In Session 1, subjects hit 8.7 ± 1.0 cm to the right of the screen center. The variability across individual subjects' trials was larger than the variability across subjects' mean values, and was higher for the highest (2.2 cm) and lowest (2.5 cm) image rates, than for the other image rates (all 2.1 cm), F(4, 20) = 2.90, p = 0.05. In Session 2, subjects hit further to the right the faster the targets were moving. They hit 6.8 ± 1.2 cm, 7.5 ± 1.9 cm, 8.1 ± 2.2 cm, 8.4 ± 2.7 cm, 9.4 ± 2.9 cm, and 9.7 ± 3.1 cm to the right of the screen center for targets moving at 10, 25, 40, 55, 70, and 85 cm/s, respectively, F(5, 35) = 6.07, p = 0.0004. For all velocities, the target was 5 cm to the right of the screen center 625 ms after it appeared, so this difference just means that subjects took longer than 625 ms to tap the screen. Individual subjects' standard deviations across trials also tended to increase with velocity. The standard deviations were 1.2, 1.6, 2.0, 2.2, 1.9, and 2.2 cm for targets moving at 10, 25, 40, 55, 70, and 85 cm/s, F(5, 35) = 3.24, p = 0.02.
In Session 3, not at all surprisingly, the point at which targets were intercepted depended on the finger's starting point, F(9, 63) = 20.2, p < 0.0001. Subjects hit furthest to the left (1.7 ± 2.2 cm to the right of the screen center) when starting at the top left and furthest to the right (10.2 ± 3.7 cm to the right of the screen center) when starting at the lower right. The standard deviation across trials was about 2.4 cm. In Session 4, performance was quite consistent across subjects and conditions, despite a lower consistency across trials than in previous sessions. Subjects hit 8.1 ± 1.5 cm to the right of the screen center, with a mean individual standard deviation across trials of 4.4 cm. The large variability across trials is undoubtedly the result of combining the data for the two different starting positions.
In Session 5, subjects were no longer free to choose where to hit the target, so the variability in where the target was hit was obviously smaller than in the previous sessions. Nevertheless, subjects hit slightly further to the right the shorter the presentation time, F(2, 14) = 5.73, p = 0.02. The center of the interception region was 5 cm to the right of the screen center. Subjects hit 5.4 ± 0.3 cm to the right of the screen center when the presentation time was 600 ms. They hit 5.1 ± 0.1 cm to the right of the screen center for the other two presentation times. The mean individual standard deviation across trials was also larger for the 600-ms presentation time (0.5 cm) than for the other two presentation times (both 0.4 cm), F(2, 14) = 8.14, p = 0.005. It was also larger for targets moving at 40 cm/s (0.5 cm) than for ones moving at 30 cm/s (0.4 cm), F(1, 7) = 8.26, p = 0.02, and larger when there was 100 ms within which to hit the target (0.5 cm) than when there was only 50 ms within which to do so (0.4 cm), F(1, 7) = 23.6, p = 0.002.
In Session 6, the position at which subjects hit the screen depended on the condition,
F(3, 21) = 4.13,
p = 0.02. For most conditions it also varied considerably across subjects (
Figure A1). Subjects hit close to the center of the interception region (10 cm to the right of the screen center) when such a region was indicated (10.3 ± 0.3 cm; black curves). Most subjects hit further to the right when starting on the right (14.2 ± 9.4 cm; green curves). Most hit further to the left in the baseline condition (5.6 ± 6.2 cm; blue curves). The average position was also further to the left when starting close to the target's path (6.8 ± 5.5 cm; red curves), but this is mainly because two subjects tapped about 2 cm to the
left of the screen center in this condition. These subjects quickly moved their finger to the left rather than waiting for the target to arrive below the finger. The other six subjects tapped close to the finger's starting point. The average temporal precision of the two subjects who tapped further to the left was not worse in this condition than in the baseline condition, whereas it was worse for the subjects who waited for the target to reach their finger (so that on average, precision was poorer;
Figure 2.6,
baseline vs.
starting just above the screen). Not only the variability across subjects but also the individual standard deviation across trials was by far the smallest when an interception region was indicated (0.7 cm, as opposed to 1.5 cm when starting just above the screen, 2.3 cm in the baseline condition, and 3.2 cm when starting on the right),
F(3, 21) = 12.6,
p < 0.0001.