Fig. 5. (A ) Individual Bayesian estimates and populations estimate (thick line ) of the analgesic responses to morphine-6-glucuronide (M6G). (B ) Individual Bayesian estimates and populations estimate (thick line ) of the placebo contribution to M6G analgesia. Placebo data were analyzed with a cubic spline with two knots of the form: current(t) = a0+ a1× t + a2× t2+ a3× t3. A significant population placebo effect could not be demonstrated at the P < 0.01 level. Continuous lines = men; broken lines = women.