Ayahuasca

Seeing with the eyes shut: Neural basis of enhanced imagery following ayahuasca ingestion

This study (n=10) details the neural correlates of ayahuasca-induced visual experiences and suggests that such experiences can be as visually intense and real as natural vision.

Authors

  • Carvalho, F. M.
  • Cecchi, G. A.
  • Crippa, J. A.

Published

Human Brain Mapping
individual Study

Abstract

The hallucinogenic brew Ayahuasca, a rich source of serotonergic agonists and reuptake inhibitors, has been used for ages by Amazonian populations during religious ceremonies. Among all perceptual changes induced by Ayahuasca, the most remarkable are vivid “seeings.” During such seeings, users report potent imagery. Using functional magnetic resonance imaging during a closed‐eyes imagery task, we found that Ayahuasca produces a robust increase in the activation of several occipital, temporal, and frontal areas. In the primary visual area, the effect was comparable in magnitude to the activation levels of natural image with the eyes open. Importantly, this effect was specifically correlated with the occurrence of individual perceptual changes measured by psychiatric scales. The activity of cortical areas BA30 and BA37, known to be involved with episodic memory and the processing of contextual associations, was also potentiated by Ayahuasca intake during imagery. Finally, we detected a positive modulation by Ayahuasca of BA 10, a frontal area involved with intentional prospective imagination, working memory and the processing of information from internal sources. Therefore, our results indicate that Ayahuasca seeings stem from the activation of an extensive network generally involved with vision, memory, and intention. By boosting the intensity of recalled images to the same level of natural image, Ayahuasca lends a status of reality to inner experiences. It is therefore understandable why Ayahuasca was culturally selected over many centuries by rain forest shamans to facilitate mystical revelations of visual nature.

Unlocked with Blossom Pro

Research Summary of 'Seeing with the eyes shut: Neural basis of enhanced imagery following ayahuasca ingestion'

Introduction

Ayahuasca is a traditional Amazonian brew combining Psychotria viridis (a source of DMT) and Banisteriopsis caapi (b-carbolines that act as monoamine oxidase inhibitors). De Araujo and colleagues frame Ayahuasca's psychopharmacology as a synergistic interaction in which orally inactive DMT becomes psychoactive when MAO is inhibited, producing autonomic and pronounced psychological effects including vivid, internally generated visual experiences commonly called "seeings". Earlier neuroimaging work has shown overlap between brain regions engaged by visual perception and by mental imagery, but the involvement of primary visual cortex (V1, Brodmann area 17) in imagery remains controversial, and the neural basis of Ayahuasca-induced vivid imagery has not been characterised. This study set out to identify brain systems underlying the heightened imagery that accompanies Ayahuasca ingestion. Using BOLD (blood oxygenation level dependent) functional MRI, the investigators compared brain activity during three conditions—viewing natural images, closed‑eyes imagery of those images, and viewing scrambled versions of the images—before and after oral Ayahuasca. They sought to determine which visual, temporal and frontal areas are potentiated during imagery after ingestion, whether primary visual cortex is engaged, and how neural changes relate to measured psychological effects.

Methods

Ten frequent Ayahuasca users were recruited (mean age 29 years, range 24–48, five female); one participant was excluded for excessive head movement, leaving nine subjects in the final analyses. Each subject completed two fMRI sessions: a baseline scan before drinking the tea and a second session beginning 40 minutes after ingestion, the timepoint at which subjective effects were expected to be robust. Psychiatric ratings (Brief Psychiatric Rating Scale, BPRS; and Young Mania Rating Scale, YMRS) were administered at 0, 40, 80 and 200 minutes post‑intake to track psychological changes. The administered dose was 120–200 mL of Ayahuasca (approximately 2.2 mL/kg). Chemical analysis of the batch by GC/MS found 0.8 mg/mL DMT and 0.21 mg/mL harmine; harmaline was below the detection threshold. Functional MRI was acquired on a 1.5 T scanner using EPI BOLD sequences (TR 3,000 ms; TE 60 ms; flip angle 90°; FOV 220 mm; matrix 128×128; slice thickness 5 mm; 16 slices). High‑resolution anatomical images were also acquired. The experimental paradigm used a block design with three conditions per block (21 s each): natural image viewing (unique images of people, animals or trees), a closed‑eyes imagery task in which subjects mentally recreated the just‑seen image, and presentation of a scrambled version of the image as baseline. Each session comprised seven blocks (total scan 441 s). Retinotopic mapping (phase‑encoded rotating checkerboard) was performed in a separate session to delineate V1–V3 boundaries. Preprocessing (motion correction, slice timing correction, 4 mm spatial smoothing, high‑pass filtering at 0.01 Hz, linear trend removal) and statistical analyses were performed in BrainVoyager QX. A general linear model (GLM) with a two‑gamma hemodynamic response function was used; group analyses used a fixed‑effects GLM with subject predictors. Multiple comparisons were controlled using the false discovery rate (FDR) with q(FDR) < 0.05; clusters smaller than 50 mm3 were excluded. Beta values were extracted from significant ROIs, and averaged BOLD time courses were computed across trials and subjects. Functional connectivity was assessed via delayed correlations between area time series: links required correlation Cij > 0.5 and to survive FDR correction (p < 0.05); a link was labelled directed if the peak delay exceeded 5 TRs (15 s).

