Credit: Tamara Kulikova/Alamy Stock Photo

The hippocampus has been hypothesized to contribute to multiple cognitive functions through parallel processing pathways within its microcircuits. Cembrowski et al. now provide further support for this model, demonstrating that the mouse hippocampal subiculum contains two discrete populations of pyramidal neurons that make distinct contributions to spatial working memory.

Previous work in rats suggested that ‘subclasses’ of subiculum pyramidal cells (subPCs) have different physiological properties; however, these subPC subclasses have not been characterized in detail. Here, the authors used retrograde labelling to show that, in mice, two spatially separate populations of subPCs project to distinct downstream targets: those in the proximal subiculum projected to a set of targets that included the nucleus accumbens and prefrontal cortex, whereas those in the distal subiculum projected to a different set of targets that included the retrosplenial cortex and ventral hypothalamic nuclei. Electrophysiological analyses demonstrated that the two populations also exhibit different firing properties, with proximal subPCs displaying a regular spiking phenotype and distal subPCs displaying a bursting firing pattern.

Population RNA sequencing revealed distinct gene expression patterns in the proximal and distal subPC populations and, using in situ hybridization, the authors found that a transcriptional boundary divides these two discrete populations. Single-cell RNA sequencing confirmed that subPCs cluster into transcriptionally distinct populations that correspond to either the proximal or distal population of subPCs.

a transcriptional boundary divides these two discrete populations

To determine the relationship between the molecular phenotypes of subPCs and their connectivity, the authors injected tracers into regions that are upstream or downstream of the subiculum, including the entorhinal cortex and the amygdala. They discovered that the proximal and distal subPCs exhibit different patterns of afferent and efferent connectivity. Local circuits within these regions also differed: neuropeptide Y-expressing interneurons were specifically enriched in the proximal subPC region.

These connectivity differences imply that different subPC populations may have specific functions in cognitive processing. To examine this possibility, the authors investigated the effects of chemogenetically silencing each subPC population on the performance of mice in a spatial working memory task. Silencing the proximal subPCs had no effect on performance; however, silencing the distal PCs specifically impaired the animals’ ability to encode new spatial working memories.

These findings indicate that subPCs can be divided into a least two subclasses that are molecularly, spatially and functionally distinct, providing support for the concept of separate and parallel processing pathways within the hippocampus.