Prediction models for similarity neuro images (FDG-PET) with 2-class classification in Python

Question

I am looking for classification models that are specified for brain image (neuro-scans) such as FDG-PET scans. I have a dataset that has images and the class assigned to each one of them (2 classes - Healthy patient, Diseased patient). I am heavily invested in the field of neuroscience and eager to seek new algorithms that could help analyze the images, however I am currently stranded on the island without any models left (and more importantly classifiers) to try out.

I have tried out Nilearn and FERM libraries but that's about it, received somewhat decent results (up to 75% CA with CV=20 on n=1000, where majority classification is 58%, there is an improvement). Is there any library that you could recommend to me to try out? This can be non-neuro oriented classification model classifiers as well.

The only requirement is that it's using supervised learning framework and that it's available as a Python library.