Computer Automation of Mental Health Measurement, Diagnosis and Therapy

点击翻译按钮获取中文摘要

We have entered an era in which the knowledge and expertise of medical professionals can begin to be automated through innovations in computer software and hardware. The long-term vision of this research is the creation of an all-software replacement for medical diagnosis and treatment. Such technology would contribute to a significant reduction in suffering and the effects of poverty in the world. In the proposed research, we attempt to take the first steps of research and development on tools and interventions that aid in the solution of existing medical problems. Our strategy is to collaborate with like-minded medical professionals who are also researchers, and find ways to automate aspects of measurement, diagnosis and therapy with their expert assistance. One of the most under-served areas of medicine that could benefit the most from the `helping hand' of computer automation is mental health diagnosis and therapy. This is motivated by the grossly insufficient number of Psychiatrists in this country available to deal with the high incidence of mental health disorders. This proposal describes several projects on monitoring, diagnosis and therapy for mental health conditions. The first project explores a method of tracking a patient's mental health state, using data generated from their smartphone. The goal is to infer, from that data, the severity of a mental health disorder such as social anxiety, generalized anxiety, depression or bipolar disorder. Such tracking could enable the prevention of such episodes from worsening, and to help determine, much more quickly, the efficacy of drug or psychotherapy treatments. This project has begun, with a collaborator who is both a clinician and research Psychiatrist. We have developed a mobile application that provides many streams of data from the patient-owner of the smartphone. This work has already raised many questions on the algorithmic and clinical side, as well as issues in data privacy, that must be answered or solved to be able to infer mental health state from smartphone data. A second major project concerns addiction, which often goes hand-in-hand with mental health issues. One common addiction is tobacco smoking, and we collaborate with a nicotine addiction clinician/researcher. We propose the development of a `chatbot' (a computer that speaks to humans) that uses known-good clinical techniques to help a smoker make the decision to quit. This is motivated by the fact that 94% of all current smokers are ambivalent about quitting, and need to make this decision before they will be in any way motivated to try to quit. The full proposed research program has several other projects, including one to measure a psychometric through game play, another to help children on the autism spectrum plan structured playdates, another to help people quit smoking once they've taken the decision to do so, and a final one to automate a measurement of cognition.