Fifty people (Mean = 26.92 year, SD = 6.23) with moderate to severe stuttering met the inclusion criteria and were enrolled in the study. No drop-out occurred and all participants completed all intervention and post-assessment sessions. In order to determine inter-rater reliability, two experienced raters independently counted the SS% of all speech samples and Cohen's kappa was calculated. The strong intraclass correlation coefficient (ICC) for the inter-rater agreement (ICC = 0.95, p < 0.05) indicate high rating reliability.
Demographic and baseline characteristics of the participants are reported in Table 1. Group differences of quantitative demographic variables were explored by Student´s t-tests (for independent samples). The results showed no significant differences between the respective variables at baseline between the two groups (p > 0.05 for all variables). Additionally, the standardized mean difference (SMD) was calculated to investigate the mean differences of primary and secondary outcomes between the two groups.
Comparison of the primary outcome measure SS% for the baselines and immediately, one week, and six weeks after the treatment resulted in a significant main effect of group (F(1,49) = 65.95, p = 0.001, =0.85 ; Fig. 1). The results also showed that the interaction between time and group was significant (F (1, 48) = 74.41, p = 0.001, = 0.77). As compared to baseline, SS% values decreased in the anodal tDCS group for the post intervention time points 3, 4, and 5 (immediately, one week and six weeks after intervention), and in the sham group for the time points 3 (immediately after intervention). For the between group comparisons, SS% were significantly smaller in the anodal tDCS group at time points 3, 4, and 5 (Table 2).
The results of the SSI-4 (one of the secondary outcome measures) showed a significant main effect of group (F (1, 49) = 59.62, p = 0.001, = 0.52; Fig. 2), and a significant time × group interaction as well (F (1, 48) = 123.55, p = 0.001, = 0.72). As compared to baseline, SSI-4 scores reduced in the anodal tDCS group for the post intervention time points 3, 4, and 5 (immediately, one week and six weeks after intervention), and in the sham group for the time points 3 (immediately after intervention). For the between group comparisons, SSI-4 score were significantly smaller in the anodal tDCS group at time points 3, 4, and 5. These findings are shown in Table 2. Additionally, the results of the SSI-4 subscores showed a significant main effect of group for frequency (F (1, 49) = 35.85, p = 0.001), and also a significant time × group interaction (F (1, 48) = 34.85, p = 0.001). As compared to baseline, frequency values reduced in the anodal tDCS group for the post intervention time points 3, 4, and 5 (immediately, one week and six weeks after intervention), and in the sham group for the time point 3 (Table 3). For the between group comparisons, frequency scores were significantly smaller in the anodal tDCS group at time points 3, 4, and 5. Moreover, for duration, a significant main effect of group (F (1, 49) = 19.78, p = 0.001), and also a significant time × group interaction (F (1, 48) = 32.49, p = 0.001) was revealed. As compared to baseline, duration scores were reduced in the anodal tDCS group for the post intervention time points 3, 4, and 5 (immediately, one week and six weeks after intervention), while in the sham group no statistically significant differences were obtained for the post intervention time points (Table 3). For the between group comparisons, duration scores were significantly smaller in the anodal tDCS group at time points 4 and 5 (one week and six weeks after intervention). For physical concomitants, the results showed a significant main effect of group (F (1, 49) = 25.95, p = 0.001), and a significant time × group interaction (F (1, 48) = 19.68, p = 0.001). As compared to baseline, physical concomitants scores were reduced in the anodal tDCS group for the post intervention time points 3, 4, and 5 (immediately, one week and six weeks after intervention), while in the sham group, no statistically significant differences in post intervention time points were present (Table 3). For the between group comparisons, no significant differences were found in physical concomitants scores between the anodal tDCS and sham groups at the post intervention time points.
The other secondary outcome measure was the OASES score, which was assessed at the same five assessment points (one week and immediately before intervention at baseline, and immediately, one week and six weeks after intervention). We found no statistically significant main effect of group (F (1, 49) = 0.28, p = 0.60, = 0.006; Fig. 3), but a significant interaction between time and group (F (1, 48) = 67.20, p = 0.001, = 0.58). As compared to baseline, OASES scores decreased in the anodal, and sham tDCS groups for the post intervention time points 3 and 5 (immediately and six weeks after intervention). For the between group comparisons, no significant differences were found in the overall OASES scores (Table 2). The results of the OASES subscores showed a significant main effect of group in general information (F (1, 49) = 103.99, p = 0.001), and also a significant time × group interaction (F (1, 48) = 58.6, p = 0.001). As compared to baseline, general information scores decreased in the anodal tDCS group for the post intervention time points 3, 4, and 5 (immediately, one week and six weeks after intervention), and in the sham group for the time points 3 and 5 (Table 3). For the between group comparisons, general information scores were significantly smaller in the anodal tDCS group at time point 5 (six weeks after intervention). Additionally, the findings showed a significant main effect of group in reactions to stuttering (F (1, 49) = 43.98, p = 0.001), and a significant time × group interaction (F (1, 48) = 107.31, p = 0.001). As compared to baseline, reactions to stuttering values decreased in the anodal tDCS group for the post intervention time points 3 and 5 (immediately and six weeks after intervention), while in the sham group, there were no statistically significant differences between baseline, and post intervention time points (Table 3). For the between group comparisons, no significant differences were found in reactions to stuttering scores between the anodal tDCS and sham groups at post intervention time points. Moreover, there was a statistically significant main effect of group in communication in daily situations (F (1, 49) = 43.10, p = 0.001), and also a significant time × group interaction (F (1, 48) = 76.04, p = 0.001). As compared to baseline, communication in daily situations values decreased in the anodal tDCS group for the post intervention time points 3 and 5 (immediately and six weeks after intervention), while for the sham group, no statistically significant differences of post intervention time points were obtained (Table 3). For the between group comparisons, no significant differences were found in communication in daily situations values between the anodal tDCS and sham groups at post intervention time points. Finally, the results showed a significant main effect of group in quality of life (F (1, 49) = 38.51, p = 0.001), and also a significant time × group interaction (F (1, 48) = 72.35, p = 0.001). As compared to baseline, quality of life scores decreased (in this scale, lower values are indicating higher quality of life) in the anodal, but not sham tDCS group for the post intervention time points 3 and 5 (immediately and six weeks after intervention) (Table 3). For the between group comparisons, no significant differences were found in quality of life scores between the anodal tDCS and sham groups at post intervention time points.
The results of the adverse effects questionnaire analysis showed no significant main effects of group (F (1, 49) = 2.76, p = 0.07), or time (F (1, 5) = 1.78, p = 0.17), and no significant time × group interaction as well (F (1, 48) = 6.46, p = 0.06). Adverse effects were limited to mild and very mild symptoms such as itching and burning sensation at the application sites.
In order to investigate successful blinding, a chi-square test was applied. The findings indicated no significant difference between the two groups with respect to guessing the intervention condition (χ 2 (1, N = 50) = 0.08; p = 0.77).