DREAMS is a Python-based framework for generating automated model cards for deep learning models applied to electroencephalography (EEG) data. Unlike existing EEG data analysis frameworks that focus solely on preprocessing techniques or deep learning model development, DREAMS focuses on structured documentation and model interpretability. It provides structured documentation in YAML format, incorporating domain-specific metadata, preprocessing details, performance metrics, and uncertainty quantification. It documents model performance, dataset bias, and interpretability limitations through case studies on emotion classification tasks using the FACED dataset and abnormal EEG classification tasks using the TUH Abnormal dataset, thereby increasing transparency. It provides visualized performance metrics, dataset alignment details, and model uncertainty estimates, and is open sourced, so that it can be widely adopted in healthcare AI, research, and ethical AI development.
