A topographic protomap in mouse subplate neurons precedes the onset of sensory experience

Travis A Babola; Minzi Chang; Rachel Jun; Marina Cardoso de Oliveira; Jessica E Messier; Yunru Chen; Lillian Choi; Dave R Gutierrez; Patrick O Kanold

doi:10.1038/s41467-026-77454-8

. 2026 Sep 9;17:10667. doi: 10.1038/s41467-026-77454-8

A topographic protomap in mouse subplate neurons precedes the onset of sensory experience

Travis A Babola ^1,^#, Minzi Chang ^1,^#, Rachel Jun ¹, Marina Cardoso de Oliveira ¹, Jessica E Messier ², Yunru Chen ¹, Lillian Choi ¹, Dave R Gutierrez ¹, Patrick O Kanold ^1,^3,^✉

PMCID: PMC13655162 PMID: 42860424

Abstract

Within sensory systems, thalamic projections to layer 4 (L4) neurons establish precise neocortical topography. During development, thalamic axons first synapse onto subplate neurons (SPNs) beneath the cortical plate, however, how thalamic inputs are organized within the subplate is unknown. To examine SPN organization and activity, we generated an inducible Cplx3-CreER mouse driver line to examine SPNs during development. Before sensory onset, we found that Cplx3⁺ SPNs in the auditory and visual cortices exhibited topographically organized spontaneous activity. Stimulating thalamocortical afferents evoked robust responses in auditory cortex SPNs, indicating functional thalamic input, while cochlear ablation or enucleation dramatically reduced spontaneous activity among Cplx3⁺ SPNs, indicating that peripherally-generated activity propagates to the subplate. After ear canal opening, Cplx3⁺ SPNs responded reliably to sound with coarse tonotopy. Thus, Cplx3⁺ SPNs receive topographically organized input and integrate spontaneous and sensory-driven activity, which may shape refinement of nascent brain circuits.

Subject terms: Development of the nervous system, Cortex, Visual system

The authors find that subplate neurons receive organized eye- and ear-driven activity before sensory experience and maturation of layer 4 inputs in mice. They persist, receive thalamic inputs, and retain tonotopic responses, suggesting a role in sensory computation.

Introduction

The discovery of the cortical sensory “homunculus” revealed exquisite topography in the mature brain, with each part of the body allocated precisely ordered space in the cortex¹. Across sensory systems, thalamocortical axons (TCAs) establish this organization by projecting to the cortex in a manner that preserves the organization of the stimulus attributes they encode. In the visual cortex (VC), neurons encode the visual field within retinotopic maps^2,3. Within the somatosensory cortex (SSC), the cortical volume contains maps of the body surface, including the prominent whisker barrels in rodents^4,5. Similarly, neurons in the auditory cortex (AC) are arranged tonotopically according to sound frequency^6–9. In all cases, stimulus features of greater behavioral relevance occupy disproportionately larger cortical representations¹⁰.

The formation of mature sensory maps requires the accurate topographical relay of sensory information from the periphery, through the thalamus, and finally to layer 4 (L4) cortical neurons. However, TCAs do not connect directly to L4 upon arrival to the cortex. Instead, they transiently pause and form functional synapses within the subplate, a pioneer zone beneath the cortical plate that contains a developmentally transient population of neurons known as subplate neurons (SPNs)^11–15. After this “waiting period,” TCAs invade the cortical plate. In rodents, this process occurs during late embryogenesis, with functional maturation of TCA synapses with L4 neurons occurring over the first two postnatal weeks in primary auditory and somatosensory cortices^16,17.

This early developmental period coincides with the emergence of sensory-independent spontaneous activity that originates in the sensory peripheral organs and propagates to the cortex in a topographic manner, a process critical for the proper refinement of sensory maps^18–21. In the visual and auditory systems, retinal waves and bursts of action potentials within cochlear inner hair cells propagate through developing sensory centers to reach the appropriate topographic regions within the cortex before the eyes and ears open. Notably, observations of this phenomenon in AC and VC relied on widefield calcium imaging in mice expressing pan-neuronal GCaMP^22,23, which revealed topographically organized activity at the mesoscale but lacked cellular resolution within cortical layers. Thus, the specific circuitry by which this activity propagates from thalamocortical afferents to the cortex, prior to the establishment of mature L4 synapses, remains unknown.

SPNs are well-positioned to organize and guide TCAs to their final targets²⁴. Indeed, prior studies demonstrated that ablation of SPNs before TCAs arrive prevents these axons from entering the cortical plate, while later ablation disrupts the topography of sensory maps, impairing the formation of ocular dominance columns in the VC and whisker barrels in the SSC^25–30. These findings suggest that SPNs play a critical role in establishing the early topographic organization of TCAs in the cortex. Because this organization emerges in the cortex prior to ear and eye opening^22,23, this supports a model in which the TCAs and SPNs exhibit topography before sensory onset. Indeed, local field potential recordings in the subplate zone of developing ferrets suggest topography of functional responses³¹, but whether these responses ultimately originate from SPNs is unknown. Moreover, the patterns of endogenous SPN neural activity during early developmental time points is unexplored, primarily due to the lack of specific genetic tools to label SPNs.

Here, we generated a Cplx3-CreER mouse line via CRISPR/Cas9-mediated insertion of CreER into the endogenous Cplx3 locus, which encodes a highly specific molecular marker of SPNs and adult L6b neurons^32,33. Following tamoxifen administration, this line selectively labels SPNs, enabling functional characterization of this early-born neuronal population. Using widefield and two-photon imaging of the AC before ear canal opening, we observed that SPNs are activated by cochlea-driven spontaneous activity relayed via thalamic input. These responses were organized within distinct topographic domains that resembled future tonotopic maps, which we collectively term a protomap. After ear canal opening, SPNs exhibited robust responses to sounds and were organized into coarse tonotopic maps. In the developing VC, SPNs display spontaneous activity patterns consistent with retinal waves prior to eye opening. Together, these findings reveal a conserved role for SPNs in encoding early sensory spatial organization and suggest that these neurons may provide a scaffold for the emergence and refinement of mature cortical maps in L4 and beyond.

Results

Cplx3-CreER mouse line enables specific labeling of SPNs

Identifying the functional response properties of SPNs requires selective labeling of these neurons. Although a variety of molecular markers have been identified in developing SPNs^34–36, genetic tools that selectively label SPNs remain limited, as existing Cre lines fail to express robustly in SPNs across sensory systems or exhibit off-target expression^33,34. Complexin 3 (CPLX3) expression is highly specific to SPNs and is absent in other brain regions^32,33, and thus provides an ideal target to gain genetic access to SPNs. Using CRISPR-Cas9 targeting of the Cplx3 locus, we inserted a sequence encoding a self-cleaving T2A peptide followed by CreER at the 3′ end of the coding region, allowing for tamoxifen-inducible recombination in Cplx3-expressing neurons (Fig. 1a).

Fig. 1 — a CRISPR/Cas9 knock-in strategy targeting the *Cplx3* locus to insert a T2A-CreER sequence. b–d CPLX3 immunostaining (green) and tdTomato expression (magenta) in sagittal and coronal sections from P60 *Cplx3-CreER;lsl-tdT* mice, showing labeled neurons in the AC, VC, and SSC. e Magnified view of the AC region from (c), tdT channel only (magenta). Left: cell bodies and axons ascending to layer 1 (L1); right: overexposed image to demonstrate that SPNs project to L1 with no evident terminal fields in middle layers. For other cortical areas, see Supplementary Fig. 2. f Representative images at P8 and P60 showing co-localization of tdT and CPLX3 across cortical areas and ages. Quantification of labeling efficiency, showing the proportion of CPLX3⁺ neurons that are tdT⁺ (g) and tdT⁺ neurons that are CPLX3⁺ (h) in AC, VC, and SSC at both ages. Data are mean ± SD from n = 5 mice for P8 and n = 7 mice for P60. For each mouse, values represent the average of 4–5 matched coronal sections per cortical area. Individual data points are shown.

To validate the specificity and efficiency of this line, Cplx3-CreER mice were crossed with a Cre-dependent tdTomato reporter line (Ai9, Rosa26-CAG-LSL-tdTomato; lsl-tdTomato), and tamoxifen was administered at postnatal days (P)1 and P2. Examination of sagittal and coronal sections at P60 revealed dense cell body labeling in L6b throughout the entire anterior-posterior axis (Fig. 1b) and in major sensory areas (auditory, visual, and somatosensory cortices) (Fig. 1c, d, and Supplementary Video 1). In other sensory subcortical regions, sparse tdTomato-positive (tdT⁺) cells were detected in the visual lateral geniculate body and superior colliculus, coinciding with sparse endogenous CPLX3 protein, but was absent in the auditory midbrain (inferior colliculus, IC) and thalamus (medial geniculate body, MGB; Supplementary Fig. 1a). Additional labeled cell bodies were observed in the endopiriform nucleus (Fig. 1b and Supplementary Fig. 1b) consistent with known CPLX3 immunopositivity^37,38.

