Beyond that, my reading is that the situation is highly similar to other situations of looking at effects of drug exposure. Likely the majority of casual users (and even not so casual users) of drugs are not going to notice specific long term consequences. Some are and they are going to have various confidences in attributing it to their use of any particular drug. Or for that matter attributing a fraction of their problem to drug exposure. (by way of example, those with a family history of alcoholism score worse on certain neuropsych tests that abstinent alcoholics do poorly on. take a family history positive alcoholic and ask what fraction of his impairment is due to the genetic risk and what fraction to the actual alcohol exposure). This is why we need step by step animal research studies to test the hypotheses. but you knew that…

with respect to “long-term delayed neurotoxicity” there’s probably someone else around here (ahem) who can give you a better read. The stereotypical “serotonergic neurotoxicity” paradigm data suggest that the simple hypothesis (massive serotonin disruption) isn’t the answer to the lasting cognitive problems.
Although the Ricaurte 2002 debacle was technically fubar, the underlying principle of “partial depletions putting one down the road to disaster” still has merit and is basically untested to date. even in a rat a full aging study is not fun. ain’t nobody going to do this in one of the usual primate species with decades-long lifespans.
I find Giorgi et al 2005 of interest because it gets into a new area for MDMA, found no monoaminergic effects and the idea of temporal lobe regions being altered in a way that influences seizure threshold leads to some rather obvious hypotheses about cognition. memory in particular. the prior obsession with “serotonin neurotoxicity” as the animal model has delayed branching out into other areas. the focus was on what “worked” rather than what might really address the underlying clinical problem, if there was one. don’t get me started…