Effects of the South American psychoactive beverage ayahuasca on regional brain electrical activity in humans: a functional neuroimaging study using low-resolution electromagnetic tomography

Riba and colleagues introduce ayahuasca as a traditional South American plant tea combining Banisteriopsis caapi, which supplies monoamine oxidase (MAO)–inhibiting ß-carbolines, with Psychotria viridis, a source of the short-acting psychedelic N,N-dimethyltryptamine (DMT). Because the ß-carbolines inhibit MAO, orally ingested DMT becomes bioavailable and produces acute perceptual, cognitive and mood effects that have been documented subjectively and in scalp electroencephalography (EEG) as dose-dependent reductions in spectral power and acceleration of centre-of-gravity frequency. Previous functional imaging work with related indoleamine psychedelics (for example, psilocybin) has shown metabolic increases in prefrontal regions, but the regional cortical generators of ayahuasca-induced EEG changes had not been assessed using source-localisation techniques available at the time. The present study therefore set out to map the intracerebral sources of ayahuasca-induced changes in spontaneous EEG using low-resolution electromagnetic tomography (LORETA). The authors aimed to identify which cortical and paralimbic regions show altered current density across classical EEG frequency bands after a single oral dose of encapsulated, freeze-dried ayahuasca, and to relate those regional effects to concurrently measured subjective psychedelic effects.

Eighteen healthy volunteers (15 males, 3 females; mean age 25.7 years) took part in a double-blind, placebo-controlled, randomised crossover trial. Eligibility required prior experience with psychedelic drugs on at least five occasions and absence of current or past neurological or psychiatric disorder; volunteers underwent a structured psychiatric interview (DSM-III-R), trait anxiety screening and a physical examination including ECG, laboratory tests and urinalysis. Two of the participants had prior exposure to ayahuasca; all reported prior cannabis and cocaine use. Experimental sessions were at least one week apart and volunteers abstained from medications, illicit drugs and from alcohol, tobacco and caffeine before sessions. The active intervention was a lyophilised ayahuasca preparation characterised by HPLC analysis: the freeze-dried powder contained 8.33 mg DMT, 14.13 mg harmine, 0.96 mg harmaline and 11.36 mg tetrahydroharmine per gram. Individual doses were administered to deliver 0.85 mg DMT per kg body weight, encapsulated in gelatin capsules; placebo capsules contained lactose. Capsules were given around 10:00 a.m. with 250 ml water, and the number of capsules was balanced between sessions so volunteers ingested identical numbers on active and placebo days. EEG was recorded with 19 scalp electrodes positioned according to the international 10/20 system, referenced to averaged mastoids, together with vertical and horizontal electrooculograms. Vigilance-controlled, eyes-closed three-minute recordings were obtained at baseline (including -15 min) and at 30, 60, 90, 120, 180, 360 and 480 min after dosing. Signals were band-pass filtered (0.3–30 Hz), digitised at 100 Hz, and subjected to a two-step artefact correction (ocular regression and automatic time/frequency rejection), followed by visual inspection. Spectral analysis was performed on six 5-second artefact-free epochs per recording; data were filtered into seven frequency bands: delta (1.5–6 Hz), theta (6–8 Hz), alpha-1 (8–10 Hz), alpha-2 (10–12 Hz), beta-1 (12–18 Hz), beta-2 (18–21 Hz) and beta-3. LORETA was used to estimate three-dimensional intracerebral current density (power) from the scalp cross-spectral matrix. The solution space was constrained to cortical gray matter and hippocampus using a digitised Talairach-based atlas; voxels with >33% probability of being gray matter were included, yielding 2,394 voxels at 0.343 cm3 resolution. Statistical analysis of LORETA power used paired-sample t tests on log-transformed data at each voxel and frequency band, with a nonparametric single-threshold permutation correction (Holmes method) based on 5,000 randomizations to control for multiple comparisons. The omnibus test identified time points with maximal suprathreshold voxels (used to select main analysis times), and anatomical distributions of suprathreshold voxels were tabulated by lobe, gyrus and Brodmann area. Subjective effects were measured with a Spanish version of the Hallucinogen Rating Scale (HRS), with six subscales (somaesthesia, affect, volition, cognition, perception and intensity); HRS scores after ayahuasca and placebo were compared using paired t tests.

