Neurophysiology of selective attention in the macaque visual system

Abstract

I studied the neurophysiological basis of selective attention in macaque subjects by measuring the effect of attention upon visually-evoked neuronal ensemble responses in different brain visual areas. Subjects were trained to perform an intermodal behavioral task that was adapted from human event-related potential (ERP) studies. In this task, auditory and visual stimuli were presented simultaneously and in counterphase. Frequently-presented "standard" stimuli and infrequently-presented "deviant" stimuli were presented in both sensory modalities.;Laminar profiles of event-related potentials (ERP), current source density (CSD), and multiunit activity (MUA) evoked by standard visual stimuli were recorded while subjects performed the tasks. Responses evoked by standard visual stimuli were compared across conditions when the auditory versus visual discriminations were performed.;Attention effects upon neuronal responses were quantified by computing a Modulation Index (MI) of the CSD, the ratio of the difference response to the sum response with respect to responses evoked by attended versus ignored visual stimuli. The average MI showed a gradient of attention effects up levels of the visual system (LGN = .002; V1 = .028; V2 = .105; V4 = .162), showing that physiology becomes increasingly more susceptible to the effects of attention at higher levels of the visual system.;The timing of attention effects was quantified by computing MIs over restricted intervals of the stimulus-evoked response. This analysis showed that the initial neuronal responses in most areas were generally insensitive to attention, while later phases of the response modulated to a much greater degree.;The largest and most consistent attention effects were recorded in area V4. The net transmembrane current flow difference in V4 was computed by computing a direction index (DI) for each site. An average DI of -0.09 +/- .02 in V4 over the 100--300ms poststimulus interval indicates that attention reduced the net transmembrane current flow evoked by visual stimuli during that interval, which, in turn, suggests a reduced net postsynaptic potential (PSP) amplitude due to attention.;In summary, (1) Attention has progressively greater effects upon neuronal responses at higher levels of the visual system from the LGN to V4; (2) Ventral stream visual areas modulate stronger than dorsal stream areas; (3) The temporal pattern of the stimulus-evoked response may be characterized by an automatic, attention-independent stage, followed by a controlled, attention-dependent stage; (4) The largest attentional modulation occurs in intermediate and higher areas of the ventral stream (V4 and IT); (5) The timing and laminar distribution of attention effects across visual areas suggest that attention effects in lower visual areas are mediated by feedback from higher visual areas; and (6) In areas that show the largest amplitude attention effects, physiology suggests that attention disinhibits neuronal responses. (Abstract shortened by UMI.).