AI and the Disordered Brain — piece one
I spent years treating stress like weather that blew in from outside.
Quit the work. Cut the teaching. Cleared the calendar. Waited for my body to get the memo.
It did not get the memo.
The thing that would not shut off was not the job. It was the copy of the job still running after the job was gone. A thought I had already thought, playing again, and my pulse answering it like it was new.
This is the first piece in a series I am calling AI and the Disordered Brain. Not AI as a productivity tip. AI as the thing that can catch a mind that will not sit still long enough to type.
There is a research name for that copy. Perseverative cognition. In 2006, Jos Brosschot, William Gerin and Julian Thayer argued that the mind’s running copy of a stressor is enough. Not a metaphor. A functional substitute. Your nervous system cannot tell a memory-of-threat from a threat.
That is why the come-down never quite comes down.
A 2016 analysis by Cristina Ottaviani and colleagues pooled sixty studies of healthy people — not patients — worrying and ruminating with wires on. Blood pressure ran higher. Heart rate ran higher. Cortisol ran higher. Heart-rate variability ran lower. Heart-rate variability is the brake. Low variability means the part that calms you down has less say.
The sizes are modest, and that honesty matters. Cortisol moved in the same neighborhood as a mild lab stress task. Heart rate moved less than a genuine startle. This is not being chased. It is the copy of being chased, left running.
A later set of lab analyses — twenty-eight studies — split the brake three ways, and only when people were compared to themselves. Resting: lower. Reactivity: lower. Recovery: lower. Recovery is the come-down.
The event ending is not the response ending.
I notice it in the size of the thought. A small loop burns off. A big one — wanting to move, needing it now — can last for days. I cannot sit. I cannot think. I pace. It becomes an overwhelming drive to the point where I no longer can think and function very well. Getting it out helps. Leaving it in does not.
Night is not an off switch.
Worry during the day shows up in the body during sleep. After a five-minute speech task, people who kept turning it over took longer to fall asleep — they said so, and a wrist monitor said so too. Evening rumination tracks late sleep and thin sleep. One 2025 paper found that on higher-stress days, racing thoughts at bedtime cost total sleep and time to fall asleep, and heart rate did not explain it. The mind staying on was enough.
You can look fine on a watch and still be paying for a conversation that ended on Tuesday.
I have a night show for this. I go to bed when it is time to sleep. If my brain will not shut off, rain sounds do nothing. A documentary with an even-voiced presenter will — interesting enough that the brain is not just racing racing racing, even enough that I doze off without noticing. Some nights that works. Some nights I am up at ten with the loop still going, coming out of sleep and coming out of it again, and I cannot even think my way to the documentary.
The next day collects the bill. Tired. I do not cook, or I do not eat. If I do not eat, the ruminations are worse. Missed meals compound it. Tired, then overstimulated, then another bad night. That cycle is why I built a program around the small leaks. Did you eat. The simple things.
I am not a scientist. I read the papers because I was trying to understand why I could not sit still after the room was already quiet.
What it feels like, for me, is not “anxiety” as a headline. Most days it is fire. I cannot sit. Too much firing at once. Other days it is pressure, not a headache, more like my skull is a size too small for what is happening inside it. It shows up in a workout. It shows up in a video game. I miss the block in front of me because a thought arrived and I have nowhere to put it.
There is no clean study that measures that pressure in the head. Anyone who tells you otherwise is stretching something. What the research will stand behind is a body held in a low-level defensive posture — jaw, neck, scalp, shoulders — consistent with the heart, the brake, the cortisol, the slow come-down. The pressure is my report. It fits. It is not a finding.
One afternoon I sat with Grok Live and did twelve brain dumps in a row. I kept thinking I was done. I was not done. Each time I emptied one, another was already waiting. On twelve, the pressure dropped. Vision cleared a little. Same shape as eating after you did not know you were hungry. That is the first time I felt a tool keep up with the speed of the thoughts.
