Results

Like in Experiment 1, statistical analyses on CRs and RTs were addressed with ANOVAs using the same within-subject factors. We first used group (bilateral tinnitus (BT), right tinnitus (RT), left tinnitus (LT) patients) and tinnitus severity (mild, moderate and severe) as between-subject factors.

The ANOVA performed on CRs revealed that all tinnitus participants responded more accurately to S2 after a standard S1 (CRs = 89.91, SD = 11.04) than after a deviant S1 (CRs = 87.30, SD = 12.25), F(1, 27) = 17.04, p < .001. However, this effect of stimulus S1 type was varying with the ISI, as shown in the stimulus S1 by ISI interaction, F(2, 54) = 25.51, p < .0001. This interaction is illustrated in Figure 5. The contrast analyses revealed that tinnitus participants performed more accurately to S2 after a standard than after a deviant S1 for the ISI 100, F(1, 54) = 84.37, p < .0001, and the ISI 150, F(1, 54) = 10.82, p < .002; but, no difference was observed between responses after deviant and after standard S1 for the ISI 200, F(1, 54) < 1.

The ANOVA performed on RTs showed that all tinnitus participants responded faster to S2 after a standard S1 (RTs = 736 ms, SD = 125.50) than after a deviant S1 (RTs = 760 ms, SD = 136.33), F(1, 27) = 18.412, p < .001. Moreover, the longer the ISI, the shorter the RTs to S2, F(2, 54) = 50.744, p < .0001. However, this effect of stimulus S1 type was modulated first by the ISI, as revealed by the stimulus S1 by ISI interaction, F(2, 54) = 4.954, p < .02. This interaction is illustrated in Figure 6.

In this figure, it seems that tinnitus participants responded faster to S2 after a standard than after a deviant S1 for each ISI, that was confirmed by contrast analyses: ISI 100, F(1, 54) = 16.61, p < .0002; ISI 150, F(1, 54) = 44.51, p < .0001; ISI 200, F(1, 54) = 5.02, p < .03. Second, the effect of S1 type was varying with the ear of the task, as revealed by the stimulus S1 by ear interaction, F(1, 27) = 5.884, p < .03. This interaction is illustrated in Figure 7. We performed contrast analyses to compare the difference between RTs after a standard or a deviant S1 for each ear. They revealed that this difference was larger when the task had to be performed in the right ear (LE/RE) than when it had to be performed in the left ear (RE/LE), F(1, 27) = 5.88, p < .03, although both differences reached the significance level.

The aim of this study was to test the influence of the tinnitus signal on attention capture according to the side of perceived tinnitus. More precisely, less attention capture was hypothesized when subjects had to perform the task in the tinnitus ear; therefore, we performed further analyses of variance on CRs and RTs. The group (right or left tinnitus patients) was the between-subject factor, the ear of the task (tinnitus ear versus non-tinnitus ear), the stimulus S1 type, and the inter-stimulus interval were the within-subject factors.

The analysis performed on CRs revealed that tinnitus participants responded more accurately when S2 was presented into the tinnitus ear – S1 in the non-tinnitus ear (CRs = 90.03, SD = 10.10) than when S2 was presented into the non-tinnitus ear – S1 in the tinnitus ear (CRs = 86.31, SD = 12.83), F(1, 18) = 5.186, p < .04. Moreover, the same effect of stimulus S1 type, F(1, 18) = 8.703, p < .01 and the same interaction between stimulus S1 type and ISI, F(2, 36) = 13.036, p < .0001 were found. The ANOVA performed on CRs also revealed that patients suffering from severe tinnitus were less accurate (CRs = 82.64, SD = 13.13), than those suffering from mild tinnitus (CRs = 92.55, SD = 6.54), who were themselves less accurate than patients suffering from moderate tinnitus (CRs = 97.01, SD = 2.84), F(2, 16) = 4.805, p < .03.

The ANOVA performed on RTs showed only that participants responded faster to S2 after a standard (RTs = 738, SD = 108.78) than after a deviant (RTs = 763, SD = 117.56) stimulus S1, F(1, 18) = 23.851, p < .001, and that the longer the ISI, the faster the response, F(2, 36) = 30.959, p < .0001.