Deep Learning (DL) has proved to be a promising methodology for classification, recognition, prediction and end-to-end tasks. Recently it has proved its high potential in healthcare applications demostrating to be a succesful approach for physiological signals analysis. Therefore, in this paper we survey the lastest scientific research on deep learning in physiological signal data. The purpose of this paper is to indentify the lastest contributions on this field and overview lastest deep learning approaches.