Talk title: Simultaneous learning of noise and volatility and its association with symptoms of psychiatric illness

Abstract

Altered learning under uncertainty is proposed to underpin symptoms of psychiatric illness, including psychosis, depression, and anxiety. Previous studies have reported changes in two types of uncertainty – noise and volatility – in people with psychiatric illness, with changes in volatility believed to be most important in psychosis. However, most experiments do not require participants to estimate both types of uncertainty simultaneously. As a result, apparent variation in volatility processing might reflect misidentified abnormalities in processing noise, and vice versa. Here, we used a new online gamified task to test simultaneous learning about noise and volatility in two general population samples. Participants had to catch space junk to repair their spaceship and go back to Earth. They learned where the space junk would appear by trial-and-error. Successful performance required them to arbitrate whether trial-wise variation in the space junk location was the result of random fluctuations (i.e., noise), or whether there had been an unexpected change in the task’s outcome contingencies (i.e., volatility). Model-free analyses showed that psychotic-like and anxious symptoms were linked to worse estimation of both noise and volatility, while depressive symptoms were associated with improved noise estimation. Computational modelling revealed more detailed alterations; showing that individuals differed in how they used their noise belief in different situations. Specifically, people with high psychotic-like symptoms failed to reduce their reliance on noise when the environment became less noisy and more volatile. This could affect their ability to update their beliefs when the environment changes. These findings suggest that we can measure the complex nature of uncertainty learning, and that this may help us get a better understanding of the exact nature of observed deficits in people with psychiatric symptoms and illnesses.