What is hypnosis. Hypnosis, a physiological modification of egoic consciousness, is a dynamic process deriving from a mental representation auto- or hetero-induced. We previously demonstrated that this mental image, active in a psychosomatic frame, produces effects not only at a psychic level (the field of application of psychotherapy), but also at a physical level, e.g. cardiovascular (Casiglia et al., 2007, 2012, 2015, 2016, 2018b; Facco et al., 2011) , splanchnic (Tikhonoff et al., 2018) , immunological (Wood et al., 2003) , neurological (Casiglia et al., 2018a) and so on. This process, never clearly defined, has been called in different ways, such as ideomotor effect, ideoplasia, ideodynamism, and plastic monoideism (Casiglia, 2012) . The more the mental image is strong and concrete, the more the hypnotic state is deep.
Is hypnotic state consciously, critically and voluntarily accepted? Although not universally shared, the prevalent belief is that hypnosis (“hypnotic trance”) is consciously experienced, and hypnotism (the action of producing hypnosis) is voluntarily and critically accepted by the participant. Nevertheless, the few studies previously aimed at finding the neural processes associated with hypnotic modified consciousness were disappointing (Faymonville et al., 2000) , and the need of clear demonstration that hypnosis is a conscioius and voluntary process is compelling.
The present study uses functional magnetic resonance imaging (fMRI) to verify if conscious and voluntary quality of hypnosis can be experimentally demonstrated employing the means, methods and instruments that are typical of human physiology and medical semiotics.
Why hypnosis during a task? As the imaging obtained in neutral hypnosis outside a task does not give definite information (McGeown et al., 2009; Mazzoni et al., 2013) , a specific task, represented herein by hypnotic focused analgesia (HFA), was applied in the present research.
Aim of the study. The study described herein is aimed at clarifying whether hypnosis, when realized in the frame of a task, is characterized by activation of brain areas that are associated with consciousness and free will. The final aim of the study is to clarify if, based on brain imaging, it is plausible to state that the participants adhere to hypnosis consciously and voluntarily.
The study group was represented by twenty young, healthy, highly hypnotizable participants (Harvard Group Scale of Hypnotisability score ≥ 9) (De Pascalis et al., 2000) , represented by 5 men and 15 women aged 30.3 ± 9.7 years, whose educational level was 18.1 ± 2.6 years of schooling. They accepted to participate because of scientific curiosity and in order to offer an opportunity to clarify a hot topic. High hypnotizability was requested by the protocol, which was aimed at demonstrating variations of activation of brain areas in response to hypnosis. Subjects free from clautrophobia were selected, as permanence in the fMRI tunnel for at lest one hour was required.
2.2. Safety and Ethics
Participants were judged fit for hypnosis on the basis of anamnestic questionnaires, confidential interview with the principal investigator and, when necessary, the Minnesota Multiphasic Personality Inventory 2 Restructured Form (Locke, 2013) . This procedure was aimed at screening the participants with a borderline personality more prone to develop side effects during hypnotic dissociation.
The study was approved by the Ethics Committee of the
Preliminary session. All participants preliminarily underwent individual hypnotic induction through verbal suggestions. The aim of this preparatory procedure was to establish an interpersonal rapport between the operator and the participant, in order to favour a rapid and valid monoideism during the following experimental setting. The voice of an expert hypnotist guided each participant towards focusing their attention on a single idea, excluding any other external or internal stimuli. Hypnotic induction consisted of a brief enumeration coupled with suggestions of eyelid heaviness and staring at a point. The verification of hypnosis was based on some signals, such as arm levitation, the easing of facial tension, a dropped lower jaw with a slight opening of the mouth, and the slowing down of breathing rate. This phenomenological approach was sufficient to ascertain the presence of deep hypnosis. The analysis of these signals enabled the hypnotist to verify whether the participants were really hypnotized, and to maintain or modify this condition by means of continuous appropriate suggestions. A more accurate verification of hypnosis was performed using hypnotic focused analgesia (HFA) as a marker when the participants performing a task in the fMRI scanner during the subsequent experimental setting.
During the preliminary session, a post-hypnotic command was left to each participant in order to obtain rapid and deep hypnotic trance during the experimental setting. The aim of this command was to reduce the time needed for further inductions. To this aim, the hypnotic conditioning was given to reach hypnosis immediately when receiving the command “please, [name], relax”. The effectiveness of this conditioning was tested immediately before the end of the session. The participant was then de-hypnotized and sent home.
The task model. A task model was conceived with the dual aim to commit participants to apply themselves, and to be used as a marker of goodness of deep hypnotic trance.
