What to track for chronic headaches — and what's just noise

Everyone who's tried to figure out their headaches has done some version of tracking. You notice a headache, think back to what you ate, how you slept, whether you were stressed — and file away a mental hypothesis. "I think it was the cheese." "Maybe I didn't drink enough water."

The problem is that human memory is terrible at this. You remember the headache day. You don't reliably remember the three non-headache days where you ate the same cheese and slept the same amount. You notice the trigger that confirms your theory and miss the times it didn't pan out. Psychologists call this confirmation bias, and it's the reason decades of mental tracking rarely produce actionable answers.

Systematic tracking works differently. Not because it's exhausting or complicated, but because it captures both headache days and non-headache days with equal attention — giving AI the comparison data it needs to distinguish real patterns from coincidences.

The seven things that matter most

Headache pattern-finding depends on parallel variables tracked consistently. The American Migraine Foundation recommends tracking for 8-12 weeks for comprehensive pattern identification, but meaningful preliminary analysis can start after 2-3 weeks. Here's what to prioritize:

Headache details. When it started, when it peaked, when it resolved. Which side. What it felt like (throbbing, pressure, stabbing). Severity on a 1-10 scale. Associated symptoms — nausea, light sensitivity, sound sensitivity, aura, neck stiffness. These details help distinguish headache types, and different types have different triggers.

Sleep. Hours and quality (good/fair/poor). What time you went to bed and woke up. Sleep disruption is one of the most consistently documented headache triggers across every headache type. Crucially, both too little and too much sleep are triggers — and so are changes in sleep schedule. Research in Therapeutic Advances in Neurological Disorders found the relationship is bidirectional: poor sleep triggers headaches and headaches disrupt sleep.

Stress. A daily score (1-10) or even high/medium/low captures enough. But also note stress changes — a high-stress week followed by a relaxed weekend can trigger let-down migraines. The pattern isn't just "high stress = headache" but "stress transition = headache." If you can note what the stress is about, even briefly, that helps distinguish types of stress that affect you from those that don't.

Food and drink. What you ate, roughly when. You don't need macro breakdowns — "pasta with tomato sauce around noon" is enough. Pay particular attention to timing: when you ate relative to when the headache started. Also track meals you skipped — fasting and irregular meal timing are well-documented triggers. Caffeine and alcohol deserve specific attention: note type, amount, and timing.

Medications and supplements. Everything you take, every change. This is critical for two reasons. First, some medications have headache as a side effect — oral contraceptives, vasodilators, PPIs, and overuse of pain medication itself. Second, tracking acute medication use (how many days per month you take painkillers, triptans, etc.) is essential for identifying medication overuse headache, one of the most common and most treatable causes of chronic daily headache. Also tell Iris about your supplements — some can affect headaches positively or negatively, and it's worth having AI review what you're taking.

Weather and environment. Barometric pressure changes, temperature shifts, altitude, bright light, strong smells. You don't need to log the weather manually — AI can cross-reference your headache dates with weather data for your location. But note environmental exposures: fluorescent lighting, strong perfume, smoke, altitude changes during travel.

Hormonal cycle. If you menstruate, cycle phase is one of the most powerful headache predictors. Research in Current Pain and Headache Reports found roughly 60% of women with migraine report menstrual-related attacks, typically in the 2 days before through 3 days after the start of menstruation. Tracking cycle phase alongside headaches can confirm or rule out this factor quickly.

What you can skip

You don't need to track water intake unless you genuinely suspect dehydration (and even then, a week of focused tracking is usually enough to confirm or rule it out). You don't need to log every physical activity. You don't need a detailed food diary with portion sizes and nutritional content — the category level ("had pizza," not "had 2.5 slices of thin-crust pizza with 180g of mozzarella") is what matters for pattern detection.

More tracking isn't better tracking. The goal is enough parallel data that AI can test multi-factor combinations, without so much detail that you burn out in a week.

Making it sustainable

The biggest risk in headache tracking isn't missing a variable — it's stopping after five days because it felt like a chore.

The end-of-day voice note. At bedtime, just talk for two minutes. "Woke up at seven, felt okay. Had coffee and toast. Meeting was stressful around eleven. Skipped lunch, had a late sandwich around two. Headache started around four, right side, throbbing, maybe a six. Took ibuprofen. Better by evening. Slept okay last night, maybe seven hours." The Tracking Agent structures all of that into entries — headache with timing and severity, meals, stress, medication, sleep. You don't fill out forms. You just ramble through your day.

Conversational logging with Iris. If voice notes aren't your style, Iris can run the sprint as a back-and-forth in chat: she reaches out at the times you've agreed on and asks the few things the investigation needs — sleep last night, headache today, stress, what you ate, what you took. You answer in a sentence or two and the Tracking Agent structures it from there.

Don't track on a bad day if you can't. If you're in the middle of an attack, lying in the dark, the last thing you need is a tracking obligation. Log what you can remember afterward. A slightly imprecise entry from the next morning is infinitely more useful than no entry at all.

When a tracking sprint makes sense

You don't need to track your headaches forever. Most of the time, you and Iris are just talking — about what you're noticing, what your neurologist said, what you're trying. Tracking enters the picture when there's a question worth answering: is the new preventive working? Is the cycle phase pattern as strong as it feels? Is medication overuse pulling the daily headache forward?

That's when Iris may suggest a tracking sprint — a focused stretch of consistent logging for two or three weeks, aimed at the question on the table. When the sprint ends, the daily logging ends with it.

Consistency beats precision (when you're in a sprint)

Three weeks of daily tracking — even rough daily tracking — gives AI enough repetitions to test whether a pattern is real. If poor sleep appeared before 8 of your 10 headaches but also appeared on 15 non-headache days, that's a weak signal. If poor sleep plus skipping meals appeared before 7 of 10 headaches and only 2 non-headache days, that's a strong signal.

AI needs that comparison. It needs the boring non-headache days as much as the dramatic headache days. That's why "I only track when I get a headache" doesn't work — you're missing the control group.

The American Migraine Foundation's 8-12 week recommendation is ideal, but don't let the perfect be the enemy of the good. Two to three weeks of consistent data is enough to start identifying your strongest patterns. You can always keep going.

When to analyze

Don't analyze too early. One week of data almost always shows patterns that are coincidences. Wait for at least two weeks, ideally three. Then ask Iris to run the analysis: "Look at my headache entries alongside my sleep, stress, food, and medication data from the last three weeks and find the strongest trigger combinations." The Data Analyst examines every possible pairing and identifies which combinations hold up across multiple instances.

And if the first analysis is inconclusive — it sometimes is — that's not failure. It means you need more data points, or you need to track an additional variable. The investigation continues.

References

Headache diary: a useful tool in the management of primary headache — The Journal of Headache and Pain, 2011. Evidence for structured headache diaries improving management.
Sleep and migraine: a bidirectional relationship — Therapeutic Advances in Neurological Disorders, 2018. Sleep disruption as trigger and consequence.
Menstrual migraine: an updated review — Current Pain and Headache Reports, 2019. Hormonal patterns and attack timing.
Perceived trigger factors of migraine — Cephalalgia, 2019. Why multi-variable tracking outperforms single-trigger tracking.
Medication overuse headache: a critical review — The Journal of Headache and Pain, 2014. Medication tracking as essential diagnostic tool.