It is just derived from the explicit state transition diagram above. When is p' true (i.e. does the transition go to the p-state)? It is precisely true if and only if the transition starts in p, and there is ¬a. The latter is written as p & ¬a. The if-and-only-if is just a biimplication.