The 6,7Li MAS NMR spectra of lithium ions in paramagnetic host materials are extremely sensitive to number and nature of the paramagnetic cations in the Li local environments and large shifts (Fermi contact shifts) are often observed. The work presented in this paper aims to provide a rational basis for the interpretation of the 6,7Li NMR shifts, as a function of the lithium local environment and electronic configuration of the transition metal ions. We focus on the layered rocksalts often found for LiMO2 compounds and on materials that are isostructural with the K2NiF4 structure. In order to understand the spin-density transfer mechanism from the transition metal ion to the lithium nucleus, which gives rise to the hyperfine shifts observed by NMR, we have performed density functional theory (DFT) calculations in the generalized gradient approximation. For each compound, we calculate the spin densities values on the transition metal, oxygen and lithium ions and map the spin density in the M-O-Li plane. Predictions of the calculations are in good agreement with several experimental results. We show that DFT calculations are a useful tool with which to interpret the observed paramagnetic shifts in layered oxides and to understand the major spin-density transfer processes. This information should help us to predict the magnitudes and signs of the Li hyperfine shifts for different Li local environments and t2g vs eg electrons in other compounds.< Réduire