Higher-magnification images of AC revealed dense L6b labeling, with sparse processes extending through the cortical layers and terminating as dense projections in layer 1 (L1) (Fig. 1e and Supplementary Fig. 2; for other cortical areas see Supplementary Fig. 2). Notably, these projections did not appear to terminate in intermediate cortical layers, in contrast to the terminal arborizations in L5 observed in Drd1a-Cre;lsl-tdTomato SPNs³⁹, which likely represent a distinct functional subgroup of SPNs. To confirm that this labeling patterning was strictly dependent on tamoxifen induction, we performed sham (oil) injections. In these controls, there were no detectable tdT⁺ neurons at P8 (Supplementary Fig. 3), confirming the chemical inducibility of this line.

We analyzed AC, VC, and SSC at P8 and P60 to compare across sensory regions (Fig. 1f). To assess laminar specificity, we quantified all the tdT⁺ and CPLX3-positive (CPLX3⁺) cells within the imaging field of view (Supplementary Fig. 4a, b). The area encompassing 90% of CPLX3⁺ cells from the white matter border was defined as subplate/layer 6b (see “Methods”). The span of this region (measured from the white matter border to the 90% superficial limit) remained consistent across all sensory areas at both P8 (AC = 32.0 ± 5.8 µm; VC = 29.4 ± 4.4 µm; SSC = 55.7 ± 6.2 µm, n = 5 mice) and P60 (AC = 34.0 ± 3.9 µm, VC = 29.3 ± 3.4 µm; SSC = 50.5 ± 3.7 µm, n = 7 mice), consistent with prior studies^33,40. Within the defined subplate/L6b, at P8, over 90% of tdT⁺ cells were CPLX3⁺ (Fig. 1g; mean ± SD, AC = 95.5 ± 3.3%, VC = 93.3 ± 7.0%, SSC = 90.7 ± 2.8, n = 5 mice). This high degree of overlap persisted into adulthood (P60), with similarly strong co-expression levels (AC = 95.3 ± 2.1%, VC = 90.8 ± 5.1 %, SSC = 85.5 ± 6.3, n = 7 mice).

We next assessed the efficiency of the transgene induction by quantifying the proportion of CPLX3⁺ neurons that expressed tdT. At P8, more than 70% of CPLX3⁺ neurons were tdT⁺ (Fig. 1h; AC = 73.4 ± 4.8%, VC = 76.0 ± 6.8%, SSC = 82.7 ± 7.9%), with similar proportions observed at P60 (AC = 77.9 ± 4.4%, VC = 86.1 ± 3.6%, SSC = 88.4 ± 4.6%). To determine the representation of these neurons within the broader subplate population, we co-stained tissue sections with NeuN antibodies to determine the proportion of tdT⁺ cells within the defined subplate/L6b region. At P8, tdT⁺ neurons constituted 50% of all NeuN⁺ neurons within the subplate/L6b region in AC, 59% in VC, and 74% in SSC (n = 2 mice; Supplementary Fig. 5a, c). This distribution was similar at P60, with 54% of NeuN⁺ neurons within the subplate/L6b region expressing tdT in AC, 56% in VC, and 72% in SSC (n = 4 mice; Supplementary Fig. 5b, c). Together, these results demonstrate that Cplx3-CreER labeled neurons (hereafter referred to as Cplx3⁺ SPNs) are efficiently and inducibly labeled, representing a major and stable subset of the L6b population that persists into adulthood.

Cplx3⁺ SPNs mature early and receive functional thalamic input before and after ear canal opening

To assess whether the Cplx3⁺ SPNs exhibit properties consistent with previously reported SPNs^41,42, we examined the intrinsic electrophysiological properties of Cplx3⁺ SPNs labeled in Cplx3-CreER;lsl-tdTomato mice. We focused on two-time windows, before (P6–P9) and after (P15–P18) ear canal opening, which marks the onset of low-threshold hearing (~P10⁴³). We prepared acute thalamocortical slices containing AC and auditory thalamus (MGB) (Supplementary Fig. 6a). Cplx3⁺ SPNs fired action potentials (APs) in response to depolarizing current injections (Supplementary Fig. 6b). Over the examined developmental period, neurons exhibited more hyperpolarized resting membrane potentials, decreased input resistance, and reduced membrane time constants (Supplementary Fig. 6c). During the same period, the minimum current required to elicit an AP (rheobase) increased, AP threshold became more negative, AP amplitude increased, and the firing rate as a function of current injected decreased (Supplementary Fig. 6c, d). These shifts parallel the electrical maturation observed in other auditory areas during this period^44,45, which facilitates the high-fidelity temporal precision required for mature auditory processing. Our results demonstrate that genetically defined Cplx3⁺ SPNs undergo a rapid postnatal physiological transformation consistent with reported trends in anatomically targeted, but not genetically defined, subplate populations^41,42 and indicate that the electrical maturation of Cplx3⁺ SPNs coincides with the onset of functional hearing.

To determine if Cplx3⁺ SPNs receive functional thalamocortical input, we performed electrical stimulation of the auditory thalamocortical tract in acute slices (Fig. 2a, b). Among the recorded Cplx3⁺ SPNs, ~74% of cells exhibited reliable evoked responses characterized by short latencies (<10 ms) and minimal temporal jitter (<1 ms; Fig. 2c, d). These properties remained consistent before and after ear canal opening, indicating that the primary response is driven directly by thalamic input, rather than arising from probabilistic polysynaptic activation. These evoked responses were abolished by the co-application of the glutamate receptor antagonists CNQX (10 μM) and APV (50 μM), and recovered following washout (Fig. 2e, f), confirming these thalamocortical inputs are glutamatergic.

Fig. 2 — a Schematic of in vitro recording. Whole-cell recordings from tdT⁺ SPNs in AC of *Cplx3-CreER;lsl-tdTomato* mice (P6–18) while electrically stimulating medial geniculate body (MGB) afferents. b Infrared (IR) and fluorescent images showing targeted recording from a tdT⁺ neuron in acute thalamocortical slices. Representative of n = 9 slices from 8 mice. c Example evoked excitatory postsynaptic currents (EPSCs) traces from one neuron at P9 (left) and P16 (right). Insets show magnified trace. d Onset jitter (SD of onset latency) for individual neurons (biological replicates). Data are mean ± SD. P6–9: n = 8 neurons from 4 mice (black); P15–18: n = 8 neurons from 4 mice (red). e EPSCs evoked by MGB stimulation at baseline, during AMPA/NMDA receptor blockade, and after washout. Light traces: individual neurons; dark traces: mean. P6–9: n = 8 neurons from 4 mice (black); P15–18: n = 8 neurons from 4 mice (red). f Charge transfer across pharmacological conditions. Data are mean ± SD; points represent neurons (biological replicates). P6–9: n = 8 neurons from 4 slices from 4 mice; P15–18: n = 8 neurons from 5 slices from 4 mice. Repeated measures ANOVA with two-sided paired t-tests and FDR correction. P6-9: baseline vs. drug, p = 0.0002, drug vs. washout p = 0.0066, washout vs. baseline p = 0.0027. P15–18: baseline vs. drug p = 0.0060, drug vs. washout p = 0.0060, washout vs. baseline p = 0.091. g, h Optogenetic stimulation schematic. AAV1-hSyn-ChR2(H134R)-EYFP injected into auditory thalamus at P1; recordings at P7–18. i Representative slice and high-resolution images showing MGB injection site at P9. ChR2-expressing thalamic fibers (green); Cplx3⁺ SPNs (magenta). Scale bar, 200 µm. Representative of n = 7 slices from 4 mice (P7–9) and n = 4 slices from 2 mice (P15–18). j Schematic showing tetrodotoxin (TTX) application to isolate monosynaptic responses. k Example optically evoked EPSCs (oEPSCs) from a Cplx3⁺ SPN at P9. Light traces: individual sweeps; dark trace: mean. l Histogram of oEPSC amplitudes. Dashed line indicates mean. P7–9: mean ± SD: 46.8 ± 43.1 pA, n = 13 neurons from 7 slices from 4 mice (black); P15–18: 110.6 ± 59.0 pA, n = 9 neurons from 4 slices from 2 mice (red). A: anterior, L: lateral. **: p < 0.01, ***: p < 0.001.

