0_3564 Periodic temporal fluctuations in the frequency range below 10 Hz characterize signals related to human basal cerebral hemodynamics and metabolism and the study of these fluctuations can reveal interesting physiological and pathological information. Functional near-infrared spectroscopy (fNIRS) can noninvasively monitor cerebral hemodynamic signals both at rest and during tasks. So far, only the continuous wave (CW) and frequency domain (FD) fNIRS modalities were employed to investigate cerebral oscillations, while the time domain (TD) fNIRS has been hindered by lower signal-to-noise ratio when operating at the high (>10Hz) sampling rate required for being sensitive to these oscillations. The authors aim to fill this literature gap by using a recently developed TD fNIRS device able to perform in-vivo measurements at acquisition rates up to 20 Hz. To support the interpretation of TD fNIRS data acquired in the in-vivo studies, a simulation study was conducted. The human head was modelled either as a homogeneous medium or as a two-layer medium. The solutions of the diffusion equation for these two geometries were employed to simulate two datasets of TD fNIRS signals (i.e., photon distribution of time of flight, DTOF) with a source-detector distance of 4 cm. Known and periodic perturbations of the concentrations of oxy- (O₂Hb) and deoxy- (HHb) hemoglobin in the modelled medium were imposed, determining changes in the absorption coefficient and hence in the DTOF. The homogeneous slab model was used to determine the effect of multiple measurement parameters (acquisition time, sampling rate, and average photon count rate) on the technique sensitivity to oscillatory phenomena. This allowed to establish guidelines for defining the best experimental protocol for in-vivo measurements. The bilayer model was designed to mimic the heterogeneous structure probed during brain monitoring applications and was used to evaluate the technique sensitivity to detect and separate oscillations occurring at different depths using two different approaches. Firstly, a time windowing of the DTOFs was performed using 10 gates of variable width, and the power spectral density of the signal was analysed for each. Secondly, the mean partial pathlength method (MPPM) was used to retrieve the concentrations of O₂Hb and HHb in the two layers of the medium. Overall, the results showed that the TD fNIRS allows for the detection and depth-localization of periodic fluctuations in the concentrations of O₂Hb and HHb within the probed medium. The gating approach demonstrated outstanding ability in detecting these oscillations even in the presence of very low count rates, along with the ability to isolate the behaviour of the more superficial layer of the medium. Meanwhile, the MPPM demonstrated to correctly reconstruct the O₂Hb and HHb signals coming from the two layers, with good sensitivity and no cross-talk between the layers.