,1 [10,8]1,01,01,018,6 [21,5]15,6 [12,6]20,9 [15,6]1,01,01,0M20013,0 [7,9]12,0 [4,7]12,2 [7,2]1,01,01,013,9 [7,0]14,5 [7,4]14,8 [9,5]1,01,01,0M25018,3 [10,3]18,3 [10,8]13,7 [9,3]1,01,01,017,9 [9,2]17,2 [11,1]17,4 [10,8]1,01,01,0M30014,7 [10,1]16,9 [11,8]15,4 [10,8]1,01,01,018,2 [10,5]19,6 [11,4]18,8 [13,5]1,01,01,0M3B16,2 [9,9]14,3 [5,4]14,4 [5,4]1,01,01,018,1 [14,4]16,0 [6,2]17,5 [10,1]1,01,01,0 GRADM10019,0 [13,2]18,5 [10,8]17,7 [8,0]1,01,01,019,6 [15,1]20,1 [13,0]20,6 [14,0]1,01,01,0M15040,8 [43,6]35,1 [37,4]32,7 [29,7]1,01,01,032,3 [30,1]30,9 [23,7]28,4 [21,9]1,01,01,0M20022,6 [10,4]23,8 [17,5]23,3 [9,8]1,01,01,021,5 [12,4]25,6 [13,5]23,4 [14,4]1,01,01,0M25025,4 [9,4]26,0 [12,4]28,8 [15,7]1,01,01,031,5 [14,3]33,2 [15,1]28,9 [12,0]1,01,01,0M30022,2 [12,4]27,1 [19,0]24,5 [13,0]1,01,01,027,3 [12,3]28,5 [10,6]26,3 [9,9]1,01,01,0M3B21,2 [9,0]20,1 [9,8]19,3 [6,7]1,01,01,021,9 [7,1]20,4 [5,7]19,9 [6,9]1,01,01,0 Stability of the evoked responses in source and sensor spaces during the critical consciousness (CC) conditions. Table 3Stability of the evoked responses in source and sensor spaces during the hypnosis conditions.Left hemisphereRight hemisphereAmplitudes (mean [std]) P-value (post-hoc with FDR)Amplitudes (mean [std]) P-value (post-hoc with FDR)[nA.m for sources, fT/cm for sensors] [nA.m for sources, fT/cm for sensors] AreaROIMean HYP1Mean HYP2Mean HYP3HYP1-HYP2HYP1-HYP3HYP2-HYP3Mean HYP1Mean HYP2Mean HYP3HYP1-HYP2HYP1-HYP3HYP2-HYP3 Response to standard stimuli Auditory cortexM10027,1 [18,6]24,4 [19,2]30,2 [20,7]0,200,70 0,02 26,7 [17,9]23,1 [14,2]25,5 [14,4]1,001,001,00M15024,7 [13,4]23,4 [15,2]25,9 [16,9]1,001,001,0032,5 [24,6]26,9 [24,1]31,3 [24,1]1,001,001,00M20035,1 [24,8]32,4 [23,4]36,6 [23,8]1,001,001,0038,6 [23,8]33,2 [24,1]37,5 [22,9]1,001,001,00M25023,1 [15,6]25,5 [15,9]22,5 [19,1]1,001,001,0025,7 [11,0]24,6 [13,7]24,0 [10,6]1,001,001,00M30025,5 [17,8]30,4 [21,1]31,2 [21,0]1,001,001,0026,4 [11,4]25,9 [11,6]27,2 [11,6]1,001,001,00M3B14,5 [7,2]16,8 [7,0]15,5 [8,8]1,001,001,0017,6 [7,3]18,2 [10,1]18,9 [8,0]1,001,001,00 ACCM1005,5 [3,4]6,1 [4,5]5,5 [3,7]1,001,001,007,2 [5,2]6,6 [5,0]5,9 [3,5]1,001,001,00M1506,4 [3,0]6,9 [4,4]6,2 [3,7]1,001,001,007,5 [4,7]8,0 [4,9]6,9 [3,9]1,001,001,00M2008,0 [3,8]7,9 [3,2]7,2 [4,9]1,001,001,007,6 [5,1]7,0 [3,6]7,7 [7,1]1,001,001,00M2507,4 [4,1]6,1 [3,4]6,0 [5,1]1,001,001,007,3 [4,4]6,0 [3,7]8,5 [5,7]1,001,001,00M3005,6 [3,8]6,8 [4,1]6,3 [3,9]1,001,001,006,4 [2,9]6,9 [3,8]7,6 [6,0]1,001,001,00M3B5,3 [2,7]6,2 [3,3]4,5 [1,9]1,001,001,005,7 [2,9]6,1 [2,6]5,8 [2,3]1,001,001,00 GRADM1009,5 [6,5]9,8 [6,7]11,2 [7,8]0,600,100,109,7 [6,1]9,8 [5,6]9,0 [5,2]1,001,001,00M15010,5 [7,0]9,6 [4,7]10,8 [6,7]1,001,001,0010,7 [8,8]10,5 [9,3]11,2 [8,2]1,001,001,00M20011,6 [6,8]10,7 [5,9]11,5 [8,6]1,001,001,0012,2 [6,7]11,5 [5,8]11,7 [6,2]1,001,001,00M2509,6 [5,6]9,3 [6,2]10,2 [9,1]1,001,001,0011,0 [5,2]10,9 [5,5]9,1 [4,3]1,001,001,00M3008,8 [5,0]9,1 [5,5]9,7 [8,3]1,001,001,0010,9 [4,0]10,2 [4,5]9,7 [3,3]1,001,001,00M3B6,7 [2,7]6,6 [1,6]6,8 [2,1]1,001,001,006,8 [2,0]6,4 [2,8]6,7 [2,2]1,001,001,00 Response to deviant stimuli Auditory cortexM10030,5 [16,9]25,4 [17,3]37,9 [20,3]0,100,10 0,02 34,7 [33,1]45,4 [33,1]31,4 [24,8] 0,04 0,40 0,02 M15049,9 [35,5]50,8 [39,9]55,8 [34,0]1,001,001,0054,9 [52,6]54,3 [54,1]57,5 [56,1]1,001,001,00M20041,1 [26,3]32,5 [22,8]40,9 [24,6]1,001,001,0043,7 [36,0]46,8 [31,6]42,6 [29,8]1,001,001,00M25049,2 [30,0]51,8 [25,4]52,5 [42,2]1,001,001,0049,7 [28,9]48,0 [30,7]59,4 [35,4]1,001,001,00M30048,0 [28,2]43,7 [32,9]51,0 [31,7]1,001,001,0046,6 [29,7]57,6 [38,6]53,8 [33,3]1,001,001,00M3B30,9 [10,4]34,4 [17,5]35,9 [18,7]1,001,001,0033,7 [21,2]35,9 [24,5]40,6 [26,7]1,001,001,00 ACCM1009,7 [5,2]12,5 [7,2]11,0 [6,4]1,001,001,0011,8 [7,1]11,8 [7,0]10,8 [5,8]1,001,001,00M15016,4 [7,2]16,3 [10,0]13,2 [10,0]1,001,001,0016,6 [10,3]14,8 [9,5]15,7 [11,2]1,001,001,00M20013,3 [7,4]12,3 [5,8]13,1 [10,2]1,001,001,0014,1 [8,0]11,0 [5,6]13,9 [7,8]1,001,001,00M25014,6 [9,1]13,3 [5,2]16,6 [9,1]1,001,001,0015,1 [8,9]12,8 [4,7]11,3 [5,7]1,001,001,00M30012,3 [6,8]15,9 [7,5]12,9 [6,4]1,001,001,0015,2 [7,8]17,1 [10,9]13,0 [6,6]1,001,001,00M3B10,3 [3,6]15,3 [5,6]13,5 [7,0] 0,02 0,080,5013,0 [6,8]15,3 [6,7]12,5 [4,0]1,001,001,00 GRADM10014,3 [6,3]13,6 [5,6]18,4 [8,5]0,50 0,03 0,02 15,3 [7,7]16,1 [10,1]15,3 [8,6]1,001,001,00M15021,0 [19,7]20,1 [14,1]22,9 [16,1]1,001,001,0022,7 [15,8]21,2 [17,9]22,5 [14,2]1,001,001,00M20016,5 [8,4]16,3 [9,2]17,6 [9,2]1,001,001,0018,6 [8,5]18,5 [7,9]21,1 [8,4]1,001,001,00M25018,8 [7,5]21,8 [8,4]22,7 [14,6]1,001,001,0020,5 [9,6]20,2 [10,9]22,7 [9,3]1,001,001,00M30017,3 [8,7]18,5 [7,1]21,4 [11,4]1,001,001,0017,5 [7,5]20,7 [11,3]18,7 [8,5]1,001,001,00M3B14,2 [5,2]16,4 [4,8]18,6 [7,6]0,10 0,02 0,2015,4 [5,2]15,8 [5,1]16,8 [5,2]1,001,001,00 Stability of the evoked responses in source and sensor spaces during the hypnosis conditions. At the sensor level over time, the same analysis confirmed the decreasing trend for the CC conditions and the increasing amplitude for the hypnosis conditions for the late components (see Table 2 and Table 3). Source analysis of brain activity Effect of the state of the subjects in the auditory cortex The evoked brain response to both standard and deviant stimuli in the auditory cortex revealed a smaller amplitude when subjects were in the HYP or DIS conditions than when they were in the CC conditions. This difference was observed in both hemispheres, even though the signal topology was the same. The auditory evoked brain response did not allow us to distinguish between the HYP and DIS conditions, except for a short period in the left hemisphere after both stimuli, where the amplitude of the response in the DIS condition was observed to be smaller than in the HYP condition. Continuous time analysis of the cerebral evoked signal indicated that the difference between the CC condition and the other conditions appeared 100 milliseconds after the auditory stimulus (standard or the deviant), in both hemispheres. This difference was statistically significant (P ≤ 0.05) (see Fig. 2 and Tables 4, 5, 6, and 7). As with the deviant responses, the analysis of the MMN revealed a decrease in amplitude in the HYP and DIS conditions compared to the CC one, and a greater M150 amplitude in the HYP condition than in the DIS condition in the left hemisphere (see Fig. 3). The statistical analysis did not identify any additional modulations related to HYP and DIS conditions. Average auditory ERFs and MMN for 20 subjects in the right and left auditory regions of the brain (circled in blue in the brain plot). The blue, green, and red responses correspond respectively to the CC, DIS, and HYP conditions. The blue and red areas correspond to the auditory stimulus duration. The green area on the graphs corresponds to the time steps where the rm-ANOVA showed significant differences between the three conditions CC, DIS, and HYP. The bars above the curves in the green area indicate the two conditions that showed significantly different values after post hoc analysis with FDR correction. Table 4Auditory evoked response component comparisons in the left hemisphere after a standard stimulus.Amplitudes (mean [std]) P-value (post hoc with FDR correction)[nA.m for sources, fT/cm for sensors]AreaROIMean CCMean DISMean HYPCC-DISCC-HYPDIS-HYPAuditory cortexM10028,0 [21,1]22,6 [15,6]28,8 [21,1]111M15040,0 [35,5]19,8 [10,6]24,4 [14,0]0,0510,0990,307M20035,5 [21,6]26,6 [19,9]35,3 [24,2]111M25035,0 [20,3]23,2 [13,0]25,3 [16,3]0,1160,1440,736M30043,3 [27,6]20,0 [12,8]27,4 [18,5] <0,01 0,0820,282M35029,9 [25,3]17,3 [9,9]20,4 [13,9]0,0670,1450,506M40024,4 [20,3]9,6 [5,4]15,5 [8,5] 0,031 0,1490,057M45020,0 [11,7]10,4 [5,6]15,9 [9,6] 0,045 0,3070,083M50017,9 [12,7]8,0 [4,8]11,2 [7,0] 0,038 0,0820,237M55017,8 [11,8]7,6 [4,8]12,7 [6,3] 0,018 0,063 0,019 ACCM1006,2 [3,9]4,9 [3,1]6,2 [4,7]111M1509,3 [7,1]6,0 [3,7]6,3 [3,6]0,1490,1020,842M2008,7 [4,4]4,5 [2,6]7,3 [5,3] 0,022 0,468 0,021 M25010,9 [6,7]4,7 [2,5]6,3 [4,0] 0,019 0,018 0,359M3008,5 [3,7]5,2 [3,3]6,5 [4,5] 0,045 0,2670,29M3506,9 [4,0]5,1 [2,4]5,9 [3,9]111M4007,1 [4,0]4,5 [2,8]5,8 [3,9]0,0710,3080,303M4507,8 [5,1]4,5 [2,7]5,9 [3,9]0,0680,2480,266M5006,6 [4,3]4,6 [3,1]4,1 [3,0]111M5506,6 [4,1]4,7 [2,6]4,3 [2,1]0,1820,0850,741GRADM10010,3 [7,0]8,4 [5,5]11,6 [7,3]0,1140,285 0,036 M15015,2 [15,5]8,8 [5,7]9,6 [5,8] 0,021 0,0820,561M20012,1 [6,2]8,1 [4,9]12,2 [8,3]0,0820,9710,073M25013,7 [7,7]8,5 [3,6]10,0 [6,6] <0,01 0,048 0,294M30014,0 [7,1]7,8 [4,5]9,7 [6,8] <0,01 0,0570,22M35010,9 [5,1]6,6 [2,8]8,3 [4,0] 0,021 0,1450,149M40010,1 [5,3]5,9 [1,9]7,1 [3,0] 0,018 0,1430,237M4509,3 [4,1]5,9 [2,1]6,7 [3,5] 0,026 0,1450,458M5009,2 [4,8]6,2 [2,0]7,0 [2,8]0,0670,2040,338M5508,1 [3,9]5,7 [1,9]6,3 [3,3] 0,036 0,230,567 Auditory evoked response component comparisons in the left hemisphere after a standard stimulus. Table 5Auditory evoked response