Discussion

The main goal of Experiment 3 was to determine whether an external noise, chosen in such a way to physically resemble tinnitus, could influence the attention capture observed in an oddball protocol. The results showed that the presence of a frequency-deviant stimulus automatically captured attention, as revealed by lower performance to target-stimuli after the presentation of the deviant than after the presentation of the standard stimuli. Moreover, the attention capture was greater when the interval between the to-be-ignored stimulus S1 and the target S2 was smaller, that is, less detrimental effects of the preceding task-irrelevant deviant S1 on the processing of relevant S2 occurred with the longest ISI (150 and 200 ms). This was already reported in previous study [Schröger, 1996]. On the contrary, participants responded more accurately to target-stimuli after the presentation of the deviant than after the presentation of the standard when the ISI was the longest. This could be associated with the “inhibition of return” phenomenon [Posner, 1980]. According to Posner, this phenomenon is based on the necessity to inhibit the spatial localization that had already been explored. As the standard to-be-ignored stimulus did not capture attention participants could have, with the longest ISI, directed their attention on the target-side, then explored the to-be-ignored side, and finally tended to return to explore the target-side that may be prevented by the inhibition of return of attention.

Eventually, the results did not reveal any effect of our « tinnitus-simulation » on the attention capture induced by the presence of a frequency-deviant stimulus in a homogeneous set of to-be-ignored stimuli. Indeed, the deviant always captured attention, whatever the side of the perceived tinnitus-simulation. This suggests that the preattentive system underlying deviant detection may not have been affected by the presence of a continuous noise in one ear. Thus, the tinnitus-simulation did not itself capture and automatically direct attention on the ear where it was perceived. A possible explanation could be that the intensity of the chosen tinnitus-like signal might be insufficient or inadequate to cause disturbance in the processes necessary for performing the auditory categorization task. Indeed, such a simple and faint noise might be irrelevant for the task, and should consequently be part of the background noise. However, it is worth to notice that the tinnitus simulation was selected to be as closer as possible to the real tinnitus situation. Moreover, like in Experiment 2 (i.e., like the real tinnitus situation), we controlled for perceptual differences due to the presence of an additional noise in one ear, by increasing, if necessary, the intensity of stimuli to reach the same level of perceived intensity for each participant and each ear. Thus, if this tinnitus-like noise was not able to impair informational processing because of its physical characteristics, why could the tinnitus signal ?