The authors of the founding deep-network critical-period paper explicitly declined to claim their network is a valid model of neurobiological information processing

The conclusion of Achille, Rovere & Soatto (2019) pre-empts the reading that their result closes a biology→deep-learning loop. The authors state their aim plainly: "Our goal in this paper is not so much to investigate the human (or animal) brain through artificial networks, but rather to understand fundamental information processing phenomena, both in their biological and artificial implementations. It is also not our goal to suggest that, since they both exhibit critical periods, DNNs are necessarily a valid model of neurobiological information processing, although recent work has emphasized this aspect."

The hedge is load-bearing because it comes from the same paper that produced the animal-timing match (claim-deep-nets-have-critical-learning-periods-timed-like-animals). The authors observe that a deep network and an animal share the onset/length signature of a critical period, and in the same breath decline to license the inference that the network is therefore a model of the brain. They frame both systems as instances of a shared information-processing phenomenon rather than one being a model of the other — which is the reading the mechanism note develops (claim-critical-periods-arise-from-information-plasticity-not-biology).

This sits directly on the vault's backpropagation-gap thesis (backpropagation-gap): shared representational or dynamical phenomena between brains and nets do not imply shared mechanism, and the identification should not be read in the strong direction the authors themselves refuse. It is also the epistemic brake on the loop-closure thread — a quantitative human-timing match, were it ever made (question-deep-net-critical-period-predicts-human-amblyopia-timing; claim-no-dnn-model-has-matched-human-critical-period-timing), would have to argue past the founding authors' own disclaimer, not merely extend their result.