Results

Psychological ratings showed significant increases on both scales after ingestion. Scores on the BPRS and YMRS rose from baseline, with significant changes at 40 and 80 minutes post‑intake (P = 0.036 for both scales, Wilcoxon test corrected for four timepoints). All subjects reported subjectively enhanced vividness of the imagery task after drinking the tea. Contrasting imagery after intake versus imagery before intake (IA > IB) revealed statistically significant BOLD increases (q(FDR) < 0.05) across a broad network. Activated regions included bilateral precuneus (BA7, 18, 19, 31), cuneus, lingual gyrus, fusiform gyrus (BA19, BA37), middle occipital gyrus, parahippocampal gyrus (BA30), posterior cingulate, superior temporal gyrus (BA22, 42), and frontal areas including superior/middle frontal gyri and inferior frontal gyrus (BA47). Retinotopic mapping confirmed that early visual areas V1–V3 showed significant modulation, and the specific BA17 locus affected corresponded to cuneus and lingual gyrus (peripheral field representations). Averaged percent BOLD signal time courses in ROIs showed a marked increase during the imagery condition after Ayahuasca ingestion. Before intake, imagery produced smaller visual cortex responses than the scrambled (visual stimulation) baseline; after intake, imagery responses rose to amplitudes comparable with the natural/scrambled visual stimulation condition. This pattern was consistent across occipital (BA17, BA19, BA7), temporal (BA30, BA37) and frontal (BA10) ROIs. Notably, BA10 already exhibited a positive BOLD response during pre‑intake imagery and was further potentiated after ingestion. Correlation analyses relating neural and psychological changes identified a single significant association: BA17 activation during post‑intake imagery correlated with BPRS scores at 40 minutes (Spearman’s rho = 0.78, P = 0.037; two‑tailed, corrected for seven brain areas). No significant correlations were found between BPRS and other regions, nor between YMRS and any region. Functional connectivity analysis showed altered temporal coordination after ingestion. For imagery post‑intake, BA17 assumed a more leading role, preceding BA7 and BA37, while maintaining leadership relative to BA19 and BA30. BA10, by contrast, shifted from a leader to a follower with respect to BA7 and BA30 during post‑intake imagery. BA19 tended to lag behind BA10 and BA17 in both natural and imagery conditions after ingestion. Overall, the connectivity pattern for post‑intake imagery resembled a superposition of pre‑intake natural‑image and imagery features, with strengthened fronto‑occipital temporal ordering centred on BA17.

