This is not correct at all.

The sourcing policy says: “If available, academic and peer-reviewed publications are usually the most reliable sources on topics such as history, medicine, and science.”

When did they change that policy?

Peer-reviewed research papers have always been allowed in citations. The nuance that evolved over time is this "For example, a paper reviewing existing research, a review article, monograph, or textbook is often better than a primary research paper." Meaning, you can cite a research paper, but something more secondary that summarizes a body of scholarship would be better.

Individual research papers that haven't been reproduced often present conflicting results with one another, especially in fields with poor quality research like nutrition. Experts often run into this issue when they try to cite their own research or a narrow set of papers in a given field, especially when recent research conflicts with prior scientific consensus. It's why tptacek's comment above is apropos.

They never did. You can't use YOUR OWN work as a source, but you can use primary sources.

> You can't use YOUR OWN work

Bad rule, I'd just use an alt or not my own name.

Which is also against the rules and undermines the systems in place to make the encyclopedia more reliable.