While low jitter is a hallmark of direct, monosynaptic responses following electrical stimulation^46–48, we sought to conclusively isolate direct thalamic connectivity using a targeted optogenetic approach. We injected AAV1-hSyn-hChR2(H134R)-EYFP into the auditory thalamus at P1 (Fig. 2g) and performed targeted whole-cell patch recording of Cplx3⁺ SPNs coupled with optogenetic stimulation between P7 and P18 (Fig. 2h). During these experiments, we verified the MGB injection site and observed that auditory thalamic fibers expressing ChR2-EYFP were readily present within the subplate (Fig. 2i). Notably, these labeled fibers exhibited sharp anatomical boundaries that were highly consistent with the known location of the AC⁴⁹, further confirming the specificity of our thalamocortical targeting (Fig. 2i). To isolate monosynaptic responses and block polysynaptic transmission from neighboring cells, we applied tetrodotoxin (1 µM) and 4-aminopyridine (100 µM) to the ACSF (Fig. 2j). Under these conditions, brief photostimulation (3–5 ms) evoked robust inward excitatory postsynaptic currents in Cplx3⁺ SPNs (Fig. 2k). Across the ages recorded, about 85% of cells exhibited light-evoked responses (Fig. 2l). Collectively, these results indicate that Cplx3⁺ SPNs receive functional glutamatergic input via thalamocortical projections, supporting an active role for these neurons in the propagation of ascending thalamocortical activity within developing, nascent circuits.

Spontaneous activity in Cplx3⁺ SPNs reflects cochlear-driven input before ear canal opening

Before ear canal opening, spontaneous activity arising in the cochlea propagates through the auditory brainstem, midbrain, and thalamus to reach the cortex, where it entrains neurons that later respond to similar sound frequencies^50–52. While widefield imaging of mice expressing pan-neuronal GCaMP revealed topography within AC at this stage of development²², the cellular and laminar organization of activity within immature cortical circuits remains unknown. Because Cplx3⁺ SPNs receive thalamic input before ear canal opening (Fig. 2), we hypothesized that their activity in vivo would reflect this ascending auditory drive (Fig. 3a). To test this, we developed a dual-labeling strategy to simultaneously visualize SPN activity and ascending auditory input via the inferior colliculus (IC; Fig. 3b), which serves as a reliable proxy for peripheral activity at this developmental stage, as its spontaneous activity is abolished by acute cochlear ablation²². Because Cplx3 is not expressed in the IC (Supplementary Fig. 1b), we generated Cplx3-CreER;lsl-GCaMP6s (TIGRE-LSL-GCaMP6s; Ai162D) mice to selectively express GCaMP6s in SPNs and injected the IC with AAV9-hSyn-Cre-P2A-dTomato to induce GCaMP and tdT expression there.

Fig. 3 — a Schematic of the auditory pathway showing topographically organized spontaneous activity from the cochlea to the inferior colliculus (IC) and ventral medial geniculate body (MGBv). We investigated whether this activity propagates to SPNs and maintains its ordered structure. b Experimental setup schematic. hSyn-Cre-P2A-dTomato virus was injected into the IC of *Cplx3-CreER;lsl-GCaMP6s* mice to express GCaMP6s in IC neurons, enabling visualization of activity in IC neurons and AC SPNs at P8. c Widefield fluorescence image showing AC, visual cortex (VC), and IC. Representative of n = 10 mice. d ΔF/F traces from regions of interest (ROIs) indicated in (c) showing strong temporal correlations between IC and AC signals, but not with VC. e Correlation map using IC signal as a seed trace; warm colors indicate high positive correlation; cool colors indicate low or negative correlation. High correlation values within the neutral density (ND) filter region reflect optical reflections at the filter edge. Same orientation as (c). f Quantification of correlation coefficients in AC and VC using the IC signal as the seed. Data are mean ± SD with individual points representing mice (biological replicates); n = 10 mice. Statistics: one-way repeated measures ANOVA (F(3, 36) = 87.6, p = 1.31 × 10⁻¹⁶, η_p² = 0.88) followed by two-sided paired t-tests with FDR correction. AC vs. AC shuffled: p = 3.59 × 10⁻⁷; VC vs. VC shuffled: p = 0.33; AC vs. VC: p = 2.36 × 10⁻⁷. A: anterior, L: lateral. ***: p < 0.001, n.s.: not significant.

Following the installation of a cranial window, widefield calcium imaging at P8 revealed robust spontaneous activity throughout the field of view (FOV; Supplementary Video 2), which encompassed the left posterior cortex and the IC (Fig. 3c). Regions of interest (ROIs) drawn around the AC, IC, and VC revealed highly correlated activity between the IC and AC, but not between the IC and nearby VC (Fig. 3d and Supplementary Video 2). To further validate this specificity, we employed a seed-based correlation analysis using the IC signal as a reference. This unbiased approach confirmed functional correlation within the lateral (auditory) cortex, while the medial (visual) cortex lacked correlation with the IC (Fig. 3e).

To control for spurious correlations due to high cortical activity levels, we performed autocorrelation analyses and found that the IC-AC correlation at zero lag exceeded the 95th percentile of autocorrelated values, unlike the IC-VC correlation (Fig. 3f). While these correlations suggest a functional link between IC and Cplx3⁺ SPNs, we wanted to verify that the observed activity originated from Cplx3⁺ SPNs rather than off-target neuronal subtypes, which could arise if our viral injection spread beyond the IC. To analyze this spread, we performed post-hoc histology on mice used for imaging, which confirmed a complete lack of viral-mediated tdT within the cortex (Supplementary Fig. 7a–c). Within the same sections, we observed robust tdT expression in the auditory thalamus (MGB), as expected from ascending IC projections (Supplementary Fig. 7c). Moreover, we found no evidence of GCaMP6s expression or spontaneous calcium activity in the cortex of lsl-GCaMP6s mice following IC-targeted AAV injections (Supplementary Fig. 7d, e). These control experiments demonstrate that cortical calcium signals in our experimental group are independent of viral infection within the IC. Together, these data indicate that the activity of Cplx3⁺ SPNs reflects cochlear-driven activity before the ear canals open.

Early subplate activity reveals a diffuse protomap of tonotopy in auditory cortex

To understand how SPN activity in the AC is structured, we examined the spatial organization of subplate activity in the AC at P8 using widefield calcium imaging (Fig. 4a). Imaging revealed discrete fluorescence events across the AC (Fig. 4b, c, and Supplementary Video 3). Using unbiased k-means clustering of fluorescence activity revealed a circular, spatially coherent domain centered on the AC (Fig. 4d, e), closely matching the circular structure observed during simultaneous IC and AC imaging (Fig. 3e). Because AC has a stereotyped tonotopic organization at later ages, we next asked whether early spontaneous activity in Cplx3⁺ SPNs also shows topography. When correlation seeds were placed along the future tonotopic axis of primary auditory cortex (A1, posterior to anterior-medial; Fig. 4f, g), several hallmarks of mature topography were observed. First, seeds within future low-frequency regions of A1 (Fig. 4a, f, g, cyan) exhibited high intra-regional correlation along the ventral edge of the cortex, which contains the low frequency areas for secondary auditory cortex (A2) and the anterior auditory field (AAF), while exhibiting lower correlations with future high frequency regions. Second, seeds corresponding to future middle-frequency regions of A1 emerged as an interstitial stripe (Fig. 4f, g; yellow). Finally, we observed robust long-range correlations between future high-frequency A1 and corresponding high-frequency zones within the AAF, A2, and the dorsal posterior (DP) region (Fig. 4f, g; magenta).

Fig. 4 — a Schematic of widefield calcium imaging in AC before ear canal opening. Right: representative imaging field with superimposed tonotopic axis indicating future A1 frequency regions. Adapted from ref. ⁸¹. b Time-lapse images showing spontaneous activity in lateral cortex. Representative of n = 12 mice. c Time-color projection of 40 s of spontaneous activity showing a spatially coherent domain encompassing the AC. d Left: K-means clustering defines the auditory cortex. Middle: cluster seed reveals high local correlation. Right: individual masks (0.85 correlation threshold) aligned to centroid (n = 12 mice). e Composite correlation map averaging aligned maps across mice (n = 12). f Seed-based correlation maps from points along the presumptive tonotopic axis reveal organization indicative of mature topography. Example from one mouse. g Group average of correlation maps aligned to k-means cluster centroid (n = 12 mice). h Individual seed-based correlation maps and i population-averaged composite map, as in (f, g), in P8–10 *Thy1-GCaMP6s* mice (n = 5 mice). j Representative maps of average fluorescence and k-means clusters (inset) before and after bilateral cochlear ablation. Difference map shows activity loss localized to AC. Representative of n = 4 mice. Cochleae adapted from ref. ²². k Composite map showing average fluorescence decrease following ablation (n = 4 mice). l Percent change in activity within and outside of the AC ROI. Data are mean ± SD; individual mice shown in gray (n = 4 mice; black: before, red: after ablation). Statistics: two-sided paired t-test with Bonferroni correction. Within AC: p = 0.0078; outside AC: p = 0.84. m Representative activity peaks (points) relative to AC mask (colored area) from one mouse. Representative of n = 4 mice. n Heatmaps showing the spatial probability distribution functions (PDFs) of spontaneous peaks. Individual maps aligned to k-means cluster centroid (n = 4 mice). o Frequency of fluorescence peaks (events/min) within the AC ROI versus surrounding cortex. Data are mean ± SD; individual mice in gray (n = 4 mice; black: before, red: after ablation). Statistics: two-sided paired t-test with Bonferroni correction. Within AC: p = 0.0026; outside AC: p = 1.0. D: dorsal, P: posterior. **: p < 0.01, n.s.: not significant.

