So you have the boundary. Divergent thinking gets a boost from walking. Convergent thinking does not, and may actually suffer. And the size of the benefit depends partly on how well your internal feedback system is running. That feels like a workable map. You know when to walk and when to sit still. You know the mechanism. You feel like you understand the tool. But there is something the map does not show you. And it is the thing that matters most for whether any of this is actually useful in your life. Think about the last time you sat down to work on something genuinely difficult — not a routine task, but something that required real creative thinking. Something where you needed new ideas, not just execution. Now think about what your mental state actually was when you sat down. Were you in a calm, open, curious state? Or were you already carrying something — a deadline that felt too close, a conversation that had not gone well, a low-grade background hum of pressure that you had been living with for days? Because here is what the walking-creativity research does not fully account for. The mechanism we have been tracing — the loosening of the prediction system, the lowering of the threshold for new associations — that mechanism assumes a particular baseline. It assumes a brain that is running in what you might call a generative mode. A brain that is, at some level, safe enough to explore. When the brain is under sustained threat, it does not run in generative mode. It runs in defensive mode. And those are not just different points on the same dial. They are different operating systems. This is not a metaphor. The predictive brain that Barrett and Friston describe is constantly making a prior-level decision before it even gets to the question of what to think about: is this a situation that requires protection, or a situation that allows exploration? Under chronic stress — not acute, sharp stress, but the low-level persistent kind that a lot of people carry around as a baseline — the prediction system is already primed toward threat detection. It is scanning for what could go wrong. It is narrowing, not opening. And when you take that brain for a walk, something interesting happens. The walk still does what it does physiologically. The vestibular system still fires. The hippocampus still gets stimulated. But the prediction system that is supposed to use that input to generate new associations is already running a different program. It is not in the mood to make unexpected connections. It is in the mood to check for danger. Suppose you are someone who has been under real pressure for several months. Not a crisis, just the sustained weight of too much to do and not enough certainty about how it resolves. You go for a walk. You come back. And instead of the insight you were hoping for, you have spent forty minutes rehearsing the same three worries you walked in with. The walk did not fail. The mechanism did not break. But the baseline state you brought to the walk determined what the mechanism had to work with. This is the part that the simple story misses. The simple story says: walking boosts creativity. The more complete story says: walking boosts creativity in a brain that is available for that kind of processing. And availability is not a given. It is a function of what state the system is already in when you start. The recurring anchor in this episode — that moment of insight that arrives during a walk, seemingly out of nowhere — is real. But notice what it requires. It requires a particular quality of mental looseness. A willingness to let the mind drift without immediately pulling it back to the problem. That quality is not just a personality trait. It is a state. And states are produced by conditions. Research on interoception and stress points in a consistent direction here: when the body is under sustained pressure, the relationship between internal signals and how the brain interprets them becomes less reliable. The signal is still there. But the interpretation is filtered through a system that is already primed to read ambiguity as threat. That is not the same as saying stress destroys interoceptive accuracy — the evidence is more nuanced than that, and the relationship runs in both directions. But it does suggest that the quality of the feedback loop between body and brain is not fixed. It shifts with the conditions you are living in. Which means the sixty-percent figure from the walking study is not just bounded by task type. It is bounded by the state of the person doing the walking. And the state of the person is shaped by things that happen long before they put on their shoes. This is where the model gets genuinely complicated. Because if the baseline state is the real variable — if what determines whether the walk works is not the walk itself but the conditions you bring to it — then the question is not just how to use the tool. The question is how to change the conditions under which the tool operates. And that is a much harder problem, but also a much more interesting one. Because it turns out there are specific, evidence-grounded ways to shift that baseline. Not permanently, not magically, but meaningfully. And some of them are simpler than you would expect.