Two Maps, One Territory? Parietal vs Hippocampal Representations

Though widely separated in terms of both neuroanatomical location and evolutionary development, there are surprising parallels between parietal cortex and the hippocampus:

- Both structures are important for spatial cognition, although parietal cortex is thought to maintain a "self-centered" map of the environment (i.e., where locations are represented relative to the direction of gaze) whereas hippocampus may maintain a "world-centered" or allocentric map (i.e., where locations are represented with respect to landmarks or surrounding geometry).

- Both structures are thought to represent conjunctive associations. Parietal cortex may represent particular stimulus-response mappings, whereas hippocampus may bring together the various contextual elements of an episodic memory.

- These structures may be disynaptically connected, interfacing through the parahippocampal cortex

- Both structures are thought to be important for memory, although over different time scales: parietal cortex is a central component of the default "task network" and may be intrinsic to working memory operations; hippocampus, obviously, is central to long-term memory.

Why would two anatomically distinct regions evolve to perform functions that are so similar?

One hypothesis, specified in a new computational model by Byrne, Becker & Burgess, suggests that allocentric representations in the hippocampus need to acquire input from the motor system in order to perform "path integration". Parietal cortex may be optimally placed to provide this form of information, given its proximity to motor cortex and its egocentric representations.

In support of this claim, Byrne et al. review evidence showing that place cell firing is altered after lesions to associative parietal cortex in rats, as though egocentric information must be passed from parietal to medial temporal lobe areas in order to trigger place cell firing.

Other evidence suggests that hippocampus and parietal cortex may bidirectionally interact to support a variety of memory and navigation operations. For example, some work indicates that the p3b scalp electrical response is actually the signature of hippocampo-parietal interactions, perhaps in order to perform a memory or context "updating" operation (which is probably the dominant hypothesized function of the P3 response).

This sort of hippocampo-parietal framework seems to have inspired Byrne et al.'s model, which successfully simulated hemispatial neglect, path integration, and single-cell recording data of parahippocampal place cells. It's worth checking out.