Figure 3.

A shift of the deviation as a function of the trial in each patient (filled circle: neglect; unfilled circle: no neglect). The patients started with 10 bisections performed with a white noise. During this period a transient trend was observed in both patients, followed by a plateau. The neglect patient then performed the task with the pitch alternation of “Low, High, Low, High, … ” for the first half 10 trials and “High, Low, High, Low, … ” for the later half 10 trials. The no-neglect patient performed the task with the reversed order of alternation: “High, Low, High, Low, … ” for the first half 10 trials and “Low, High, Low, High, … ” for the later half 10 trials. During this period the no-neglect patient exhibited a stable performance while the neglect patient showed a sustained improvement of her bisection bias what turned into an over-compensation, i.e., leftward bisection bias. Then, after a week, the neglect patient performed 10 bisections with the white noise again, showing a sustained effect in the neglect patient. The mean deviations of line bisection (standard neuropsychological test without auditory stimulation, horizontal orientation, 200 mm in length) in the neglect patient in 4 and 5 months before the experiment are also plotted.