In polymer gel dosimetry, magnetic resonance imaging (MRI) is used to determine the spin-spin relaxation rate (R2) which in turn can be correlated with absorbed dose to provide a map of the spatial distribution of the absorbed dose in the irradiated dosimeter. High accuracy, precision and reproducibility of these dose maps are essential. Moreover, for dose verification around brachytherapy sources used for intravascular brachytherapy, a high spatial resolution is required (typically 0.01-0.1 mm). To achieve these microscopic resolutions, strong imaging gradients are applied. The Brownian motion of water molecules in the presence of these strong magnetic field gradients causes an attenuation of the MR signal. When using a multiple spin-echo sequence, this may result in a significant deviation in the measured R2. The diffusion-related change in R2 at high resolutions was investigated experimentally and correlated with predictions that were obtained numerically and algebraically. Diffusion weighting is determined by the self-diffusion coefficient D, and imaging parameters, quantified by the b-factor. The b-factor was calculated for a multiple spin-echo sequence for different gradient strengths and gradient pulse durations. The variations in R2 that were observed when changing the matrix size and slice thickness are explained. It is shown that a linear correlation between the matrix size and the variation in R2 is based on the diffusion weighting caused by the read-out gradients and slice selective gradients. In conclusion, the essence of taking into account molecular self-diffusion to quantify variations in the measured dose-R2 response when using high-resolution MRI in polymer gel dosimetry is emphasized.