Thank you very much, this is exactly the starting point I needed. I'll keep following the citations trail into more recent years, but your second link from 2007 reinforces my feeling that we're still very far from understanding the mechanisms:

> Second, within LIPC, we found a gradient in which a more dorsal-posterior region was involved in SR, a mid region was involved in both SR and EE, and a more ventral-anterior region was involved in EE, but only when SR was high.

To me, these are merely clues about how the high-level pieces fit together, and there's a long road to actually understanding the neural correlates of memory.