Quantitative Aversive Cognitive Testing (QACT): a new toolkit for digital healthcare

We never stop learning when we suffer, but we all learn in unique ways. Adaptive learning helps us cope with adversities, whereas maladaptive learning could hinder coping and lead to more suffering. Indeed, maladaptive aversive learning is thought to be associated with both mental health disorders and chronic pain. However, we currently lack tools to quantify aversive learning at scale. We have developed a suite of game-like cognitive tests which helps us tease apart individual differences in aversive learning, which can be performed by anyone, in the comfort of their home, using a touch based device. These games consist of an armed bandit task, an aversive generalisation task and a motivated motor control task. Each game lasts ~10 minutes. Researchers can directly use these tasks to collect their own data as well as build their own tasks. Furthermore, we have developed a set of data simulation and (hierarchical bayesian) model fitting tools, which, in addition to verifying model assumptions, allows us to fit existing and new models to the collected behavioural data. We provide complementary tools to visualise and compare the fitted parameter distributions between healthy and patient populations. Having piloted these tools on data from anxiety and chronic pain patients, a more extensive data collection is underway. Project link for further details: https://aversionscience.org