To determine how the spatial precision of early subplate topography compares to that of the broader cortical population, we compared Cplx3⁺ SPN activity with recordings from P8 to 10 Thy1-GCaMP6s mice, which express GCaMP6s pan-neuronally and exhibit highly structured correlation maps (Fig. 4h, i). Notably, the Thy1-GCaMP6s maps were remarkably similar in their global organization, with the same functional areas and topographic hallmarks identifiable across both mouse lines. However, in general, Cplx3⁺ SPN activity was broader and less sharply defined than activity in the Thy1-GCaMP6s line, suggesting a more diffuse representation at this early developmental stage.

Our observation of highly correlated activity between the IC and AC (Fig. 3) strongly suggests that spontaneous activity in Cplx3⁺ SPNs is driven by ascending peripheral input. To causally test this hypothesis, we performed widefield imaging of Cplx3⁺ SPN activity before and after bilateral cochlear ablations within the same animals (Fig. 4j–o). This manipulation resulted in a localized attenuation of average fluorescence within the AC (%ΔF: mean ± SD, −9 ± 2%; paired two-sided t-test on average ΔF/F values, t(3) = 8.14, p = 0.004, Cohen’s d = 7.64, 95% CI = [0.001, 0.003]), while activity in surrounding regions remained unaffected (3 ± 6%, t(3) = − 0.93, p = 0.41; Fig. 4j–l). Similarly, event-wise peak detection revealed a dramatic reduction (>90% drop from 4.9 ± 0.8 to 0.4 ± 0.4 events per minute; paired two-sided t-test, t(3) = 11.8, p = 0.0013, Cohen’s d = 7.64, 95% CI = [3.28, 5.72]) in localized activity within the AC (Fig. 4m–o). These data indicate that activity of Cplx3⁺ SPNs within AC is largely driven by peripheral input prior to ear canal opening.

Divergent SPN-layer 1 projections mask the subplate protomap

Widefield imaging in Cplx3-CreER;lsl-GCaMP6s mice captures activity predominantly from upper cortical layers; therefore, the observed diffuse representation in widefield may reflect divergent projections in L1 from SPN somata or Cplx3⁺ processes from other cortical areas, masking a potentially more spatially defined representation at the somatic level. To investigate this possibility, we performed two-photon imaging across cortical depths in P8–10 Cplx3-CreER;lsl-GCaMP6s mice to capture both SPN somata as well as their projections to more superficial cortical layers (Fig. 5a, b). Robust spontaneous calcium signals were observed in L1 (~50 µm below the pial surface) (Fig. 5c), consistent with SPN axonal projections^32,53,54, and in SPN somata (~700–800 µm below the pial surface) (Fig. 5f). Activity in L1 was spatially widespread and less confined within discrete domains, with largely synchronous activity across the anterior-posterior axis (Fig. 5c, d, and Supplementary Video 4) and highly overlapping correlation maps across the FOV (Fig. 5e). In contrast, SPN somata exhibited sparse but highly correlated activity, with discrete events along the anterior-posterior axis (Fig. 5f, g, and Supplementary Video 4) and spatial maps containing more discrete domains (Fig. 5h). Correlation maps from L1 were more homogeneous, while SPN somata exhibited more structured, spatially segregated correlation domains. To quantify this observation, we computed pairwise Pearson correlations on spatially binned ROIs and examined how correlation strength decayed with distance (Fig. 5i and “Methods”). Correlation coefficients were higher in L1 and declined more gradually with distance compared to SPN somata (Fig. 5j). Together, these findings suggest that the dense arborization of axons in L1 partially contributes to the observed diffuseness of the neonatal AC map in widefield imaging (Fig. 4f, g), while the underlying somata exhibit a relatively more compact topographic representation.

Fig. 5 — a Schematic of two-photon imaging in *Cplx3-CreER;lsl-GCaMP6s* mice to monitor axonal activity in L1 (~50 µm from pial surface) and somatic activity in L6b (~750 µm depth). b Representative images and magnified insets from two-photon imaging sessions (~1 mm² FOV). Right: imaging stacks downsampled for analysis. Boxes show anterior-posterior orientation of ROIs for (d, g). Representative of n = 11 mice (axons) and n = 5 mice (somata). c Time-lapse images of spontaneous activity in Cplx3⁺ SPN axons. c–h From the same mouse. d ΔF/F traces for ROIs oriented along the anterior-posterior axis. e Correlation map of axonal activity from seed ROIs (colored dots). f–h Similar to (c–e), except for Cplx3⁺ SPN somata within subplate. i Schematic showing how correlation coefficients were calculated as a function of distance from a central ROI. See “Methods.” j Left: 95th percentile correlation coefficient versus distance between pixel pairs after downsampling (100 × 100 pixels). Right: normalized data showing steeper decay of spatial correlation with distance in SPN somata. Data are group mean with 95% confidence interval (n = 11 mice for axons, n = 5 mice for somata). Statistics: two-way ANOVA with main effects of distance (F(10, 165) = 77.2, p = 6.04 × 10⁻⁵⁷) and layer (F(1, 165) = 126.5, p = 3.78 × 10⁻²²), no interaction (F(10, 165) = 1.67, p = 0.091). Post-hoc two-sided unpaired t-tests with FDR correction showed axons vs. somata differed significantly at all distance bins (all p < 0.02; see Source Data for exact values). A: anterior, D: dorsal, P: posterior.

Retinal waves entrain Cplx3⁺ SPNs in visual cortex

To determine whether the early SPN topography generalizes across sensory systems, we extended our imaging analysis to the VC (Fig. 6a), which exhibits peripherally generated retinal wave activity during this developmental period^23,55. At this stage of development, these waves are characterized by long-duration, propagating wavefronts that traverse the retinotopic space⁵⁶. Before eye opening (P8–10), we observed long-duration retinal wave-like activity propagating across VC (Fig. 6b). We also examined our previous AC/IC datasets (Fig. 3) and identified instances where superior colliculus and VC activity co-occurred (Fig. 6c), indicating that some VC SPN activity is driven by the retina. To examine the spatial structure of spontaneous activity within VC, we performed seed-based correlation similar to that shown in Fig. 4. These correlation maps revealed highly organized topography that, when merged, formed the characteristic triangular shape of primary VC (Fig. 6d, e)²³. To better characterize this activity, we examined the temporal dynamics by placing a grid of ROIs across V1 (Fig. 6f). Raster plots of calcium transients revealed two general types of activity: longer, propagating waves and shorter, non-propagating events (Fig. 6g). To confirm that the retinal wave dynamics observed via widefield imaging accurately reflected the activity of SPNs rather than off-target signals or neuropil contamination, we performed two-photon imaging of Cplx3⁺ SPN somata within the VC and observed that somatic activity recapitulated the propagating waves observed in widefield imaging (Fig. 6h, i, and Supplementary Video 5), validating the Cplx3⁺ SPN origin of the widefield events.

To determine the extent VC SPN activity is driven by the periphery, we performed paired widefield imaging sessions of mice before and after bilateral enucleation (Fig. 6j). Following the loss of retinal input, propagating waves were effectively abolished. Specifically, the frequency of long-duration events (>8 s) dropped precipitously (Fig. 6k). Despite this near-total loss of propagating waves, substantial spontaneous activity persisted in the form of transient, short-duration events, the frequency of which remained stable post-enucleation (Fig. 6k). These findings demonstrate that, while retinal input is required to observe retinal waves within SPNs, a portion of SPN activity in early postnatal V1 is generated by local or extra-retinal sources.

