top of page


The phenomenon of dissociations appears more complex than it seems at first. It is very simple to attribute the dissociations to psychiatric pathologies, however, in certain cases it has been seen that the brain organization works in a very different way from what could be expected.

These two studies bring us closer to a strange reality.


Julio Fernando Peres, Alexander Moreira-Almeida, Leonardo Caixeta, Frederico Leao, Andrew Newberg. Published: November 16, 2012


Despite the growing interest in pathological and non-pathological dissociation, few researchers have focused on spiritual experiences involving dissociative states such as mediumship, in which an individual (the medium) claims to be in communication or under the control of the mind of a deceased person Our preliminary study investigated psychography, in which supposedly «the spirit writes through the hand of the medium», in search of possible associations with specific alterations in brain activity.

We examined ten healthy psychologists: five less experienced mediums and five with substantial experience, ranging from 15 to 47 years of automatic writing and from 2 to 18 psychographs per month, using computed tomography to emit individual photons to scan the activity while the subjects wrote , both in dissociation states of trance and non-trance. The complexity of the original written content they produced was analyzed for each individual and for the sample as a whole. Experienced psychologists showed lower levels of activity in the left culm, left hippocampus, left lower occipital gyrus, left anterior cingulate, right superior temporal gyrus, and right precentral gyrus during psychography compared to their normal (non-trance) writing.

The mean scores of complexity for the psychographed content were superior to those of the control writing, both for the complete sample and for the experienced mediums.

The fact that the subjects produced complex content in a state of trance dissociation suggests that they were not simply relaxed, and relaxation seems an unlikely explanation for the activation of brain areas specifically related to the cognitive processing that takes place.


Dissociation is typically defined as the lack of normal integration of thoughts, feelings and experiences in consciousness and memory. The idea that traumatic experiences cause dissociative symptoms is a recurrent theme in the clinical and neuroimaging literature, and some of the cognitive phenomena associated with dissociation seem to depend on the emotional context or attention. Although non-pathological dissociation is quite common in the general population, dissociative experiences are mainly studied as a risk factor for dissociative pathology.

It has been shown that spirituality and religiosity are highly prevalent in patients with schizophrenia and dissociative symptoms. However, the various methodological issues and discrepancies between the studies developed so far make it difficult to articulate a comprehensive framework for brain activity and cognitive mechanisms in pathological and nonpathological dissociation.

Although the nature of the mind and its relation to the brain remains one of the most challenging topics for science, the assumptions made in this regard are the cornerstones that guide therapeutic interventions. This study addresses important theories that support creativity and includes religious and spiritual experiences.The American Psychiatric Association noted the need to conduct more research in this field by recognizing the non-diagnostic (non-pathological) category of «spiritual and religious problems» in the DSM-IV, therefore, healthy forms of dissociation can be distinguished from those pathological

The medium, a spiritual phenomenon that has often been reported throughout the history of mankind, is defined as an experience in which an individual (the medium) claims to be in communication or under the control of a person’s mind. deceased or another non-material being. The mediumistic experiences are usually dissociative, such as motor, sensory or cognitive automatisms (for example, hearing spirits or reports of bodily movements or thoughts caused by spirits) and an identity or alternative possession. Therefore, it is not surprising that the study of mediumistic experiences was crucial for the development of ideas concerning unconscious and dissociative processes.

Pierre Janet’s classic dissociation study of 1889 examined several mediums; Carl Jung’s doctoral thesis was a case study, and William James conducted a meticulous research on the medium Leonore Piper. There has been a tendency to divide dissociation into two broad categories: detachment (a sense of separation from self or world) and compartmentalisation (inability to deliberately control actions or cognitive processes that would normally be susceptible to such control). Although sometimes also involves detachment, mediumship is generally related to the subtype of compartmentalization.

Psychography is one of the many possible dissociative forms of mediumistic expression. The «mediums of writing» or the psicógrafos affirm that they write under the influence of the spirits, and some psychographed writings have had a great impact in different communities around the world. The most important and prolific psychographic medium in Brazil, Chico Xavier, whose education ended in elementary school, produced more than 400 automatic writing books covering a wide range of styles and themes, selling several million copies, and all the profits of Copyright donated to charities.

