Dissociable roles of human frontal eye fields and early visual cortex in presaccadic attention -- evidence from TMS

Shortly before each saccadic eye movement, presaccadic attention improves visual sensitivity at the saccade target1-5 at the expense of lowered sensitivity at non-target locations6-11. Some behavioral and neural correlates of presaccadic attention and covert attention -which likewise enhances sensitivity, but during fixation12-are similar13. This resemblance has led to the debatable13-18 notion that presaccadic and covert attention are functionally equivalent and rely on the same neural circuitry19-21. At a broad scale, oculomotor brain structures (e.g., FEF) are also modulated during covert attention22-24 - yet by distinct neuronal subpopulations25-28. Perceptual benefits of presaccadic attention rely on feedback from oculomotor structures to visual cortices29,30 (Fig. 1a); micro-stimulation of FEF in non-human primates affects activity in visual cortex31-34 and enhances visual sensitivity at the movement field of the stimulated neurons35-37. Similar feedback projections seem to exist in humans: FEF+ activation precedes occipital activation during saccade preparation38,39 and FEF TMS modulates activity in visual cortex40-42 and enhances perceived contrast in the contralateral hemifield40. We investigated presaccadic feedback in humans by applying TMS to frontal or visual areas during saccade preparation. By simultaneously measuring perceptual performance, we show the causal and differential roles of these brain regions in contralateral presaccadic benefits at the saccade target and costs at non-targets: Whereas rFEF+ stimulation reduced presaccadic costs throughout saccade preparation, V1/V2 stimulation reduced benefits only shortly before saccade onset. These effects provide causal evidence that presaccadic attention modulates perception through cortico-cortical feedback and further dissociate presaccadic and covert attention. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=127 SRC="FIGDIR/small/529691v2_fig1.gif" ALT="Figure 1"> View larger version (37K): org.highwire.dtl.DTLVardef@1567165org.highwire.dtl.DTLVardef@2ecec9org.highwire.dtl.DTLVardef@10a00adorg.highwire.dtl.DTLVardef@1b29dfd_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 1.C_FLOATNO Background and experimental design. (a) Feedback connections originally established in non-human primates29,30,36,90,104,105 assumed to underly perceptual correlates of human presaccadic attention. FEF: Frontal eye field, SC: Superior Colliculus; (b) Response gain effect of presaccadic attention on the contrast response function: presaccadic attention scales the response by a multiplicative gain factor, resulting in an increase of the asymptotic response dmax (the maximal response achieved at high contrast)9,48. (c) Determining occipital (V1/V2) stimulation sites (Exp.1 & 2a) and stimulus placement for both experiments via phosphene mapping: Observers were stimulated laterally around the occipital pole until they perceived a phosphene (in the contralateral visual field), then drew its outline on the screen. The center of the phosphene drawings (stimulated region) and the symmetric region (not stimulated) were used for stimulus placement in the main experiment, where we applied sub-phosphene-threshold TMS using identical coil positioning. (d) Determining rFEF+ stimulation sites (Exp.2b). rFEF+ (yellow ROI) was localized on each individual observers anatomy (two exemplary observers shown here) via a probabilistic topography atlas76 and verified by anatomical landmarks (junction of the precentral and superior frontal sulcus; light gray areas)64,65,67. (e) Presaccadic orientation discrimination task. After a fixation period, a central direction cue (black line) appeared. Observers were instructed to make a saccade to the indicated target marked by placeholder dots. Note that the saccade was equally likely directed to the stimulated and to the symmetric region / hemifield. 100ms after cue onset, a tilted test Gabor patch was presented at either the saccade target (valid; 50%) or at the opposite location (invalid), randomly intermixed; a vertical Gabor was presented at the other location. Importantly, stimuli were presented during saccade preparation, i.e., while gaze was still at the screen center. After saccade offset, a response cue (white dots on placeholder) indicated the location at which the test Gabor had appeared, and observers reported its orientation. In the neutral condition (separately blocked) the cue pointed to both placeholders and observers kept fixating (supplemental Video S1 demonstrates the trials sequence of each condition). (f) Trial timeline. Observers received double-pulse TMS (50ms inter-pulse interval) locked to stimuli onset (Exp.1) or at various times during saccade preparation (0-200ms relative to cue onset; Exp.2). Whereas grating contrast was varied to measure contrast response functions in Exp.1, contrast was fixed to 46% in Exp.2. (g) Experimental conditions. The test was equally likely presented at the stimulated region (test stimulated) or in the opposite hemifield (not stimulated). Moreover, the test was equally likely presented at the saccade target (valid) or opposite of it (invalid). Note that the fixation condition (neutral) was only tested in Exp.1. C_FIG