component comparisons in the right hemisphere after a standard stimulus.Amplitudes (mean [std]) P-value (post hoc with FDR correction)[nA.m for sources, fT/cm for sensors]AreaROIMean CCMean DISMean HYPCC-DISCC-HYPDIS-HYPAuditory cortexM10031,1 [25,1]18,7 [11,4]26,6 [19,6] 0,038 0,3090,062M15050,0 [42,2]25,7 [22,0]29,7 [24,1] 0,019 0,045 0,267M20039,0 [23,3]32,4 [23,4]35,0 [21,5]111M25038,0 [16,9]21,3 [12,8]27,1 [14,0] 0,017 0,1140,269M30043,1 [22,6]24,0 [12,3]26,7 [11,3] <0,01 0,045 0,567M35039,0 [22,2]17,1 [11,2]21,4 [11,3] <0,01 0,024 0,173M40037,3 [26,0]15,3 [9,5]19,6 [15,0] 0,014 0,038 0,337M45030,5 [20,3]17,3 [10,4]18,9 [16,8]0,0710,0990,784M50029,0 [17,6]13,4 [8,2]18,7 [12,5] <0,01 0,04 0,143M55021,6 [14,7]12,4 [6,1]14,7 [10,2]0,0710,1020,548ACCM1006,6 [4,0]5,6 [3,0]7,5 [6,9]111M1509,6 [8,6]8,5 [5,7]6,7 [4,2]111M2008,0 [5,4]6,7 [5,4]7,6 [5,2]111M2509,6 [6,8]6,3 [3,7]7,0 [5,0]111M3007,8 [4,7]5,0 [2,3]7,2 [4,5]111M3507,1 [4,2]4,3 [2,3]7,0 [4,2]0,0610,9320,102M4006,7 [4,6]4,5 [1,7]7,0 [4,4]0,1020,8910,052M4506,8 [4,5]4,7 [3,5]6,9 [4,4]111M5006,5 [2,9]4,9 [2,4]5,4 [3,1]111M5505,1 [2,7]4,9 [2,2]5,5 [3,3]111GRADM10010,4 [8,0]7,8 [5,4]9,6 [5,3] 0,05 0,5580,116M15019,3 [14,8]10,3 [10,1]10,8 [8,9] <0,01 <0,01 0,738M20013,9 [6,1]9,4 [7,5]12,3 [5,1] 0,017 0,2480,086M25012,9 [5,1]7,9 [3,7]11,0 [4,9] 0,022 0,3150,052M30013,3 [6,0]7,8 [3,4]10,2 [3,9] 0,022 0,0920,071M35011,6 [4,0]6,8 [2,9]8,5 [4,0] 0,017 0,0660,122M4009,8 [3,9]6,3 [3,3]7,8 [4,0]0,0540,3070,2M4509,5 [4,1]7,2 [3,6]6,4 [2,8]0,1730,0650,564M5008,4 [2,9]5,7 [2,6]6,4 [3,4]0,0430,1930,521M5507,7 [3,3]5,8 [2,1]5,9 [2,1]0,070,0710,899 Auditory evoked response component comparisons in the right hemisphere after a standard stimulus. Table 6Auditory evoked response component comparisons in the left hemisphere after a deviant stimulus.Amplitudes (mean [std]) P-value (post hoc with FDR correction)[nA.m for sources, fT/cm for sensors]AreaROIMean CCMean DISMean HYPCC-DISCC-HYPDIS-HYPAuditory cortexM10031,1 [25,1]18,7 [11,4]26,6 [19,6]111M15050,0 [42,2]25,7 [22,0]29,7 [24,1] <0,01 0,052 <0,01 M20039,0 [23,3]32,4 [23,4]35,0 [21,5] 0,038 0,1430,293M25038,0 [16,9]21,3 [12,8]27,1 [14,0] 0,019 0,038 0,587M30043,1 [22,6]24,0 [12,3]26,7 [11,3]0,0620,1850,222M35039,0 [22,2]17,1 [11,2]21,4 [11,3]111M40037,3 [26,0]15,3 [9,5]19,6 [15,0]111M45030,5 [20,3]17,3 [10,4]18,9 [16,8]111M50029,0 [17,6]13,4 [8,2]18,7 [12,5]0,0620,3730,145M55021,6 [14,7]12,4 [6,1]14,7 [10,2]111ACCM1006,6 [4,0]5,6 [3,0]7,5 [6,9]111M1509,6 [8,6]8,5 [5,7]6,7 [4,2]111M2008,0 [5,4]6,7 [5,4]7,6 [5,2]111M2509,6 [6,8]6,3 [3,7]7,0 [5,0]111M3007,8 [4,7]5,0 [2,3]7,2 [4,5]111M3507,1 [4,2]4,3 [2,3]7,0 [4,2]111M4006,7 [4,6]4,5 [1,7]7,0 [4,4]111M4506,8 [4,5]4,7 [3,5]6,9 [4,4]111M5006,5 [2,9]4,9 [2,4]5,4 [3,1]111M5505,1 [2,7]4,9 [2,2]5,5 [3,3]111GRADM10010,4 [8,0]7,8 [5,4]9,6 [5,3]111M15019,3 [14,8]10,3 [10,1]10,8 [8,9] <0,01 0,045 0,149M20013,9 [6,1]9,4 [7,5]12,3 [5,1]0,096 0,043 0,979M25012,9 [5,1]7,9 [3,7]11,0 [4,9] 0,022 0,036 0,267M30013,3 [6,0]7,8 [3,4]10,2 [3,9]0,1560,20,293M35011,6 [4,0]6,8 [2,9]8,5 [4,0]111M4009,8 [3,9]6,3 [3,3]7,8 [4,0]111M4509,5 [4,1]7,2 [3,6]6,4 [2,8]111M5008,4 [2,9]5,7 [2,6]6,4 [3,4]111M5507,7 [3,3]5,8 [2,1]5,9 [2,1]111 Auditory evoked response component comparisons in the left hemisphere after a deviant stimulus. Table 7Auditory evoked response component comparisons in the right hemisphere after a deviant stimulus.Amplitudes (mean [std]) P-value (post hoc with FDR correction)[nA.m for sources, fT/cm for sensors]AreaROIMean CCMean DISMean HYPCC-DISCC-HYPDIS-HYPAuditory cortexM10031,1 [25,1]18,7 [11,4]26,6 [19,6]0,17 0,024 0,63M15050,0 [42,2]25,7 [22,0]29,7 [24,1] 0,038 0,028 0,945M20039,0 [23,3]32,4 [23,4]35,0 [21,5]111M25038,0 [16,9]21,3 [12,8]27,1 [14,0]0,130,0520,637M30043,1 [22,6]24,0 [12,3]26,7 [11,3] 0,045 0,1820,143M35039,0 [22,2]17,1 [11,2]21,4 [11,3]0,082 0,037 0,915M40037,3 [26,0]15,3 [9,5]19,6 [15,0]111M45030,5 [20,3]17,3 [10,4]18,9 [16,8]111M50029,0 [17,6]13,4 [8,2]18,7 [12,5]0,0690,0630,945M55021,6 [14,7]12,4 [6,1]14,7 [10,2]0,2070,0690,543ACCM1006,6 [4,0]5,6 [3,0]7,5 [6,9]111M1509,6 [8,6]8,5 [5,7]6,7 [4,2]111M2008,0 [5,4]6,7 [5,4]7,6 [5,2]111M2509,6 [6,8]6,3 [3,7]7,0 [5,0]111M3007,8 [4,7]5,0 [2,3]7,2 [4,5]111M3507,1 [4,2]4,3 [2,3]7,0 [4,2]111M4006,7 [4,6]4,5 [1,7]7,0 [4,4]111M4506,8 [4,5]4,7 [3,5]6,9 [4,4]111M5006,5 [2,9]4,9 [2,4]5,4 [3,1]111M5505,1 [2,7]4,9 [2,2]5,5 [3,3]111GRADM10010,4 [8,0]7,8 [5,4]9,6 [5,3]111M15019,3 [14,8]10,3 [10,1]10,8 [8,9]0,1020,1650,766M20013,9 [6,1]9,4 [7,5]12,3 [5,1]111M25012,9 [5,1]7,9 [3,7]11,0 [4,9] <0,01 0,017 0,559M30013,3 [6,0]7,8 [3,4]10,2 [3,9] 0,017 0,05 0,419M35011,6 [4,0]6,8 [2,9]8,5 [4,0] 0,022 0,0570,567M4009,8 [3,9]6,3 [3,3]7,8 [4,0]111M4509,5 [4,1]7,2 [3,6]6,4 [2,8] 0,017 0,1450,443M5008,4 [2,9]5,7 [2,6]6,4 [3,4] 0,038 0,0720,842M5507,7 [3,3]5,8 [2,1]5,9 [2,1]111 Auditory evoked response component comparisons in the right hemisphere after a deviant stimulus. Average auditory ERFs and MMN for 20 subjects in the right and left dorsal anterior cingulate cortex (purple in the brain plot). The blue, green, and red responses correspond respectively to the CC, DIS, and HYP conditions. The blue and red areas correspond to the auditory stimulus duration. The green area corresponds to the time steps where the rm-ANOVA returned significant differences between the three conditions CC, DIS, and HYP. The bars above the curves in the green area indicate the two conditions that had significantly different values after post hoc analysis with FDR correction. Effect of the state of the subjects in the dorsal anterior cingulate cortex (dACC) The brain response at the dACC also showed smaller amplitudes in the DIS and HYP conditions than those measured in the CC condition. Although smaller than in the auditory cortex, a difference was recorded in both hemispheres after both stimuli (see Fig. 3). We detected no differences between HYP and DIS. The statistical analysis of the MMN did not highlight any difference between the three conditions. Sensor analysis of temporal data Topomap of the evoked responses At the sensor level, the evoked response was mainly identified in temporal and parietal locations, whatever the condition or stimulus. The comparison of the CC and HYP/DIS conditions also localized in the temporal areas, as shown by the post hoc tests of the evoked fields measured by the gradiometers in Fig. 4. In addition to the temporal modulation, some smaller modulations appeared in the parietal areas (particularly at M150 and M300 for the standard stimulus). In contrast, the difference between the HYP and DIS conditions tended to be small and mostly co-localized with the CC-HYP differences (see Fig. 4). Finally, based on analysis of our topomaps, the difference between the conditions when using the deviant stimulus was found to be greater than when using the standard stimulus. Those statistically significant areas appeared as more compact patches on the topomaps. Post hoc analysis of the topomap for M150, M250, and M300 components produced with data from all 20 subjects after applying FDR corrections. Only two comparisons are shown (CC versus HYP and DIS versus HYP) for each stimulus type [standard (top) and deviant (bottom)]. The color scale reflects the differences in amplitude between the two conditions compared. Black dots show the gradiometer-pair locations; white circles correspond to gradiometer-pairs where differences between the two conditions were statistically significant (permutation test with FDR correction). The topomaps highlight the predominance of temporal and parietal areas in the discrimination of conditions. Effect of the state of the subjects on their AEF Our main findings indicated that HYP and DIS are linked, associated with a decrease in late amplitudes in gradiometer measurements. However, no strong differences in amplitudes were observed that could distinguish between the HYP and DIS. The analysis of the MMN showed only smaller M250 amplitude in DIS than in CC. The detailed results are presented in Fig. 5 and in the and in the Tables 4, 5, 6, and 7. Average auditory ERFs and MMN measured on sensors 023 (left) and 134 (right) for 20 subjects. The sensors used for the analysis are highlighted in the yellow sensor layout (blue dots). The blue, green, and red responses correspond respectively to the CC, DIS, and HYP conditions. The blue and red areas correspond to the auditory stimulus duration. The green area of the graph corresponds to the time steps where the rm-ANOVA indicated that differences between the three conditions CC, DIS, and HYP are significant. The bars above the curves in the green area indicate the two conditions for which post hoc analysis with FDR correction revealed significant differences. Discussion The aim of this study was to determine whether the hypnotic process produces a modulation of the cerebral evoked response after an auditory stimulation that could be used in clinical practice as a biomarker to monitor hypnosis. While our analyses revealed robust neurophysiological signatures associated with altered attentional processing, a key observation was the challenge in ‘distinguishing hypnotic from distraction conditions’ based on evoked responses alone. The challenge of distinguishing DIS from HYP aligns with Erickson’s view of hypnosis as an intensified form of everyday absorption (Erickson 1970), where the boundary between states is fluid, and prompts a critical consideration: ‘Does clinical practice actually require such a distinction?’ The clinical goal is generally to induce focused immersion via hypnosis or distraction (e.g. VR) without differentiating between these states. In a recent work, Rousseaux (Rousseaux et al. 2025) found minimal outcome differences between hypnosis and VR for reducing pain and anxiety, favoring VR. This suggests that their shared attentional mechanisms may outweigh their neurophysiological distinctions in applied settings. In this study, the Stanford scores, the correlation with the participants’ subjective perceptions, and the hypnotherapists’ subsequent reviews confirmed the presence of hypnotic trances during the experiments. However, a key limitation was our inability to continuously monitor wakefulness levels during conditions. Although Stroop tasks were administered periodically to verify alertness, participants’ closed eyes, while ecologically valid for hypnosis, precluded visual assessment. The distribution of the scores obtained replicated the expected population distribution (De Saldanha da Gama et al. 2012). This alignment with large-scale data suggests our findings may generalize beyond highly suggestible individuals (Jensen et al. 2017). Our analyses revealed significant condition-dependent modulations in the cerebral structures involved in processing sensory signals and at the sensor level, in both hemispheres. More precisely, the evoked response was modulated with a decrease in ERF amplitude, mainly in late components, from M100 to M3b. The M100 component exhibited no evidence of cross-condition modulation (CC/DIS/HYP). However, within CC, deviant stimuli elicited a robust M150 response, aligning with the expected oddball effect (Näätänen et al. 2007). As the brain’s response to a sensory stimulus reflects both the perception of information and its processing, the early components of the ERF are linked to direct perception, while the subsequent components are associated with decision-making and response-related processes (Luck et al. 2000). The presence of an early M300 component in our MMN results, peaking between 250–300 ms post-stimulus onset, seems to align with findings from García-Larrea et al. (1992), who reported a P250 response to unexpected, non-targeted auditory stimuli in EEG recordings. García-Larrea et al. suggested that the P250 (or M250) component is linked to a stimulus classification process. However, in our case, this suggestion does not fully explain the apparent conservation of the late component while the M150 is reduced. In addition, while some significant differences appear in the latest components (M250, M300), distinguishing between distracted and hypnotic conditions remains challenging. The oddball paradigm is typically used to elicit a response from a deviant or a target stimulus. (Franz et al. 2020; Munoz Musat et al. 2022). However, our results also demonstrate modulations of responses to standard stimuli, not previously reported by those studies. The explanation for this difference may lie in their protocol implementation. As it seems that their implementation does not include competing auditory stimulation during the measurement of the oddball paradigm (the distractor itself or the voice of the hypnotherapist), the discrepancy in the modulation of responses to standard tones may be associated with attentional processes. Compared to the first CC condition (start of our tests), we observed a decrease in amplitude of the ERF during the third CC condition. This finding correlates with habituation, where the brain minimizes the impact of similar repeated stimulations (Ross et al. 2020). A lower amplitude of auditory ERP during hypnosis has previously been reported with EEG (Franz et al. 2020; Munoz Musat et al. 2022), but under experimental conditions where deafness to the target sounds