Ten participants with self-reported normal hearing were ... self-reported normal-hearing listeners (12 female, 4 male) participated with a mean age of 23.93 years.
Supplementary Data Supplementary Table 1. Supplementary analyses for Fig. 1d presenting mean P(T1H) and bootstrapped Bonferroni-corrected (1 - α) × 100% confidence intervals for different context intervals. Context interval (lowest, highest interval) -6, -5 -5, -4 -4, -3 -3, -2 -2, -1 -1, 0 0, 1 1, 2 2, 3 3, 4 4, 5 5, 6 * CI does not overlap with .5
Mean (CI) .63 (.35, .82) .69 (.44, .88) .78 (.56, .92)* .85 (.44, 1) .78 (.42, .1) .45 (.17, .77) .34 (.10, .69) .28 (.03, .64) .12 (.01, .29)* .09 (.00, .25)* .17 (.03, .40)* .34 (.11, .67)
Supplementary Table 2. Supplementary analyses for Fig. 1f presenting mean P(Bias) and bootstrapped Bonferroni-corrected (1 - α) × 100% confidence intervals for different numbers of context of tones. # Context tones 1 2 3 4 5 6 7 8 9 10 * CI does not overlap with .5
Mean (CI) .73 (.61, .83)* .84 (.71, .93)* .91 (.80, .96)* .92 (.85, .95)* .94 (.87, .97)* .94 (.88, .97)* .95 (.90, .98)* .95 (.91, .99)* .96 (.93, .98)* .95 (.91, .98)*
Supplementary Table 3. Supplementary analyses for Fig. 2b presenting mean P(Bias) and bootstrapped Bonferroni-corrected (1 - α) × 100% confidence intervals for different context tone durations. Duration of C (ms) 5 10 20 40 80 160 320 * CI does not overlap with .5
Mean (CI) .53 (.46, .64) .49 (.44, .53) .60 (.56, .66)* .76 (.56, .90)* .81 (.64, .93)* .90 (.80, .98)* .89 (.75, .98)*
1
Supplementary Table 4. Supplementary analyses for Fig. 2d presenting mean P(Bias) and bootstrapped Bonferroni-corrected (1 - α) × 100% confidence intervals for different gap durations between context and test. C- T1 delay (s) 0.5 1 2 4 8 16 32 64 * CI does not overlap with .5
Mean (CI) .94 (.86, .98)* .93 (.89, .98)* .92 (.87, .97)* .86 (.81, .93)* .79 (.68, .88)* .78 (.68, .87)* .73 (.63, .84)* .64 (.49, .79)
Supplementary Table 5. Supplementary analyses for Fig. 2f presenting mean P(Bias) and bootstrapped Bonferroni-corrected (1 - α) × 100% confidence intervals for different component densities. # Components/octave 0.32 0.53 1.05 2.10 4.20 8.40 * CI does not overlap with .5
Mean (CI) .91 (.84, .95)* .96 (.93, .98)* .98 (.94, 1.00)* .96 (.84, 1.00)* . 87 (.74, .96)* .62 (.51, .73)*
2
1
P(Bias)
0.75 0.5 n=1
0.25
n=20 75
0
0 #
0
0.25 0.5 0.75 P(Correct test)
1
#
0 75 Supplementary Figure 1. Related to online experiment with large sample size. Results for the online Experiment 3, for N=100 listeners. The P(Bias) is plotted a function of P(Correct), the proportion correct on test trials consisting of unambiguous pitch changes of 1 semitone. The size of the marker represents the number of listeners. No context effect corresponds to P(Bias)=0.5; chance for pitch direction identification corresponds to P(Correct)=0.5. Dotted lines mark these values, as well as the 0.75 proportion in each case and the identity diagonal. Count histograms are also displayed for P(Bias) and P(Correct).
3
P(Bias w.r.t probe)
1
4 Hz 7 Hz 10 Hz
0.75 0.5 0.25 0
b 1 P(Bias w.r.t probe)
a
35 55 75 Probe tone duration (ms)
0.75 0.5 0.25 0
0.5 1 2 4 Sequence duration (s)
Supplementary Figure 2. Psychophysics relating to neural correlates of bias. A pilot experiment was run to determine the acoustical parameters of the probe tones that would interfere minimally with the bias due to the context sequence. Ten participants with self-reported normal hearing were tested (6 females, mean age M = 26 y, SD = 2.8) in a sound-attenuated booth (J.W. Manny Inc., Eckel Sound Rooms). Sounds were played using an Intel HD Audio sound card with 44.1 kHz output sample rate and 24-bit sample depth, delivered diotically using Sony MDR-V700 headphones. Stimuli were as in the main MEG experiment, except for the parameter varied for the pilot experiment. In all panels error bars represent the 95% confidence intervals of the mean across participants. (a) We first varied the duration of each probe tone and the probe tone presentation rates. Tone duration was the main determinant of perceptual bias, with a main effect of tone duration (H(2)=16.2, p.05, N=10). We selected the 4-Hz presentation rate and 35-ms duration and performed a second control experiment, where we varied the duration of probe sequences. (b) Sequence duration had an effect on behavioral bias (H(3)=14.89, p