A study on the mental health of 115 spiritual mediums found that the subjects had high socio-educational levels, showed a low prevalence of psychiatric disorders and were well adjusted socially compared to the general population.

His experience of mediumship was different from dissociative identity disorder.However, few studies have investigated the neural substrates underlying the dissociative states of consciousness related to religious experiences. In a previous study of glossolalia neuroimages, a trance-like state with vocalizations that sound like language but lacking a clear linguistic structure, the subjects were found to have reduced activity in the left caudate nucleus and the right prefrontal cortex, along with an increase in activity in the upper core. parietal lobes. Neurofunctional research on sensitive experiences, such as religious ones, requires specific methods that do not adversely affect the performance of volunteers.

As in the study of glossolalia, the present study used single photon emission computed tomography (SPECT) to measure regional cerebral blood flow (rCBF), which is closely related to brain activity. We use the SPECT neuroimaging method for this study because it allows researchers to maintain an adequate environment without distracting / anxiogenic effects for subjects who perform complex tasks that require silence and concentration. As far as we know, there have been no previous studies on the association between the claimed dissociative media states and the specific alterations of BCF.

Based on previous research on related practices, such as meditation and prayer, we focus primarily on the prefrontal cortex and anterior cingulate gyrus, since both are known to be involved in the brain’s attention network. In addition, these areas are involved, together with Broca’s area, in the production of speech. We also find evidence of changes in thalamic activity in limbic structures such as the hippocampus, and the upper temporal region is involved in several processes, including language reception. The precentral gyration may be involved in motor function related to writing. Therefore, our hypothesis-based analysis focused on these regions.

We study the neurophysiological nature of dissociative mediumship in psychography, as measured by changes in rCBF. During psychography, individuals write legible structured narratives, but often claim not to know the content or grammar of the written text. The present study aims to determine if this type of dissociative trance state is associated with specific alterations in brain activity that differ from those found when writing normally, that is, not in a state of dissociative trance. Given that the psychographed contents present complexity and planning, our a priori hypothesis was that the areas involved in the cognitive processes when writing consciously, such as the reasoning and the content of the planning, would show a similar activation during the medium trance.


We examined 10 Brazilian psychologists who had been writing automatically for 15 to 47 years, producing 2 to 18 psychographies per month, who we divided into 5 «less expert mediums» and 5 with «substantial experience». All were white, right-handed, with good mental health, and currently do not use psychiatric drugs. The criteria used to describe the mediums as ‘experienced’ consisted in having practiced mediumship for at least 20 years and produced at least 10 psychographs per month at the beginning of the study.

The 10 mediums adjusted well in terms of their family, social and professional life, and regularly helped people who had lost loved ones. None of them were paid for their mediumistic activity, which they consider part of their mission to help people. They all reported spiritual experiences in childhood or adolescence. Both groups had the same average age: experienced (48 + -9.8 years) and less experienced (48.6 + -6.7). The mediums ‘experienced’ had practiced mediumship for 37.4 + -8.8 years with an average of 15.6 + -2.2 experiences of psychography per month, against the records of ‘less experts’ of 22.4 + -14.8 years and 4.8 + -3.0 times respectively .

The number of participants needed to determine the statistical power of the study was based on previous research related to glossolalia. Several mental health inventories and qualitative assessments of subjective experience were administered. Depressive symptoms were assessed using the Beck Depression Inventory (BDI), anxiety symptoms using the Beck Anxiety Inventory (BAI), current and past mental disorders using the Appendices for Clinical Evaluation in Neuropsychiatry (SCAN) ). Borderline personality disorder and history of child abuse were based on data from the Dissociative Disorder Interview Program (DDIS) and psychiatric morbidities were evaluated using the Psychiatric Self-Reporting Questionnaire (SRQ).

The local Human Research Ethics Committee in Brazil and the Institutional Review Board of the University of Pennsylvania authorized the study, and all participants signed informed consent forms.