was induced as part of the hypnotic suggestions. Here, we used a hypnosis technique based on reminiscence of a pleasant memory, without suggesting that subjects ignore sounds. We measured a similar decrease in brain response to standard and deviant sounds. This similarity may reveal a common mechanism between suggested deafness and neutral hypnosis. As indicated above, the decrease in amplitude occurred during the late phase of ERF, and affected the M150, M200, M250, and M3b components. De Pascalis (2024) describes hypnosis as a dynamic and complex interaction between sensory processing, memory systems, attentional mechanisms, and higher-order cognitive functions. The fact that in our study, the change in ERF signal during hypnosis affected only the late components of this response supports the idea that hypnosis alters secondary processing of the sensory stimulus, especially attentional one. For Luck et al. (2000), the amplitude of the event-related components is related to the degree of attention paid to the stimulus. Therefore, the decrease in amplitude of M150 may be interpreted as related to an attenuation of processes linked to detecting external stimuli. Moreover, for the dACC source, we observed a smaller amplitude of the M3b component in the HYP conditions compared to the CC condition. Many authors (Sergent et al. 2005; Naccache et al. 2016; Munoz Musat et al. 2022) have suggested that a higher amplitude in the M3b (or P3b) component of the dACC is associated with conscious perception of a stimulus. In light of this, the reduced response to deviant stimuli observed during hypnosis suggests a diminished conscious perception of these stimuli. Importantly, in this study, the modulation of auditory responses induced by hypnosis was observed for all subjects, regardless of their Stanford scores. This is of particular interest, as most studies up to now specifically selected their participants among only the most suggestible people. As a result, the biomarker proposed here is likely to be reliable in a real-world scenario, as it seems to be independent of hypnotic suggestibility. However, the relationship between cortical responses and different levels of suggestibility will now need to be confirmed using a larger group of subjects. In our study, the HYP and DIS conditions differed significantly from the critical consciousness condition. A notable reduction in ERF amplitude was observed 100 ms after the auditory stimulus. Although it was difficult to distinguish between distraction and hypnosis because the overall changes produced were very similar, we identified some elements specific to each of these two conditions. Thus, after both stimuli, in the left hemisphere, a short period was identified during which the amplitude of the response in the DIS condition was smaller than the response in the HYP condition. The difficulty of distinguishing between the distraction and hypnosis conditions raises questions about whether the decreased brain response to sound is linked to distraction or to the hypnotic process itself. The underlying question is whether a phase of deep hypnosis was effectively reached. While some authors have questioned the need to induce hypnosis (Callara et al. 2023), we believe that our experimental protocol enabled us to ensure actual hypnosis. Indeed, we only analyzed the MEG signals for recordings from subjects in whom the hypnotic trance had been independently confirmed by three hypnotherapists, indicated by involuntary movements, and confirmed by the participants’ responses to the debriefing questionnaire and phenomenological evaluations. Additionally, other authors have previously reported a decrease in EEG response amplitude during hypnotic trance (Franz et al. 2020). Another possible explanation for the relatively similar DIS and HYP conditions, and limitation in this study, relates to our paradigm. First, the differences between these conditions may have been masked by the small population size. Secondly, the fixed order of conditions, with DIS always preceding HYP, may have introduced order effects. This design choice was intentional: to mitigate the risk of auto-hypnosis during the DIS condition, we exclusively recruited hypnosis-naïve participants and avoided randomization in the absence of empirical data to determine an appropriate washout period between conditions. The method used to induce distraction could also be