Figure 1.

A low dose of baclofen accelerates instrumental learning. (A) Schematics of the gambling task. Subjects selected either the upper or lower of two visual symbol stimuli presented on a computer screen and subsequently observed the outcome. The correct symbol of the gain pair was associated with 80% probability of winning 1 CHF, the correct symbol of the loss pair was associated with 20% of loosing 1 CHF, the neutral pair served as control. ms, Milliseconds. (B) Observed mean of behavioral choices over three concomitant sessions of 20 trials per condition (gain and loss pair) for the 20-mg baclofen, 50 mg baclofen and control group. The learning curves show the proportion of correct choices for each trial (1 means correct symbol choosing for the gain pair, 0 means correct symbol choosing for the loss pair). Trial per trial comparison between the 20-mg baclofen and the control group showed statistical significance for trial 5, 7, 8, 9, 10, and 11 (*p < 0.05). Trial points were smoothed starting from trial 1 using Gaussian algorithm. Neutral condition data not shown. ±SEM means standard error of mean.