Digital Phenotyping is an indirect method of monitoring and assessing human behavior by measuring and analyzing how a person uses the digital devices around them. This is most commonly done through the use of mobile phone sensor data, but other devices can be used as well such as automobile data or other smart sensors in the home or hospital. Digital phenotyping is sometimes used as an alternative to wearable sensors in cases where wearable sensors are not feasible or possible. Digital phenotyping with mobile phones dates back to the Funf project (http://www.funf.org/) at the MIT Media Lab in 2011, shortly after the emergence of smart phones. The term "digital phenotyping" first appeared in academic journals in the paper by John Torous et al in JMIR Mental Health, in May 2016.
Mobile Path, developed in our group, is one of many digital phenotyping platforms that have been created in recent years as a research tool for conducting clinical studies and as a tool for health monitoring (e.g. to monitor the recovery of cancer patients) or deploying behavioral interventions (e.g. supporting substance use recovery). MobilePath uses some unique approaches and algorithms for monitoring physiology, providing clinical features for use with machine learning algorithms.
One of the reasons our group created its own digital platform is to have full control over privacy. MobilePath was designed with Privacy in mind and we place highest importance on preserving user privacy. Our software does not record the user voice, and we do not capture images or screenshots. For all of our projects, the software is used only with the full consent of the user/patient. As part of our innovation, if the user gives permission to record GPS or phone number data, all these data are scrambled and randomized so that all data remains anonymous.