Tactile spatial attention enhances gamma-band activity in somatosensory cortex and reduces low-frequency activity in parieto-occipital areas

Abstract

We investigated the effects of spatial-selective attention on oscillatory neuronal dynamics in a tactile delayed-match-to-sample task. Whole-head magnetoencephalography was recorded in healthy subjects while dot patterns were presented to their index fingers using Braille stimulators. The subjects' task was to report the reoccurrence of an initially presented sample pattern in a series of up to eight test stimuli that were presented unpredictably to their right or left index finger. Attention was cued to one side (finger) at the beginning of each trial, and subjects performed the task at the attended side, ignoring the unattended side. After stimulation, high-frequency gamma-band activity (60-95 Hz) in presumed primary somatosensory cortex (S1) was enhanced, whereas alpha- and beta-band activity were suppressed in somatosensory and occipital areas and then rebounded. Interestingly, despite the absence of any visual stimulation, we also found time-locked activation of medial occipital, presumably visual, cortex. Most relevant, spatial tactile attention enhanced stimulus-induced gamma-band activity in brain regions consistent with contralateral S1 and deepened and prolonged the stimulus induced suppression of beta- and alpha-band activity, maximal in parieto-occipital cortex. Additionally, the beta rebound over contralateral sensorimotor areas was suppressed. We hypothesize that spatial-selective attention enhances the saliency of sensory representations by synchronizing neuronal responses in early somatosensory cortex and thereby enhancing their impact on downstream areas and facilitating interareal processing. Furthermore, processing of tactile patterns also seems to recruit visual cortex and this even more so for attended compared with unattended stimuli.

Figure 1.

The spatial-selective delayed-match-to-sample task. A, The Braille patterns used. B, An example sequence is shown with the timing of sample, cue, and test presentations and of the response. Sequences could have any length between one and eight test stimuli.

Figure 2.

Planar gradients of evoked fields (for details, see Materials and Methods). Left column, Time courses of activity for left-hand stimulation in individual sensors as marked in the topographies in the middle column. Stimulus onset is at 0 s. Blue traces are for the unattended stimuli, and red traces are for the attended stimuli. Shown are z scores indicating the deviation from baseline level. Middle column, Grand-averaged topographies during epochs of interest for left- and right-hand stimulation (left and right panels, respectively). Right column, Time courses of individual sensors for stimulation of the right hand for attended (blue) and unattended (red) stimuli (z scores).

Figure 3.

A, Topography of the planar gradiometer representation of evoked fields of one example subject in the window between 350 and 600 ms after stimulus onset. B, Topography of the axial gradiometer representation of the same data. C, Correlation coefficient of all sensors with the sensor marked with a black dot in B.

Figure 4.

A, B, Time–frequency resolved power change after stimulation compared with baseline, calculated separately for unattended and attended stimuli over combined somatosensory channels as marked in the topographies to the left and to the right(C, D).C, Topography of stimulus-induced power change for left finger stimulation in the time–frequency windowasmarked in B. D, Topography of stimulus-induced power change for right finger stimulation in the same time–frequency window. All data are shown in planar gradients.

Figure 5.

Source analysis of the gamma response as an effect of stimulation versus baseline. Stimulation of the left finger resulted in activation of regions corresponding to right primary somatosensory cortex (right) and vice versa (left). L, Left; R, right.

Figure 6.

A, B, Topography of the squared M50 component, the first peak of the evoked field, 50 ms after stimulus onset. The evoked field was squared to allow better comparison with the frequency-domain results, because power is also a squared amplitude value. C, D, Axial gradiometer topographies of the gamma-band enhancement after left and right finger stimulation. Shown are z values between poststimulus and baseline periods as in Figure 4.

Figure 7.

A, B, Time–frequency resolved power change after stimulation compared with baseline, calculated separately for unattended and attended stimuli over combined somatosensory channels as marked in the topographies to the left and to the right (C–F). C, Topography of beta suppression for left finger stimulation. Averaged over the early (0.1–0.4 s) time–frequency window as marked in A. D, Topography (planar gradients) of beta rebound for left finger stimulation. Averaged over the late (0.5–0.8 s) time–frequency window as marked in A. E, F, Topographies of beta suppression and rebound, respectively, for right-hand stimulation. Same time–frequency windows as in C and D, respectively. All data are shown in planar gradients.

Figure 8.

Source analysis of beta suppression as an effect of stimulation versus baseline. Stimulation of the left finger resulted in suppression of beta-band activity in regions corresponding to right primary sensorimotor cortex (left) and vice versa (right). L, Left; R, right.

Figure 9.

Source analysis of beta-rebound as an effect of stimulation versus baseline. Stimulation of either finger resulted in bilateral rebound in sensorimotor areas with a slight dominance over the right hemisphere. L, Left; R, right.

Figure 10.

A, B, Time–frequency resolved power change after stimulation compared with baseline calculated separately for unattended and attended stimuli over combined occipital channels as marked in the topographies to the left and to the right (C, D).C,Topography of the attentional effectonalphasuppression in the marked time–frequency window. Note that this is different from the previous figures in which stimulation effects were shown. D, Topography of the attentional effect on alpha suppression in the marked time–frequency window. All data are shown in planar gradients.

Figure 11.

Source analysis of alpha suppression as an effect of stimulation versus baseline. Stimulation of either finger resulted in bilateral suppression of alpha-band activity in parieto-occipital cortex. L, Left; R, right.

Figure 12.

Left column, Time courses of the stimulation-induced power changes for different frequency bands in the group of channels (contralateral to stimulation) as indicated in Figures 4, 6, and 9 separately for right- and left-hand stimulation. Right column, Scatter plots showing individual subjects' data of stimulation-induced power changes in the respective frequency bands for unattended and attended stimuli as measured over the depicted channels and time–frequency windows (see Figs. 4, 6, 9).

Figure 13.

Source analysis of the effect of attention on gamma-band activity. Attention to the left finger resulted in enhanced gamma-band activity in right somatosensory cortex (left) and vice versa (right). L, Left; R, right.

Figure 14.

Source analysis of the effect of attention on the beta rebound. Attention to the left finger resulted in a reduced beta rebound in right sensorimotor cortex (left) and vice versa (right). L, Left; R, right.

Figure 15.

Source analysis of the effect of attention on alpha suppression. Stimulation of either finger resulted in bilateral suppression of alpha oscillations in parieto-occipital and occipital cortex.