Discussion

De Araujo and colleagues interpret the findings in pharmacological and systems terms. They propose that Ayahuasca’s serotonergic action across widespread cortical receptors underlies the observed potentiation of visual and associative areas during intentional imagery. While occipital responses to actual visual stimulation (natural image vs scrambled) were not increased by Ayahuasca—possibly because sensory‑driven activity approaches a physiological ceiling—the key effect was a large increase in BA17 activation specifically during closed‑eyes imagery after ingestion, bringing imagery‑evoked activity to levels comparable to perception with eyes open. The authors emphasise the selective relationship between BA17 activation and the degree of psychotic‑like perceptual change as measured by the BPRS, arguing for a tight link between primary visual cortex engagement and the subjective emergence of seeings. Nonprimary visual areas (BA7, BA18, BA19) also showed strong modulation; these regions have been implicated in pathological hallucinations and REM‑sleep dreaming, suggesting common network features between Ayahuasca seeings, hallucination phenomena and dreaming. Increased responses in parahippocampal cortex and retrosplenial regions (BA30, BA37) are interpreted as engagement of episodic memory and contextual‑association circuitry, potentially supplying mnemonic content to visual areas to construct elaborate internal scenes. Frontal modulation of BA10 is discussed in relation to intentional prospective imagination and processing of internally generated information; BA10 was unique in showing positive imagery‑related BOLD before ingestion and being further potentiated afterward. Connectivity results are used to argue that Ayahuasca alters temporal ordering across regions, notably increasing BA17’s capacity to lead other cortical areas during imagery, which the authors suggest supports the idea that seeings may be initiated in primary visual cortex and then recruit memory and frontal systems. They note an important limitation: the study did not counterbalance session order (baseline before, drug after), and while they consider it implausible that order alone accounts for the strong effects observed, this remains an uncontrolled factor. The authors also comment that subjects were experienced Ayahuasca users and that canonical hemodynamic responses were observed in other tasks in the same participants, supporting the interpretability of the BOLD signals. Overall, the investigators conclude that Ayahuasca transiently boosts the intensity of recalled images to percept‑like levels by engaging an extensive occipital–temporal–frontal network involved in vision, memory and intention, which helps explain the vivid, reality‑like quality of the reported seeings.

View full paper sections

INTRODUCTION

Ayahuasca, a central pillar of Amerindian traditional medicine, has been traditionally used by rain forest populations during religious ceremonies. Since the 1980s, Ayahuasca-based religions have been spreading to urban centers in South as well as in North America. In 1992, the Brazilian government approved a resolution legalizing the use of Ayahuasca within a religious context, and similar action was taken by the United States Supreme Court in 2006. Ayahuasca is served as a tea prepared as a decoction of a bush (Psychotria viridis) and a liana (Banisteriopsis caapi). Psychotria viridis is a rich source of the psychedelic substance N,N-dimethyltryptamine (DMT), whereas Banisteriopsis caapi contains b-carbolines such as harmine, harmaline, and tetrahydroharmine, which are potent monoamine oxidase inhibitors (MAOi). The synergistic interaction of these alkaloids determines the psychotropic action of Ayahuasca. DMT is a serotonergic agonist that acts mainly on 5-HT 2A and 5-HT 2C receptors, but in itself it is not orally active, since it is inactivated by MAO. However, the inhibition of MAO by b-carbolines allows DMT to be psychoactive when ingested. MAOi also contribute directly to the neuropharmacological effects of Ayahuasca by increasing extracellular levels of 5-HT. The effects of Ayahuasca begin $ 30-40 min after oral intake, and last up to 4 hours. Autonomic responses include increases in cardiac and respiratory rates, blood pressure, temperature, and pupil diameter. Ayahuasca users report many psychological effects, which include changes in self-perception, spatio-temporal scaling], and sensory hallucination. Among all perceptual changes induced by Ayahuasca, the most remarkable one are vivid visual hallucinations called mirac ¸o ˜es (seeings). Individuals report a variety of scenarios, as exuberant as a colored dream. Such seeings appear in different forms and context, ranging from simple to complex situations, from the vision of an animal to a long conversation with somebody unknown. However, despite their visual nature, Ayahuasca seeings are all internally generated, without the need of external stimuli]. Although enhanced imagery occupies center stage in Ayahuasca-based rituals, very little is known about the neural mechanisms underlying such seeings. The ability to generate visual mental imagery has long been a subject of research. Apart from the debate that this theme has provoked among psychologists and philosophers], we will herein consider mental imagery as the action of mentally summoning a visual representation of the world. Previous studies using functional Magnetic Resonance Imaging (fMRI) have investigated the neural basis of imagery (reviewed in] and demonstrated a substantial overlap among brain regions involved with imagery and percep-tion. The involvement of associative visual areas during imagery has been consistently reported in the literature, but the participation of the primary visual area (V1; Brodmann area 17, BA17) remains controversial. To investigate the neural basis of imagery induced by Ayahuasca, we used BOLD (Blood Oxygenation Level Dependent) fMRI during three sequential conditions: natural image of pictures, imagery of the same pictures, and natural image of scrambled versions of the same pictures. The same paradigm was applied before and after the oral intake of Ayahuasca.

