The objective of this research is to leverage multimodal wearable time series data to advance psychophysiological monitoring across a variety of behavioral health contexts. Mental health and substance use disorders remain global health crises, affecting nearly 1 billion individuals across the world. Currently, the main methods for psychiatric screening rely on surveys and self-report questionnaires, which are infrequently measured and may suffer from recall bias. As such, there is a need for better tools to monitor changes in psychophysiological state across a variety of clinical contexts to augment screening tools, improve symptom management, and enable adjunctive closed-loop therapies downstream. To this end, objective physiological biomarkers, especially those extracted from wearable sensors and recorded continuously over time, represent promising data streams for characterizing and predicting psychophysiological dynamics. Our first aim focuses on the characterization of brain-heart dynamics during episodic memory processing in healthy adults using neural and cardiomechanical signals. The second aim of this work focuses on the analysis of cardiorespiratory biomarkers extracted from a multimodal wearable sensing patch during varying levels of acute opioid withdrawal severity. The final and proposed aim of this research will focus on personalized (participant-specific), data-driven modeling of withdrawal severity using over 2000 hours of cardiorespiratory and human activity recognition time series data.