Solvent-based processes have shown technical advantages over thermal techniques for recovery of heavy oil and bitumen. The success of these processes relies on accurate computation of molecular diffusion coefficient which determines how fast a solvent penetrates into oil. Concentration profile measurements of solvent in oil are used for the determination of the molecular diffusion coefficient. Although numerous experimental techniques have been proposed, the accurate estimation of this parameter is still a topic of debate in the literature. In this work, 1-D nuclear magnetic resonance imaging (MRI) is employed to obtain diffusivity data for a toluene heavy oil system. Diffusion of toluene in heavy oil was monitored for 20 days at a controlled temperature of 35 degrees C and ambient pressure. Over time, toluene diffusion into oil leads to changes in spatial distribution of T-1 and T-2 that affect the received signal. This serves as the basis of the solvent and heavy oil concentration estimation. Consequently, concentration profiles were established by converting the MRI signals to concentration values. This conversion was achieved by creating samples with known concentrations of heavy oil toluene and measuring their response in the same environment and parameter settings. A concentration-dependent diffusion coefficient was obtained from concentration profiles. The results show that relaxation based 1D MRI is an accurate and robust tool to obtain diffusivity data in complex fluids such as heavy oil.