Flexible compression of sensory evidence makes human decisions more resilient to noisy neural computations
Categorizing ambiguous stimuli requires accumulating sensory evidence over time through noisy neural computations. Theories disagree on whether this evidence is accumulated in its native sensory space or in an abstract and compressed category space that retains only decision-relevant information. Here we designed a visual categorization task in which revealing decision categories before or after…
When confronted with ambiguous stimuli, the brain must accumulate sensory evidence over time via noisy neural computations to make a categorization decision. Scientists have debated whether this evidence is stored in the original sensory space or a more abstract, compressed category space that retains only the most relevant information for the decision-making process.
To investigate this, researchers developed a visual categorization task that allowed them to reveal the decision categories before or after a sequence of stimuli. This manipulation determined which space the brain used to accumulate evidence without altering the optimal decision-making strategy. Human participants demonstrated increased accuracy and reduced noise in their inferences when they could accumulate evidence in the category space.
This finding was confirmed by measuring brain activity with magnetoencephalography (MEG) recordings, which showed flexible representations of evidence compressed onto decision categories during these conditions. Furthermore, recurrent neural networks (RNNs) trained on the same task adopted the same strategy, making their decisions more resilient to noisy computations.
These findings suggest that the human brain conserves computational resources by focusing on decision-relevant information, mitigating the cost of noisy inferences.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
