Figure 1. Experimental approach.

(a) Four principle phases of the pointer-alignment trials. (b) Photograph of the apparatus. (c) The left panel shows a top-down schematic view of the dial and pointer along with an exemplar fingerprint superimposed on the contact surface for scale purposes. The six panels on the right show the six edge lengths. The edge that spanned the entire area contacted by fingertip was termed the infinite edge and the 0 mm edge refers to raised dot stimulus. (d) Cross-sectional and side views of the edges. (e) Normal force, pointer position and rotation velocity shown for six superimposed exemplar trials with the six initial dial orientations. Data aligned on initial touch (vertical line). Dashed horizontal lines represents the target ±2° zone. The resultant pointer position was measured when the rotation velocity fell below 10°/s (red dots). Gray segments of the traces represent final adjustments of the orientation with the shutter glasses opened to allow visual guidance of the movement during the final adjustment of the pointer into the target zone when required.

Figure 1—figure supplement 1. Participants quickly learned the tactile pointer-alignment task.

The horizontal axis shows the actual sequence of tactile pointer-alignment trials performed by each participant, starting with the infinite edge practice block (Pr), followed by three experimental blocks for each of the infinite, 8, 4, and 2 mm edge length. Note that data from the 1 mm edge length and the raised dot are not shown for clarity. The vertical axis represents the absolute alignment error. The black line was obtained by averaging the absolute alignment error trial by trial across participants after first filtering each participant’s data with a symmetrical moving median filter comprising three consecutive trials. The shaded area represents the standard error of the mean. Most of the performance improvement took place in the first 10 trials of the practice block. A repeated measures ANOVA restricted to the infinite edge length, including the practice and three experimental blocks shows a significant effect of block on absolute alignment accuracy (F_3.27 = 3.14, p=0.04). Post hoc examinations indicated that the alignment accuracy during the practice block differed from the three test blocks (p<0.002 for all three comparisons; Tukey HSD test) but that there were no significant differences between the test-blocks (p>0.75 for all three comparisons). A repeated measures ANOVA with edge length (0 – Infinite) and experimental block (1, 2, 3) as factors failed to indicate an effect of block on the absolute alignment accuracy (median value during the block) and there was no significant interaction between block and edge length.