Intermodal selective attention in monkeys. II: physiological mechanisms of modulation.

Of all areas studied in the accompanying study, attention effects were most consistent and well resolved in V4. In this study, to define some of the anatomical circuits and neural processes underlying the influence of attention, we examined the laminar distribution and physiology of attention effects in V4 and in two lower areas, V1 and V2. Laminar event-related potential (ERP), current source density (CSD) and multiunit activity (MUA) profiles allowed identification of processes occurring in the local ensembles, as well as their sequence and laminar distribution. These methods also permitted us to analyze the brain processes reflected in attention-sensitive components of the surface ERP. As outlined in the previous study, the first robust modulation by attention occurred in V4 during the 100-300 ms poststimulus interval. This is the time frame of the net refractoriness which follows the net local excitatory response to luminance increment. Over this interval, attention reduced CSD amplitudes and increased action potential firing rates, findings consistent with disinhibition as a mechanism for attention in V4. Similar effects were observed during the 100-300 ms time frame in V2. In V4, attention had no effect on the initial excitatory response at the depth of lamina 4, but it did produce large modulations in supragranular and deep laminae, origins of both feedforward and feedback projections. Attentional modulation in V2 was later than in V4 and concentrated in extragranular laminae, with no modulation of the initial layer 4 response. Attentional modulation in V1 was smaller and still later than that in V2 and was focused in the supragranular laminae. In this paradigm, attention did not modulate either the response in lateral geniculate nucleus (LGN) or the initial excitation in lamina 4C of V1. The timing of effects across areas and the laminar distribution of effects within areas indicate that attention effects are mediated by feedback projections. Moreover, our findings suggest that attention may increase the perceptual salience of stimuli by reducing stimulus-evoked refractoriness and/or inhibition in cortical ensembles. Finally, attentional modulation of transmembrane current flow in V4 produced a sustained negative deflection in the laminar ERP profile, that was manifested in the ERP over the occipital surface. This posits a mechanism for the 'selection negativity', a scalp ERP effect noted under similar experimental conditions in human subjects.