This study tested various sources of changes in respiratory sinus arrhythmia (RSA). Twenty-two healthy participants participated in three experimental conditions (mental stress, relaxation, and mild physical exercise) that each consisted of three breathing parts (normal breathing, breathing compressed room air, and breathing compressed 5% CO2-enriched air). Independent contributions to changes in RSA were found for changes in tonic vagal modulation of heart rate, central respiratory drive (i.e., PaCO2), respiratory depth, and respiratory frequency. The relative contributions to changes in RSA differed for mental stress and physical exercise. It is concluded that uncorrected RSA will suffice to index within-subject changes in tonic vagal modulation of heart rate in most situations. However, if the central respiratory drive is expected to change, RSA should ideally be corrected for changes in PaCO2, respiratory depth, and respiratory frequency.