Breaking respiratory airflow into time-localized parametric components reveals sub-breath dynamics that standard metrics miss, enabling better detection of breathing changes under cognitive stress.
This paper presents a new method to analyze breathing patterns by breaking down airflow signals into simple, interpretable components (half-sine, Gaussian, and beta shapes) rather than treating breath as a single unit. The approach captures fine details within each breath—like timing and coordination—and improves detection of cognitive fatigue by 30% compared to traditional breathing metrics.