Animals with noise exposure were divided into two groups: those with no significant gap detection in gap-carrier bands 8 to 18 kHz – “tinnitus group“, and those with significant gap detection at all carrier bands – “no-tinnitus group“. **(A & B)**: Gap-PPI and **(C & D)**: noise pulse-PPI. **(A & C)**: The normalized startle (startle with the gap or noise pulse / startle without the gap or noise pulse) is shown for the two noise-exposure groups and the control group after the noise/sham exposure. The dotted line designates a normalized startle of 1, i.e., the startle amplitude without gap or noise pulse. **(B & D):** Dot plot of the absolute startle without the gap (black symbols) and with the gap (white symbols). The mean and 95 % confidence intervals are shown as dotted horizontal lines below and above the mean. The mean + 95 % confidence interval of the noise-floor is given for each group as a line plot (black: control, white: tinnitus, gray: no-tinnitus). Black bars, circles: sham exposure control group (n=4 animals), white bars, triangles: tinnitus group (n=4 animals), gray bars, squares: no-tinnitus group (n=3 animals). Normalized startle responses were derived from the mean of all trials over four days of one animal with the gap normalized to the mean of all trials of one animal without the gap. Stars mark significance in two-way repeated measures ANOVA (p<0.05, for details see text).

## PubMed Commons