SUBJECTS

Ten frequent Ayahuasca users participated in the study (mean age: 29 years, from 24 to 48 years, 5 female), after informed consent was obtained from all subjects in accordance with the guidelines approved) by the Human Research Committee and the Ethics Committee of the University of Sao Paulo. One subject was excluded from analyses due to uncorrectable amounts of head movement, leaving 9 participants (5 women) in the final dataset.

AYAHUASCA DOSAGE

Each subject drank 120-200 mL of Ayahuasca (2.2 mL/kg of body weight). The Ayahuasca batch used in the experiment contained 0.8 mg/mL of DMT and 0.21 mg/mL of harmine. No harmaline was present in the batch, at the chromatography detection threshold of 0.02 mg/mL. To quantify these amounts, a 1 mL sample of the Ayahuasca batch taken by the subjects was homogenized with sodium acetate buffer solution pH ¼ 9, extracted with 5 mL of diethyl ether in a shaker (20 min), and centrifuged at 3,000 rpm for 15 min. The organic phase was collected and evaporated under a nitrogen stream. The residue was dissolved in 1 mL of methanol and 1 lL was analyzed by gas chromatography/mass spectrometry (GC/MS). GC/MS analyses were performed using a Varian CP3800 gas chromatograph coupled to a Varian Saturn 2000 ion trap mass spectrometer (Varian Inc.). A capillary column (DB-5MS 30 m  0.25 mm i.d.  0.25 lm film thickness (Agilent) was used. The chromatographic conditions were as follows: injector temperature was 250 C in the splitless mode, oven temperature program was 80 C for 1 min ramped at 5 C/min to 220 C and held for 10 min, and then to 300 C for 5 min. Helium at a flow rate of 0.8 mL/min was used as carrier gas.

MR IMAGE ACQUISITION

All images were acquired in a 1.5T scanner (Siemens, Magneton Vision). The functional dataset was acquired using r Seeing With the Eyes Shut r r 2551 r EPI-BOLD like sequences, and comprises 147 brain volumes with the following parameters: TR ¼ 3,000 ms; TE ¼ 60 ms; flip angle ¼ 90 , FOV ¼ 220 mm; matrix ¼ 128 Â 128, slice thickness ¼ 5 mm, and number of slices ¼ 16. High spatial resolution images were obtained from typical Gradient Recalled Echo sequences, constituted of 156 sagital slices, covering both hemispheres, with a 1 mm 3 of voxel resolution.

EXPERIMENTAL DESIGN

Each subject performed two fMRI sessions. In the first session, subjects were scanned before Ayahuasca intake. Immediately after the first scan, subjects drank the tea. Psychiatric scales were applied at intervals of 0, 40, 80, and 200 min after Ayahuasca intake. The second fMRI session began 40 min after intake, when subjects were engaged in the same set of tasks of the first session. Figureshows a diagram of the experimental design (one session). In each fMRI session, subjects were submitted to three conditions in a block design (21 sec per condition; 3 conditions per block, 7 blocks per session). In the first condition (natural image), subjects passively viewed images of people, animals or trees (one different image per block, no repetition). The second condition (imagery) consisted of an imagery task, during which subjects were asked to close their eyes and mentally generate the same image they had just seen. During the last condition (scrambled image) subjects passively viewed a scrambled version of the image previously presented in the natural image condition. The scrambled image served as baseline for the analysis. Each scanning session lasted 441 sec (7 blocks  3 conditions  21 sec). The stimulation protocol was designed in Presentation V R software (Version 0.60, Neurobehavioral Systems). Images were projected onto a translucent screen and reached the eyes of the subjects by reflection on a mirror system adapted to the head coil.

