Uncertainty and expectancy deviations require cortico-subcortical cooperation.

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Learning in uncertain environments requires minimizing uncertainty by estimating probable outcomes.
• Distinguishing unexpected but valid outcomes from errors is critical for effective learning.
• Subcortical structures play a key role in probabilistic inferential learning and regulating expectations.

Purpose of the Study:

• To investigate the role of subcortical structures in probabilistic inferential learning.
• To determine how subcortical activation patterns differ based on the nature of unexpected events (low probability, ambiguous, incongruent).
• To understand the contribution of subcortical regions to cognitive control in dynamic environments.

Main Methods:

• A 7-Tesla functional magnetic resonance imaging (7T-fMRI) study was conducted.
• Participants performed a novel task involving learning to predict outcomes of visual sequences with varying probabilities.
• Analysis focused on subcortical activation patterns in response to unambiguous, ambiguous, and incongruent contextual evidence.

Main Results:

• Increased activation in the dorsomedial striatum and thalamus was observed for less probable sequences.
• Ambiguous sequences elicited greater activation in the red nuclei.
• Incongruent sequences showed distinct subcortical activation in the dorsolateral and posterior dorsomedial striatum.

Conclusions:

• Different subcortical structures are involved in regulating uncertainty and expectancy deviations.
• These findings highlight the importance of subcortical regions for learning predictions and maintaining flexible cognitive control.
• The study demonstrates a neural basis for adapting to and learning from unexpected environmental events.