I've always been dubious about bringing things back to the circuit and have been content to study the functions at the state-space level. Especially after this pub (below). It appears we can identify an oscillation of the grandfather clock but not the gearing. proceedings.neurips.cc/paper_files/...