Laminar fMRI and computational theories of brain function

Area of Science:

• Neuroscience
• Computational Neuroscience
• Neuroimaging

Background:

• Functional MRI (fMRI) at the resolution of cortical layers (laminar fMRI) is a recent advancement.
• Laminar fMRI provides insights into neurophysiological mechanisms and brain function theories.
• Hierarchical Bayesian theories, such as predictive coding, propose specific computational roles for cortical layers.

Purpose of the Study:

• To review predictive coding and related hierarchical Bayesian theories.
• To summarize predictions of these theories regarding layered cortical computations.
• To examine how laminar fMRI can test these predictions and discuss challenges.

Main Methods:

• Review of predictive coding and hierarchical Bayesian theories.
• Analysis of predictions for layer-specific cortical computations.
• Assessment of laminar fMRI's capability to test these predictions.

Main Results:

• Laminar fMRI offers a novel approach to investigate layer-specific brain computations.
• The technique can noninvasively test hypotheses from hierarchical Bayesian brain theories in humans.
• Methodological challenges in applying laminar fMRI need consideration.

Conclusions:

• Laminar fMRI holds significant potential for testing computational theories of brain function.
• It enables noninvasive investigation of layer-specific information processing.
• Future applications may include clinically useful computational assays for layer-specific processing.