From the second plot the slope started with average slope -.01MMSE/day to >-.01MMSE/day average that is for all in enviorment of increasing concentration of the drug. but this totally obscures the effect of positive scores of the high concentration Strong as those decilning do "a better job at it" (decilne more rapidly then the Strong improve). That is the distortion of taking the average in this particular case.
I haven't looked closely at this. But where are you getting the "6 patients they purposely omitted"? That just seems out of left field. What evidence do you have for that?
Market Data 
Markets 
AI
