Purpose:
This study aimed to examine, in an exploratory secondary analysis, whether signed airflow was linearly associated with horizontal impulse balance in three trained middle-distance runners and whether the accompanying multivariate movement-pattern responses were similarly organised.

Material and methods:
Three trained male runners (22–28 years) completed one 3-min trial per condition in randomised athlete-specific sequences. Thirteen completed athlete-condition records covered airflow velocities of –7, –5, 0, +3, and +5 m · s^–1. Twenty consecutive valid steps from the final minute were averaged for each condition. A pooled within-athlete linear slope for the propulsive-to-braking impulse ratio was tested with a conditional exact randomisation procedure.

Results:
The impulse ratio decreased by 0.0665 per 1 m · s^–1 progression from headwind towards tailwind (95% conditional randomisation interval: –0.0819 to –0.0517; exact p = 0.000014). Individual slopes were negative in all three athletes. Propulsive impulse decreased and braking impulse increased, both associations persisting after Holm correction; temporal-spatial outcomes showed no adjusted statistical association. The response-vector analysis was exploratory and hypothesis-generating.

Conclusions:
Signed airflow was associated with a directionally consistent redistribution of horizontal impulse in the three investigated athletes. These observations do not establish a population effect or a stable individual characteristic and require confirmation with repeated trials and larger samples.