Sleep disorders and nightmare disturbances challenge clinical assessment due to subjective self-reporting and lack of quantitative precision. This work presents DreamSense, a multi-task deep learning framework using electroencephalography (EEG) biomarkers that classify dream states on three independent tasks: occurrence detection, emotional sentiment, and nightmare intensity. Via systematic...