The predictive power of the brain
A new study from Birkbeck’s Department of Psychological Sciences looks at how ‘predictions’ about the outcomes of actions influence sensory processing in the brain.
It is thought that the brain monitors our actions – for example, pressing a doorbell – by sending predictions from action systems to areas involved in perceiving what we are doing, and that these predictions change the way the input – like the sound of the bell – is processed.
New research from Birkbeck’s Action Laboratory, within the Department of Psychological Sciences, investigates what is actually happening in our brains when predicting the outcomes of our physical actions. Led by Dr Clare Press and Dr Daniel Yon, the study examined the brain activity of volunteers through MRI scanning while they performed finger movements and watched similar movements being performed by an avatar hand.
They were looking to understand how predictions influence sensory processing in the brain. They trained pattern classifiers to determine which outcome observers were looking at based on visual brain activity. These classifiers were found to do a better job of determining the observed outcomes when the outcomes were in line with expectations about what would happen, based on the actions participants performed themselves. This better performance was found to result from a reshaping of patterns in the visual brain, where unexpected sensory activity was suppressed. This suggests that predictions during action help visual brain areas to develop better sensory models of the actions we are performing, which will help us to perceive our environment accurately.
Dr Yon and Dr Press explain: “People have generally thought that we use action predictions to silence expected sensory signals, leading to better perception of unexpected events. This was thought to provide an explanation for why we can’t tickle ourselves. However, our study suggests instead that we use predictions during action to improve perceptual processing of what we expect - which will generally lead to a more veridical picture of the outside world, since what we expect is more likely to be true. For example, if you are reaching for a cup of coffee in a dark kitchen before sunrise your perceptual experience will be more accurate if you perceive what is likely (your hand moving towards the cup) rather than what isn’t (for example, a snake slithering across the table).”
“More generally, our study fits with a growing appreciation that the brain doesn’t just receive information from the outside world, but is involved in dynamically predicting what will happen next. These processes are essential to shaping our perceptual experience when the sensory world is ambiguous and rapidly changing.”
The research is published in Nature Communications and was funded by the Leverhulme Trust.
Further Information
- Action Laboratory
- Department of Psychological Sciences
- Dr Clare Press
- Dr Daniel Yon
- Courses in Psychological Sciences