Feature specific neuronal oscillations in cortical layers

The particular role of cortical oscillations has been a long-debated topic that resulted in a variety of theoretical frameworks. Oscillatory activity in the α band has been associated with sensory processing, attention as well as other cognitive functions, while γ band oscillations is thought to be related to stimulus feature processing. Current theoretical frameworks rely on the separation of the cortical architecture into layers. Recently, methodological advancements have allowed to test layer specific frameworks on the role of oscillations in cortical computations in healthy human participants. Using EEG-fMRI, we have investigated for the first time both, stimulus feature specificity (line orientation) and the relationship between the laminar BOLD activity and α and γ band oscillations. We find γ oscillations to be positively correlated with feature-specific signals in superficial layers as predicted by the literature, but we found a deep layer contribution as well. Furthermore we found a layer (and frequency) dissociation within the α band for general, feature unspecific, processes and a feature related process. The power of the α -band correlated negatively with feature unspecific neural activity in all cortical layers. We further found that high frequency α oscillations were specifically related to stimulus feature specific BOLD signal in deep and superficial layers. More interestingly, we also observed a general modulation effect for negative BOLD signal deflections in line with the inhibitory role of α during visual attention in superficial layers. Those findings support the association of γ band oscillations with visual feature processing and further point towards the involvement of multiple α oscillations in more general and feature related processes.