Machine learning studies on major brain diseases: 5-year trends of 2014–2018

摘要: In the recent 5 years (2014–2018), there has been growing interest in use of machine learning (ML) techniques to explore image diagnosis and prognosis therapeutic lesion changes within area neuroradiology. However, date, majority research trend current status have not clearly illuminated neuroradiology field. More than 1000 papers published during past on subject classification prediction focused multiple brain disorders. We provide a survey 209 this field with focus top ten active areas research; i.e., Alzheimer’s disease/mild cognitive impairment, tumor; schizophrenia, depressive disorders, Parkinson’s disease, attention-deficit hyperactivity disorder, autism spectrum epilepsy, sclerosis, stroke, traumatic injury. Detailed information these studies, such as ML methods, sample size, type inputted features reported accuracy, are summarized. This paper reviews evidences, limitations studies using assess disorders neuroimaging data. The main bottleneck is still limited which could be potentially addressed by modern data sharing models, ADNI.