Bipolar Disorder (BD) is a chronic psychiatric condition characterized by recurring manic and depressive episodes that severely impair a person’s social, occupational and interpersonal functioning. Traditional psychiatric assessment relies almost exclusively on clinical interviews and subjective patient reports, which are highly susceptible to recall bias and often lead to diagnostic delays of several years. Digital phenotyping offers a different approach by utilizing wearable sensors for the continuous, passive, and objective quantification of physiological and behavioral patterns in every day settings to support faster diagnosis. This thesis, aims to address these challenges through two primary research tracks. The first objective is to evaluate the comparability and interchangeability of the Empatica E4 and EmbracePlus wearable devices both at the raw signal and extracted features levels, with the purpose of seamlessly employing data recorded with both devices.