What is settled: worry and rumination travel with a higher heart rate, higher blood pressure, higher cortisol, and a softer brake, in healthy people. Recovery is slower when the copy keeps running. Sleep tracks the mind staying on.
What is open: how much of that leftover activation happens without you noticing. Whether this is the road from chronic stress to illness, or one road among several. How much of it is ordinary overthinking and how much is the far end of the range.
What this is not: a panic attack. Panic is a specific thing. This is prolonged low-grade activation. Nothing here diagnoses anybody, including me.
I used to think writing would empty it. Notebooks in every room. Writing is too slow for the way this brain moves, and once a sentence is on paper I can see it — I cannot picture an org chart in my head — so the page becomes a map, and then I chase every road on the map. Dopamine. Tabs. Information I do not need.
Spoken AI is the first thing that let me empty the loop without turning the loop into a project. I cannot see it, so I cannot chase it. I spew the idea and I can leave it. That is a brain dump: I am not ready to work yet. My brain has not emptied everything out yet and I am not ready to concentrate. Paper is for when I am ready to work. Spoken AI is for emptying.
That is not a protocol. There is no trial that says talk to a chatbot and your cortisol behaves. It is one woman’s workaround. Medication and a doctor were part of how I even got still enough to notice the loop. I am not handing either of those out.
The door this series walks through is not “use AI.” It is: match the tool to how a disordered brain actually fires, and how it types — or cannot type fast enough. A brain that writes will want a page. A brain that outruns a keyboard needs something that can listen.
Next in this series is one article: the systems I actually work with, in their own voices. What each one is for. What kind of thinker it fits. What you should not use it for. That piece is an interview, not a review.
After that, a separate piece for people who do not live in this house: how to match a system to a brain, in plain language, without the roster.
If you see yourself in this: I feel you.
I took this to a doctor and a therapist. That is my route. Not advice.
The alarm was not in the world.
The alarm was in the house.
The house is me.
Sources
- Brosschot JF, Gerin W, Thayer JF. The perseverative cognition hypothesis: a review of worry, prolonged stress-related physiological activation, and health. Journal of Psychosomatic Research. 2006;60(2):113-124.
- Brosschot JF. Markers of chronic stress: prolonged physiological activation and (un)conscious perseverative cognition. Neuroscience and Biobehavioral Reviews. 2010;35(1):46-50.
- Ottaviani C, Thayer JF, Verkuil B, Lonigro A, Medea B, Couyoumdjian A, Brosschot JF. Physiological concomitants of perseverative cognition: a systematic review and meta-analysis. Psychological Bulletin. 2016;142(3):231-259. doi:10.1037/bul0000036
- Zoccola PM, Dickerson SS, Lam S. Rumination predicts longer sleep onset latency after an acute psychosocial stressor. Psychosomatic Medicine. 2009;71(7):771-775. doi:10.1097/PSY.0b013e3181ae58e8
- Watkins ER, Roberts H. Reflecting on rumination: consequences, causes, mechanisms and treatment of rumination. Behaviour Research and Therapy. 2020;127:103573. doi:10.1016/j.brat.2020.103573
- Verkuil B, Brosschot JF, Gebhardt WA, Thayer JF. When worries make you sick: a review of perseverative cognition, the default stress response and somatic health. Journal of Experimental Psychopathology. 2010;1(1):87-118.
- Shaif NAS, Lim J, Reffi AN, Chee MWL, Massar SAA, Ong JL. Sleep reactivity amplifies the impact of pre-sleep cognitive arousal on sleep disturbances. Journal of Sleep Research. 2026;35(3):e70220. doi:10.1111/jsr.70220 (epub 2025 Oct 14)
- Lim C, Mueller E, O’Brien W. Perseverative cognition and vagally mediated heart rate variability in laboratory studies: a series of meta-analyses. International Journal of Psychophysiology. 2026;224:113368.