The article summary is incorrect. MR Spectroscopy (MRS) is used today to measure molecules inside the brain. Resolution is not great for 3D MRS in clinical applications (due to the tradeoff between SNR and resolution, acquisition times are slow), but it's more than high enough to distinguish between different regions of the brain. And it's very common to perform single-voxel imaging and only get the spectroscopy for a given piece of tissue - for example, where a tumor is located.
MRS easily detects metabolites and ratios, like choline, NAA, as well as things like lipids, and alcohols. It requires expensive scanners, but it works and is used routinely in brain imaging today. The article mentions something that does not work clinically, and is being demonstrated in a lab with a piece of meat. The technology in the article is not a "first step" to understanding molecules in the brain, because we already have that technology today with MRS.