Journal of Physics: Complexity
January 22, 2024
Caroline L. Alves, Manuel Ciba, Thaise G. L. de O. Toutain et al.
5 citations
Machine learning applied to EEG data reveals that the psychedelic DMT disrupts the balance between functional segregation and integration in cortical brain networks, making brain activity more entropic. Complex network measures such as closeness centrality best capture these changes, achieving 89% AUC in classifying brain states before and after DMT inhalation. Key connectivity differences involve the temporal and central cortex and the precentral gyrus and lateral occipital cortex, the latter presumably related to emotional, visual, and mystical experiences. Larger communities and longer average path lengths occur under DMT, supporting the view that psychedelics increase brain entropy.
bioRxiv (Cold Spring Harbor Laboratory)
December 11, 2025
Caroline L. Alves, Fernanda Palhano-Fontes, Thaise G. L. de O. Toutain et al.
Ayahuasca alters conscious experience, and this study identifies EEG markers of its network-level effects using machine learning and complex-network analysis. In a randomized, double-blind, placebo-controlled trial with naïve ayahuasca users, resting-state EEG was recorded before dosing, 2 hours after, and 4 hours after. Connectivity was estimated with sliding windows; optimal classification performance occurred at 60–70 seconds (AUC and accuracy = 0.93). Network analysis revealed a bilateral decrease in eigenvector centrality (weaker hub influence), increased degree heterogeneity in the right hemisphere, and reduced global efficiency in the left. Posterior-left connections weakened acutely, while right temporal–central coupling transiently strengthened. The findings suggest that hub-centric shortcuts weaken, routing communication through more distributed, less efficient pathways with right-lateralized expression.
medRxiv
June 16, 2022
Caroline L. Alves, Thaise G. L. de O. Toutain, Joel Augusto Moura Porto et al.
preprint
A machine-learning method using support vector machines classified EEG data from volunteers before and after inhaling the psychedelic DMT. Complex network measures derived from brain connectivity achieved 89% AUC, outperforming raw connectivity matrices. Key distinguishing features included connections between temporal and central cortex regions (TP8-C3) linked to finger movements, and between precentral gyrus and lateral occipital cortex (FC5-P8) potentially related to emotional and mystical experiences. Closeness centrality was the most important network measure. DMT increased community size and average path length, disrupting the balance between functional segregation and integration, supporting the idea that cortical activity becomes more entropic under psychedelics.
PLoS One
January 1, 2026
Kleber Monteiro Pinto, Thaise G. L. de O. Toutain, Hugo Saba et al.
Mediums in a trance state showed distinct brain electrical activity compared to when they wrote in a normal waking state, with increased power in theta, alpha, beta1, and beta2 frequency bands. The quality of texts produced during trance was higher than the mediums' own waking texts, but there was no overall difference in text quality between mediums and a control group who simulated mediumistic writing. The findings indicate that complex language production is preserved across groups, while trance is associated with unique spectral brain dynamics during writing.