Fig 1.
Experimental design with stimulation sites, regions of interest.
(A) Experimental design: sham controlled crossover within-subject design with three sessions consisting of cTBS over left (1) aIFG or (2) pMTG or (3) sham stimulation over aIFG or pMTG followed by resting-state functional magnetic resonance imaging and a semantic task. (B) Regions which have been shown to be associated with semantic control processes in a prior study [21] are employed as cTBS stimulation sites (left aIFG and pMTG) and regions of interest (left aIFG, aMTG and pMTG) for resting-state functional connectivity analysis. Anatomical labels according to AAL2-Toolbox for SPM12. Numbers refer to coordinates in MNI space. Brain renderings show the left hemisphere.
Fig 2.
(A) Mean reaction times increase with semantic control demands but do not differ between stimulation sites. (B) Error rates are highest for anomalous sentence endings without significant influence of stimulation site (n = 19, *p < .05, Bonferroni corrected).
Fig 3.
Whole brain resting-state functional connectivity of aIFG, aMTG and pMTG.
Resting-state functional connectivity of left (A) aIFG, (B) pMTG, (C) aMTG and (D) their overlap of 19 healthy subjects after sham continuous theta burst stimulation. Panels (A)-(C) show a left-lateralized but rather unspecific pattern, while (D) reveals a specific left-lateralized fronto-temporal network corresponding to known semantic nodes. Renderings are based on thresholded T-maps with p(FWE) < .05 on the cluster-level and a cluster-forming threshold of p < .001 (uncorrected). Black circles indicate regions of interest.
Fig 4.
ROI-to-ROI functional connectivity and its correlation with reaction times.
(A) ROI-to-ROI resting-state functional connectivity of 19 healthy subjects after sham continuous theta burst stimulation reveal significant connectivity between all pairs of ROIs (*p < .05, Bonferroni corrected). (B) Linear correlation analysis reveals higher resting-state functional connectivity (after sham cTBS) between aIFG and pMTG (shown in red in panel A) to be associated with faster normalized reaction times for unexpected sentences (n = 18, p < .05, Bonferroni corrected). Solid line matches best linear fit with 95% confidence interval indicated by dashed lines.
Table 1.
Correlations between functional connectivity and behavior.