Automatic quality control of brain T1-weighted magnetic resonance images for a clinical data warehouse

•We propose a framework for the automatic QC of 3D T1w brain MRI for a clinical data warehouse.•We manually labeled 5500 images to train/test different convolutional neural networks.•The automatic approach can identify images which are not proper T1w brain MRIs (e.g. truncated images).•It is able to identify acquisitions for which gadolinium was injected.•It can also accurately identify low quality images.