questioned. Here, we relied on the concentration required to listen to a story to achieve this. We chose distraction based on the spoken word in the hope of differentiating between the effects of passive listening and listening to suggestions. However, if the listener’s attention did not stay focused on the story, they may experience mind wandering. This shift in attention from external to internal processing, along with the engagement in states of internal absorption, shares phenomenological similarities with hypnotic trance (Jamieson 2009). The inconsistent findings of studies on the link between auditory responses, particularly MMN, and hypnosis studies (Jutai et al. 1993; Hiltunen et al. 2019; Jamieson et al. 2005b) may partially stem from shared methodological challenges in distinguishing hypnosis from distraction. For instance, the absence of effect in Hiltunen et al. work could be attributed to their use of visual distractors during measurements, potentially adding distraction effects during the control conditions. Similarly, Jamieson et al. reported an increase in MMN responses during hypnosis, but it is unclear whether hypnotic communication was maintained throughout the deep phase of hypnosis. Finally, Jutai et al. employed pre-recorded audio, which, while standardized, may not fully capture the interactive, dynamic nature of clinical hypnosis. Franz et al. (2020) used a different approach to create distraction, but nevertheless reported a similar difficulty in distinguishing between the two conditions based on amplitude of ERP in response to an auditory oddball scenario based on EEG recordings. The explanation for this similarity between conditions may lie in how the hypnosis is defined. According to Elkins et al. (2015), hypnosis is defined as a state of consciousness involving focused attention and reduced peripheral awareness characterized by an enhanced capacity to respond to suggestions. Conversely, distraction can be achieved by focusing on a specific mental task that diverts attention from external stimuli. Based on these definitions, the two situations share one thing: focused attention. For Erickson (1970), the founder of modern hypnosis, the hypnotic state of consciousness is a natural state used in everyday life, based on extreme concentration on a subject or situation, which has the effect of distancing the subject from their immediate environment. For example, this everyday trance state is achieved by people who are immersed in a book or film to the extent that they become completely abstracted from their environment. When induced by a third party, this state can reach a greater depth, culminating in a deep hypnotic trance. This view of hypnosis presupposes the existence of a unique cerebral mode of operation, the intensity of which can vary. If this mechanism is responsible for the modification of the ERF/ERP signal during formal hypnosis, then it could also be triggered when listening to a story, if the listener is fully concentrated on the story being told. This approach to hypnosis seems to be supported by the results of the experiment by Franz et al. (Franz et al. 2020), who reported a progressive decrease in the amplitude of the ERP signal between the control and the distraction conditions and a further decrease with the formal hypnosis. The fact that these authors were unable to use this measurement to distinguish between true and simulated hypnosis may be explained by the level of concentration required by the subjects to simulate the deep phase of hypnosis. Moreover, if such a marker of hypnosis were used in clinical practice, it would not necessarily be essential to distinguish between distraction and hypnosis, since both approaches help the patient to distance themselves from their environment. In addition to the source localization analysis, we present a single-sensor analysis in this MEG study. While a previous analysis of the same dataset demonstrated the potential for using ERFs to predict hypnotic trance, regardless of whether the stimuli are standard or deviant (Sid-Ahmed et al. 2025), we believe that this type of sensor analysis is crucial to estimate the feasibility of monitoring a subject’s state in a constrained environment like an operating theater. However, this method shows weaker statistical significance, which we suspect is due to the variability in the positioning of the sensors relative to each subject’s head. Although