Cplx3⁺ SPNs do not exhibit responses to sounds prior to ear canal opening

Single-unit recordings in anesthetized ferrets and mice have shown that neurons in the subplate respond to sounds before ear canal opening^31,57. To determine if Cplx3⁺ SPNs exhibit sound-driven responses prior to ear canal opening in mice, we performed two-photon imaging of SPN somata within the primary AC in unanesthetized P9–10 Cplx3-CreER;lsl-GCaMP6s mice (Fig. 7a). Mice were presented with a 1-s long, 8 kHz tone at 90 dB SPL (sound pressure level, 50 repetitions, 30 s interstimulus interval; Fig. 7a). Despite prevalent spontaneous activity that sometimes overlapped with the trial period, the neuronal population lacked consistent time-locked, sound-evoked responses. Of the 1902 neurons examined (n = 3 mice), only 28 neurons (1.47%) showed uncorrected p-values < 0.05 (mean ± SD: 0.027 ± 0.005). This frequency of “responsive” neurons is less than one-third of the 5% false-positive rate (95 neurons) expected under the null hypothesis, indicating that SPNs do not respond robustly to sound. To ensure this lack of response was not an artifact of signal processing, we compared p-value distributions using two analytical frameworks: a “corrected” condition with neuropil subtraction (α = 0.7, ϵ = 20) and an “uncorrected” condition (α = 0, ϵ = 0) (Supplementary Fig. 8a and “Methods”). Across both conditions, p-value distributions were approximately uniform, indicating a lack of sound-driven activity beyond the false-positive rate (Supplementary Fig. 8b, c). Even with a permissive false discovery rate (FDR) threshold of 0.2, no neurons remained significant after correction. Averaging across the entire FOV also revealed no significant tone-evoked signal (Fig. 7c). In marked contrast, after ear canal opening (P13–14), we observed robust and reliable tone-evoked responses; of the 3713 neurons examined, 1062 showed uncorrected p-values < 0.05 (n = 3 mice; mean p ± SD: 0.010 ± 0.001), of which ~65% remained significant after FDR correction (q = 0.05) (Fig. 7d–f).

Fig. 7 — a Representative image of SPNs before hearing onset and, right, corresponding ΔF/F traces during the imaging session. Vertical lines represent tone onset times (8 kHz, 90 dB SPL, 1 s long). Note the presence of robust spontaneous activity. ROIs represent the 15 neurons with highest significance (lowest p-values) when comparing baseline to sound-triggered average activity. b ΔF/F traces from SPNs in P9–10 *Cplx3-CreER;lsl-GCaMP6s* mice in response to 8 kHz tones (1 s tone duration, 90 dB SPL, 50 trials, 30 s ISI); 15 neurons with lowest p-values are shown. c Trial-averaged traces from individual neurons (gray) and population mean (black). Right, stimulus-triggered average image following tone onset. Representative of n = 6 mice. d, e Same as (b, c), from P13–14 mice after ear canal opening. e Is representative of n = 3 mice. f Population average before (n = 6 total mice, with n = 3 *Cplx3-CreER;lsl-GCaMP6s* and n = 3 *Cplx3-CreER* mice injected with hSyn-FLEX-GCaMP8s) and after (n = 3 *Cplx3-CreER;lsl-GCaMP6s* mice) ear canal opening. Gray: individual mice; black: population mean. g Percentage of sound-responsive neurons to 8 kHz and white noise before and after ear canal opening. Individual mice shown (circles: *Cplx3-CreER;lsl-GCaMP6s* mice; triangles: P8–9 *Cplx3-CreER* mice with hSyn-FLEX-GCaMP8s). Statistics: two-way ANOVA with main effects of age (F(1, 14) = 414.2, p = 8.49 × 10⁻¹²), stimulus type (F(1, 14) = 18.2, p = 0.00078), and significant interaction (F(1, 14) = 37.1, p = 2.78 × 10⁻⁵). Post-hoc one-sided one-sample t-tests vs. zero with FDR correction: before ear canal opening, 8 kHz: p = 0.22; white noise: not tested (all zeros). After ear canal opening, 8 kHz: p = 0.025; white noise: p = 0.0030. Unpaired two-sided t-tests with FDR correction comparing before vs. after: 8 kHz: p = 0.00025; white noise: p = 1.24 × 10⁻⁷.

Although cortical responses to pure tones (<10 kHz) and bone-conducted sounds have been reported prior to ear canal opening^57–59, we reasoned that broadband noise, which contains energy across a wider frequency range, might more effectively activate SPNs. However, presentation of white noise at 90 dB SPL failed to evoke detectable responses before ear canal opening, despite robust responses to the same stimulus after ear canal opening, which recruited a larger fraction of neurons than 8 kHz tones (Fig. 7g). Analysis of trial-to-trial reliability further supported this lack of responsiveness; the few “significant” neurons at P9–10 rarely exceeded a threshold of 2 SD above the baseline mean across repetitions (Supplementary Fig. 8d), contrasting with the high reliability observed at later stages (Fig. 7d). To increase detection sensitivity, we injected AAV9-hSyn-FLEX-GCaMP8s, which boasts improved signal-to-noise ratios and faster response dynamics than GCaMP6s⁶⁰, into the AC of Cplx3-CreER mice at P1. Despite the increased sensitivity of this calcium sensor, p-value distributions remained flat (Supplementary Fig. 8b), and we still failed to detect tone-evoked activity at P9–10 (Fig. 7g, triangles). While we cannot exclude the possibility of weak or sparse responses to sound falling below our detection threshold, these data indicate that Cplx3⁺ SPNs do not exhibit robust responses to sound before ear canal opening in mice. Rather, peripherally-generated spontaneous activity largely coordinates the activity patterns of SPNs before low-threshold sensory onset (Figs. 4 and 6).

Cplx3⁺ SPNs exhibit tonotopic organization after ear canal opening

Given the topographic organization of SPN activity observed prior to ear canal opening (Fig. 4) and robust responses to 8 kHz and white noise stimuli after ear canal opening (Fig. 7), we hypothesized that Cplx3⁺ SPNs would exhibit tonotopy. To test this, we used two-photon imaging of Cplx3⁺ SPNs in A1 after ear canal opening (Fig. 8a, b). We observed that Cplx3⁺ SPNs responded to tones in a frequency-dependent manner, with greater sensitivity to mid-range frequencies and reduced responses at the low and high limits of the mouse hearing range (Fig. 8c). For each sound-responsive SPN, we extracted its best frequency (BF, the frequency eliciting the strongest overall response) and characteristic frequency (CF, the frequency with largest amplitude response at the lowest sound level) and plotted their distributions. These frequency distributions displayed a modestly strong peak at 16 kHz, near the peak of mouse hearing sensitivity (Fig. 8d). Presentation of tones revealed a coarse topography: low frequency tones activated neurons in posterior regions, with higher frequencies progressively recruiting more anterior areas (Fig. 8e and Supplementary Video 6). To visualize this apparent tonotopic organization, we generated CF maps, where each neuron is represented by a circle, and the color of the circle indicates its CF (Fig. 8f). Overlaying maps from all animals by aligning each map to the centroid of low frequency (<8 kHz) neurons revealed a coarse gradient of frequency tuning aligned along the primary tonotopic axis of primary AC (Fig. 8g). To determine how SPN tonotopy compared to the thalamorecipient layer (L4), we performed a similar set of experiments in Thy1-GCaMP6s mice. L4 neurons exhibited more spatially compact responses to tonal stimuli and a sharper clustering of characteristic frequencies compared to SPN responses (Fig. 8h–j). To compare the spatial organization between SPNs and L4, we calculated the absolute distance of each neuron to the centroid of the low-frequency region; neurons with higher CFs were consistently located farther away in both Cplx3⁺ SPNs and L4 populations (Fig. 8k). However, local CF variability within 100 μm circular neighborhoods, as measured by taking the standard deviation of CFs within the neighborhood, was significantly greater in SPNs, indicating a coarser microstructure (Fig. 8l). Additionally, the overall spatial spread of tone-evoked responses was broader across frequencies in SPNs compared to L4 (Fig. 8m and Supplementary Video 6). Together, these findings indicate that Cplx3⁺ SPNs exhibit a nascent and spatially diffuse tonotopic map that may guide the emergence of precise frequency organization in upper cortical layers.

Fig. 8 — a Schematic of widefield imaging in AC of *Cplx3-CreER;lsl-GCaMP6s* mice. Adapted from ref. ⁸¹. b Left: representative widefield responses to 4 kHz (cyan) and 32 kHz (magenta) tones. Right: two-photon image of SPNs in the same FOV. Representative of n = 11 mice. c Averaged ΔF/F responses across sound-responsive SPNs to tones. Gray: individual mice; black: population mean (n = 11 mice, 1372 total sound-responsive neurons). d Distribution of best (gray) and characteristic (CF, black); n = 11 mice, 1372 total sound-responsive neurons. e Representative tone-evoked activity maps for low (4 kHz, cyan), mid (11.3 kHz, yellow), and high (45.3 kHz, magenta) tones. Representative of n = 11 mice. f CF map for the imaging session in (e). Each dot represents a sound-responsive neuron; color indicates CF. g CF map across SPNs from all animals (n = 11 mice), aligned on the centroid of low-frequency responsive neurons. h–j Same as (e–g) for L4 neurons imaged at 300 μm depth in *Thy1-GCaMP6s* mice (n = 6 mice). k Distance from each neuron to the centroid of low-frequency responsive neurons (≤8 kHz). Data are mean ± SD; individual mice shown as points (SPNs: n = 11; L4: n = 6 mice). Linear mixed-effects model with significant frequency effect and interactions, but no layer effect. Post-hoc: two-sided t-tests with FDR correction. See Source Data for full statistics. l Local CF dispersion within 100 μm diameter regions. Data are mean ± SD; individuals shown as points (SPNs: n = 11, L4: n = 6 mice). Two-sided Mann–Whitney U = 12.0, p = 0.039. #: A subset of mice showed CF maps dominated by a small number of frequencies, resulting in low dispersion values. m Spatial spread (SD) along the anterior-posterior axis (SD_AP). Data are mean ± SD; individual mice shown as lines (SPNs: n = 11, L4: n = 6 mice). Linear mixed-effects model with two-sided t-tests and FDR correction (low: p = 0.0028; mid: p = 4.94 × 10⁻⁵; high: p = 4.94 × 10⁻⁵). A: anterior, P: posterior, D: dorsal, L: lateral. *: p < 0.05, **: p < 0.01, and ***: p < 0.001.

