What you describe in your question does not match your title, but let's assume you are interested in the question in the title. My field is related to medicine and in general to the question of whether ethics approval is necessary or not, experiments on oneself are not treated differently than experiments on any other human being. Which means, you do need ethics approval. According to this article from the Harvard Gazette
Some [institutional review boards] have categorically excluded investigators from their own experiments; others have considered the issue case by case.
Moreover, self-experimentation is very likely to give useless results (there are notable exceptions) because of the obvious bias present when subject and experimenter are the same. Think about all the efforts we put into designing double-blind studies, placebo control groups, etc. not to mention you'd end up with a N of 1, hardly enough to draw meaningful conclusions.
I can't find the exact document but I remember that the NIH formally disregard publications where the experiments were conducted on the authors themselves.