Fig 1.
Procedure and trial design of the different tasks that the participants conducted in the scanner.
(a) Shows the four runs that composed the procedure. The first three runs (i.e., Pre, Speaker Familiarization, and Post) corresponded to the experiment while the fourth run corresponded to the Independent Functional Localizer for voice and speech-sensitive regions. (b) Depicts the trial structure of the speech perception task that participants conducted in the Pre and Post runs. Each block was composed of four trials. Trials followed an ABX design; participants had to respond if the third stimulus of each trial was a repetition of either the first or second stimulus presented during that trial. (c) Portrays the run in which participants were familiarized with one of the speakers. It began with the presentation of a recording of one of the speakers featured in the speech perception task. This recording was a first-person narration which contained autobiographical information of a fictional character. When this recording finalized, five statements concerning the identity of the fictional character were presented. Participants had to judge whether these affirmations were true or false. Note that, despite the text of the Familiarization appearing in the English language in the figure, the Familiarization was conducted in the native language of the participants, Spanish. (d) Portrays the three conditions that composed the Independent Functional Localizer for voice and speech-sensitive regions: spectrally rotated speech, single-speaker, and multi-speaker. During the Independent Functional Localizer participants were instructed to press a button when a pure tone was presented.
Table 1.
Performance in the speech perception task.
Table 2.
Functional connectivity results of the seed-to-whole brain analyses.
Fig 2.
Functional connectivity results of the interaction contrast.
(a) Speech perception from a familiar speaker led to a change in functional connectivity strength between the independently localized Voice-Sensitive Region (VSR), employed as seed region, and the right Superior Frontal Gyrus (rSFG; t = 5.48; p = .004 FDR-corrected). Glass brain shows the approximate location of the voxel that exhibited the strongest change in functional connectivity strength of the rSFG and VSR. Functional connectivity results were considered significant at a voxel-height threshold of p < .001 and cluster-wise threshold of p < .05 FDR-corrected. A = anterior. R = right, P = posterior, L = left. (b) Bar-plots (AU: Arbitrary Units) display the functional connectivity strength between the VSR and rSFG in each condition. Participant-specific values for each condition are represented by the scatterplots overlayed with the respective conditions. Values plotted in bar plots were extracted with the REX toolbox [62] as implemented in the CONN toolbox.
Table 3.
Activity results at the whole brain level.
Fig 3.
Activity results at the whole brain level.
(a) The BOLD activity recorded from the right Supramarginal Gyrus (rSMG) exhibited a significant interaction for speaker familiarity and run (t = 6.90; p = .004 FWE corrected at the whole brain level). (b) Bar plots show the parameter estimates obtained from the peak voxel in each condition. Scatterplots overlayed with the respective conditions show participant-specific values. At the group level, speaker familiarity led to BOLD activity in this region increasing 0.17 AU in Post relative to Pre (i.e., Post/Fam – Pre/Fam)”.
Fig 4.
Activity results of the interaction contrast in the independently localized regions of interest.
Small volume correction of the interaction contrast revealed significant interactions in both independently localized regions of interest: the Voice-Sensitive Region (VSR), located in the right temporal pole (a), and in the Speech-Sensitive Region (SSR), located in the left posterior superior temporal sulcus (b). Both results were significant (VSR: t = 4.73, p = .001, FWE corrected; SSR: t = 5.02; p = .001, FWE corrected) at p < .05 FWE corrected at the peak level for the ROI and Bonferroni corrected for the two ROIs (i.e., p < .025 FWE corrected). Bar plots (c) show the parameter estimate obtained from the peak voxel of each ROI in each condition at the group level. Scatterplots overlayed with the bar plots show participant-specific values for the respective conditions. At the group level, speaker familiarity (i.e., Post/Fam – Pre/Fam) increased the BOLD activity in the VSR and the SSR by .02 and .04 AU, respectively. (d) Exploratory analysis revealed an association between the accuracy participants attained in the speech recognition task (RAU: Rationalized Arcsine Units) and the interaction contrast parameter estimate extracted from the peal voxel of the SSR. This association was marginally significant after Bonferroni correction for the two ROIs (t = 3.36; p = .025 FWE corrected).