Functional brain networks can be studied through homological cycles—topological objects that capture mesoscopic structure in weighted correlation networks. A new method, homological scaffolds, compactly represents these cycles and makes them amenable to standard network analysis. Applied to resting-state fMRI data from 15 healthy volunteers given placebo or psilocybin, the homological structure of brain activity changed dramatically after psilocybin, producing many transient, low-stability cycles and a few persistent ones absent under placebo.
Loss of consciousness from propofol sedation reduces long-range temporal correlations in frontothalamic brain activity and weakens the link between functional connectivity and anatomical structure. A model based on phase transitions in complex systems reproduces these patterns and also explains the cortex's reduced sensitivity to external stimuli during unconsciousness. The findings suggest that these neural changes are universal across different causes of unconsciousness.