Concentration of magnetic resonance imaging (MRI) contrast agents (CA) cannot be measured directly and is commonly determined indirectly using their relaxation effect. This requires knowledge of the relaxivity of the used CA. Quantitative perfusion studies involve measurement of CA concentration during a bolus passage. Relaxation is commonly converted into concentrations assuming identical relaxivities for tissue and blood (1,2). We show that the relaxivity of blood pool CA depends significantly on both host tissue and pulse sequence. We developed a comprehensive model using Monte Carlo simulations and analytical theory to describe this effect and created a program for fast, realistic simulation of transverse relaxation in perfused tissues. We performed simulations for gray matter and white matter. The predicted values of the relaxivity are in agreement with the relaxivity found in animal studies (3,4) as demonstrated in (5). It was previously found (6) that the perfusion measurements using dynamic susceptibility contrast inherently overestimate cerebral blood flow and volume. In view of the present result, this is attributed to the significant difference in the relaxivity of the CA in brain tissues compared to blood. The standard data analysis method of perfusion data neglects this difference, (1,2) and due to the linearity of the deconvolution in the standard data processing this difference becomes the overestimation factor in the blood flow and volume. (1) Østergaard, L et al., MRM 36, 1996. (2) Østergaard, L et al., MRM 36, 1996. (3) Johnson, KM et al. MRM 44, 2000. (4) Bjørnerud, A et al. MRM 47, 2002 (5) Kjølby, BF et al, MRM 56, 2006. (6) Kiselev, VG, MRM 46, 2001.