The use of deep-learning based magnetic resonance image reconstruction methods is a rapidly developing area. Such techniques hold promise to solve significant clinical challenges in terms of improvements in image quality and decreases in acquisition time for patient comfort, increased accessibility, and decreased cost. Neuroimaging is the use-case for this year’s fastMRI challenge and clinically relevant: MRI is the best way to image the brain with excellent soft tissue contrast that other imaging modalities lack and brain MRI is the number one most common type of MRI performed. In addition to quantitative metrics for reconstruction quality, in medical applications, it is important incorporate expert reader review into the evaluation process. We will review the evaluations of six subspecialty neuroradiologists who compose the 2020 fastMRI challenge expert panel and discuss the broader context of deep learning-based approaches to MR image reconstruction.