Can visual information encoded in cortical columns be decoded from magnetoencephalography data in humans?

TL;DR

This study demonstrates that magnetoencephalography (MEG) can decode orientation-specific information encoded in cortical columns in humans, revealing the potential to access fine-scale neural information noninvasively.

Contribution

The paper provides empirical evidence and theoretical modeling showing that MEG signals contain decodable information at the cortical column level in humans.

Findings

MEG signals contain orientation-specific information as early as 50ms after stimulus onset.

Controlling for confounds confirms the specificity of the decoded information.

Theoretical modeling supports the empirical findings.

Abstract

It is a principal open question whether noninvasive imaging methods in humans can decode information encoded at a spatial scale as fine as the basic functional unit of cortex: cortical columns. We addressed this question in five magnetoencephalography (MEG) experiments by investigating the encoding of a columnar-level encoded visual feature: contrast edge orientation. We found that MEG signals contained orientation-specific information as early as ~50ms after stimulus onset even when controlling for confounds, such as overrepresentation of particular orientations, stimulus edge interactions, and global form-related signals. Theoretical modeling confirmed the plausibility of this empirical result. An essential consequence of our results is that information encoded in the human brain at the level of cortical columns should in general be accessible by multivariate analysis of electrophysiological signals.