A vast number of drug molecules are unable to cross the blood-brain barrier, which results in a loss of therapeutic opportunities when these molecules are administered by intravenous infusion. To circumvent the blood-brain barrier, local drug delivery devices have been developed over the past few decades such as reverse microdialysis. Reverse microdialysis (or retrodialysis) offers many advantages, such as a lack of net volume influx to the intracranial cavity and the ability to sample the tumour’s micro-environment. However, the translation of this technique to efficient drug delivery has not been systematically studied. In this work, we present an experimental platform to evaluate the performance of microdialysis devices in reverse mode in a brain tissue phantom. The mass of model drug delivered is measured by computing absorbance fields from optical images. Concentration maps are reconstructed using a modern and open-source implementation of the inverse Abel transform. To illustrate our method, we assess the capability of a commercial probe in delivering methylene blue to a gel phantom. We find that the delivery rate can be described by classical microdialysis theory, except at low dialysate flow rates where it is impacted by gravity, and high flow rates where significant convection to the gel occurs. We also show that the flow rate has an important impact not only on the overall size of the drug plume, but also on its shape. The numerical tools developed for this study have been made freely available to ensure that the method presented can be used to rapidly and inexpensively optimise probe design and protocol parameters before proceeding to more in-depth studies.