RETINOTOPIC MAPPING

We used a flickering white and black checkerboard formed into a ray-shaped configuration subtending 22.5 (1/8) of polar angle. The disk segment started at the right horizontal meridian and slowly rotated anticlockwise for a full cycle of 360 . Each mapping session consisted of thirteen repetitions of a full rotation, which lasts 67 sec. Retinotopy of polar angle was revealed with cross-correlation analysis. The BOLD signal was correlated with an ideal response function that assumes the first 1/8 of a stimulation cycle as being the reference section (corrected for a hemodynamic delay). Areas activated at particular polar angles were revealed through selecting the lag value that resulted in the highest cross-correlation value for a particular voxel. Lag values were colorcoded and used to compute the cross-correlation maps, to find the retinotopic boundaries of early visual areas V1, V2, and V3.

IMAGE PROCESSING

Images were processed in Brain Voyager QX 1.9 (Brain Innovation, Maastricht, The Netherlands). Data analysis consisted of preprocessing steps, which included, 3D head motion correction (sinc interpolation), slice scan time correction (sinc interpolation), spatial smoothing (4.0 mm FWHM, 3D Gaussian filter), a high pass filtered at 0.01 Hz, and linear trend removal. The contrasts, in each condition, were evaluated using a general linear model (GLM) taking into account the hemodynamic response function (modeled by a two-gamma function) and baseline state. Group differences were analyzed using a fixed effect GLM with separate subject predictors, and statistical threshold was set taking into account a correction for multiple comparison based on the false discovery rate (FDR). Within and between multi-subjects corrections were made by setting q(FDR) <0.05. After transformation into Talairach space, the group analysis model included three orthogonal contrasts: mental imagery before intake versus baseline, mental imagery before intake versus mental imagery after intake and real imagery before intake versus real imagery after intake. Clusters of activation were then segregated into regions of at least 50 mm 3 .

IMAGE ANALYSES

The anatomical MRI of all subjects were investigated by an experienced neuroradiologist, and all the reports were classified as normal. The analysis of fMRI data was based on the GLM. The hemodynamic response function was modeled based on the conventional difference between two gamma functions. Unless otherwise specified, the GLM analysis was based on a fixed effect (FFX) assumption. Statistical maps were corrected for multiple comparison based on the FDR, and a threshold of q(FDR <0.05) was set for all group-level contrasts. Contrast maps were overlaid onto a coregistered anatomical image. The individual beta values used in the co-variance analyses were extracted from all statistically significant voxels in the ROI. Moreover, BOLD signal averages were computed for imagery as well as natural image conditions, both before and after potion intake. The strategy used is similar to computing EEG evoked potential averages along a series of trials. To generate BOLD signal time course averages Brain Voyager QX 1.9 (Brain Innovation, Maastricht, The Netherlands) was used. In brief, the averaged signal was computed for the selected ROI based on the average of all trials of particular conditions of interest (imagery before and after, natural image before and after), encompassing a baseline period (scrambles image). These time courses were then averaged across subjects.

FUNCTIONAL CONNECTIVITY ANALYSIS

This analysis was used to capture temporal correlations between areas, as a simplified snapshot of the dynamical state of the system. The correlations are represented by a graph, with nodes corresponding to areas, and links to functional connections. The time traces fed into the algorithm were, for each area, the fold activation over noise. Area-to-area correlations were computed as The maximum correlation, and the delayed at which it was met, were computed as and T ij ¼ arg max s c ij ðsÞ . A link was considered to be present whenever C ij > 0:5. A link was considered to be present if the P-value of the correlation survived FDR correction at 0.05; with only one exception, all the links were found to be significant. If the delay at which this threshold was met exceeded 5 TR's (i.e., 15 sec), then the link was considered to be directed, with the direction pointing from the ''leader'' to the ''follower''; otherwise, the link was considered as undirected, that is, no temporal precedence relationship could be determined. The choice of 5 TR's was based on the auto-correlation of the activations, which on average drop significantly for this time delay (see Supporting Information for details on this analysis).