The task was represented by a cold pressor test (CPT) (Peckerman et al., 1991, 1994) , consisting in producing a painful stimulus by immersing the participant’s left hand in a plastic tub containing icy water at 0˚C temperature. The test, not described here but reported in details elsewhere (Sevre and Rostrup, 1999; Casiglia et al., 2007) , produces severe ischemic pain and a potent increase of the sympathetic discharge (Casiglia et al., 2015) . CPT has already been used in the frame of experimental hypnosis by our research group (Casiglia et al., 2007, 2015, 2016) and by other Authors (Freeman et al., 2000) . As water only is needed for CPT, this method is particularly fit to produce pain in fMRI setting, where cables, needles or other magnetic objects are not allowed.
Experimental session. After some days, the participants underwent the experimental setting, consisting in a sequence of fMRI scans during the task “CPT pain in HFA” and “CPT pain out of hypnosis” (control). fMRI was chosen because it is non-invasive, is advantageous in terms of temporal and spatial resolution, and does not expose the participant to radiations. It takes advantage of the atomic magnetic properties, measuring their resonance signals through magnetic fields and radio waves. In practical terms, it evaluates the hemodynamic variation in the brain areas involved in the task.
In this setting, hypnotic induction was obtained in a few seconds by means of the previously left hypnotic conditioning. This allowed inducing the participants in the scanner, when the physical presence of the hypnotist was impossible as he was outside the fMRI room and only communicated with the participant via an intercom.
Each session lasted about one hour. Once the participant was inside the scanner, their head was immobilized to minimize movements. Good comprehension of the procedure was ascertained, and headphones were placed on their ears to ensure a connection with the external environment. Experimental instructions were presented on a light-grey background display and projected via the Nordic Neurolab fibre optic visual presentation device mounted over the scanner head coil. Visual acuity was adjusted to individual needs. The chronological sequence of these commands was executed through the use of the program E-Prime 2. The experiment followed a box car design with an alternation of 120 s of “immerse” epochs and of 60 s of “rest” epochs.
Brain scanning took place at the Fondazione Ospedale San Camillo in
To address the experimental question, echo planar single shot T2* weighted MRI images (repetition time, TR = 2 s, echo time, TE = 50 ms, flip angle = 90˚, voxel dimensions 3.28 ×3.28×5.00mm , field of view 240 mm ). 252 volumes of 30 contiguous axial slices were acquired in ascending order in each run. Each run was preceded by 30-s dummy scans to allow the scanner to reach equilibrium, for a total scanning time of 59 minutes and 32 s. Four runs were acquired (two in hypnosis, both for the pattern “rest” and “immerse”; two in non-hypnosis, both for the pattern “rest” and “immerse”).
Analytical phase: functional image pre-processing, analysis and modelling. The Statistical Parametric Mapping 12 (SPM12) package was used in a Matlab environment on a Linux interface both for image pre-processing and statistical analysis.
The pre-processing phase included slice timing, realignment, normalization and spatial smoothing. For each voxel, slice-timing correction examined the time course and slightly moved it by interpolating with each other the points that were really detected in order to give the amount of time that would have been obtained if each voxel had been recorded exactly at the same time. The algorithm for slice-timing correction used an interpolation synchronized between time points, obtained through a Fourier’s transform of the signal of each voxel. During pre-processing, all volumes of each participant were corrected by the slice-timing using the 15th image as the reference slice (ascending acquisition). Slices were then realigned to their own mean according to the SPM12 protocol and then resliced through a 4th degree B-spline interpolation to correct the residual movement due to signal changes. The realigned scans were normalized in a first attempt to superimpose each brain to the common template. The filter resting time (REST, part of the SPM12 toolbox) was applied after normalization but before spatial smoothing with a 6 × 6 × 6 mm width at half maximum isotropic Gaussian kernel to compensate for any residual variability after spatial normalization. Movement parameters were included as regressors in the analysis.
The analytical phase was composed of a first- and a second-level analysis. During the first-level analysis, all the runs of the participants were put together in a multiple analysis using a general linear model. Image data were high-pass filtered with a set of discrete cosine basis functions with a cut-off period of 200 s. This passage generated a series of contrasts where (echo-planar imaging) EPI sequences during HFA were compared to EPI sequences without HFA.
One-sample t-tests conducted on the component images, with a threshold at p < 0.05 family-wise error (FEW) cluster-level, identified the various areas. This resulted in the acquisition of x, y, z coordinates in the Minnesota Neurological Institute (MNI) space using models of standard brains from the Montreal Neurological Institute; the coordinates of the significant correlations were converted in Talairach coordinates with the Matlab function mni2tal and later identified using the Talairach Daemon Client program. Differences were found comparing the pattern “immerse/rest” during HFA and during non-hypnosis. Images were presented according to the neurological convention, where right indicates the right hemisphere and left the left hemisphere.
