Conor Feehly writes that neuroscientists Lisa Feldman Barrett (Northeastern) and Earl Miller (MIT) propose a new framework for brain categorization in which the nervous system acts as a prediction engine rather than a passive filing cabinet.
Their model, published in Nature Reviews Neuroscience, proposes categories are projected onto the world based on the body's energetic needs, with the limbic core providing predictive signals that interact with compressed sensory input to guide behavior.
- 90% of synaptic connections in the visual cortex facilitate feedback signaling rather than feedforward, suggesting the brain is structurally biased toward prediction.
- The framework views the nervous system as a bow tie with two funnels, compressing sensory and physiological signals towards the limbic core.
Their model, published in Nature Reviews Neuroscience, proposes categories are projected onto the world based on the body's energetic needs, with the limbic core providing predictive signals that interact with compressed sensory input to guide behavior.
- 90% of synaptic connections in the visual cortex facilitate feedback signaling rather than feedforward, suggesting the brain is structurally biased toward prediction.
- The framework views the nervous system as a bow tie with two funnels, compressing sensory and physiological signals towards the limbic core.
Researchers tracked hippocampal reward representation in mice learning a task, finding it shifts from encoding reward to preceding features, supporting a predictive role for the hippocampus in learning.
