Fig 1.
Experimental stimuli and design.
(A) Long-term average spectra for the target instruments, arranged in order of decreasing amplitude of high-frequency energy content: trumpet (pink), oboe (yellow), trombone (green), and tuba (blue). Spectra of instruments playing G4 are depicted in the top panel; spectra of instruments playing C4 are depicted in the bottom panel. (B) Pitch-timbre pairings in Consistent blocks: instruments with darker timbres (trombone, tuba) played the lower note (C4), and instruments with brighter timbres (trumpet, oboe) played the higher note (G4). (C) Pitch-timbre pairings in Reversed blocks: instruments with darker timbres played the higher note (G4), and instruments with brighter timbres played the lower note (C4). Musical instrument images courtesy of iStock photos.
Table 1.
Mixed-effects modeling results for Experiment 1 (n.b., since models shared fixed effects architectures, results from both models are presented side-by-side for ease of comparison). Results from the linear mixed-effects model analyzing the logarithm of response times are listed at left (A); results from the generalized linear mixed-effects model analyzing response accuracy are listed at right (B). Block 1 (Consistent) was the default level of the factor Block, so all fixed effects are in reference to Block 1.
Fig 2.
(A) Each dot represents the mean response time for a given listener in that experimental condition; each listener’s means across conditions are connected by grey lines. The estimated marginal mean response times for each condition are depicted using black squares, with error bars denoting one standard error. Dots are colored according to each listener’s score on the musical training subscale of the Gold-MSI, with brighter (toward green) colors indicating higher scores and darker (toward black) colors indicating lower scores (see inset legend). Grey dots indicate listeners who did not complete the Gold-MSI. (B) Each dot represents the mean accuracy for a given listener in that experimental condition; each listeners’ means across conditions are connected by grey lines, with error bars denoting one standard error. The estimated marginal mean accuracy for each condition is depicted using black squares. Dots are again colored according to each listener’s score on the musical training subscale of the Gold-MSI (the same as in panel A). Asterisks denote a statistically significant influence of musical training scores on performance in that block (*p < .05, **p < .01, ***p < .001).
Table 2.
Mixed-effects modeling results for Experiment 2 (n.b., since models shared fixed effects architectures, results from both models are presented side-by-side for ease of comparison). Results from the linear mixed-effects model analyzing the logarithm of response times are listed at left; results from the generalized linear mixed-effects model analyzing response accuracy are listed at right. Block 1 (Reversed) was the default level of the factor Block, so all fixed effects either depict or are in reference to Block 1.
Fig 3.
(A) Each dot represents the mean response time for a given listener in that experimental condition; each listener’s means across conditions are connected by grey lines. The estimated marginal mean response times for each condition are depicted using black squares, with error bars denoting one standard error. Dots are colored according to each listener’s score on the musical training subscale of the Gold-MSI, with brighter (toward green) colors indicating higher scores and darker (toward black) colors indicating lower scores (see inset legend). (B) Each dot represents the mean accuracy for a given listener in that experimental condition; each listeners’ means across conditions are connected by grey lines. The estimated marginal mean accuracy for each condition is depicted using black squares, with error bars denoting one standard error. Dots are again colored according to each listener’s score on the musical training subscale of the Gold-MSI (the same as in panel A). Asterisks denote a statistically significant influence of musical training scores on performance in that block (*p < .05, **p < .01, ***p < .001).
Table 3.
Mixed-effects modeling results for Experiment 3 (n.b., since models shared fixed effects architectures, results from both models are presented side-by-side for ease of comparison). Results from the linear mixed-effects model analyzing the logarithm of response times are listed at left; results from the generalized linear mixed-effects model analyzing response accuracy are listed at right. Block 1 (Consistent) was the default level of the factor Block, so all fixed effects either depict or are in reference to Block 1.
Fig 4.
(A) Each dot represents the mean response time for a given listener in that experimental condition; each listener’s means across conditions are connected by grey lines. The estimated marginal mean response times for each condition are depicted using black squares, with error bars denoting one standard error. Dots are colored according to each listener’s score on the musical training subscale of the Gold-MSI, with brighter (toward green) colors indicating higher scores and darker (toward black) colors indicating lower scores (see inset legend). Grey dots indicate listeners who did not complete the Gold-MSI questions. (B) Each dot represents the mean accuracy for a given listener in that experimental condition; each listeners’ means across conditions are connected by grey lines. The estimated marginal mean accuracy for each condition is depicted using black squares, with error bars denoting one standard error. Dots are again colored according to each listener’s score on the musical training subscale of the Gold-MSI (the same as in panel A). Asterisks denote a statistically significant influence of musical training scores on performance in that block (* p < .05, ** p < .01, *** p < .001).
Table 4.
Mixed-effects modeling results for Experiment 4 (n.b., models shared fixed effects architectures). (A) Results from the linear mixed-effects model analyzing the logarithm of response times are listed at left; (B) results from the generalized linear mixed-effects model analyzing response accuracy are listed at right. Consistent was the default level of the factor Condition, so all fixed effects either depict or are in reference to those trials.
Fig 5.
(A) Each dot represents the mean response time for a given listener in that experimental condition; each listener’s means across conditions are connected by grey lines. The estimated marginal mean response times for each condition are depicted using black squares, with error bars denoting one standard error. Dots are colored according to each listener’s score on the musical training subscale of the Gold-MSI, with brighter (toward green) colors indicating higher scores and darker (toward black) colors indicating lower scores (see inset legend). Grey dots indicate listeners who did not complete the Gold-MSI questions. (B) Each dot represents the mean accuracy for a given listener in that experimental condition; each listeners’ means across conditions are connected by grey lines. The estimated marginal mean accuracy for each condition is depicted using black squares, with error bars denoting one standard error. Dots are again colored according to each listener’s score on the musical training subscale of the Gold-MSI (the same as in panel A). Asterisks denote a statistically significant influence of musical training scores on performance in that block (* p < .05, ** p < .01, *** p < .001).