Two interactions were significant. First, the sound type (voice, music) by stimulus type (standard, deviant) interaction (F1,34 = 4.298, P = 0.046, ηp2 = 0.112) revealed that participants responded equally fast to vocal and musical standards (F1,35 < 1), but were faster to respond to vocal, rather than musical, deviants (F1,35 = 4.913, P = 0.033, ηp2 = 0.123). Second, the naturalness (NAT, ROT) by sound type (voice, learn more music)

interaction was also significant (F1,34 = 9.464, P < 0.01, ηp2 = 0.218) due to faster RTs to vocal as compared with musical sounds in the NAT condition (F1,35 = 9.395, P < 0.01, ηp2 = 0.212). In summary, musicians were overall more accurate at the temporal discrimination task and tended to be distracted less by irrelevant timbre change. Additionally, while musicians were equally accurate in their responses to vocal and musical deviants, non-musicians were significantly less accurate and more distracted when classifying musical as compared with vocal deviants. Event-related potentials collected from both groups displayed the expected ERP components. In Figs 3 and 4, ERPs elicited by standards are overlaid with ERPs elicited by deviants, separately for NAT (Fig. 3) and ROT (Fig. 4) check details conditions. Figures 5 and 6 directly compare ERPs elicited in musicians and non-musicians for NAT (Fig. 5) and ROT (Fig. 6) sounds in order to better visualize group differences. The N1 and P3a components

are marked on the Cz site, P3b – on the Pz site, and RON – on the F8 site. Below we present ERP results separately for each of the components of interest, which is followed by a summary with an emphasis on the effect of group and its interactions with other factors. Musicians had a significantly larger N1 peak amplitude compared with non-musicians. This effect was present across all

sites in the midline analysis (F1,34 = 5.205, P = 0.029, ηp2 = 0.133), over frontal, fronto-central and central sites in the mid-lateral analysis (group by site, F4,136 = 3.729, P = 0.038, ηp2 = 0.099; group, F1,34 = 4.314–7.84, P = 0.008–0.045, ηp2 = 0.113–0.187), and over frontal and fronto-temporal sites in the lateral analysis (group by site, F3,102 = 3.701, P = 0.04, ηp2 = 0.098; group, F1,34 = 3.58–7.372, P = 0.01–0.055, ηp2 = 0.104–0.178). The BCKDHB effect of group did not interact with naturalness (group by naturalness: midline F1,34 < 1; mid-lateral, F1,34 < 1; lateral, F1,34 = 1.423, P = 0.241). Additionally, deviants elicited a significantly larger N1 peak amplitude compared with standards (stimulus type: midline, F1,34 = 86.22, P < 0.001, ηp2 = 0.717; mid-lateral, F1,34 = 130.727, P < 0.001, ηp2 = 0.794; lateral, F1,34 = 118.833, P < 0.001, ηp2 = 0.778). Lastly, there were several significant results involving the effect of hemisphere over mid-lateral and lateral sites. In mid-lateral sites, the peak amplitude of N1 was overall larger over the right than over the left hemisphere sites (hemisphere, F1,34 = 4.277, P = 0.