PSYCHOLOGICAL EFFECTS

To estimate the time course of the psychological changes associated with Ayahuasca intake, we applied two psychiatric scales at intervals of 0, 40, 80, and 200 min after Ayahuasca intake: the brief psychiatric ratings scale (BPRS) to detect psychotic symptoms, and the Young Mania Rating Scale (YMRS), to measure mania symptoms. Figureshows that all subjects experienced increases in the two psychiatric scales following Ayahuasca ingestion. The effects were significant at 40 and 80 min post-intake (P ¼ 0.036 for BPRS and P ¼ 0.036 YMRS), Wilcoxon corrected for multiple comparisons of 4 time points), with respect to baseline (T ¼ 0 r Seeing With the Eyes Shut r r 2553 r min), in agreement with the time course of psychological, electroencephalographic changes induced by Ayahuascaand single photon emission computed tomography. Importantly, all subjects reported a marked increase in the ability to perform the imagery task after tea intake, corroborating that imagined scenes become much more vivid and detailed after Ayahuasca].

BOLD SIGNAL MODULATION BY AYAHUASCA

To test for the modulation of brain activity during imagery by Ayahuasca, fMRI data from individual subjects were analyzed using the following additional contrast: imagery after intake (IA) > imagery before intake (IB). Statistically significant areas of increased BOLD signal (q(FDR) < 0.05) comprised bilateral precuneus (BA7, 18, 19, 31), cuneus, lingual gyrus, fusiform gyrus (BA 19 and 37), middle occipital gyrus, parahippocampal gyrus (BA 30), posterior cingulate gyrus, superior temporal gyrus (BA22 and 42), superior and middle frontal gyrus, and inferior frontal gyrus (BA 47). To identify which areas were specifically modulated by Ayahuasca during the imagery task, and not by a generalized, nonspecific effect of Ayahuasca across all conditions, a further contrast was built in which the main effect of the imagery task (IA > IB) was subtracted from the main effect of natural image task (natural image after intake (NIA) > natural image before intake (NIB)). To determine the extent and borders of the early visual areas (V1, V2, and V3), a phase-encoded retinotopic mapping was performed in an additional fMRI session. As shown in Figure, BOLD signal increase was most prominent in the occipital cortex, extending into inferior and mesial temporal lobe, and to portions of the frontal lobe. Statistically significant voxels (q(FDR) < 0.05) were found bilaterally in the occipital cortex, comprising the three early visual areas with their respective retinotopic representations. Statistically significant modulation was also detected bilaterally in the parahippocampal gyrus (BA 30), middle temporal cortex (BA 37) and frontal cortex (BA 10; Fig.; Table).

BOLD SIGNALS AVERAGES BEFORE AND AFTER AYAHUASCA INTAKE

To further inspect the modulation elicited by Ayahuasca, we computed the averaged percent BOLD signal changes for the same ROI that were specifically modulated by Ayahuasca during the imagery task. Figureshows the extracted mean values and standard deviations were based on all trials. These included a baseline interval (5 sec-scrambled image, presented in Fig.as a shaded gray area) and the whole periods for the following conditions: imagery before (Fig., blue line), imagery after (Fig., red line), natural image before (Fig., light gray line), natural image after (Fig., green line). As can be observed, the averaged BOLD responses for scrambled image (Fig., shaded gray area) were similar to the ones from the natural image conditions (before intake-Fig., light gray line), in all ROI from the visual cortex. This was expected since the BOLD signal in visual cortex should not be different in the scrambled condition when compared to the natural image condition. The same pattern was preserved when observing the averaged signals after Ayahuasca intake. However, the average BOLD response during the imagery condition shows a remarkable increase after Ayahuasca intake. This can be observed by comparing the traces of the imagery before (blue lines) with imagery after intake (red lines). It is important to note that the averaged BOLD signal amplitude on visual areas during the imagery task before Ayahuasca intake (blue line) is expected to be smaller than during the scrambled period, when the visual system is actually being stimulated. After Ayahuasca intake, however, the averaged signal reaches amplitudes compatible to the ones obtained during the scrambled conditions. The same pattern of positive modulation was observed in all ROI, in the occipital (BA17, BA19, and BA7), temporal (BA30 and BA37) and frontal areas (BA10). The effect of Ayahuasca in occipital areas was particularly noteworthy, because the signal amplitude after intake increased markedly during imagery, but not during natural image (see Fig.). Supported by retinotopic mapping (see Fig.), the specific BA17 location modulated by Ayahuasca corresponds to the cuneus and lingual gyrus, which are related to the peripheral visual field. Also worth mentioning is the modulation by Ayahuasca of the parahipocampal cortex and the retrosplenial cortex (BA30 and BA37) during the imagery task. These structures are important for the retrieval of episodic memories, and have recently been implicated in the processing and representation of contextual associations]. As shown in Figure, in addition to occipital and temporal areas, Ayahuasca also potentiated parts of the frontopolar cortex (BA10) known to be involved in imagery. Among all regions modulated by Ayahuasca, BA10 was the only one that showed a positive BOLD signal during the imagery task even before Ayahuasca intake (see Fig.), and it was potentiated after tea ingestion during the imagery task (see Fig.). ; x, y, and z ¼ coordinates of the center of the cluster; (r x ,r y ,r z ) ¼ respective standard deviation. The contrast subtracts the main effect of the imagery task (IA > IB) from the main effect of natural image (NIA > NIB). The Talairach coordinates correspond to the center of the cluster of activity with its respective standard deviation.

