On the advances in machine learning and complex network measures to an EEG dataset from DMT experiments

• DOI: 10.1088/2632-072x/ad1c68
• Publication Date: 2024-01-09T22:20:52Z
• Authors: Caroline L Alves; Manuel Ciba; Thaise G L de O. ...; Joel Augusto Mour...; Eduardo Pondé de ...; Christiane Thiele...; Francisco A Rodri...
• Keywords: Precentral gyrus

Abstract There is a growing interest in the medical use of psychedelic substances, as preliminary studies using them for psychiatric disorders have shown positive results. In particular, one these substances N, N-dimethyltryptamine (DMT), an agonist serotonergic that can induce profound alterations state consciousness. 


In this work, we exploratory tool to reveal DMT-induced changes brain activity EEG data and provide new insights into mechanisms action substance. We used two-class classification based on (A) connectivity matrix or (B) complex network measures derived from it input support vector machine.

We found both approaches could detect brain's automatic activity, with case showing highest AUC (89%), indicating measurements best capture occur due DMT use. 

In second step, ranked features contributed most result. For (A), differences high alpha, low beta, delta frequency bands were important distinguishing between before after inhalation, which consistent results described literature. Further, connection temporal (TP8) central cortex (C3) precentral gyrus (FC5) lateral occipital (P8) The regions TP8 C3 has been literature associated finger movements might occurred during consumption. However, cortical areas FC5 P8 not presumably related volunteers' emotional, visual, sensory, perceptual, mystical experiences consumption.

For (B), closeness centrality was crucial measure. Furthermore, discovered larger communities longer average path lengths when converse not, balance functional segregation integration had disrupted. These findings support
the idea becomes more entropic under psychedelics.

Overall, robust computational workflow developed here interpretability how (or other psychedelics) modify networks their mechanism action. Finally, same methodology applied may help interpret time series patients who consumed drugs.