Supplementary Material Accuracy in the target discrimination task The accuracy was analyzed using a 3 (location: structured, random, neutral, within-subjects) × 2 (time: first half, second half, within-subjects) ANOVA in each experiment. Experiment 1: There was no main effect of location [F(2,42) = 0.74, p = 0.48, ηp2 = 0.03], time [F(1,21) = 1.12, p = 0.30, ηp2 = 0.05], or interaction [F(2,42) = 0.20, p = 0.82, ηp2 = 0.01]. In the first half of the exposure, mean accuracy in the structured, random, and neutral locations was 0.81 (SD = 0.19), 0.79 (SD = 0.17), and 0.76 (SD = 0.12), respectively. In the second half of the exposure, mean accuracy in the same three locations was 0.83 (SD = 0.17), 0.80 (SD = 0.18), and 0.80 (SD = 0.13) respectively. Experiment 2: There was no main effect of location [F(2,42) = 0.65, p = 0.53, ηp2 = 0.03], time [F(1,21) = 0.004, p = 0.95, ηp2<0.001], or interaction [F(2,42) = 0.78, p = 0.46, ηp2 = 0.04]. In the first half of the exposure, mean accuracy in the structured, random, and neutral locations was 0.82 (SD = 0.13), 0.82 (SD = 0.14), and 0.83 (SD = 0.08), respectively. In the second half of the exposure, mean accuracy in the same three locations was 0.83 (SD = 0.15), 0.80 (SD = 0.15), and 0.86 (SD = 0.10), respectively. Experiment 3: There was a marginal effect of location [F(2,86) = 2.93, p = 0.06, ηp2 = 0.06] but no effect of time [F(1,43) = 3.63, p = 0.06, ηp2 = 0.08] or interaction [F(2,86) = 1.24, p = 0.30, ηp2 = 0.03]. In the first half of the exposure, mean accuracy in the structured, random, and neutral locations was 0.81 (SD = 0.14), 0.74 (SD = 0.16), and 0.80 (SD = 0.12), respectively. In the second half of the exposure, mean accuracy in the same three locations was 0.82 (SD = 0.13), 0.80 (SD = 0.15), and 0.82 (SD = 0.14), respectively. Experiment 4: There was a main effect of time [F(1,21) = 5.71, p = 0.003, ηp2 = 0.21] but no main effect of location [F(2,42) = 1.52, p = 0.23, ηp2 = 0.07] or interaction [F(2,42) = 0.003, p = 0.80, ηp2 = 0.01]. The accuracy was higher in the second half than in the first half. In the first half of the exposure, mean accuracy in the structured, random, and neutral locations was 0.80 (SD = 0.18), 0.74 (SD = 0.17), and 0.76 (SD = 0.13). respectively. In the second half of the exposure, mean accuracy in the same three locations was 0.84 (SD = 0.12), 0.78 (SD = 0.23), and 0.83 (SD = 0.12). respectively. Experiment 5: There was no main effect of location [F(2,42) = 0.59, p = 0.58, ηp2 = 0.02], time [F(1,21) = 1.04, p = 0.32, ηp2 = 0.05], or interaction [F(2,42) = 0.34, p = 0.71, ηp2 = 0.02]. In the first half of the exposure, mean accuracy in the structured, random, and neutral locations was 0.78 (SD = 0.16), 0.74 (SD = 0.16), and 0.78 (SD = 0.15), respectively. In the second half of the exposure, mean accuracy in the same three locations was 0.78 (SD = 0.14), 0.78 (SD = 0.17), and 0.80 (SD = 0.11), respectively. Experiment 6: There was a main effect of time [F(1,43) = 10.13, p = 0.003, ηp2 = 0.19] but no main effect of location [F(1,43) = 0.87, p = 0.42, ηp2 = 0.02] or interaction [F(2,86) = 0.23, p = 0.79, ηp2 = 0.005]. The accuracy was higher in the second half than in the first half. In the first half of the exposure, mean