TABLE I. AREAS MODULATED BY AYAHUASCA DURING THE IMAGERY TASK

r Seeing With the Eyes Shut r r 2555 r

CORRELATION BETWEEN BOLD AMPLITUDE AND PSYCHIATRIC SCALES

To investigate the relationship between neural and psychological changes, we calculated correlations between BOLD signal amplitude in different brain areas during post-intake sessions, and the values reached on the two psychiatric scales 40 min after intake, that is, immediately before subjects entered the scanner (see Fig.). A significant correlation was observed exclusively between BA17 activation and BPRS data (Spearmann's Rho ¼ 0.78, P ¼ 0.037, two-tailed, corrected for multiple comparisons of 7 brain areas; BA7, BA10, BA17, BA18, BA19, BA30, and BA37). No significant correlations were observed between BPRS individual scores and activation of any other brain region, nor between YMRS scores and BOLD signal modulation in any brain region. This indicates that BA17 activation levels during post-Ayahuasca imagery are specifically correlated with the occurrence of increased manifestation of perceptual changes, which included visual and auditory, as measured by the psychiatric scales.

FUNCTIONAL CONNECTIVITY

Is the Ayahuasca potentiation of intentional imagery accompanied by changes in the coordination of frontal, temporal, and occipital cortical areas? To address this issue, we implemented a functional connectivity analysis based on delayed correlations. The results are presented in Figure. The top row shows the full connectivity in the four conditions: imagery and natural image, pre/post intake. To understand the changes in connectivity, we sorted apart the links for areas BA17, BA10, and BA19, and showed them individually in separate rows. An interesting change observed for imagery is that BA17 becomes a leader of BA7 and BA37 after intake (Fig., second row), while maintaining its leadership with Average time courses of BOLD responses before and after Ayahuasca intake (blue for imagery before intake, red for imagery after intake, white for natural image before intake, and green for natural image after intake). Each time point corresponds to an fMRI TR ¼ 3 sec. Note the marked increase in BOLD signal during imagery following Ayahuasca intake. Shaded gray area represents baseline periods (scrambled image condition). Blue line-imagery before intake; red line-imagery after intake; light gray line-natural image before intake; green line-natural image after intake. respect to BA19 and BA30. Overall, the connectivity pattern centered on BA17 for post-intake imagery seems to be a superposition of connectivity features observed during pre-intake natural image and imagery. Other effect associated with imagery after Ayahuasca intake is the change from ''leader'' to ''follower'' for BA10 with respect to BA7 and BA30, by their turn ''followers'' of BA17. Furthermore, both for natural image and for imagery, BA19 becomes a ''follower'' area, lagging behind BA10 and BA17 (fourth row from the top).

