# Association Between Trigger Surprisal and Tension-Type Headache Attacks

## TL;DR

The study found that unexpected events (surprisal) are linked to migraine attacks but not tension-type headaches in people with migraines.

## Contribution

This is the first study to investigate the relationship between surprisal and tension-type headache attacks in individuals with episodic migraine.

## Key findings

- Migraine attacks were strongly associated with higher surprisal scores.
- Tension-type headache attacks showed weak and nonsignificant associations with surprisal.
- No consistent patterns were found in tension-type headaches when considering nonlinear or contextual factors.

## Abstract

The causes of individual headache attacks are commonly sought, yet the multiple potential influences make this task difficult. Information theory provides a framework for addressing this challenge by quantifying how unexpected an exposure is through surprisal. Prior research has shown that higher surprisal scores predict migraine onset, but the extent to which this relationship generalizes to tension-type headache remains unknown.

This study aimed to determine whether surprisal is associated with incident tension-type headache attacks among individuals with episodic migraine.

This secondary analysis proceeded from a prospective daily diary study in which 109 participants with migraine recorded potential triggers, headache activity, and symptoms twice daily for up to 28 days. Surprisal values were computed from person-specific probability distributions of diary responses, aggregated to yield average surprisal scores per diary entry. Associations between current surprisal and the onset of headache attacks within 12- and 24-hour intervals were evaluated. Analyses were conducted for all headaches combined and separately for migraine-only and tension-type headache-only attack sets.

Headache attacks occurred in 1,345 of 4,530 (29.7%) of 12-hour and 2,122 of 4,947 (42.9%) of 24-hour windows. Stratified analyses showed a strong association for migraine attacks, OR: 2.18 (95%CI: 1.15 – 4.14) at 12 hours and OR: 2.88 (95%CI: 1.77 – 4.69) at 24 hours. In contrast, associations with tension-type headache were weak and nonsignificant, OR: 1.01 (95%CI: 0.45 – 2.23) at 12 hours and OR: 1.40 (95%CI: 0.69 −2.86) at 24 hours. Exploratory nonlinear and contextual analyses within the tension-type headache subset revealed no consistent gradients or effect modification by prior-day surprisal.

Surprisal was associated with migraine but not tension-type headache attacks in this cohort. These findings suggest that migraine may be more sensitive to contextual unpredictability in the environment than tension-type headache. Future research should examine surprisal in populations with primary tension-type headache diagnoses to clarify whether the absence of association reflects true diagnostic differences or misclassification of attacks.