Biomedical Engineering Reference
In-Depth Information
Figure 4.3 for subjects 2, 4, and 5 (time/checker: 12.0-15.6 s), but the intercept
for subject 5 (123 s) is much higher than that for the other two (26 and 34 s);
this same subject also has much larger within session variability (error bars),
although the same can be seen for a few board types (i.e. sessions) in subjects 2
and 4 as well. In most subjects, both mean time and variability drop below the
regression line for the largest numbers of white squares, indicating the effect of
practice reached by the last session of this experiment. This practice effect is
also suggested by the high intercept values in subjects 1, 3, and 5.
Experiment 2
Experiment 2 was used to monitor an increase in subjects' accuracy on the
counting task. Boards were given to the test subjects in random order, for four
consecutive test sessions. In addition to the prior instructions, subjects were
allowed to start over if they felt that they had lost count; trials in which this
happened were disregarded. The results of this task are depicted in Figure 4.5.
The plots show the counting time as a function of the number of white squares.
Similar to Experiment 1, a regression line was calculated from individual data
points. In addition, 95% confidence interval bands for the regression line were
added using Origin (Ver. 7.0, OriginLab Corp Northampton, MA). As can be
seen from the time scale, all subjects have become significantly faster than in
Experiment 1: Regression line slopes range from 0.5 to 1.2 s/square. Also the
y-axis intercept is lower, at 5-17 s. Note that all five subjects now fit into these
intervals, and that subject 1 (visually impaired) performs nearly as well as the
other subjects. The regression lines provide a better fit to the mean values than
those in Figure 4.3, and the SEM values are lower, suggesting that the subjects
have a more even, i.e. practiced, performance.
Experiment 3
The purpose of Experiment 3 was to achieve proficiency in the placing task.
The subjects placed the checkers on the boards without first counting the white
fields, and boards were given to them in random order. The difficulty here was
to coordinate hand and eye only by proprioceptive information because the hand
was not directly visible in the pixelized view: Only the obstruction of the white
border or white squares by the hand was noticeable. The possibility to re-start
a trial was again provided. In this case the video image was switched off, all
checkers were removed, and the orientation of the board was changed. Unlike
in Experiment 1, subjects were always given 16 checkers. This was done to
eliminate any possible cues about the number of white squares on the board.
Figure 4.6 shows a test subject placing a checker and the corresponding pixelized
camera image as it presented itself to the test subject at that moment. Figure 4.7
indicates the approach one of the test subjects developed for placing a checker.
Other test subjects used different techniques. This experiment was more time
consuming, and the number of trials per session varied from subject to subject,
Search WWH ::
Custom Search