What the hippocampus actually does

I’ve always been attracted to a theory that, as far as I can tell, was first put forward by Tyler and DiScenna in 1986 in Behavior Neuroscience. The notion was that the hippocampus indexes cortical module associations rather than episodes. Experimental results which appear to reveal mnemonic associations in the hippocampus actually do so because of the “indexed” cortical modules (think columns) which do actually collectively (think cell assemblies) correlate with episodic memory.

The theory has been modified and amplified by many in the field over time. The conventional wisdom was that while the theory was attractive, its fatal flaw was the lack of neuroanatomical evidence for reciprocal connections between neocortex and the hippocampus. In fact, the hippocampus does receive information from the entire neocortex, but unlike the thalamus or striatum, the mapping consists of a convergent stream where many different cortical areas flow information towards the entorhinal cortex which then dumps that combined flow into the hippocampus in a reciprocal fashion (i.e. the output of the hippocampus is also the entorhinal cortex). This makes the Tyler and DiScenna’s original idea difficult to conceptualize, but not impossible.

If the hippocampus was facilitating the formation of neocortical cell assemblies in some way that acted like an index, how might such a mechanism work given the serial (rather than parallel) connectivity between the hippocampus and the neocortex via the entorhinal cortex?

I invite ideas for how this might work from loyal readers.