Discussion

The formation of cortical sensory circuits requires the precise connection between thalamic axons and cortical input layers. Before innervating L4, TCAs contact SPNs residing above white matter tracts^11–15,61, but how SPNs integrate into developing circuits and the nature of their network activity remain poorly understood. While existing tools such as Ctgf-dgCre mice offer temporal access to SPNs during this period, their utility for functional imaging is limited by off-target expression in the vasculature; conversely, other drivers like Drd1a-Cre lack expression in the auditory cortex during early postnatal development³³. Here, we developed the Cplx3-CreER mouse line, which provides the high anatomical specificity required to selectively monitor and manipulate SPNs in vivo during this critical developmental window. Using this approach, we find that these early-born neurons encode a protomap of functional cortical organization during a key stage of cortical circuit assembly.

SPNs are well positioned to organize incoming thalamic fibers during development. Upon arrival in the cortex, TCAs accumulate within the subplate for several days (~embryonic day 15–P0 in mice; post-conception weeks 15–24 in humans) before ascending into the cortical plate¹¹, a “waiting period” that may facilitate initial circuit patterning. Indeed, SPNs express a rich repertoire of axon guidance molecules (e.g., Epha3, Epha5, Sema5a, Sema5b, Sema6d, Unc5d, and Slit2) that likely stabilize and sort thalamic afferents^62–64. We found that Cplx3⁺ SPNs exhibit spatially structured spontaneous activity at P8, prior to the maturation of thalamorecipient L4 circuits¹⁶. In the visual system, SPNs exhibit retinal waves, which intrinsically propagate across retinotopic space within the developing eye (Fig. 6). In the auditory system, spontaneous activity patterns aligned with tonotopic gradients observed after hearing onset (Fig. 8). In both cases, SPN activity was inherited from spontaneous activity generated in the peripheral sense organs themselves. After hearing onset, the tonotopic map in SPNs was coarser than in L4 (Fig. 8). This coarseness likely reflects the broad, horizontally spanning morphology of SPNs^65–68, allowing them to sample across the tonotopic axis and support competitive refinement in upper layers. Together, these findings demonstrate that sensory topography is established within the subplate prior to the maturation of thalamic–L4 connections (~P7). This implies that the organization of cortical maps is established even earlier, likely governed by a synergistic combination of molecular guidance cues, broad anatomical sampling by SPNs, and early spontaneous activity.

Despite the presence of a structured protomap (Fig. 4) and functional thalamic input (Fig. 2), we did not observe robust sound-evoked responses in SPNs prior to ear canal opening (P9–10; Fig. 7). This result is unexpected due to previous reports of sound-evoked activity in SPNs at approximately the same developmental stage^31,57. This discrepancy may arise from differences in physiological state; while anesthesia suppresses background activity to boost detection of weak or sparse responses, the prominent spontaneous activity in our unanesthetized mice may have masked subtle evoked signals. Furthermore, while 8 kHz tones can activate other AC neurons at this age^57,69, it is possible that the stimulus set used here (8 kHz tone and broadband stimuli) did not align with the specific receptive field properties of Cplx3⁺ neurons.

Other biological and technical constraints may also account for these differences. For instance, species-specific maturation rates may account for discrepancies with studies performed in ferrets³¹. Additionally, imaging SPNs at depth (~750 µm below the pial surface) hinders photon collection and results in a diminished signal-to-noise ratio, which may limit the sensitivity required to detect weaker physiological events, such as single action potentials. Nevertheless, the immediate emergence of robust responses following ear canal opening suggests that any earlier sound-evoked activity is considerably weaker. Given the presence of functional thalamic input (Figs. 2 and 3) and the dependence of patterned SPN activity on the cochlea (Fig. 4), our findings indicate that Cplx3⁺ SPNs within the AC are primarily entrained by peripherally-generated spontaneous activity, rather than sensory-evoked activity, prior to ear canal opening.

This entrainment aligns with a core developmental principle whereby structured, spontaneous activity refines circuits prior to sensory experience. Indeed, disruption of early activity in the visual^18,19, somatosensory⁷⁰, and auditory system²⁰ results in coarser topography in these systems. This manifests as overlapping receptive fields in the superior colliculus, diffuse barrel structures in the SSC, and compression of the tonotopic map in the IC and AC, respectively, indicating the conserved role of spontaneous activity in refining topographic maps across sensory modalities.

As the first targets of thalamic input, SPNs exhibit robust spontaneous activity that is temporally aligned and dependent on peripherally generated electrical activity (Figs. 3, 4, and 6). Combined with their extensive anatomical (Fig. 1e) and functional projections to L1⁵⁴, these features position SPNs as potential orchestrators of cortical maturation²⁴. Layer 1 serves as a major integration zone for the apical dendrites of both pyramidal and inhibitory neurons across multiple layers, including L2/3, L5, and, notably within the AC, L4^71,72. Moreover, L1 receives diffuse input from neuromodulatory systems, long-range cortico-cortical projections, and, later, direct thalamic projections^72,73. By ramifying within L1, SPN axons may provide a substrate for modulating excitability and synchrony within and across developing columns. The density of these axonal arborizations in superficial layers also suggests a broader capacity to coordinate populations of immature neurons across layers via their apical dendrites during the early postnatal period, when synaptic connectivity is sparse and labile.

While the Cplx3-CreER line provides a window into postnatal maturation, it is less suitable for investigating SPNs during embryogenesis. Low Cplx3 expression before birth⁶³ and the technical risks of prenatal tamoxifen administration^74,75 preclude the use of Cplx3-CreER mice for examining the earliest subplate-TCA interactions. However, because Cplx3⁺ neurons persist into adulthood as L6b (Fig. 1), this line enables investigation of their evolutionary conserved functions. Unlike humans and carnivores, where the subplate is a vastly expanded but largely transient structure^13,76,77, rodents exhibit minimal loss of SPNs⁷⁸. Our results confirm that Cplx3⁺ SPNs persist in mice, providing a tractable model to examine how Cplx3⁺ neurons in L6b contribute to mature cortical computation, plasticity, and higher-order processing⁷⁹. This persistent population may therefore represent an additional substrate for noncanonical thalamocortical pathways that directly activate deep cortical layers⁸⁰. The relevance of this pathway is underscored by the fact that SPNs and L6b neurons are enriched with genes associated with neurodevelopmental and psychiatric disorders, such as autism and schizophrenia³⁴; consequently, its dysfunction could have profound implications for higher-order processing. Within this framework, the early SPN protomap may serve not just as a transient scaffold, but as an early manifestation of a functional deep-layer circuit that integrates thalamic input into maturity.

Together, these observations support a model in which SPNs form a protomap of cortical organization and act as a hub for integrating thalamic drive and distributing it across nascent cortical networks. By facilitating spatially restricted, temporally correlated activation across layers and columns, SPNs may help establish or sculpt the foundational connectivity patterns necessary for sensory cortical map formation and plasticity.

Methods

Contact for reagent and resource sharing

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Patrick Kanold (pkanold@jhu.edu).

Experimental model and subject details

Both male and female mice (Mus musculus) of postnatal days (P) 6–P60 were used for all experiments and randomly allocated to experimental groups. All mice were healthy and were only used for experiments detailed in this study. Mice were housed on a 12 h light/dark cycle and were provided food ad libitum. All experiments and procedures were approved by the Johns Hopkins Institutional Care and Use Committee (protocol: MO23M146).

