DMT alters cortical travelling waves

N,N-Dimethyltryptamine (DMT) is a serotonergic psychedelic that produces a rapid, short-lived but intensely immersive visual and perceptual state. Earlier EEG and MEG studies of ayahuasca and of intravenously administered DMT have consistently reported broadband decreases in oscillatory power—particularly in the alpha band (8–12 Hz)—and increases in signal diversity; decreases in alpha power have correlated with the intensity of visual phenomenology. Separately, non-drug studies have described spatio-temporal patterns of cortical activity known as travelling waves: forward (occipital-to-frontal) waves predominate during visual perception, whereas backward (frontal-to-occipital) waves predominate during eyes-closed rest. These travelling waves have been linked to predictive-coding accounts in which direction of propagation relates to bottom-up prediction errors and top-down predictions respectively. Alamia and colleagues set out to test whether DMT alters the amount and direction of cortical travelling waves during eyes-closed resting conditions, and whether any such changes relate to the vivid visual imagery produced by the drug. The primary hypothesis was that DMT would disrupt the physiological balance between top-down and bottom-up information flow in favour of increased bottom-up signalling, manifesting as decreased backward waves and increased forward waves similar to those observed during actual visual stimulation. The study therefore aimed to quantify travelling-wave directionality and frequency under placebo versus intravenously administered DMT and to relate these measures to real-time and post-hoc subjective reports of intensity and visual imagery.

This analysis reused a previously collected dataset. Thirteen healthy participants (6 female; the extracted text does not clearly report mean age) completed two sessions: a placebo session and a DMT session. DMT was administered intravenously in ascending doses across participants (reported doses: 7 mg to 20 mg, distributed across participants as 7 mg to 3 participants, 14 mg to 4 participants, 18 mg to 1 participant, and 20 mg to 5 participants). Participants were medically and psychiatrically screened; standard exclusion criteria (age under 18, no prior psychedelic experience, history of psychiatric illness, excessive alcohol use) and pre-session drug/pregnancy/breathalyser checks were applied. During each session participants rested semi-supine with eyes closed; eye closure was monitored visually. EEG was recorded with a 32-channel BrainProducts system at 1,000 Hz. Signals were filtered (high-pass 0.1 Hz, anti-aliasing low-pass 450 Hz, band-pass 1–45 Hz), epochs with artefacts removed visually, and ICA used to remove ocular and cardiac components. Data were re-referenced to the average of electrodes. Travelling-wave quantification used a sliding 1-second window with 0.5-second overlap. For each window the authors formed a time–electrode 2D map from five midline electrodes (Oz, POz, Pz, Cz, FCz) and computed a 2D Fast Fourier Transform (2D-FFT). The maximum values in the upper and lower quadrants of the 2D-FFT were taken as raw measures of forward (FW) and backward (BW) wave power respectively. To obtain a null distribution that preserves temporal spectral content but destroys spatial directionality, electrode order was shuffled 100 times and the surrogate FW and BW values averaged. The net amount of FW and BW waves was then expressed in decibels (dB) as the log-ratio of true to surrogate values; this metric indexes the amount of wave power beyond that expected from temporal fluctuations alone and can be tested against zero to assess significance. Frequencies of maximal peaks from the 2D-FFT were analysed to assess band-specific changes. Statistical analyses comprised separate Bayesian ANOVAs for FW and BW measures with factors pre/post injection and drug (DMT versus placebo), treating subjects as random factors. Minute-by-minute analyses used t-tests at each time bin against zero with False Discovery Rate (FDR) correction. Correlations were computed between minute-by-minute intensity ratings (collected every minute for 20 minutes post-injection) and wave amplitudes, and between post-session visual imagery questionnaire items (Visual Analogue Scale) and wave measures. The authors note a reduced number of datapoints (n=12) for some correlational analyses as reported in the extracted text.

