Signal contributions to heavily diffusion-weighted functional magnetic resonance imaging investigated with multi-SE-EPI acquisitions

Diffusion-weighted (DW) functional magnetic resonance imaging (fMRI) signal changes have been noted as a promising marker of neural activity. Although there is no agreement on the signal origin, the blood oxygen level dependent (BOLD) effect has figured as one of the most likely sources. In order to investigate possible BOLD and non-BOLD contributions to the signal, DW fMRI was performed on normal volunteers using a sequence with two echo-planar acquisitions after pulsed-gradient spin-echo. Along with the changes to the signal amplitude (δS/S) measured at both echo-times, this sequence allowed changes to the transverse relaxation rate (δR₂) to be estimated for multiple b-values during hypercapnia (HC) and visual stimulation (VS). δS/S and δR₂ observed during HC were relatively insensitive to increasing b-value. On the other hand, δS/S demonstrated a clear dependence on b-value at both echo-times for VS. In addition, δR₂ during the latter half of VS was significantly more negative at b=1400s/mm² than for the time-courses at lower b-value, but δR₂ during the post-stimulus undershoot was independent of b-value. The results have been discussed in terms of two models: the standard intravascular-extravascular model for fMRI and a three-compartment model (one intra- and two extravascular compartments). Within these interpretations the results suggest that the majority of the response is linked to changes in transverse relaxation, but possible contributions from other sources may not be ruled out.