Mean time course of the BOLD signal was calculated by averaging the responses to each condition across the four repetitions. We then computed the half-maximum decay time as the time lag from the block onset

to the time when the activation reached half of the peak value (we used linear interpolation to extract this time point, because in most cases the response reached half the maximum in Y-27632 in vitro between samples). Results Participants performed the auditory detection task easily and with high accuracy (>90%) providing confirmation of attention maintenance throughout the experiment. In order to compare the efficacy of the two baselines (SCN, Reversed), we first Inhibitors,research,lifescience,medical calculated the likelihood of detecting significant activation in the language network per individual using each baseline. The identification rate of core regions of the speech processing network (LIFG, bilateral pSTS, bilateral aSTS) was significantly higher in the Speech versus SCN contrast (93%) than in the Speech versus Reversed contrast

(55%) (χ2 (1,59) = 20.58; Inhibitors,research,lifescience,medical p < 0.0001) (see Table 1). Table 1 Identification rate Inhibitors,research,lifescience,medical of core speech processing regions The above analysis considers each ROI as an all-or-none value (activation passes the threshold or not). To further quantify the difference between the two baselines, we compared the mean cluster size for each contrast across all anatomical locations (Fig. 2). An analysis of variance (ANOVA) produced a significant main effect of baseline condition (F (1,11) = 63.8; p < 0.001), with larger clusters elicited by the Speech versus SCN contrast compared with Inhibitors,research,lifescience,medical the Speech versus Reversed contrast (mean volumes: 452

mm3 and 101 mm3, respectively). Post hoc t-tests confirmed that the Speech versus SCN contrast elicited larger clusters of activation in each region (p < 0.001). We also observed Inhibitors,research,lifescience,medical a significant main effect of ROI location (F (4,44) = 5.3; p < 0.002), reflecting larger clusters in bilateral posterior and anterior STS compared with LIFG across baselines. Finally, we observed a significant interaction between baseline condition and ROI location (F (4,44) = 4.2; p < 0.006), revealing a more pronounced cluster size difference between the Speech versus SCN contrast and the Speech versus Reversed contrast in bilateral pSTS regions. Figure 2 Cluster size comparison. Clusters were defined by contrasting Speech versus Reversed (dark gray) and Speech versus SCN (light gray), within the almost anatomical boundaries of Left IFG, bilateral pSTS, and bilateral aSTS (all defined individually at a threshold … Next we calculated, for each subject, the overlay between speech and each of the baseline conditions, as well as contrast maps that directly compare the spatial distribution of signals using each baseline condition. Figure 3 shows such overlay maps in four individual participants centered on bilateral pSTS. These representative maps demonstrate best the overall findings.