Mental health has become a global issue with growing numbers of cases. Digital phenotyping in mental healthcare provides a highly effective, scaled, cost-effective approach to handling global mental health problems. Monitoring and diagnosis of mental health through mobile technologies are at an early stage. This paper proposes a novel architecture for a mobile application design based on speech and behavioural analysis that can lead to evidence-based monitoring outside of clinical settings. The application is designed to monitor the overall mental health status of individuals based on mood, stress, behaviour, and personality. It proposes to integrate objective mental health assessment from smartphone data with subjective assessments via periodic, short, self-reported, standardized questionnaires. The solution proposes explainable assessments for individual users and clinicians to build user trust and system credibility. This research provides insight into how digitally assisted mental health monitoring can be implemented in future mobile digital technologies using passive sensing.
