Maybe more like the perception of reality is triggering emotions from a memory of a similar reality. In more extremem cases yes, it brings back the memory too. I'd argue (and this is just me) that emotions are more distinct than perceptions and memories, and are more like annotations to both. But I'm getting way outside of any clinical model I've read about. We want traumatic events to get resolved and stored without the high emotional content that distorts current perceptions.