Expectation vs. Reality: study maps the neurons that tell the difference

- EN - DE- FR- IT
FMI researchers mapped specific populations of neurons (green) that turn on in t
FMI researchers mapped specific populations of neurons (green) that turn on in the brain’s outer layer when mice experience a mismatch between expectations and reality. Inhibitory neurons are shown in red. Image credits: Sean O’Toole/FMI.

When our expectations differ from reality, specific sets of brain cells are activated. Working in mice, FMI researchers have characterized these neurons based on their gene-expression patterns, paving the way for a better understanding of some of the neuronal deficits associated with neuropsychiatric conditions.

Anyone who has ever opened a Danish butter cookie tin only to find a sewing kit knows it: we often face a mismatch between our expectations and reality. When this happens, specific sets of neurons located in the cortex — the brain’s outer layer — turn on. Malfunctions in these brain cells, called prediction-error neurons, may also lead to psychosis — an altered perception of reality that characterizes conditions such as schizophrenia.

Prediction-error neurons come in two flavors: positive prediction error neurons turn on when the brain detects an unexpected sensory cue, whereas negative prediction error neurons start to fire when a predicted sensory signal ceases to exist. Researchers led by Sean O’Toole, a postdoctoral fellow in the Keller lab, set out to map these neurons in the visual cortex of mice. Their hypothesis was that neurons that have different gene-expression signatures may serve different functions.

To test this hypothesis, the researchers analyzed gene expression in the mice’s prediction-error neurons as the animals performed behavioral tests. During the tests, the animals learned to navigate a virtual-reality maze while running on a treadmill. To trigger a negative prediction error mismatch, the researchers briefly stopped the visual stimulus as the animals ran on the treadmill. To elicit a positive prediction error mismatch, they presented unexpected visual stimuli to the animals.

When the mice experienced mismatches between their expectations and reality, prediction-error neurons started to fire. The researchers found that neurons that respond to a negative prediction error mismatch show different gene-expression patterns compared with neurons that respond to a positive prediction error mismatch. The team also developed a method to target these neurons based on their gene-expression signatures — work that could help other researchers better understand the function of prediction-error neurons and their role in brain conditions.

To this day, only a limited number of therapies are available for people with neuropsychiatric conditions such as schizophrenia, and most of these therapies have severe side effects. By deciphering the diversity of neuronal cell types, the researchers say, neuroscientists may be able to advance towards more effective therapeutic approaches.

Sean M. O’Toole, Hassana K. Oyibo, and Georg B. Keller Molecularly targetable cell types in mouse visual cortex have distinguishable prediction error responses Neuron (2023)

About the first author

Sean O’Toole was born in Boston, Massachusetts. After completing a PhD in Neuroscience at Brandeis University, he joined the FMI in 2017. In his free time, he enjoys exercising and playing video games. This summer, he became the proud father of a daughter.