Figure 4. The quality of fit of the optimal pharmacodynamic model (#4) to the values for adductor pollicis twitch tension (left panels) and laryngeal adductor twitch tension (right panels). The x axis displays time (in minutes) after administration of rapacuronium. The y axis shows the difference between the measured twitch tension and the value predicted by the population pharmacokinetic model (top panels) or the post hoc fit (bottom panels). Each line represents a value from a single volunteer; the duration of neuromuscular monitoring differed among the volunteers. If the model fit the data perfectly, all lines would lie horizontally at 0.0. The improved quality of fit of the post hoc values compared with those from the population model is expected, because the post hoc model permits interindividual variability but the population model does not.