Neuroimaging procedures

RCBF was measured using SPECT during psychography (writing in a dissociative trance state) and the data were compared with those collected during normal conscious or non-trance writing (the control task). Both writing assignments were carried out in a quiet and low light environment. Volunteers were asked to do psychography in the same way as in their regular activity as mediums. They all followed the same procedure: they sat in the chair where they would perform their tasks, prayed, closed their eyes and concentrated. Usually, they were in a trance state a few minutes later, they took a pencil and began to write. The mediums reported entering into a trance state very easily and silently. For non-trance writing, in the same place,

After the task of psychography, all subjects were asked if they had reached the mediumistic state (contact with a deceased person), and they were also asked to rate their level of mediumistic experience from 1 «poorly achieved» to 4 «successfully. accomplished». The order of tasks was random among the subjects to avoid the sequence effect and the controlled interval between tasks ensured the distinction between trance and non-trance states for psychography and control writing, respectively.

The use of SPECT images for the purposes of this study allowed the evaluation of the trance state itself. The rCBF SPECT studies are performed in such a way that the scans reflect what is happening at the time of the radioactive marker injection, which was during control writing tasks or psychography instead of after. The subjects are scanned afterwards, but the distribution of the marker is not reversible once it is injected and collected in the brain. This allows you to obtain images of your own trance state.

The subjects began writing in the room and wrote for 10 minutes, at which time they were injected through the IV cannulas (inserted in their left arms) with 7 mCi of 99m Tc-ECD. After writing for another 15 minutes, an investigator instructed him to stop writing and they were taken to the SPECT scanner for a 40-minute scan.

The images were acquired in a triple-head scanner (Trionix Research Laboratory) using high-resolution fan beam collimators. The projection images were obtained at three-degree angle intervals in a 128 × 128 matrix (3.56 mm × 3.56 mm pixel size) at 360 °. The SPECT images were reconstructed using the filtered back projection, followed by a low pass filter and a 1st order Chang attenuation correction.

After the first exploration of the writing task, the subjects returned to the room to perform the second task (psychography or control). After observing that they performed the second task for 10 minutes, they were injected in the same way with 25 mCi of 99m Tc-ECD, without disturbing them. Then, the subjects continued to perform the second task for 15 minutes, and the session ended. Each subject was scanned (second scan of writing tasks) for 40 minutes using the same image parameters as above. The phenomenological experience of the mediums during the task of control and psychography was evaluated through a semi-structured interview just after the acquisitions of image scanning.

Analysis of complexity of written content.

After writing for 25 minutes without interruption, the written content was evaluated by a PhD in Brazilian Language and Literature with extensive experience in the qualification of essays sent for the entrance exams to the university using the Analytical Evaluation, which weighs several characteristics or components. of effective writing to provide a deep quality rating and writing skills. The writing evaluated involved approximately 350 words related to the period in which the brain was impregnated with a marker. This analysis was masked (blind) so that the analyst did not know to which group each volunteer belonged. The following criteria were used to analyze the written content: (i) punctuation, (ii) selection of lexical elements and spelling, (iii) concordance of verbs and names, and placement of pronouns, (iv) development of the subject, (v) structure of prayer and articulation between parties, and (vi) consistency. Scores varied from 1 to 4 for each criterion as follows: (1) poor, (2) regular, (3) good, and (4) very good. The content scores for the two groups were compared using the Wilcoxon Signal Classification Test.


Although the subjects studied reported apparent delusions, auditory hallucinations, personality changes and other dissociative behaviors, they did not present mental disorders and were able to use their mediumistic experiences to help others. The structured clinical interviews exclude the current psychiatric illness.

None of the subjects, except one with previous signs of borderline personality disorders, showed clear signs of current mental disorders of Axis I or II. All the subjects stated that they felt very comfortable during the study and that they had successfully reached their usual trance state during the psychographic task (4 ‘successfully achieved’), and this evaluation was made shortly after the task of psychography. All reported being in their regular / wakefulness state during the control task. Seven found that writing for the control task was easy, and the three who mentioned some difficulty reported that, in general, it was difficult for them to write written texts in their daily lives.

