Hemodynamic response based mixture model to estimate micro- and macrovasculature contributions in functional MRI

A multicomponent model reflecting the temporal characteristics of micro- and macrovasculature hemodynamic responses was used to fit the time-course of voxels in functional MRI (fMRI). The number of relevant components, the latency of the first component, the time-separation among the components, their relative amplitude and possible interpretation in terms of partial volume contributions of micro- and macrocomponents to the time-course data were investigated. Analysis of a reversing checkerboard experiment revealed that there was no improvement in the filing beyond two components. Using a two-component model, the fractional abundances of the micro- and macrovasculature were estimated in individual voxels. These results suggest the potential of a mixture-model approach to mitigate partial volume effects and separate contributions of vascular components within a voxel in fMRI.