Skip to main content
. 2020 Jun 3;11:2786. doi: 10.1038/s41467-020-16448-6

Fig. 4. Results of online fusion and preference experiments in Western participants.

Fig. 4

a Mean fusion of musical intervals. Here and in (b), bars plot mean across participants, error bars plot SEM, and insets show histograms of average responses for individual participants (same conventions as in Fig. 3). b Mean rated pleasantness of musical intervals. c Mean fusion vs. mean pleasantness for individual musical intervals (scatter plot of quantities plotted in (a) and (b)). Error bars plot 95% confidence intervals around the mean. rs value is the correlation between the mean fusion and mean pleasantness across intervals. Unison was omitted from correlation as it does not consist of two notes. d Individual differences in fusion. Left panels: mean fusion ± SEM for consonant and dissonant intervals, averaged across all participants (top) or for individual participants (bottom). Consonant and dissonant intervals were all those colored blue and brown, respectively, in (a)–(c). A Wilcoxon signed-rank test was used to compare the two conditions, ****p < 0.0001. Right panel: test–retest reliability of the difference in fusion for consonant and dissonant intervals, computed from two splits of each participant’s trials. Here and in (e), reliabilities were Spearman–Brown corrected to best estimate reliabilities of measures derived from full experiment. e Individual differences in consonance preferences. Left panels: mean pleasantness ± SEM for consonant and dissonant intervals, averaged across all participants (top) or for individual participants (bottom). A two-tailed paired t-test was used to compare the two conditions, ****p < 0.0001. Right panel: test–retest reliability of the difference in pleasantness for consonant and dissonant intervals (their “consonance preference”), computed from two splits of each participant’s trials. f Consonance preference vs. consonant–dissonant fusion difference for individual participants. Each dot represents a participant. Here and in scatter plots in (d) and (e), the x-coordinates of individual dots were jittered by a small amount (randomly drawn from U[−0.01, +0.01]) to mitigate the visual effect of dot overlap.