In this work, we use a deep learning framework for simultaneous classification and regression of Parkinson disease diagnosis based on MR-Images and personal information (i.e. age, gender). We intend to facilitate and increase the confidence in Parkinson disease diagnosis through our deep learning framework.