Running direction in the hippocampus is encoded by rate modulations of place field activity but also by spike timing correlations known as theta sequences. Whether directional rate codes and the directionality of place field correlations are related, however, has so far not been explored and therefore the nature of how directional information is encoded in the cornu ammonis remains unresolved. Here, using a previously published dataset that contains the spike activity of rat hippocampal place cells in the CA1, CA2 and CA3 subregions during free foraging of male Long-Evans rats in a 2D environment, we found that rate and spike timing codes are related. Opposite to a place field's preferred firing rate direction spikes are more likely to undergo theta phase precession and, hence, more strongly impact paired correlations. Furthermore, we identified a subset of field pairs whose theta correlations are intrinsic in that they maintain the same firing order when the running direction is reversed. Both effects are associated with differences in theta phase distributions, and are more prominent in CA3 than CA1. We thus hypothesize that intrinsic spiking is most prominent when the directionally modulated sensorymotor drive of hippocampal firing rates is minimal, suggesting that extrinsic and intrinsic sequences contribute to phase precession as two distinct mechanisms.

Hippocampal place cell sequences have been hypothesized to serve as diverse purposes as the induction of synaptic plasticity, formation and consolidation of long-term memories, or navigation and planning. The interrelation of sequence replay during offline states such as sleep or consummatory behaviors and online theta sequences during running and navigation is highly debated. Offline sequences are inherently 1-dimensional, whereas online sequences reverse with running direction and thus reflect the 2-dimensional topology of space, which poses a fundamental and unresolved inconsistency. Here, we propose a computational model of cornu ammonis 3 (CA3) and dentate gyrus (DG), where sensorimotor input drives the direction-dependent online sequences within CA3, and the intrahippocampal CA3-DG projections produces prospective intrinsic sequences. The model thereby suggests that sequence propagation on multiple 1-D manifolds underlies a relational code that contains stable signatures for the encoding of spatial memories and that could be used for prospective planning.

Hippocampal place cell sequences have been hypothesized to serve as diverse purposes as the induction of synaptic plasticity, formation and consolidation of long-term memories, or navigation and planning. The interrelation of sequence replay during offline states such as sleep or consummatory behaviors and online theta sequences during running and navigation is highly debated. Offline sequences are inherently 1-dimensional, whereas online sequences reverse with running direction and thus reflect the 2-dimensional topology of space, which poses a fundamental and unresolved inconsistency. Here, we propose a computational model of cornu ammonis 3 (CA3) and dentate gyrus (DG), where sensorimotor input drives the direction-dependent online sequences within CA3, and the intrahippocampal CA3-DG projections produces prospective intrinsic sequences. The model thereby suggests that sequence propagation on multiple 1-D manifolds underlies a relational code that contains stable signatures for the encoding of spatial memories and that could be used for prospective planning.