The omnibus voxel-wise analysis across time showed a steep rise in suprathreshold voxels at 60 min after dosing and a pharmacodynamic peak at 90 min; LORETA comparisons versus placebo were therefore computed at those two time points. At 60 min, statistically significant decreases in intracerebral power were most prominent in the alpha-2 band, with 459 suprathreshold voxels. These alpha-2 reductions were distributed across parietal (135 voxels), temporal (170), occipital (79) and limbic (69) lobes, with only minimal involvement of left frontal and sublobar regions. The spatial pattern at 60 min concentrated around the temporo-parieto-occipital junction, notably the angular and supramarginal gyri, precuneus, superior and middle temporal gyri and fusiform gyrus; limbic suprathreshold voxels were mainly in the cingulate and parahippocampal gyrus. Smaller delta-band decreases at 60 min were detected in a left occipito-temporal border region (15 voxels, BA 19/37/21). Theta decreases at 60 min narrowly missed significance (lowest t = -3.58, p = 0.0522), localising to BA 24 in medial frontal cortex. At the 90-min peak, significant power reductions extended to the delta, theta and beta-1 bands. Delta-band decreases comprised 471 suprathreshold voxels located in parietal (135), temporal (263), occipital (70) and limbic (3) lobes, again centred on the temporo-parieto-occipital junction (BA 37, BA 19, BA 39). Theta-band decreases (128 suprathreshold voxels) formed three non-confluent regions: medial and superior frontal gyri (BA 6, 8), the anterior cingulate, and a left temporomedial cluster including fusiform and parahippocampal gyri and the uncus. Beta-1 reductions (139 suprathreshold voxels) were mainly parietal (122 voxels), occurring in two non-contiguous regions that included posterior cingulate/precuneus/superior parietal lobule and right supramarginal/angular gyri (BA 39). Alpha-2 power was reduced at 90 min relative to placebo but did not reach statistical significance at that time point. Subjective effects followed a time course that paralleled the LORETA findings: volunteers reported initial effects by 15 min, greater effects between 30 and 45 min, a steep rise at 60 min and maximal subjective intensity between 90 and 120 min, with effects subsiding and disappearing by 360 min. The HRS, administered at 240 min, nevertheless showed statistically significant increases in all six subscales after ayahuasca compared with placebo. Reported phenomena included somatic sensations (pins and needles, increased skin sensitivity, physical comfort), a range of visual and auditory experiences including synesthetic reports by five subjects, altered time sense, accelerated thought, reduced ability to sustain focussed attention and positive mood changes; none of the participants experienced full psychosis-like symptoms or loss of awareness that effects were drug-induced.

Riba and colleagues interpret the LORETA findings as evidence that ayahuasca produces time-dependent reductions in intracerebral EEG power that implicate posterior association cortices and medial paralimbic regions. The principal effects were decreases in slow rhythms (delta and theta) and reductions in alpha-2 and beta-1 power, with maximal spatial extent around the temporo-parieto-occipital junction and with theta decreases in medial frontal and temporomedial areas. These electrophysiological changes accompanied a constellation of perceptual, cognitive and mood effects typical of psychedelics, and the temporal alignment between EEG source changes and subjective peak effects supports a pharmacodynamic relationship. The authors relate the spectral pattern to proposed neuropharmacology: diminished delta and theta power is seen with psychostimulants and psychedelics and is broadly compatible with an arousing or excitatory action, whereas the alpha-2 suppression observed here is a feature shared with several perception-altering drugs and distinguishes ayahuasca from prototypical stimulants that often enhance alpha. Comparisons with PET and SPECT studies of psilocybin and mescaline reveal partial concordance, notably in temporomedial and cingulate involvement, but the present EEG-source results did not show the same degree of frontal metabolic increases reported in some metabolic imaging studies; the authors suggest this divergence could arise from differences in the physiological variables measured (electrical current density versus blood flow or glucose metabolism) and from differences in subjective response profiles. Given the exploratory design, the investigators propose hypotheses about functional implications: decreases in unimodal sensory association cortices may underlie modality-specific perceptual alterations, while effects in heteromodal association cortex, cingulate and temporomedial structures could account for higher-order changes in attention, emotion and memory. The authors acknowledge the exploratory nature of the work and note that the volunteers' prior psychedelic experience may have influenced the predominantly positive character of subjective reports. They conclude by recommending further studies that directly examine ayahuasca effects on the cognitive and emotional domains implicated by the present source-localisation results, to better characterise the neural basis of its complex psychological effects.