Skip to contentSkip to filters

We use cookies for analytics.

Predictive Coding with Cortical Cell Types and Connectivity

Hannah Choi

School of Mathematics, Georgia Institute of Technology

Recording

1 h 29 min

Abstract

The brain processes sensory information efficiently by generating an internal model of the environment that is continuously updated through prediction errors and the suppression of expected input, a process known as predictive coding. Experimental evidence supports a cortical implementation of predictive coding, but it remains unclear how the process is shaped by connectivity across the cortical hierarchy, cortical layers and diverse neuronal subtypes. This talk presents recent work investigating the cortical mechanisms underlying predictive coding.

A biologically grounded network model with realistic connectivity among cortical cell types reveals context-dependent functional connectivity that reflects how predictions and prediction errors are communicated across cortical circuits. Mapping the algorithmic components of predictive coding onto neuronal subtypes yields experimentally testable predictions about cell-type-specific responses to expected and unexpected stimuli. The work shows how different computational objectives, including prediction-error minimization, energy efficiency and reward maximization, shape cell-type-specific responses to distinct forms of expectation violation: contextual, absolute and omission novelty. By integrating a multi-objective normative framework with a biological circuit model, it offers a unified account of how predictive coding, energy efficiency and reinforcement learning map onto cell-type-specific mechanisms.

Topics