Evidence of a link between learning errors and psychotic-like experiences


In a nutshell: Poor predictive models may hinder the brain’s ability to perceive our environment correctly, potentially leading to psychotic-like experiences.

View Paper Abstract

Evidence of a link between learning errors and psychotic-like experiences

We take in sensory information – sights, sounds and smells – to understand what’s happening around us at any given time. But the human brain also uses this information to predict the future. It learns what is regular and irregular in our sensory environment. Based on this ‘statistical learning’, the brain can build predictive models.

When things are stable and predictable, these models help us to process information more efficiently. But they also help us to react more quickly and flexibly when circumstances become unpredictable or volatile.

Impaired statistical learning and prediction can affect how the brain processes sensory information. Problems with the brain’s predictive ability have been linked to psychosis. Psychotic symptoms, which include delusions and hallucinations, are a hallmark of mental health disorders such as schizophrenia.

But psychotic-like experiences are not limited to people with mental health disorders. Healthy people can have them, too.

An international team of researchers has now found a connection between impaired predictive processes and psychotic-like experiences in healthy people. The study was led by Brain Function CoE investigators Ilvana Dzafic and Marta Garrido from the University of Melbourne.

The researchers tested a group of participants using the auditory oddball task. In this task, surprising sounds are embedded in a sequence of predictable sounds. Participants are asked to judge the probability of certain sounds being played – a method of statistical learning. At the same time, their brain activity was recorded using electroencephalography (EEG).

The participants also completed a questionnaire about any psychotic-like experiences, such as hearing or seeing things that are not really there, or feeling paranoid.

The researchers found that participants with higher levels of psychotic-like experiences were worse at judging the probability of sounds.

Using the EEG data, they measured the brain’s response to unpredictable events – which reflects the difference between the predictive model and reality. Their results showed for the first time that errors in statistical learning were directly linked to less accurate predictive models.

The researchers replicated their findings in a different group of participants. This step was important, as the fields of psychology and neuroscience are in the midst of a so-called ‘replication crisis’. This crisis arose from the discovery that many scientific results could not be replicated successfully, suggesting that they may not be accurate.

Next steps:
The researchers would like to examine whether psychotic symptoms appear as a result of a failure in statistical learning. They also hope to understand what happens in the brain during statistical learning when psychotic symptoms worsen. They plan to do this in a longer-term study of a larger group of young people who have sought help for mental health disorders.

Dzafic, I., Randeniya, R., Harris, C.D., Bammel, M., & Garrido, M.I. (2020). Statistical learning and inference is impaired in the non-clinical continuum of psychosis. Journal of Neuroscience, JN-RM-0315-20. doi: 10.1523/JNEUROSCI.0315-20.2020

Republish this article:

We believe in sharing knowledge. We use a Creative Commons Attribution 4.0 International License, which allows unrestricted use of this content, subject only to appropriate attribution. So please use this article as is, or edit it to fit your purposes. Referrals, mentions and links are appreciated.