Intelligent real-time therapy: Harnessing the power of machine learning to optimise the delivery of momentary cognitive-behavioural interventions

Background: Experience sampling methodology (ESM) [Csikszentmihalyi, M. & Larson, R. (1987). Validity and reliability of the experience-sampling method. Journal of Nervous and Mental Disease, 175(9), 526-536] has been used to elucidate the cognitive-behavioural mechanisms underlying the development and maintenance of complex mental disorders as well as mechanisms involved in resilience from such states. We present an argument for the development of intelligent real-time therapy (iRTT). Machine learning and reinforcement learning specifically may be used to optimise the delivery of interventions by observing and altering the timing of real-time therapies based on ongoing ESM measures. Aims: The aims of the present article are to outline the principles of iRTT and to consider how it would be applied to complex problems such as suicide prevention. Methods: Relevant literature was identified through use of PychInfo. Results: iRTT may provide an important and ecologically valid adjunct to traditional CBT, providing a means of balancing population-based data with individual data, thus addressing the knowledge-practice gap [Tarrier, N. (2010b). The cognitive and behavioral treatment of PTSD, what is known and what is known to be unknown: How not to fall into the practice gap. Clinical Psychology: Science and Practice, 17(2), 134-143] and facilitating the delivery of interventions in situ, thereby addressing the therapy-real-world gap. Conclusions: iRTT may provide a platform for the development of individualised and multifaceted momentary intervention strategies that are ecologically valid and aimed at attenuating pathological pathways to complex mental health problems and amplifying pathways associated with resilience.