DISCUSSION

The effects of Ayahuasca reported here are putatively mediated by the activation of serotonergic receptors widely distributed in the brain, including the several cortical areas modulated during post-Ayahuasca imagery. Interestingly, Ayahuasca did not enhance occipital BOLD signal during the natural image condition in comparison with the scrambled image condition. Perhaps activity in BA17 and other visual areas reaches a ceiling when subjects see with open eyes, irrespective of Ayahuasca ingestion. In contrast, the activation levels of BA17 during imagery were very different depending on whether Ayahuasca had been previously administered or not. While BA17 activation was very low during pre-Ayahuasca imagery, post-intake imagery was concomitant with very high BA17 activity, comparable to BOLD signal amplitude during natural image condition (see Fig.). Our study did not control for potential effects related to the order of the sessions with and without Ayahuasca ingestion. Although this question was not addressed in our study, it seems very not parsimonious to attribute the remarkable BOLD signal modulations described here to an order effect. Importantly, the effects in BA17 were positively and quite selectively correlated with the individual scores reached on the BPRS psychiatric scale (Spearman's Rho ¼ 0.78, P ¼ 0.037, two-tailed, corrected for multiple comparisons of 7 brain areas), pointing to a tight relationship between neural changes in BA17 and psychotic effects caused by Ayahuasca intake (see Fig.). In addition to BA17, Ayahuasca had a strong effect in other brain areas related to vision. The nonprimary visual areas strongly modulated by Ayahuasca (BA7, BA18, and BA19) are known to be activated during psychopathological hallucinationsas well as during normal dreaming, within rapid-eye-movement (REM) sleep. Although these areas showed BOLD signal time courses similar to that of BA17 (see Fig.), no statistically significant correlation was found between their signal modulation during imagery and the psychiatric scales used here (see Fig.). The activity of cortical areas BA30 and BA37, known to be involved with episodic memory retrieval and the processing of contextual associations, was also potentiated by Ayahuasca intake during imagery. Within this framework, Ayahuasca engages cortical regions necessary for the integration of separate visual elements into a whole scene. This suggests that the seeings induced by the tea are associated with an endogenous engagement of mnemonic circuits, possibly feeding visual areas with the content of the Ayahuasca seeings. Activation of BA 10 during mental imagery has already been reported], but its exact role in mental imagery is yet to be characterized. Previous studies have Figure. BOLD signal correlations between brain areas before and after Ayahuasca intake. The labels correspond to the same Brodmann areas defined in Figure. Arrow direction indicates that the source precedes the target by at least 5 TR's (15 sec); undirected links indicate that there is no such temporal delay between the areas. The top row shows the full connectivity for the four conditions. The second, third, and fourth rows from the top correspond to links involving only BA17, BA10, and BA19, respectively. shown that BA10 activity correlates with the amount of intentional effort involved in self-awareness and the imagination of future events. Furthermore, neuroimaging and lesion studies reported that BA10 plays an important role in prospective memory. It has been recently proposed that this region is involved with the temporal direction of an imaginary event, and there is evidence of its engagement when internally generated information needs to be evaluated. At present, it is believed that working memory depends on the interaction of BA10 and the dorsolateral prefrontal cortex (DLPFC): While the former processes information from internal sources, the latter is concerned with information generated externally. It is noteworthy to mention that although all subjects studied were experienced Ayahuasca users, their hemodynamic responses have a canonical shape. Such observation comes from another experiment conducted on the same subjects when performing a classical verbal fluency task, before and after Ayahuasca ingestion. The results show a consistent and expected response in expressive language centers, such as Broca's area (BA 44), with a classical hemodynamic response function. In this study, subjects were asked to intentionally imagine visual scenes. The broad range of neuroanatomical sites significantly affected by Ayahuasca during intentional imagery underlies the remarkable psychological changes produced by the tea. The findings from the connectivity analysis indicate that Ayahuasca intake strongly alters fronto-occipital relationships, producing marked changes in the temporal ordering of events across several brain regions. In particular, Ayahuasca intake is accompanied by an increased capacity of BA17 to lead other cortical areas during imagery. The functional prevalence and temporal precedence of BA17 during post-Ayahuasca imagery suggest that the seeings caused by Ayahuasca ingestion, robust even with the eyes shut, may in fact be initiated in the primary visual cortex. Ayahuasca-induced seeings have been traditionally used within religious contexts to give access to a deeply meaningful internal world. Altogether, our results indicate that these seeings stem from the activation, during voluntary imagery, of an extensive network of occipital, temporal, and frontal cortical areas respectively involved with vision, memory, and intention. By boosting the intensity of recalled images to the same level of natural image, Ayahuasca lends a status of reality to inner experiences. It is therefore understandable why Ayahuasca was culturally selected over many centuries by rain forest shamans to facilitate mystical revelations of visual nature.

Study Details

Your Library