Today, most studies of cognitive processes using functional MRI (fMRI) experiments adopt highly flexible stimulation designs, where not only the activation amount but also the time course of the measured hemodynamic response is of interest. The measured signal only indirectly reflects the underlying neuronal activation, and is understood as being convolved with a hemodynamic modulation function. An approach to better allow inferences about the neuronal activation is given by modeling this convolution process. In this study, we investigate this approach and discuss computational models for the hemodynamic response. An analysis of a recent fMRI experiment underlines the usefulness of this approach.
Frithjof Kruggel, D. Yves von Cramon