it remains difficult to distinguish between distraction and hypnosis, the decrease in the amplitude of the M150, M250, M300, or M3b could potentially serve as a universal biomarker for shifting from critical consciousness to an altered mode of attentional processing, of the type sought during hypnotherapy. This main experimental result will need to be confirmed in further study of the evoked responses associated with an auditory oddball scenario with a larger population. In these studies, the modulation of the evoked responses by a suggestion of deafness should be compared to non-directive hypnosis in the same person. The results can be used to assess our hypothesis that it is enough to be in a hypnotic trance for deafness to be induced. Conclusion The present MEG study shows that the brain responses evoked by simple auditory stimuli are modulated by the state of consciousness of the subject, especially in later latencies, regardless of the subject’s Stanford score. The use of an oddball paradigm revealed that a stronger response and stronger modulation were obtained for deviant stimuli than for standard stimuli. This effect was found with both gradiometer measurements and source calculations in the auditory cortex. M150, M250, and M300 amplitudes were lower for subjects in a non-directive hypnotic trance or in conditions leading to auditory distraction than in subjects in a critical consciousness state. Although harder to distinguish, differences in amplitude in late latencies were also observed between hypnotic trance and distraction, with distraction producing a lower amplitude. The estimation of brain activity in the dACC showed a reduction in M300 amplitudes when subjects were in hypnotic trance. This result supports previous observations from (Munoz Musat et al. 2022) and the hypothesis that the dACC may play a crucial role in the conscious perception of auditory stimuli. Supplementary Material Acknowledgements We would like to thank Caroline Sandre-Ballester for her administrative support for the clinical protocol, the clinical team at Clinatec for technical support, and Maighread Gallagher for proofreading the article. Contributor Information Hafid Sid-Ahmed, CEA, Leti, Clinatec, Univ. Grenoble Alpes, F-38000 Grenoble, France. Julie Alayrangues, CEA, Leti, Clinatec, Univ. Grenoble Alpes, F-38000 Grenoble, France. Lilia Langar, Grenoble University Hospital, Univ. Grenoble Alpes, F-38000 Grenoble, France. Nelly Richard, Grenoble University Hospital, Univ. Grenoble Alpes, F-38000 Grenoble, France. Véronique Albaladejo, Grenoble University Hospital, Univ. Grenoble Alpes, F-38000 Grenoble, France. Stéphane Pezzani, Grenoble University Hospital, Univ. Grenoble Alpes, F-38000 Grenoble, France. Virginie Brun, CEA, Leti, Clinatec, Univ. Grenoble Alpes, F-38000 Grenoble, France. Daniel Anglade, Grenoble University Hospital, Univ. Grenoble Alpes, F-38000 Grenoble, France. Vincent Auboiroux, CEA, Leti, Clinatec, Univ. Grenoble Alpes, F-38000 Grenoble, France. Author contributions Hafid Sid-Ahmed (Formal analysis [equal], Software [equal], Visualization [equal], Writing—original draft [equal]), Julie Alayrangues (Investigation [equal], Methodology [equal], Software [equal], Writing—review & editing [equal]), Lilia Langar (Conceptualization [equal], Investigation [equal]), Nelly Richard (Investigation [equal]), Véronique Albaladejo (Investigation [equal]), Stéphane Pezzani (Conceptualization [equal], Funding acquisition [equal], Project administration [equal]), Virginie Brun (Funding acquisition [equal], Writing—review & editing [equal]), Daniel Anglade (Conceptualization [equal], Investigation [equal], Methodology [equal], Writing—review & editing [equal]), and Vincent Auboiroux (Conceptualization [equal], Funding acquisition [equal], Methodology [equal], Project administration [equal], Writing—review & editing [equal]) Conflict of interest None declared. Funding This work was supported by two grants awarded by Fonds Clinatec (‘Nouvelle équipe 2018’ and ‘Edmond J. Safra 2019’) and one doctoral scholarship ‘Focus Biomarqueurs’ from CEA. Data availability The data supporting the findings of this study are available upon reasonable request from the corresponding author. References Associated Data Supplementary Materials Data Availability Statement The data supporting the findings of this study are available upon reasonable request from the corresponding author.