During psychography, all mediums reported altered states of consciousness, but to different degrees. Experienced mediums spoke of a deeper trance, with a clouded consciousness, often reporting that they were outside the body and that they had little or no awareness of the content of what they were writing.

The groups were assigned randomly, so there were no significant differences in the mean time between the scans. When using a t test analysis of the regions according to the voxel counts, there were no significant differences when the whole group was analyzed. However, subjects experienced during the control condition showed significantly greater activity in these regions (p <0.001 for all regions) than less experienced mediums. Significantly higher rCBF (p <0.01 for all regions) was shown in several areas of the brain by less experienced psychologists, particularly in the left culm, left hippocampus, left inferior occipital gyrus, left anterior cingulate, right superior temporal gyrus and right precentral gyrus during psychography compared to normal writing (without trance). The precentral gyro approach actually covers the precentral gyrus and the medial frontal gyrus, but we report the region as a function of the MNI coordinates. Experienced mediums of writing in a trance state showed rCBF consistently lower in these regions than when writing in the control condition.

The difference was significant in comparison with the less experienced (p <0.05).

The written content produced by the subjects during both types of tasks, with or without medium trance, had never been written before. The level of complexity of both types of written content (psychographed and control-task) was analyzed individually for each topic. The content produced during media writing and control usually involved ethical principles, the importance of spirituality and the union of science and spirituality. The mean scores of complexity for the psychographed content were higher than those of the control writing, both for the complete sample [16.8 (SD 3.33) vs 14.4 (SD 2.95) – p = 0.007] and for the experienced mediums [18.4 ( SD 2.30) 3.36) – p = 0.041]. For less experienced mediums, the difference was almost significant [15.2 (SD 3.63) versus 13.4 (SD 2.41) – p = 0.066].

Finally, a linear correlation analysis was performed that compared the change in the global complexity score so that the written content changed in CBF in the six regions identified as being significantly associated with the state of psychography.In general, there was a trend towards an inverse correlation between the change in complexity and the change in the CBF in each region. The correlation coefficients varied from 0.59 to 0.74 for p values ​​from 0.03 to 0.12. All correlations were reversed, so that the greatest increases in complexity were associated with a progressive decrease in HR in each region.


Our hypothesis was not confirmed for less expert psychologists, since the results showed significant changes in the rCBF in several areas of the brain during psychography compared to writing without trance. In addition, contrary to our hypothesis, experienced mediums who write dissociatively in a trance state showed a consistently lower rCBF in these regions than when writing in the control condition.

In relation to the hypnotic suggestion, some studies showed prefrontal activation, but not others, while our subjects showed lower levels of activity in the frontal care system. Although reduced frontal-parietal connectivity and frontal deactivations are observed after a hypnotic induction in highly suggestive individuals, hypnosis is phenomenologically distinct from mediumistic expressions, therefore, the two conditions are not directly comparable. In addition, the idea that hypnosis reflects dissociative states remains controversial.

Brain scanning studies of meditation have generally found an increase in frontal lobe activity and a related care network, unlike our findings for experienced mediums. Although meditative states do not necessarily involve dissociation and phenomenological expressions are very different from psychography, a recent study suggested that meditation improves the efficiency of brain functioning, so that experts’ brain activation levels are lower than those of less experienced meditators, pattern similar to that reported in the present study.

Previous research in neuroimaging has shown that writing is a complex process that requires synchronized cognitive, language and perceptual-motor skills. The complexity of the written content reflects the creativity of the author and the planning work that underlies the underlying activity in the precentral gyrus, the right superior temporal gyrus, the left anterior cingulate, the hippocampus, the culm and the occipital lobes. Damage or hypoperfusion in these regions has been correlated with severely damaged writing.

