Why does depression lead to learning disorders?

These are the findings of a study conducted over several months by a team led by DZPG neuroscientist Prof. Dr. med. Markus Ullsperger from the Institute of Psychology at the Otto-von-Guericke University Magdeburg together with colleagues from the University Clinic for Psychiatry and Psychotherapy and the German Center for Mental Health.

Using brain wave measurements (EEG) and complex mathematical computer modeling, the research team discovered that learning deficits in depressive and schizophrenic patients are caused by reduced flexibility in the use of new information.

The study has just been published in the internationally renowned journal Brain under the title "Transdiagnostic inflexible learning dynamics explain deficits in depression and schizophrenia".

"People suffering from depression or schizophrenia often suffer from cognitive impairments," says the study's first author, Dr. Hans Kirschner. Among other things, they find it difficult to understand complex information, learn, plan, or generalize a situation. "Above all, deficits in using feedback from the past to control future behavior are a core problem for those affected." These cognitive limitations are very distressing for patients and have a strong influence on the outcome of therapy, adds neuropsychologist and psychotherapist Dr. Tilmann Klein. "If we can better understand these deficits and their causes, we will be able to design therapies such as functional training in a more specific and targeted manner in the long term."

In order to find out whether the psychological and neuronal mechanisms that lead to cognitive impairments are the same in different mental illnesses, the scientists examined both patients with a diagnosis of major depressive disorder or schizophrenia and a control group of 33 people.

Experiments with animal images

The subjects were repeatedly shown pictures of animals on a screen that were associated with either a high or low probability of reward or punishment, i.e. positive or negative feedback. The test subjects had to decide whether they wanted to bet on the animal or not and thus either win or lose 10 points. If they didn't bet, they didn't win or lose anything. But then they saw whether they would have won or lost if they had bet. "During the rounds, the subjects had to decide whether it was worth betting and risking a loss, or whether it was better not to bet and lose nothing," explains Dr. Kirschner.

"The procedure can be compared to a game of roulette," explains the neuroscientist. "If you bet, you win or lose. But if you don't bet, you can still see where the little ball lands and figure out what would have happened if you had bet. The difference in our study is that the subjects were actually able to learn, because over time you eventually learn whether an animal is rewarded or punished on average, and then you can either always bet on that animal and thus maximize your profits or minimize your losses."

Optimal learning in this task would therefore mean that the test subjects pay more attention to the feedback - i.e. the gains or losses of an animal - at the beginning of the learning process, says Kirschner. "After they get a feeling for the probability of an animal winning, they ignore misleading feedback, for example that a picture that usually has a high probability of losing also wins from time to time."

While healthy subjects behaved in the same way, groups of patients suffering from depression or schizophrenia were more influenced by random errors. "Imagine a basketball player shooting at a basket," explains neuroscientist Dr. Kirschner. "The bad player rarely scores and would not make the team. The good player scores often, but not always, but you would still put him on the team. However, both patient groups in the study would also replace the good player immediately after a bad shot". The EEG showed that both patient groups had a reduced neural representation of reward expectation. "This means that the scoring rate of a good basketball player is stored weaker in the brain and is overwritten more quickly when the player does not score.”

Dr. Hans Kirschner summarizes the results of the study by saying that it greatly expands our knowledge of cognitive impairments in patients with a diagnosis of schizophrenia or depression. "In particular, we have also been able to demonstrate the benefits of computer models in which we attempt to describe complex learning mechanisms mathematically and implement them as computer simulations." In this way, learning behavior that is difficult to predict can be simulated and compared with the behavior of test subjects in specific tasks. "With this approach, we will be able to quantify and characterize learning deficits in a more differentiated way in the future. And a better understanding of these deficits will in turn allow us to further develop current therapies for depression and schizophrenia in a targeted manner. We hope that our research will benefit many patients in the future and help them to cope better in everyday life."

Original publication: Transdiagnostic inflexible learning dynamics explain deficits in depression and schizophrenia. Kirschner, H., Nassar, M.R., Fischer, A.G., Frodl, T., Meyer-Lotz, G., Froböse, S., Seidenbecher, S., Klein, T.A., Ullsperger, M. Brain. 2024 Jan;147(1): 201–214.

Source: DZPG