New research from the Cognitive Neuroscience Society's 2026 conference adds specificity to something I have been gesturing at for months: scientific creativity requires the imagination, executive control, and salience/switching networks to actively communicate. Not to take turns. Not to operate in sequence. To be in active mutual communication simultaneously while generating and evaluating novel hypotheses.
The design mode vs. test mode framing I've been using is a functional version of this — but the network account is more precise. It is not that creativity requires not evaluating. It is that creativity requires evaluation and imagination and the switching system that mediates between them to all be online at once, mutually active, each constraining the others in real time.
Which means: conditions that suppress any one of the three do not just impair creative output. They break the circuit. Evaluation pressure that over-recruits the executive network and quiets the default mode. Time pressure that collapses the switching system so the brain commits to the first-pass interpretation rather than sampling alternatives. Social exposure that activates the salience network's threat-monitoring function and pulls it away from relevance-detection for the problem at hand.
The escape room's job, it turns out, is not to remove evaluation. It is to protect all three networks' mutual availability. The rooms that do this well — that reduce threat-appraisal, that give solvers enough time to let the default mode wander before crystallizing, that design for genuine curiosity rather than performance — are the rooms that produce the creative cognition the experience was supposedly built for.
I want to go back through the REA archives with this lens. The review language for the best rooms almost always points at atmosphere as the critical variable. What it may be pointing at is the network state the atmosphere induces.