In particular, experienced mediums showed higher complexity scores, suggesting that planning for psychographed content was more sophisticated than for written content when it was not in a dissociative mediumistic trance. The greater complexity of the text that involves more creativity and planning work during psychography would presumably require more activity in the central precentral gyrus, the right superior temporal gyrus, the left anterior cingulate, the left hippocampus, the left culm, and the left inferior occipital gyrus that the task of control less complex, but this was not the case, especially for experienced mediums.

Findings about the lower activity levels of the left hemisphere and the higher activity of the right hemisphere have been reported in pathological expressions of dissociative and psychotic experiences. Unlike our volunteers, patients with schizophrenia had lower blood flow levels in the left hemisphere regions, while areas of higher flow may reflect the need to turn to the right hemisphere to compensate for deficits in the left hemisphere networks . In addition, BCS abnormalities in the anterior, precentral, temporal, and culminating cingulate could predict the development of psychosis in high-risk subjects with a subsequent transition to psychosis. The anterior cingulate is involved in the attention system along with emotional regulation, learning, memory, error detection, conflict monitoring, strategy planning and empathy. The decrease in the activity of the anterior cingulate, the precentral gyrus, the superior temporal gyrus and the hippocampus in experienced mediums may partly explain the absence of focus, self-consciousness and consciousness during the dissociative state observed in psychography. Despite several similarities with brain activation related to schizophrenic patients, the subjects who participated in the present study did not present schizophrenia or any other mental illness. This finding underscores the importance of further research on the differential diagnosis between pathological and nonpathological dissociation.

We try to maintain the greatest possible similarity between the groups in order to better compare their brain functions. The differences observed in CBF may be related to their different levels of experience, but may also reflect differences in anxiety, effort or efficiency. For example, studies have shown that anxiety is associated with increased uptake in the right ventral frontal cortex and the insula / left area. Therefore, some of the changes we observed may have reflected anxiety, although none of the subjects reported particularly high levels of anxiety or stress.

Studies of cognitive skills have revealed two general patterns of changes in brain activity. Several studies have found that experts and non-experts show greater activity in different regions. The level of activity in the area of ​​the fusiform face (FFA) when experts identify objects such as cars or birds, predicts performance in a measure of behavior of the experience performed outside the scanner. However, studies have found that some regions of the brain show increases while others show decreases during the task performed by experts.

Observations of the calculations made by experts have reported an increase in activity in the medial frontal gyrus, the parahippocampal gyrus, the anterior cingulate gyrus and the right-middle occipito-temporal junction, as well as the left paracentral lobe

Other studies of arithmetic expertise have shown larger regions of greater activity.These results suggest that experts use different or more extensive brain pathways.However, other studies suggest that more skilled subjects make more efficient use of regions and brain activity. Under these circumstances, less brain activity is observed in cognitive tasks. On the other hand, those who struggle to perform cognitive tasks often have to recruit more areas of the brain as a compensatory mechanism.

The results of the present study suggest that the level of experience may have an important effect on brain function. There was a tendency towards an inverse correlation between the change in complexity and the alteration of the CBF in each region. Given that these correlations were reversed, the implication is that greater increases in complexity were associated with a progressive decrease in HR in each region. This interesting finding that takes into account the complexity of psychographed texts deserves future research and elucidative hypotheses. It could be speculated that these findings were related to improvisation musical performances, in which the activity decreased in some areas of attention has been involved in a change induced by training to the inhibition of attention driven by stimuli.

In addition, a recent study showed that alcohol consumption, which decreases frontal lobe activity, seems to improve creativity. However, the states of musical improvisation and alcohol consumption are quite peculiar and different from psychography. Future research is needed to thoroughly compare psychography with other similar states and to clarify more precisely the relationship between frontal lobe function and the depth, intensity and complexity of the written content produced in this interesting mediumistic state.

In general, the fact that experienced mediums had less rCBF than less experienced mediums may be due to having more years of practice and doing more psychographs per month. However, considering the high complexity scores of the experts for their psychographed content, it is not clear whether the decrease in brain activity is related to a more efficient brain function during the task, or the influence of other variables.