Transgenic animal models

The Cplx3-CreER knock-in mouse line was generated by inserting a T2A-CreERT2 cassette into the 3′ end of the endogenous Cplx3 locus using CRISPR/Cas9-mediated homologous recombination. Briefly, a guide RNA targeting the final exon of Cplx3 was selected (against ATCCTAGAGGAGAATCCAGG; IDT), and a donor plasmid was constructed containing a self-cleaving T2A peptide sequence followed by CreERT2, a Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element (WPRE), and a bovine growth hormone polyadenylation signal, flanked by ~500 bp homology arms (synthesized by Twist Biosciences). The targeting construct and sgRNA/Cas9 ribonucleoprotein complexes were microinjected into fertilized zygotes derived from C57BL/6J mice. Injected embryos were transferred into pseudopregnant foster females. Correct targeting was confirmed by PCR genotyping and sequencing of founder animals. Founders were bred to C57BL/6J mice for at least 3 generations to establish a stable line and minimize off-target effects. Mice were heterozygotes for all experiments. Functional validation was performed by crossing to loxP-flanked reporter strains and assessing recombination following tamoxifen induction (Fig. 1 and Supplementary Figs. 1–5).

Other lines used include Ai9 (lsl-tdTomato, JAX #007909) and Ai162 (lsl-GCaMP6s, JAX #031562). To induce recombination in Cplx3-CreER mice throughout this study, 60 µg of tamoxifen (30 µL of 2 mg/mL dissolved in corn oil; ~60–100 mg/kg) was injected into the milk sac at P1 and P2. All mouse lines were maintained on a C57BL/6J background. To preserve auditory function, all mice used in hearing and sound-evoked experiments were crossed to the B6.CAST-Cdh23^Ahl+ line (JAX #002756), which carries the Ahl+ allele to prevent age-related hearing loss⁸¹. Genotyping was performed by Transnetyx or in-house PCR using published primer sequences.

Terminology

To maintain consistency and clarity in the context of our developmental studies, we use the term subplate neurons (SPNs) to refer to neurons labeled in the Cplx3-CreER mouse line when experiments occurred before P21 and as layer 6b thereafter. This terminology reflects the developmental lineage and location of these neurons beneath the cortical plate in the early postnatal neocortex, consistent with prior literature identifying Cplx3 as a highly specific marker of the subplate population³⁶. We recognize, however, that the precise nomenclature for these neurons at later stages of development remains an area of active discussion⁸².

Histological preparation and analysis

Mice aged P8–P60 were anesthetized with isoflurane and transcardially perfused with 4% paraformaldehyde (PFA; Electron Microscopy Sciences, Cat No. 15714) in 0.1 M phosphate-buffered saline (PBS, pH 7.4; Sigma, Cat. No. P4417). Brains were removed and sectioned coronally at 50 μm using a vibratome (Leica VT1000S). Coronal or sagittal sections were collected in 24-well plates containing 0.1 M PBS and stored at 4 °C prior to staining.

Coronal sections encompassing the auditory (AC), visual (VC), and somatosensory (SSC) cortices were selected based on anatomical landmarks from the mouse brain atlas (Franklin & Paxinos, 3rd edition) and criteria from prior studies^33,83 (Supplementary Fig. 4a). Medial sagittal sections encompassing the olfactory bulb, superior colliculus, and inferior colliculus were selected for staining. Free-floating sections were stained using standard immunofluorescence protocols. After blocking in a 0.1 M PBS solution containing 5% goat serum (SeraCare, Cat. No. 5560-0007) and 0.3% Triton X-100 (Sigma, Cat. No. T9284) at room temperature for 90 min, sections were incubated overnight at 4 °C with primary antibodies against chicken anti-GFP (1:500, Rockland, Cat# 600-901-379), rabbit anti-RFP (1:500, Abcam, Cat# ab124754), chicken anti-RFP (1:500, Rockland, Cat# 600-901-379), rabbit anti-Cplx3 (1:500, Synaptic System, Cat# 122302), or mouse anti-NeuN (1:1000, Milipore, Cat# MAB377). Secondary antibodies (1:500, Jackson ImmunoResearch) were then applied for 80 min at room temperature. The secondary antibodies used in this study are goat-anti-rabbit IgG, Alexa Fluor 488-conjugated (Cat# 111-545-144), goat-anti-rabbit IgG, Cy3-conjugated (Cat# 111-165-144), goat-anti-chicken IgY, Cy3-conjugated (Cat# 103-165-155), goat-anti-chicken IgY, Alexa Fluor 488-conjugated (Cat# 103-545-155), and goat-anti-mouse IgG, Cy5-conjugated (Cat# 115-175-146). Nuclei were counterstained with DAPI (5 μg/mL; Molecular Probes, Cat. No. D1306) for 3–5 min. Sections were mounted using antifade mounting medium (Vector Laboratories, H-1000).

Definition of subplate/L6b and cell count analysis

Fluorescence images were acquired on a Keyence fluorescent microscope or a Zeiss LSM 800 confocal microscope using 10× Plan APO objectives. Image brightness and contrast were uniformly adjusted for display using Fiji (ImageJ) software. For quantification, confocal image fields (638.9 µm × 638.9 µm) were analyzed using Fiji and the Cell Counter plugin, with researchers blinded to age and cortical area. Four to five sections per animal were analyzed from five animals at P8 (n = 3 for Cplx3 and RFP immunostaining; n = 2 for Cplx3, RFP, and NeuN immunostaining) and seven animals at P60 (n = 3 for Cplx3 and RFP immunostaining; n = 4 for Cplx3, RFP, and NeuN immunostaining). For RFP and CPLX3 immunostaining, cell counts were classified into three types: (i) somata expressing tdT and CPLX3, (ii) somata expressing only tdT, and (iii) somata expressing CPLX3. For RFP, CPLX3, and NeuN immunostaining, cell counts were classified into four types: (i) somata expressing tdT, CPLX3, and NeuN (ii) somata expressing only tdT and NeuN, (iii) somata expressing CPLX3 and NeuN, and (iv) those only expressing NeuN. In all cases, all counted cells were verified for DAPI staining (nuclei marker). Cell counts were performed using the z-stacks to ensure colocalization of expression within the same cell body and no over-counting by careful examination from each z-plane. In addition, cells that are located at the border with less than 70% of the cell body are not accounted for the analysis.

All counted cells with their x- and y-location information were saved in a CSV file and imported into the MATLAB software for further cell analysis. Data were averaged within and across animals and reported as mean ± standard deviation (SD). To standardize spatial comparisons, the white matter boundary was estimated in each image by averaging the y-positions of the ten lowest CPLX3⁺ cells. Distances of all labeled cells from this boundary were calculated. To prevent bias in defining the subplate/layer 6b and layer 6a boundary, a cumulative distribution of all CPLX3⁺ cells was generated for each tissue section, and the 90th percentile was used as the threshold to define the boundary. The values were then averaged within and across animals (P8: n = 5 mice; P60: n = 7 mice). Our definition of subplate/L6b was within 50–60 µm from the white matter boundary, consistent with prior studies^33,40. All labeled cells within the field of imaging were plotted from the white matter boundary (Supplementary Fig. 4b). Quantification in Fig. 1 and Supplementary Fig. 5c was restricted to subplate/L6b. To determine labeling efficiency, the proportion of CPLX3⁺/tdT⁺ double-labeled cells was computed relative to either the total CPLX3⁺ or total tdT⁺ population. To determine the proportion of CPLX3⁺ or tdT⁺ cells among the other subplate/L6b neuronal population, NeuN+ cells that were immunopositive or immunonegative for CPLX3 and tdT were counted, and the proportions were calculated, respectively.

Fluorescence intensity analysis

Fluorescence images at P8 were acquired on a Keyence fluorescent microscope using a 4× PLAN APO objective. A rectangular with a 400 µm-width, and length spanning the pial surface and the white matter, is drawn for AC, VC, and SSC (Supplementary Fig. 4a). The intensity of tdT expression was measured using the plot profile plugin function in Fiji (ImageJ). The values were then averaged within and across the animals (n = 3) and plotted as a function of cortical depth, normalized from the subplate/L6b and white matter border (0%) to the pial surface (100%).

Brain clearing

A Cplx3-CreER;lsl-tdTomato brain at P9 was perfused, and the whole brain was processed for brain clearing following published protocol⁸⁴. Briefly, the brain was transferred into SHIELD-OFF solution for 3 days and incubated at 4 °C, followed by SHIELD-ON solution at 37 °C for a day (SHIELD kit, LifeCanvas Technologies). After washing with PBS, the brain was transferred into CUBIC-L solution (10% v/v Triton X-100; 10% v/v n-butyldiethanolamine; 80% v/v dH₂O) for delipidation. The delipidation process was performed at 37 °C over 4 days, with the solution refreshed every 2 days. Next, the brain was washed with PBS and incubated in RIMS solution (∼60% w/v Nycodenz, 0.6% w/v meglumine, 1% w/v diatrizoic acid, 0.01% sodium azide, and 40% w/v urea in 20 mM PB buffer; refractive index is adjusted to 1.495) for 2 days before imaging. The sample was imaged in a horizontal plane with a Zeiss Lightsheet 7 microscope, with 5× 0.16/NA air objective and two 5× illumination objectives set orthogonal to the detection path. Images were downsampled to 20% using Fiji (ImageJ) and rotated to obtain a coronal view using an open source platform (napari.org: https://doi.org/10.5281/zenodo.3555620).

