Figure 3. Learning under high rate practice schedule (H-rate) is more labile to interference than learning under the low rate (L-rate).

Performance in all plots is measured by the initial directional deviations of the hand. (A) Effect of the overnight pause: Performance on the first and the last 5 force-field trials in each learning day (L-rate: days 1-5 in black, H-rate: days 1–2 in grey). In the L-rate performance gradually improves over the days with almost no overnight drops between consecutive learning days, whereas under the H-rate it shows a significant overnight drop (t-test, p<0.01) between the first (10°) and second day (20°). (B) Effect of the standard epoch (STD2): The plots show the aftereffects of learning on standard trials in STD2 (left) and the effect of these trials on the subsequent continuation of learning (right plot). The left plot shows that the aftereffects gradually decreased in both the L-rate (black) and the H-rate (grey). However, the right plot shows that the drop in performance between force-field trials before and after STD2 (right) was significantly larger under the H-rate than the L-rate (t-test, p<0.01). (C) Effect of washout: Performance without force-field (aftereffects) after learning under L-rate practice schedule (black) was washed out less than performance after learning under H-rate (grey), which was washed out completely. The figure shows that performance on the last 10 force-field trials of the L-rate and H-rate learning sets were similar (∼10°, ANOVA p>0.5) and resulted in similar aftereffects when switching to standard trials (∼−25°, ANOVA p>0.5). The aftereffects gradually decreased (less negative values in the plot), but from trial 15 until the end of washout the aftereffects in the L-rate were larger than in the H-rate. The aftereffects in the last 10 trials of the washout were significantly different from zero in the L-rate (t-test, p<0.01) but not in the H-rate (t-test, p>0.5).