Although we know the problems to conceptualize the trance, for the purpose of this study we use a more consensual and phenomenological definition of the trance proposed by Cardeña: a temporary alteration of consciousness, identity and / or behavior evidenced by at least two of the following :

(1) marked alteration of consciousness;

(2) reduced knowledge of the immediate surroundings;

(3) Experienced movements that are beyond one’s control.

In qualitative terms, given that there is not a single expression of mediumship, but rather important differences between people and occasions, our subjects reported on various types of «spiritual contact». The less experienced mediums were emotionally affected and reported that they felt inspired during the psychography, and that they were in a semi-conscious state (the sentences arrived as if they were dictated) in relation to the written content, while the experienced mediums said that they were «Out of their bodies» and had no control over the content «elaborated by the spirit».

The superior temporal gyrus, which contains the auditory cortex, was activated during psychography for the less experienced mediums, who listened to sentences as if they were dictated, but were deactivated in the experienced subjects, who had no conscious control over the psychographed content. The superior temporal gyrus is also involved in linguistic comprehension and is a key area related to auditory hallucination in psychotic patients.

The decrease in activity of the left prefrontal cortex, which is involved in the classification and classification of the experience may be related in part to the subjective description of the dissociative trance as reported by experienced mediums, and is consistent with the notion of writing automatic Instead of planning the written content. Language processing studies consistently show that the superior temporal cortex and the precentral gyrus are crucial for word processing and its hypoperfusion results in the selective deterioration of written work. Constant activation is expected in these areas during writing in healthy subjects. These regions were hypoactivated in the brain of the subjects experienced during psychography, and did not show the deteriorated written text that we would expect with the hypoactivation presented.

The lowest level of activity in the temporal cortex and the precentral gyrus, as well as the hippocampus and the anterior cingulate in experienced mediums, supports their subjective reports of ignorance of the written content during psychography. It should be noted that no changes were observed in the CBF in the caudate nuclei described above in the glossolalia. The subjects also showed a reduced CBF in the right prefrontal cortex, and these discrepancies may be related to different tasks related to language processing during these trance-like states.

The subjects attributed their writing in trance to «spirits». Compared to normal writing, less experienced media showed greater activation in the same areas of cognitive processing during psychography, whereas experienced media showed a significantly lower level of activation.

The less experienced had to «work harder,» as evidenced by their relatively high levels of activation of the area of ​​cognitive processing during psychography.Experienced media showed a significant reduction in rCBF changes during psychography, which is consistent with the notion of automatic (non-conscious) writing and their claims that an «external source» was planning written content.The brain regions that are known to be involved in writing planning were activated less, although the content was more elaborate than their writing without trance.

These findings are not consistent with falsification or role play, which have been offered as explanations for psychography. Neural circuits related to planning would probably be recruited to compose more elaborate texts if the subjects were faking trance states.

Conversely, studies of regions of cognitive processing involved in reasoning and written content planning showed a decrease in activity in experienced mediums, who reported that they were not aware of the psychographed content and that they had no control over it.

The subjects reported that their trance state implied a «relaxed state of mind». The state of relaxation could only explain the lower overall activity of the brain, but the fact that the subjects produced complex content in a state of trance dissociation suggests that they were not simply relaxed. In addition, relaxation seems an unlikely explanation for the activation of brain areas specifically related to the cognitive processing that takes place. As a first step toward understanding the neuronal mechanisms involved in non-pathological dissociation, we emphasize that this finding deserves further investigation in terms of both replication and explanatory hypotheses.

In non-pathological conditions, a person may benefit from these dissociative abilities, although such a disposition may become dissociative pathology after adverse / traumatic events. The absence of current mental disorders of Axis I or II in the groups is in line with current evidence that dissociative experiences are common in the general population and not necessarily related to mental disorders, especially in religious / spiritual groups. Alterations in medium blood flow differed between experienced and less experienced subjects, highlighting the diversity of the dissociative phenomenon in healthy subjects and suggest that additional research should address the criteria for distinguishing between healthy and pathological dissociative expressions in the field of mediumship.

A limitation of this study