So for attentiveness and anxiety, the 5 and 20 mcg doses both had a statistically significant effect vs. placebo, but the 10 mcg dose didn't. Should that lead me to believe that there's really a dose-response curve with two peaks near 5 and 20 and a valley in between, or should I be wondering whether they got "lucky" finding a significant result with either the 5/20mcg doses? (P-value was <1% for the significant results.) Is the former (multiple peaks in the response curve) known to happen in other drugs?