A novel automated quantification workflow was created for the analysis of neuromelanin (NM) in human post-mortem formalin-fixed paraffin embedded (FFPE) midbrain sections. After standard tissue processing, some sections were hematoxylin and eosin (H&E)-stained, and the remaining samples were left unstained. These samples were digitally captured using a slide scanner. Next, section annotations were created using the Olympus VS200 Desktop software (EVIDENT Technology GmbH, ver. 4.1.1 build 29408). These annotations are listed here as: ‘Training label for unstained NM.exp’ and ‘Training label for H&E NM.exp’. These files are Olympus VS200 Desktop-software specific and are necessary for performing automated neural training. Neural training is a specific, novel function of the TruAI feature of the Olympus VS200 Desktop software. The previously generated ‘.exp’ training label files are applied to digital images during automated neural learning sessions in the Olympus VS200 Desktop software. After several learning sessions the system sufficiently learns to recognise NM. These successful neural networks are listed here as ‘Neural network for unstained NM.nn’ and ‘Neural network for H&E NM.nn’. These networks are essential for performing automated quantification, and these .nn files are Olympus VS200 Desktop-software specific. Lastly, these neural networks are applied to scanned midbrain sections to generate raw NM data. These exported raw data can be found on: 10.5281/zenodo.14523067. To select features of interest and clean raw data, a general ‘NM cleaning script.R’ was created and used for further statistical analysis.