Why We Don't Learn from our Mistakes
Olav Krigolson
University of Victoria

We have all been told that we learn from our mistakes – but as it turns out that is simply not true. Instead, a growing body of evidence suggests that we actually learn when expectations differ from outcomes. Typically, learning theorists refer to the deviation between expectancies and outcomes as prediction errors. Over the past 20 years a multitude of neuroimaging studies support theoretical accounts of learning – neural learning signals within the human brain appear to reflect prediction errors and do not encode rewards and punishments in an absolute sense. Here, I will review work from my laboratory that has showed how electroencephalography can be used to track human learning, I will provide a unifying framework for a learning system within the medial-frontal cortex, and I will also review some of our more recent work outlining factors that do impact the learning system within our brain.