Deep Cortex Secrets: Hidden Neurons Map the Developing Brain

Deep inside the developing mammalian cortex, a thin band of early-born neurons called the subplate has long puzzled neuroscientists. These cells are among the very first neurons generated in the embryonic brain, and they serve as temporary scaffolding that guides later-born neurons to their final positions in the cortical plate. Nearby, separated from the subplate by only a sliver of tissue, sits the claustrum, a mysterious, evolutionarily conserved sheet of neurons sandwiched between the insular cortex and the striatum. Because both structures are born in the same early window, occupy adjacent deep territories, and share a number of molecular markers, many researchers have suspected that they share a common developmental origin. A new study challenges that assumption, showing that despite their similarities, the subplate and claustrum follow fundamentally different developmental scripts.

The research, led by Shalini Iyer and Mark S. Cembrowski of the University of British Columbia and published in iScience, deployed single-cell spatial transcriptomics to chart the developing mouse brain with unprecedented detail. Using the 10x Genomics Xenium platform, the team profiled whole-brain sections at four key stages: embryonic days 12.5, 15.5, and 18.5, plus postnatal day 1, with sex-balanced biological replicates at each time point. The custom gene panel combined a standard mouse brain set with 100 additional probes enriched for developing cortex, subplate, and claustral markers. The result is an atlas of more than 1.5 million spatially resolved cells, allowing the researchers to link molecular identity directly to precise anatomical location across development.

From this complete atlas, the authors extracted 448,261 cortical cells and subjected them to high-resolution clustering. The analysis captured the full sweep of cortical development, from radial glial cells and intermediate progenitors in the ventricular zone through transient migrating populations to terminally differentiated excitatory and inhibitory neurons. Within this landscape, two populations emerged as anatomically identifiable at embryonic day 15.5: the subplate and the claustrum. Both expressed well-known markers such as Nxph4, Kcnab1, Ccn2, Cplx3, and Tle4, but crucially these shared genes were biased toward the subplate, while the claustrum showed significantly higher expression of its own markers, including Nr2f2 and Cux2.

The most striking finding concerned how each population relates to broader cortical patterning. As the cortex matures, excitatory neurons diverge into deep-layer and superficial-layer identities, defined by distinct transcription factor codes. When the team mapped the subplate and claustral clusters onto this developmental landscape, the subplate sat squarely at the endpoint of deep-layer maturation, consistent with its molecular kinship to deep excitatory neurons expressing markers such as Foxp2. The claustrum, in contrast, landed at the terminus of the superficial-layer trajectory, expressing superficial-layer genes such as Cux2 and Satb2 at high levels. Differential expression analysis reinforced the split, revealing dozens of genes specific to each population, including subplate-enriched genes involved in neuronal differentiation, extracellular matrix organization, and synaptic function, and claustrum-enriched genes tied to cell adhesion, synapse formation, and neurotrophic signaling.

To validate these findings with an independent method, the researchers reanalyzed a published single-cell RNA sequencing dataset spanning the same developmental window. The complementary analysis reproduced the deep-versus-superficial divergence and went further, resolving three transcriptionally distinct subplate cell types: an early subplate-progenitor population co-expressing Eomes, a putative subplate stage, and a terminal subplate type marked by a mature suite of genes. The claustrum appeared as a single cell type defined by Nr4a2, Gnb4, Rgs6, and Lxn. Concordance between the two datasets was statistically strong, and cross-mapping the annotated reference onto the spatial data placed the claustrum precisely where anatomy predicted, confirming that the molecular divergence observed in intact tissue was not an artifact of the platform.

Tracing the subplate backward in time yielded another surprise. At embryonic day 12.5, no distinct subplate cluster appeared in the mature sense, but a unique population of radial glial cells expressed the subplate marker Nr4a2, and a separate Nxph4-expressing progenitor subgroup showed molecular similarity to hippocampal progenitors. The spatial location of this Nxph4-positive cluster shifted dramatically between day 12.5 and day 15.5, migrating from the cortical ventricular region toward the hippocampal ventricular zone. Together, these observations suggest that subplate neurons may arise from a specialized radial glial subtype, and that the earliest subplate precursors share a molecular heritage with hippocampal lineages, a connection that had not been clearly appreciated before.

Within the subplate itself, the spatial atlas resolved four subtypes that change dynamically across development: two immature types enriched at embryonic day 15.5, one transitional type, and one mature type dominating at postnatal day 1. The immature subtypes expressed genes for extracellular matrix organization and axonal guidance, consistent with early differentiating neurons building scaffolds and initial connections, while the mature subtype upregulated genes involved in synaptic function, guidance refinement, and neuromodulatory signaling. The claustrum told a different story: rather than cycling through transient subtypes, it maintained stable spatial domains, including a principal claustral domain, a claustral-subplate zone continuous with the cortical subplate, and dispersed cells extending into the insular cortex that resemble the previously described Arimatsu cells.

The claustrum’s origin story also differed sharply from the subplate’s. At embryonic day 12.5, no defined claustral structure existed, but the team detected Nr2f2-expressing cells streaming through the lateral cortical stream, a known migratory corridor. By day 15.5, these cells had apparently arrived, and Nr2f2 expression shifted from the stream into the emerging claustrum itself. This trajectory matches the unusual reversed migration pattern previously described for claustral neurons, in which cells travel through the lateral cortical stream before settling in their terminal position. The subplate, by contrast, had already reached its destination by day 12.5. The two structures, in other words, not only end up molecularly distinct but also take completely different routes and timetables to get there.

The authors emphasize that their dataset, available through an interactive web portal called CellestialCortex, complements other large-scale developmental brain atlases by adding early-embryonic coverage with spatial context and single-cell resolution. Because subplate neurons have been implicated in neurodevelopmental disorders, including autism and schizophrenia, and because both subplate-derived layer 6b neurons and the claustrum regulate sleep, arousal, attention, and salience in adulthood, understanding when and how these cell types diverge may illuminate the roots of circuit dysfunction. By revealing that shared markers can mask genuinely distinct developmental programs, the study provides a molecular foundation for dissecting the specialized roles these hidden deep-cortex neurons play in building, and then modulating, the mammalian brain.

Subject of Research: The spatiotemporal development of subplate and claustral excitatory neurons in the embryonic and early postnatal mouse cortex

Article Title: Spatiotemporal development of sparse excitatory neuronal types within the deep mouse cortex

Article References: Iyer, S., & Cembrowski, M. S. (2026). Spatiotemporal development of sparse excitatory neuronal types within the deep mouse cortex. iScience, 29(10), Article 117558. https://doi.org/10.1016/j.isci.2026.117558

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117558

Keywords: subplate neurons, claustrum, single-cell spatial transcriptomics, cortical development, mouse brain atlas, neuronal migration, layer 6b, cell types, gene expression, neurodevelopment, Xenium, CellestialCortex