An exploratory persistent-homology analysis of resting-state fMRI functional connectivity under Ayahuasca
Tales Ramos Monteiro Dos Santos, Dráulio B. Araújo, Helcio Felippe, José Garcia Vivas Miranda, Fernanda Palhano-Fontes, Raphael Silva Do Rosário, Heloisa Onias, Aline Viol, Gandhimohan M. Viswanathan, Fernando A. N. Santos
Chaos Solitons & Fractals May 30, 2026 Peer reviewed DOI: 10.1016/j.chaos.2026.118554 via OpenAlex
Summary
Ayahuasca ingestion in nine participants was examined for its effects on brain connectivity using resting-state fMRI. A nominal decrease in persistent entropy of H2 features was observed before and after ingestion, but this finding did not survive statistical correction. Additionally, analyses showed non-significant increases in temporal complexity. The study indicates that while preliminary results suggest potential changes in brain topology associated with psychedelics, further research with larger samples and placebo controls is necessary.
Study at a glance
| Design | exploratory analysis |
|---|---|
| Sample size | 9 |
| Population | participants who ingested Ayahuasca |
| Key finding | A nominal decrease in persistent entropy of H2 features was observed before and after Ayahuasca ingestion, but this effect did not survive correction for multiple comparisons. |
Abstract
Psychedelic states offer a useful setting for studying changes in large-scale brain organization. Here, we applied Topological Data Analysis (TDA) to resting-state fMRI functional connectivity from nine participants scanned before and after Ayahuasca ingestion. Vietoris–Rips filtrations were constructed from correlation-derived dissimilarity matrices, and persistent entropy was used to quantify the distribution of persistence lifetimes across homology dimensions S 0 – S 3 . In the primary absolute-correlation analysis, persistent entropy of H 2 features showed a nominal pre/post decrease ( W = 4 . 0 , p = 0 . 027 , rank-biserial correlation = 0 . 822 ). This effect did not survive correction across the four tested homology dimensions ( q FDR = 0 . 109 ) and was not reproduced when signed correlations were preserved using d ( i , j ) = 1 − r ( i , j ) . Exploratory signal-complexity analyses using Lempel–Ziv complexity and sample entropy showed descriptive but statistically non-significant increases in temporal complexity. These results should therefore be interpreted as preliminary and hypothesis-generating, particularly given the small sample size, lack of placebo control, availability of only GSR-preprocessed connectivity data, and sensitivity to the distance definition. The study suggests that persistent homology may provide a useful framework for studying psychedelic-associated changes in the higher-dimensional topology induced by functional connectivity, but replication in larger placebo-controlled datasets is required.