Skip to main content

Table 1 Risk of bias analysis criteria

From: Machine learning methods for predicting progression from mild cognitive impairment to Alzheimer’s disease dementia: a systematic review

Risk of bias Score Criteria
Database Low (0) Use of validated and widely used dataset for the study of biomarkers of Alzheimer’s disease (AD) including several years of follow-up with information of stable and progressive MCI patients
Medium (1) Use of similar database with less widespread usage
High (2) The participants were selected by the authors and no validated database was used
Validation of the classification method Low (0) The study validates the classification method with a test sample and/or an independent sample
Medium (1) It uses a different validation method
High (2) There is no validation of the classification method
Mathematical development of the algorithms Low (0) Explanation of their theoretical basis or architecture for neural networks
Medium (1) The authors refer to literature but do not develop their mathematical notation or architecture
High (2) No information about the model