Experimental Design, Task, Stimuli, and Stimulation Site
(A) Experimental design and task. Rhythmic TMS was applied in short bursts of five pulses at one of two frequencies (theta band, 5 Hz; beta band, 20 Hz) on each trial in the main experiment, in random order across trials, with 10 s intervening between successive bursts. Onset of a global/local hierarchical visual stimulus, centered at fixation, coincided with the last TMS pulse of each burst so that the critical visual display closely followed the rhythmic TMS bursts (future studies might vary this timing to examine the temporal profile of our effects). A sham TMS condition was also conducted (coil tilted at 90° over the same parietal site), in separate blocks that were randomly intermingled with active-TMS blocks.
(B) Stimulation site for one representative participant. TMS was applied over a right-hemisphere intraparietal sulcus site in our main experiment, determined by neuronavigation with Brainsight and individual anatomical MRI scans, at Tailarach coordinates 28, −51, 50 (see and ).
(C) Examples of stimuli for the global/local target blocks. In the global target blocks, observers were asked to detect the presence (versus absence) of the global letter H (versus S or D). The local distractors were all Hs, or all Ss or Ds, independent of global identity, leading to equiprobable congruent and incongruent conditions. For blurred stimuli, the global letter was more salient than the local letters; the reverse was true for nonblurred stimuli. In other blocks of trials, the same stimuli were used, but the local level was judged instead. For both tasks, salient incongruent distractors interfered the most (see and ).

Effects of Rhythmic TMS on Performance for Global or Local Target Identification
Sham-normalized (active − sham TMS at each frequency) effects of rhythmic right-parietal TMS bursts at 5 Hz (light gray bars) or 20 Hz (dark gray bars) in the global task (right) or the local task (left). Data are shown separately for distractor-salient and target-salient conditions; these show a similar pattern that is stronger with salient distractors (top). The y axis plots mean differences (for active − sham) in inverse efficiency (±standard error of the mean), so that negative values correspond to improved performance with active TMS and positive values correspond to impaired performance with active TMS. Asterisks indicate significant differences on t tests, either from the null hypothesis of no difference between active and sham TMS or between pairs of conditions as bracketed. ^∗p < 0.05; ^∗∗p < 0.01. See also , , and .