Cerebrovascular reactivity mapping (CVR), using magnetic resonance imaging (MRI) and carbon dioxide as a stimulus, provides useful information on how cerebral blood vessels react under stress. This information has proven to be useful in the study of vascular disorders, dementia and healthy ageing. However, clinical adoption of this form of CVR mapping has been hindered by relatively long scan durations of 7 to 12 minutes. By replacing the conventional block presentation of carbon dioxide enriched air with a sinusoidally modulated stimulus, the aim of this study was to investigate whether more clinically acceptable scan durations are possible. Firstly, the conventional block protocol was compared with a sinusoidal protocol of the same duration of 7 minutes. Estimates of the magnitude of the CVR signal (CVR magnitude) and the relative timing of the CVR response (CVR phase) were found to be in good agreement between the stimulus protocols. Secondly, data from the sinusoidal protocol was reanalysed using decreasing amounts of data in the range 1 to 6 minutes. The CVR magnitude was found to tolerate this reduction in scan duration better than CVR phase. However, these analyses indicate that scan durations in the range of 3 to 5 minutes produce robust data.