Several prospective studies, including the studies of MeDi [23, 33–34, 36–37], the RSF
study [29], and the study using RRR method [48] have been conducted and they are
generally considered to be more useful (although still tentative) for causal inference.
However, it needs to be shown that the association between DPs and AD or cognitive
decline can be consistently observed in different populations under different circumstances
[57]. In addition, prospective studies with longer follow-up time are preferred since they are
less likely to be subject to the possibility of reverse causality, i.e., changes in diet being
secondary to the preclinical disease process