Well, also this claim is definitely something that would represent an extraordinary advance. Any detailed correspondence between machine learning neural nets and actual neurology would be quite a step and not something neurologist or machine learning experts expect.