Estimating arterial and venous volume dynamics using spatially resolved NIRS

Non-invasive monitoring of cerebral metabolism using near-infrared spectroscopy (NIRS) offers critical insights into brain function in health and disease. Important biomarkers such as the cerebral metabolic rate of oxygen (CMRO2) and oxygen extraction fraction (OEF) require assumptions about arterial and venous blood volume fractions, typically considered constant. However, these compartments show dynamic and distinct responses to physiological stimuli, limiting the accuracy of current models. We present a method for estimating relative changes in arterial blood volume using spatially resolved NIRS, using the slope of cardiac pulsation amplitudes across source-detector distances. While absolute quantification remains challenging, relative changes allow us to compute dynamic arterial volume fractions, from which venous contributions can be derived. Assuming a baseline arterial fraction of 10%, we estimate arterial and venous hemoglobin concentrations independently over time. Nine healthy participants underwent controlled respiratory tasks, including breath-holding and hyperventilation, while cerebral and peripheral hemodynamics were recorded. In the periphery, breath-holds revealed delayed venous responses and opposing arterial and venous volume trends. Hyperventilation induced pronounced venous volume reduction with smaller arterial change, consistent with enhanced respiratory pump activity. Cerebral responses differed markedly: arterial volume increased while venous volume dropped during hyperventilation, and both compartments showed phase-shifted dynamics during breath-holds. These findings highlight the importance of modeling arterial and venous compartments separately. Our method enables more physiologically accurate interpretations of NIRS signals and lays the groundwork for improved estimates of CMRO2 and OEF. This approach may enhance the utility of NIRS in both research and clinical neuro-monitoring applications.