Couldn’t you get consent from the participants to consume hallucinogenic substances but also getting them to agree that there will be a number of trials and they won’t know which one if any of them will actually contain it?

So instead of a single experiment where half of the people get it and half don’t, the experiment will be repeated ten times with the same people over a few months, different meals each time and in total each person will get hallucinogens in their food either zero or one times.

And then you see what effects people report and if that correlates well with who was actually given the hallucinogens, and if there is a lot of placebo among the rest of the people each time that they were not given hallucinogens.

> I assume knowlege of the experiment could influence the outcome.

That's why you do randomisation and double-blinding.