Citation
Power-law statistics and universal scaling in the absence of criticality
TL;DR
Power law-like patterns possible without criticality
Challenges neural criticality hypothesis by demonstrating that alternative mechanisms (network modularity, sampling bias, specific inhibitory circuit dynamics) can generate power law-like distributions without requiring criticality.
Important counterpoint showing that power law observations don't necessarily imply specific underlying mechanisms - multiple processes can produce similar distributional patterns.
Key Findings from Touboul & Destexhe (2017)
- Power law-like patterns possible without criticality
- Network structure can generate similar distributions
- Sampling effects can create apparent power laws
- Caution needed in inferring mechanisms from distributions