Empirical Viability of Nested Markov Blanket Identification Procedures Across Scales

Determine whether existing procedures for identifying nested Markov blankets from empirical timeseries can reliably decompose distributed systems across scales in real-world empirical studies.

Background

The authors outline an empirical program to test whether a system fits their metacognitive particle framework by inferring dynamics from data and checking for the required structural properties. They note that practical procedures exist for identifying nested Markov blankets from timeseries, with a worked example using brain imaging data.

However, it remains uncertain whether these procedures can effectively decompose large, distributed systems across multiple scales in broader empirical contexts.