Attentional selection and communication through coherence: Scope and limitations

Synchronous neural oscillations are strongly associated with a variety of perceptual, cognitive, and behavioural processes. It has been proposed that the role of the synchronous oscillations in these processes is to facilitate information transmission between brain areas, the ‘communication through coherence,’ or CTC hypothesis. The details of how this mechanism would work, however, and its causal status, are still unclear. Here we investigate computationally a proposed mechanism for selective attention that directly implicates the CTC as causal. The mechanism involves alpha band (about 10 Hz) oscillations, originating in the pulvinar nucleus of the thalamus, being sent to communicating cortical areas, organizing gamma (about 40 Hz) oscillations there, and thus facilitating phase coherence and communication between them. This is proposed to happen contingent on control signals sent from higher-level cortical areas to the thalamic reticular nucleus, which controls the alpha oscillations sent to cortex by the pulvinar. We studied the scope of this mechanism in parameter space, and limitations implied by this scope, using a computational implementation of our conceptual model. Our results indicate that, although the CTC-based mechanism can account for some effects of top-down and bottom-up attentional selection, its limitations indicate that an alternative mechanism, in which oscillatory coherence is caused by communication between brain areas rather than being a causal factor for it, might operate in addition to, or even instead of, the CTC mechanism.Author summary: The ability to select some stimulus or stimulus location from all of those available to our senses is critical to our ability to navigate our complex biological and social niche, and to organize appropriate actions to accomplish our goals, namely to survive and to reproduce. We study here a possible mechanism by which the brain could accomplish attentional selection. We show that the interaction between 10 Hz and 40 Hz neural oscillations can provide increases in coherence and information transmission between computationally modelled cortical areas. Our results support the idea that synchronization between oscillations facilitates communication between cortical areas. Limitations and sensitivities of the model, however, also point to the idea that the relation between coherence and communication could go in the opposite direction in some cases, particularly when selective attention is not involved.