Institute of Psychology
Institute of Psychology
New publication in "Communications Psychology"
13 February 2025, by Uğur Turhan
How do we learn to make good decisions in an uncertain world? This paper explores how the brain updates beliefs based on incoming information—what researchers call predictive inference. Learning well means adjusting to different types of uncertainty: sometimes we’re not sure what we’re seeing (perceptual uncertainty), sometimes we’re facing known risks, and sometimes the world changes unexpectedly.
We use models from statistics and machine learning to understand how people ideally should learn in these situations. But humans often show systematic biases—they don’t always follow the ideal. Interestingly, some of these biases may not be mistakes at all, but smart shortcuts: for example, using simple strategies that save mental effort. Others may reflect prior beliefs that don’t match reality.
By understanding these patterns, we can learn more about how the brain balances accuracy with efficiency—and how these processes might go awry in mental health conditions.
Why is this important?
This research shows that learning biases aren’t always flaws—they can reflect smart adaptations to uncertainty or limited mental resources. Understanding when biases are adaptive versus problematic can help us better support learning in everyday life and offer insights into psychiatric conditions.