General trend. During the phase of neutral hypnosis, the parietal, frontal and limbic lobes were generically interested, with a prevalent involvement of right hemisphere. During the “CPT pain out of hypnosis” phase, an activation of the primary somatosensory areas [Brodmann areas (BA) 1, 2, and 3] was observed, demonstrating CPT really produced a perceived pain. During the task “CPT in HFA”, BA 1, BA 2 and BA 3 were no more involved, demonstrating the efficacy of the task and the achievement of deep hypnosis with realization of the plastic monoideism represented by HFA (Table 1).
Specific trend during hypnotic task. During the task “CPT in HFA”, the dorso-lateral pre-frontal cortex (BA 9), the genual cortex (BA 25), the dorsal anterior cingulate cortex (BA 32), the orbital portion of the inferior frontal (BA 47) were activated (Table 1). This was patent in the images obtained with the Talairach stereotactic system (Figure 1), where the activated BA 9, 32 and 47 were evident in the transverse and coronal projections, and the sagittal projection also put in evidence activation in BA 25. These areas were inactive when the task was performed out of hypnosis.
A hypothesis. The mechanisms underlying hypnosis are sustained by weak theories (Mazzoni et al., 2013) . Many authors acknowledge hypnosis is a physiological modification of consciousness, in other word a conscious state. Another cornerstone of the theories of hypnosis is that the participant exposes him/herself to the rapport deliberately, intentionally and voluntarily, without any coercion or subordination. As a matter of fact, it is generally assumed that hetero-hypnosis performed in the frame of a setting of hypnotism is accepted consciously and with a critical mind.
Discussion of results. The results of the present study demonstrate that the brain areas activated during hypnosis were just those supposed to be associated with consciousness and free will.
Table 1. Brain areas activated during a hypnotic task performed in 20 highs.
BA: Brodmann areas. ‡corrected for free water excess.
Figure 1. Functional magnetic resonance imaging of the brain of 20 highly hypnotizable volunteers in the transverse, coronal and sagittal projections. Brodmann areas that were activated during the hypnotic task are indicated by their numbers. As figure adopts the neurological convention, the sagittal projection shows a slice in the right hemisphere.
The egoic consciousness, whatever it is, is certainly non-local, not associated with the activity of “this or that” area, but probably deriving from the recursive activity of cerebral electric phenomena related to unstable neuronal connections (Casiglia et al., 2016; Solms, 2017) . Nevertheless, it is irrefutable that integrity of some brain areas is necessary for the existence of consciousness. This is demonstrated by some pathological conditions in which certain areas are damaged and consciousness fails even if the sensorium is maintained (Hannawi et al., 2015) , and by some experiments where de-activation of given brain areas by electrode stimulation disrupts consciousness (Koubeissi et al., 2014) .
Strength and limitations of the study. The strength of the study is that it was based on the objective, technological evidence represented by fMRI, and confirmed hypnosis as a real and measurable state. As we are used to do (Casiglia et al., 2012) , we considered hypnosis an on/off phenomenon. This could be seen as a limitation by some authors, not working in experimental setting but rather employing hypnosis in psychotherapy only, as they can be reluctant to consider the participants as “hypnotized” and “de-hypnotized”. On the other hand, for researchers working with deep hypnosis in the field of neurophysiology and neuroimaging, these two states are very clearly distinguishable (Casiglia et al., 2003, 2007, 2011, 2012, 2012a, 2015, 2016, 2018, 2018a. 2018b; Facco et al., 2011; Giordano et al., 2012; Tikhonoff et al., 2018; Wood et al., 2003) . Finally, different participants’ hypnotizability (De Pascalis et al., 2000) or a particular echogenetic context, demonstrated in different settings (Stolarz et al., 2004; Tikhonoff et al., 2003) , could have infuenced the results of the study, making confirmation in larger samples mandatory.
The results of the present experimental research demonstrate that: 1) hypnosis is a real and measurable state, which can be studied with the instrumental methods that are typical of human physiology; 2) hypnosis, when realized in the frame of a task, is characterized by activation of particular brain areas that are those associated with consciousness and free will, areas that are inactive when the task is performed out of hypnosis; 3) as a consequence, the concept that hypnosis is a physiological modification of normal consciousness and that the participant adheres to it voluntarily appears plausible. Further studies focusing on this topic using fMRI or other imaging techniques are mandatory to confirm this evidence.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Edoardo Casiglia and Annalena Venneri conceived the research protocol. Edoardo Casiglia, Annalena Venneri, Francesco Finatti and Federica Albertini wrote the manuscript. Annalena Venneri, Francesco Finatti, Maria Rosaria Stabile and Micaela Mitolo analysed the fMRI data. Federica Gasparotti, Valérie Tikhonoff, Antonio M. Lapenta and Enrico Facco revised the manuscript.