Normative evidence weighing and accumulation in correlated environments

Area of Science:

• Cognitive Neuroscience
• Decision Science
• Computational Neuroscience

Background:

• The brain makes decisions based on normative principles of weighing and accumulating sensory evidence over time.
• Previous research demonstrated that these principles explain adaptation to the temporal dynamics of observations.

Purpose of the Study:

• To investigate if the brain's decision-making adaptability extends to accounting for correlations between sensory observations.
• To determine if human participants can adapt to changing correlations in evidence and adjust their decision strategies accordingly.

Main Methods:

• Online human participants completed a novel visual-discrimination task.
• The task involved pairwise-correlated observations with uncued, trial-by-trial changes in correlation strength.
• Decision-making behavior was analyzed to assess the weighing and accumulation of evidence.

Main Results:

• Participants demonstrated adaptation to trial-by-trial changes in observation correlations with minimal training.
• Decisions were based on an approximately normative weighing and accumulation of evidence, reflecting the changing correlations.
• The findings indicate a robust ability to adjust evidence weighting based on observational statistics.

Conclusions:

• The brain's decision-making processes are highly adaptable, extending beyond temporal dynamics to include correlations in sensory evidence.
• This adaptability highlights the brain's capacity to dynamically adjust the weight of evidence based on statistical properties of observations.
• The study underscores the robustness of neural mechanisms for evidence integration in complex, changing environments.