Methodological pipeline.

Thirty participants answered 7 given questions (60 seconds each; speaking condition) as well as listened to audio-recordings of their own voice from previous sessions (listening condition) while MEG data were recorded. Artefacts were removed from the recorded MEG data (Abbasi et al., 2021). Individual MRIs were used to estimate source models per participant which were interpolated to a template volumetric grid. Relevant areas in speech production and perception networks were identified from Neurosynth.org platform. Fourteen corresponding anatomical parcels in HCP and AAL atlases were identified: L-FOP (1), R-PEF+6v (2), L-SCEF (3), R-SCEF (4), L-3b (5), R-3b (6), L-STS (7), L-TPOJ1 (8), L-A5 (9), R-A5 (10), L-TH (11), R-CB-Cruss2 (12), L-CB-6 (13), R-CB-6 (14). For each identified parcel, estimated source time-series were extracted. Next, using a blockwise approach, we considered the first three SVD components of each parcel as a block and estimated the connectivity between each pair of parcels using a multivariate nonparametric Granger causality approach (mGC; (Schaum et al., 2021). In this study, the connectivity results are presented using connectogram plots. In the connectograms, nodes represent the brain areas and edges represent the strength and direction of the connections between them. The thickness of the edges indicates the magnitude of the t-values, while color indicates the directionality of the connectivity. In other words, when node A connects to node B, the edge will have the same color as node A, and vice versa when node B connects to node A. Note that only significant connections are shown in the connectograms (p < .05, cluster correction). For instance in the illustrated connectogram, the purple edge between L-CB6 and R-PEF shows significant connectivity from L-CB6 to R-PEF. R=Right, L=Left.

Connectivity analysis between the sensory-motor, cerebellum, and STG.

Upper panels illustrate the strength for L-SCEF, L-A5, and R-CB6 nodes in the speaking condition. In the middle part, connectograms illustrate connections between the left sensory-motor area (left plots) and the right cerebellum (right plots) during speaking to other cortical and subcortical parcels at low frequencies (top: 7-20 Hz) and high frequencies (bottom: 60 -90 Hz). Lower panels illustrate spectrally-resolved DAI from L-3b (left spectra), LSCEF (middle spectra), and R-CB6 (right spectra) to L-A5. The results represented in the whole figure are significant connectivity patterns that passed a cluster-based permutation test (p < .05, cluster correction).

Connectivity analysis.

Significant connectivity between 14 ROIs involved in speech production and perception. A cluster-based permutation test was used to detect significant connectivity patterns. The results of statistical analysis revealed significant connectivities between different brain areas during speaking (first column of connectograms), listening (second column of connectograms), and the comparison between speaking and listening conditions (third column of connectograms) across various frequency bands. In the connectograms, nodes represent the brain areas and edges represent the strength and direction of the connections between them. The thickness of the edges indicates the magnitude of the t-values, while color indicates the directionality of the connectivity. In other words, when node A connects to node B, the edge will have the same color as node A, and vice versa when node B connects to node A. Note that only significant connections are shown (p < .05, cluster correction).

Connectivity results.

Spectrally-resolved DAI between all pairs of ROIs. A cluster-based permutation test was used to detect significant connectivity patterns. The results of statistical analysis revealed significant connectivities between different brain areas during speaking (top), listening (middle), and the comparison between speaking and listening conditions (bottom). The directionality is from y-axis to x-axis. X-axis represents frequency (Hz) and y-axis represents DAI values. Note that significant values are highlighted with increased line width (p < .05, cluster correction).

Correlation between speech-brain coupling and granger causality.

Speech-STG coupling in theta range (positively lagged: 130ms) is negatively correlated with the directional connectivity from the right cerebellum lobule VI to the left temporal areas in speaking condition in theta frequency band.