Depth dependence of coherent [Hb] and [HbO2] oscillations

The depth dependence of coherent hemodynamics may be measured in several ways with FD-NIRS. One way is to perform measurements at different source-detector separations, with data collected at longer separations being more sensitive to deeper hemodynamics than data collected at shorter source-detector separations. Another way is to perform measurements to collect different data types, with phase and dual-slope data being typically more sensitive to deeper tissue than AC amplitude and single-distance data. We considered both approaches and consistently found the results reported in Fig. 1(a) for data collected on the human forehead (in this case, for coherent hemodynamic oscillations at a frequency of 0.1 Hz):

• The amplitude ratio of deoxy- to oxyhemoglobin oscillations (|D/O| increases at greater tissue depths (i.e. in brain vs. scalp);
• The phase difference between deoxy- and oxyhemoglobin oscillations (Arg(D/O)) transitions from values close to 0 (in phase oscillations) to values close to 180 degrees (opposition of phase oscillations) from superficial (scalp) to deeper (brain) tissue.

A phasor decomposition of the measured oscillations (O_Measured, D_Measured) into blood volume (O_CBV, D_CBV) and blood flow (O_CBF, D_CBF) components, with guidance from the CHS model, allows us to interpret this result. The case of shallow tissue (scalp) (namely a smaller relative amplitude of D vs. O, & D and O oscillations almost in phase) is consistent with hemodynamics dominated by blood volume oscillations (see Fig. 1(b)). By contrast, the case of deeper tissue (brain) (namely a greater relative amplitude of D vs. O, & D and O oscillations almost in opposition of phase) is consistent with hemodynamics where blood flow provides a more significant contribution. This analysis is in line with the fact that blood volume changes are more limited in the brain than in scalp tissue because of the physical confinement of the rigid skull. This result provides important indications toward the goal of achieving specific sensitivity to cerebral hemodynamics in fNIRS and cerebral oximetry.