Studying the role of human parietal cortex in visuospatial attention with concurrent TMS-fMRI

Abstract

Combining transcranial magnetic stimulation (TMS) with concurrent functional magnetic resonance imaging (fMRI) allows study of how local brain stimulation may causally affect activity in remote brain regions. Here, we applied bursts of high- or low-intensity TMS over right posterior parietal cortex, during a task requiring sustained covert visuospatial attention to either the left or right hemifield, or in a neutral control condition, while recording blood oxygenation-level-dependent signal with a posterior MR surface coil. As expected, the active attention conditions activated components of the well-described "attention network," as compared with the neutral baseline. Also as expected, when comparing left minus right attention, or vice versa, contralateral occipital visual cortex was activated. The critical new finding was that the impact of high- minus low-intensity parietal TMS upon these visual regions depended on the currently attended side. High- minus low-intensity parietal TMS increased the difference between contralateral versus ipsilateral attention in right extrastriate visual cortex. A related albeit less pronounced pattern was found for left extrastriate visual cortex. Our results confirm that right human parietal cortex can exert attention-dependent influences on occipital visual cortex and provide a proof of concept for the use of concurrent TMS-fMRI in studying how remote influences can vary in a purely top-down manner with attentional demands.

Figure 1.

Schematic illustration and timeline of the sequence of events starting from the beginning of a block. Empty square placeholders were present in the left and the right upper quadrants (each square subtending 7° visual angle, centered at an eccentricity of 8.5° vertically and 6° horizontally) throughout the experiment in order to denote the locations where bilateral checkerboards could appear (8 × 8 checks each, 0.875° visual angle per check, with 2, 3, or 4 checks marked in red pseudorandomly for each checkerboard). Note along the timeline how functional image acquisition was interleaved with presentation of TMS bursts and bilateral visual displays, with the TMS bursts and bilateral displays coinciding.

Figure 2.

(A) Group statistical parametric T-maps shown for the contrast of active attention (red-deviant target-counting task initially considered regardless of whether this was required for the left or right visual field) versus the “neutral” attention condition (that required only a simple button press whenever bilateral visual displays appeared). The SPM for this contrast is superimposed onto the segmented and rendered brain (with the cerebellum removed plus anterior brain regions missing that were not covered by the posterior MR surface coil) of one subject (thresholded at P <0.05 FWE corrected; see Table 1 and main text for peak coordinates of activations). Wide expanses of visual cortex were modulated by the attentional task, despite equivalent visual stimuli being presented in all conditions, as was superior parietal cortex also within the scanned volume. The L/R labels in the figure identify the left versus right hemispheres, as also for panels (B) and (C). Please note that in the rotated views that reveal ventral cortex in the lower panels, posterior cortex appears upper. Anterior regions beyond the scanned volume have been removed. (B) displays the group statistical T-map of the contrast for attending left minus right, superimposed onto the rendered brain of one subject; (C) shows the reverse contrast, that is, attending right minus attending left. Thresholded at P <0.05 FWE corrected; see also Table 1 and main text. Thus, wide expanses of occipital visual cortex showed higher BOLD signal for contralateral than ipsilateral covert visual attention, as expected.

Figure 3.

(A, B) Interaction of high versus low TMS intensity with side of attention (left minus right), small volume corrected (at P <0.05 FWE) for the orthogonal contrast of attend left minus right (cf. Fig. 2B), projected onto the mean (A) coronal and (B) transversal structural scan of the subjects, with structural regions beyond the imaged functional volume (recall that a posterior surface MR coil was used for fMRI) shown in lower contrast. The interaction reveals that some regions in right occipital–temporal cortex that had displayed activation by contralateral attention (vs. ipsilateral, cf. Fig. 2B) showed a stronger version of this differential attentional modulation under high than low intensity TMS over right PPC. (C) plots the extracted data in percent signal change (relative to the session mean, which corresponds to “zero” along the y-axis) from the cluster, now including for neutral attention conditions also (though these did not contribute to the tested interaction). Note that high versus low TMS had no impact whatsoever under neutral attention but that high TMS specifically increased the differential effect of contralateral minus ipsilateral attention during the peripheral attention conditions. These results thus demonstrate a remote effect of right parietal TMS on visual cortex that is highly dependent on current attentional state. (D, E) Interaction of high versus low TMS intensity with side of attention (now right minus left), small volume corrected (P <0.05 FWE) for regions that also showed an effect of contralateral versus ipsilateral (here, right minus left) attention in that orthogonal contrast (cf. Fig. 2C). The SPM for this contrast is overlaid on the (D) coronal and (E) transversal slice of the averaged structural scan (with lower contrast for anterior regions beyond the volume imaged for fMRI). Thus, some regions in left occipital–temporal cortex that showed activation by contralateral attention (vs. ipsilateral) also showed enhancement of this differential attentional modulation under high- versus low-intensity TMS over right parietal cortex. (F) plots the extracted data in percent signal change (relative to the session mean) from the left occipitotemporal cluster, now including the neutral attention conditions also (although these did not contribute to the tested interaction). Note that high versus low PPC TMS had no impact whatsoever under neutral attention but that it increased the differential effect of contralateral minus ipsilateral attention during the peripheral attention conditions; see main text for discussion.