Baseline (pre-injection) closed-eyes recordings showed a significant predominance of backward travelling waves in both placebo and DMT sessions (Bayesian t-tests against zero: BFs10 >> 100), with no significant forward waves (BFs10 < 0.15). After intravenous DMT, the cortical dynamics changed: backward wave amplitude decreased relative to baseline but remained above zero (BFs10 = 12.6), while forward wave amplitude increased significantly above zero (BF10 = 5.4). These pre/post by drug effects were supported by separate Bayesian ANOVAs run for FW and BW measures (reported BFs10 >> 100 for factors and interactions). A minute-by-minute analysis showed that changes emerged rapidly after injection and decayed over the subsequent tens of minutes. Forward waves increased and backward waves decreased during the peak DMT effect period; the time-bin tests against zero survived FDR correction in several minutes for FW and consistently for BW (all FDR-corrected p-values < 0.05 where reported). When compared qualitatively with results from a separate experiment involving alternating visual stimulation and blank-screen periods, the temporal profiles of FW and BW changes under DMT were remarkably similar to those observed during photic visual stimulation, despite participants having their eyes closed under DMT. Examining spectral content, both FW and BW wave spectra showed a strong alpha-band presence at baseline. Following DMT, alpha-band power associated with these waves decreased markedly while delta (0.5–4 Hz) and theta (4–7 Hz) bands increased for both wave directions (all BFs10 > 100). Absolute spectral power from the 2D-FFT also decreased in the alpha band after DMT, a pattern akin to that seen with visual stimulation. Analyses of FW–BW relationships during the interval when both were significantly present (minutes 2–5 post-DMT) revealed a robust negative correlation between their net amounts on a moment-by-moment basis: FW tended to be weaker when BW was stronger and vice versa. This inverse relationship was consistent across participants and appeared independent of drug condition (difference between pre and post, Bayesian t-test BF10 = 0.225). Relating waves to subjective experience, minute-by-minute intensity ratings correlated positively with forward-wave amplitude and negatively with backward-wave amplitude; these time-resolved correlations peaked a few minutes after injection. Across participants, correlations between intensity ratings and wave amounts reached coefficients of approximately 0.4 around peak effect but did not achieve statistical significance after correction, a shortfall the authors attribute to limited sample size (n=12 for some analyses). Post-session visual imagery ratings (Visual Analogue Scale) showed a consistent positive relationship with the amount of forward waves across all 20 questionnaire items; Bayesian tests indicated strong evidence for FW associations (BFs >> 100) and no evidence for BW associations (BF = 0.41).

Alamia and colleagues interpret their findings as evidence that intravenous DMT shifts cortical spatio-temporal dynamics away from the typical eyes-closed dominance of backward travelling waves toward increased forward propagation, producing a pattern closely resembling that seen during genuine visual stimulation. They report concurrent decreases in alpha-band oscillatory power and increases in slower bands, and note that increases in forward-wave amplitude correlated with both real-time intensity ratings and post-hoc visual imagery measures. Taken together, the authors argue these observations link cortical travelling-wave directionality with conscious visual phenomenology: forward waves in particular appear to map onto vivid visual experience whether driven by external stimulation or DMT-induced endogenous imagery. Placing the results in the broader literature, the authors reference prior neuroimaging work showing occipital activation during ayahuasca imagery and earlier EEG findings of reduced alpha and increased signal diversity under psychedelics, reasoning that their travelling-wave results dovetail with these observations. They further discuss pharmacological considerations, noting that classic psychedelics, including DMT, exert actions primarily via 5-HT2A receptor agonism; this receptor has been implicated in drug-induced hallucinations and in clinical hallucinations in neurological disorders, offering a plausible receptor-level mediator of the observed cortical effects. The discussion frames the findings in predictive-coding terms: if backward waves convey top-down predictions and forward waves convey bottom-up prediction errors, then a DMT-induced reduction in backward waves and increase in forward waves is consistent with models positing that psychedelics reduce the precision weighting of priors and thereby liberate bottom-up signalling. The authors suggest that their data provide mechanistic support for this account and propose that similar travelling-wave shifts might characterise other endogenous visionary states such as dreaming or non-drug hallucinations; they recommend future work to test this conjecture and to examine other sensory modalities. The authors acknowledge limitations in the correlational analyses due to limited statistical power (n=12 in some tests) and the indirect nature of certain comparisons (for example, the visual-stimulation data derive from a different experiment with different participants and EEG setups). They therefore present their conclusions as consistent with, but not definitive proof of, the proposed predictive-coding interpretation, and they call for further studies to generalise and extend the present findings.

These are reported as the first EEG data examining human resting-state brain activity under pure DMT. Consistent with the study hypothesis, DMT shifted cortical travelling waves away from the baseline predominance of backward waves toward increased forward waves—a pattern resembling eyes-open visual stimulation—while reducing alpha-band oscillations. Increases in forward-wave amplitude correlated with both the time-resolved intensity of the DMT experience and post-session measures of visual imagery. The authors conclude that these results have implications for understanding the neural mechanisms of the psychedelic state and for theories of conscious visual perception.