The binocular advantage in visuomotor tasks involving tools

Abstract

We compared performance on three manual-dexterity tasks under monocular and binocular viewing. The tasks were the standard Morrisby Fine Dexterity Test, using forceps to manipulate the items, a modified version of the Morrisby test using fingers, and a "buzz-wire" task in which subjects had to guide a wire hoop around a 3D track without bringing the hoop into contact with the track. In all three tasks, performance was better for binocular viewing. The extent of the binocular advantage in individuals did not correlate significantly with their stereoacuity measured on the Randot test. However, the extent of the binocular advantage depended strongly on the task. It was weak when fingers were used on the Morrisby task, stronger with forceps, and extremely strong on the buzz-wire task (fivefold increase in error rate with monocular viewing). We suggest that the 3D buzz-wire game is particularly suitable for assessing binocularly based dexterity.

Keywords: binocular vision; dexterity; motor coordination; stereoscopic depth perception; visuomotor skills.

Figure 1.

Stereopair of the buzz-wire game used in the study. The image pair is suitable for cross-fusing, i.e. the left eye's view is on the right. The wooden base is 35 cm long.

Figure 2.

Performance on the Morrisby manual dexterity task, for monocular versus binocular viewing. (a–b) Number of pegs successfully completed within the 2-minute time period, averaged over all 30 subjects, with binocular viewing (“Binoc”), or monocular viewing with the dominant eye (“Dom”) or non-dominant eye (“Non-dom”). (a) Using fingers; (b) using forceps. In each case, performance is significantly better with binocular viewing than with monocular viewing using either eye. Bar height shows the mean; error bars show ±1 standard error on the mean. (c) The binocular advantage, i.e. ratio of number of pegs completed with binocular viewing to that for monocular viewing with the dominant eye, calculated for each subject individually and then averaged over all 30 subjects. Bar heights show geometric mean of the binocular advantage (equivalent to mean of log₁₀(Binoc/Dom), raised to power 10); error bars mark ±1 standard error on log₁₀(Binoc/Dom).

Figure 3.

Performance on the buzz-wire task, for monocular versus binocular viewing. (a–d) Performance on four different metrics, with binocular viewing (“Binoc”), or monocular viewing with the dominant eye (“Dom”) or non-dominant eye (“Non-dom”). In every case, binocular performance is significantly better than either monocular condition. (e) Binocular advantage, i.e. ratio of the metric in the monocular-dominant condition to that in the binocular condition, calculated for each subject individually and then averaged over all 40 subjects. Other details are as in Figure 2.

Figure 4.

Binocular advantage on the buzz-wire task, % contact time metric, plotted against stereoacuity (reciprocal of stereo threshold in arcsec). The horizontal line shows a ratio of 1 (performance same in both conditions, no binocular advantage); the vertical dashed line shows the median stereoacuity, corresponding to a threshold of 70 arcsec. The mean stereoacuity corresponded to a threshold of 49 arcsec, and the upper and lower quartiles were 100 and 34 arcsec. Data for the six young children who were tested on the “Animals” stereotest are shown with blue diamonds. There was no significant correlation (r = 0.12, n = 40).