accuracy in the structured, random, and neutral locations was 0.78 (SD = 0.18), 0.79 (SD = 0.16), and 0.82 (SD = 0.11), respectively. In the second half of the exposure, mean accuracy in the same three locations was 0.84 (SD = 0.15), 0.83 (SD = 0.16), and 0.85 (SD = 0.11), respectively. Figure 1s. Accuracy of the target discrimination task in the first and the second halves of each experiment. Error bars indicate ±1 between-subjects SEM. Inverse efficiency in the target discrimination task Inverse efficiency (RT/accuracy) was analyzed using a 3 (location: structured, random, neutral, within-subjects) × 2 (time: first half, second half, within-subjects) ANOVA in each experiment. Experiment 1: There was no main effect of location [F(2,42) = 1.91, p = 0.16, ηp2 = 0.08], time [F(1,21) = 1.89, p = 0.18, ηp2 = 0.08], or interaction [F(2,42) = 0.33, p = 0.72, ηp2 = 0.02]. Experiment 2: There was no main effect of location [F(2,42) = 1.08, p = 0.35, ηp2 = 0.05], time [F(1,21) = 0.01, p = 0.90, ηp2<0.001], or interaction [F(2,42) = 0.96, p = 0.43, ηp2 = 0.04]. Experiment 3: There was a main effect of location [F(2,86) = 6.02, p = 0.003, ηp2 = 0.12] but no main effect of time [F(1,43) = 0.001, p = 0.98, ηp2<0.001] or interaction [F(2,86) = 2.14, p = 0.12, ηp2 = 0.05]. Moreover, inverse efficiency was reliably lower in the structured location than in the random location in the first half [t(43) = 3.74, p < 0.001, d = 0.56] and in the second half [t(43) = 3.08, p = 0.003, d = 0.46]. In the first half, inverse efficiency also was lower in the structured location than in the neutral locations [t(43) = 2.38, p = 0.02, d = 0.36], and inverse efficiency was lower in the neutral locations than in the random location [t(43) = 2.12, p = 0.04, d = 0.32]. No other pair-wise comparisons were reliable [t’s(43) < 1.8, p’s < 0.08, d’s < 0.27]. Experiment 4: There was a main effect of location [F(2,42) = 4.285, p = 0.002, ηp2 = 0.17] but no main effect of time [F(1,21) = 0.59, p = 0.45, ηp2 = 0.03] or interaction [F(2,42) = 0.12, p = 0.89, ηp2 = 0.006]. Moreover, in the first half, inverse efficiency was reliably lower in the structured location than in the random location [t(21) = 2.20, p = 0.04, d = 0.47] and inverse efficiency was lower in the neutral locations than in the random location [t(21) = 2.42, p = 0.02, d = 0.52]. In the second half, inverse efficiency was marginally lower in the initially structured location than in the initially random location [t(21) = 1.73, p < 0.1, d = 0.37]. No other pair-wise comparisons were reliable [t’s(21) < 1.6, p’s < 0.14, d’s < 0.34]. Experiment 5: There was a marginal effect of location [F(2,42) = 3.20, p = 0.051, ηp2 = 0.13] but no effect of time [F(1,21) = 0.36, p = 0.56, ηp2 = 0.02] or interaction [F(2,42) = 0.48, p = 0.62, ηp2 = 0.02]. Experiment 6: There was a marginal effect of location [F(2,86) = 2.67, p = 0.08, ηp2 = 0.06], but no effect of time [F(1,43) = 2.73, p = 0.11, ηp2 = 0.06], or interaction [F(2,86) = 0.16, p = 0.44, ηp2 = 0.01]. Figure 2s. Inverse efficiency (RT/accuracy) of the target discrimination task in the first and the second halves of each experiment. Error bars indicate ±1 between-subjects SEM. (*p < 0.05, **p < 0.01, ***p < 0.001, †p < 0.1).