Electrophysiology

Brain slice preparation and recording conditions

Mice were deeply anesthetized with isoflurane and decapitated. Brains were rapidly removed and immersed in ice-cold oxygenated dissection solution (in mM): 212.7 sucrose, 2.6 KCl, 1.23 NaH₂PO₄, 26 NaHCO₃, 10 glucose, 3 MgCl₂·6H₂O, and 1 CaCl₂·2H₂O (pH 7.35–7.4, bubbled with 95% O₂/5% CO₂). To preserve thalamocortical afferents, the brain was blocked at a ~15°–20° angle (lateral side raised)^42,49. Coronal thalamocortical slices (400 µm) were cut using a vibratome (Leica VT1200) and incubated at 30 °C in oxygenated artificial cerebrospinal fluid (ACSF; in mM: 130 NaCl, 3 KCl, 1.25 NaH₂PO₄, 20 NaHCO₃, 10 glucose, 1.3 MgSO₄·7H₂O, and 2.5 CaCl₂·2H₂O; pH 7.35–7.4) for 45 min.

Recordings were performed at room temperature (21–24 °C) in slices superfused with ACSF (2–4 mL/min). Whole-cell patch-clamp recordings were obtained using borosilicate pipettes (5–11 MΩ, Sutter Instruments) filled with an internal gluconate-based solution (in mM): 108 potassium gluconate, 2 KCl, 0.2 CaCl₂·2H₂O, 10 HEPES, 1 EGTA, 1.5 Na₂ATP, 0.3 Na₂GTP, 1 MgCl₂·6H₂O, and 5 glutathione (pH 7.25, ~280 mOsm). tdT⁺ neurons were identified under fluorescence using a 40× objective and LED illumination (Thorlabs, wavelength 530 nm).

Recordings were acquired with a MultiClamp 700B amplifier (Molecular Devices), digitized with a National Instruments PCI-6259 DAQ board, and collected using WaveSurfer software (HHMI Janelia; https://doi.org/10.5281/zenodo.15103514).

Intrinsic properties

To assess intrinsic excitability, neurons were initially held at −60 mV (corrected) in voltage-clamp mode and then switched to current-clamp. A series of square current injections (500 ms duration; −50 pA to 80 pA in 10 pA steps) was applied; higher steps (up to 100 pA) were used as needed to elicit spikes. Each cell underwent 10 current injection sweeps.

Electrophysiological data were analyzed using custom MATLAB scripts. A liquid junction potential of 10 mV was corrected offline. Resting membrane potential was calculated from the 500 ms pre-stimulus period. Input resistance was calculated from I–V curves derived from hyperpolarizing current steps using linear regression. Membrane time constant (τ) was estimated by fitting an exponential function to small hyperpolarizing voltage responses (usually −5 to −10 pA steps). Rheobase was defined as the minimum current required to elicit an action potential (AP). The first AP in each sweep was used to measure action potential threshold and amplitude. AP threshold was defined as the membrane potential where the first derivative of voltage reached 10% of its maximum. Amplitude was calculated from the threshold to the AP peak. Statistical comparisons between age groups were performed using either unpaired t-tests for normally distributed data or Mann–Whitney tests for non-normal distributions, with Benjamini–Hochberg correction applied for multiple comparisons. Data are reported as mean ± SD.

Electrical thalamic stimulation and analysis

To evoke excitatory postsynaptic currents (EPSCs), a bipolar electrode (0.1 MΩ) was positioned within the auditory thalamocortical radiation, and 0.2 ms stimulation was delivered via a stimulus isolation unit (SIU-91A; Cygnus Technology). To verify that EPSCs were glutamatergic, ACSF was supplemented with 10 µM CNQX (Tocris) and 50 µM D-AP5 (Tocris) to block AMPA and NMDA receptors, respectively. Washout and recovery were monitored after 30 min in ACSF.

To quantify evoked synaptic responses following thalamocortical stimulation, all traces were filtered using a low-pass Butterworth filter. Baseline correction was performed by subtracting the mean signal from 0 to 190 ms window before the onset of electrical stimulation (at 200 ms). The threshold for an evoked response was defined as 5 × root mean square value of the baseline signal. The onset latency was defined as the time from the start of electrical stimulation to the first time point at which the signal exceeded a detection threshold and remained above it for at least 5 ms, to avoid detecting transient electrical artifacts. The standard deviation (jitter) of the latency for each recorded neuron was calculated. Recorded neurons with onset latency <10 ms and jitter <1 ms was included in Fig. 2 (P6–9: 8/12 cells; P15–18: 8/10 cells). Evoked synaptic responses were quantified by integrating the area under the baseline-corrected traces from the onset latency to 50 ms window following electrical stimulation. A similar analysis was performed for traces recorded during drug application and after ACSF washout. Traces from each recording were plotted with the average data overlayed on top. Group comparisons across ages and experimental conditions were assessed using a mixed-design ANOVA. When significant effects were observed (p < 0.05), post hoc paired t-tests with Benjamini–Hochberg correction were performed to determine the pairwise differences. Data are reported as mean ± SD.

Optogenetic thalamic stimulation and analysis

At P1, ~100 nL of AAV1-hSyn-hChR2(H134R)-EYFP (Addgene #26973-AAV1; titer ~1 × 10¹³ GC/mL) was injected into the left auditory thalamus at a stereotaxic coordinate of 0–50 µm anterior from the transverse sinus, 1300–1400 µm lateral from the midline, and 1400–1500 µm from the surface of the skull. Neonates were cryoanesthetized by placing on ice for 5–6 min until unresponsive, then positioned and stabilized under a dissecting microscope (Leica M80). A small midline scalp incision was made, and the skull was cleared of overlying tissue. A pulled glass micropipette (~20–30 μm tip diameter) backfilled with viral solution was lowered vertically into the target region. Viral delivery was performed using a Nanoject (Nanoject II or III, Drummond) at ~10 nL/s. The pipette was left in place for 30 s post-injection to minimize backflow and then slowly retracted. Following injection, the scalp was sealed with histoacryl (TISSUESEAL, LLC), and pups were allowed to recover on a warming pad before returning to the home cage.

After at least a week of recovery, auditory thalamocortical slices were prepared for slice recording. A single pulse of widefield blue light (3–5 ms; 470 nm LED, POLYGON1000, Mightex) was delivered through either a 10× or 40× objective lens centered on the recorded tdT⁺ cells. The intensity of LED was adjusted to stimulate the thalamic terminals, and the range of illumination radiance was 58–116 mW/cm² and 9–15 mW/cm² for 10× and 40×, respectively. To prevent light-evoked polysynaptic responses, 1 µM tetrodotoxin (TTX; Tocris) and 100 µM 4-aminopyridine (4-AP; Tocris) were added to the ACSF for all recordings. For analysis, all traces were filtered using a low-pass Butterworth filter and normalized with baseline correction performed by subtracting the mean signal from the beginning of the recordings to the end of the light-stimulation. The amplitude of the optically-evoked EPSCs (oEPSCs) were measured from the peak of the signal to the corrected baseline. The latency for all oEPSCs was within 10 ms window following photostimulation. At P7–P9, about 70% of recorded cells showed oEPSC (P7: 4/5; P8: 4/6; P9: 5/8 cells; 7 slices from 4 mice), while at P15–P18, the proportion was 100% (P17: 5/5; P18: 4/4 cells; 4 slices from 2 mice).

In vivo calcium imaging

Intracranial virus injection

To enable simultaneous imaging of IC neurons and AC SPNs, Cplx3-CreER;lsl-GCaMP6s pups were injected with AAV9-hSyn-Cre-P2A-dTomato (Addgene #107738-AAV9; titer ~1 × 10¹³ GC/mL) into the left inferior colliculus (IC) at P3. For experiments testing sound-responsiveness of SPNs before ear canal opening (Fig. 7), a separate cohort received injections of AAV9-syn-FLEX-jGCaMP8s (Addgene #162377-AAV9; titer ~1 × 10¹³ GC/mL) into the left auditory cortex at P2. Viral delivery in IC (50–60 nL) and AC (200 nL) were performed as described earlier, similar to injection in auditory thalamus at P1. Successful IC targeting was confirmed by in vivo imaging (Fig. 3) and post hoc histology (Supplementary Fig. 7).

Cranial window implantation

Animals (P7–P21) were anesthetized with inhaled isoflurane (4% induction and 1.5–2% maintenance). After sterilizing the scalp, a midline incision was made to expose the skull. A customized head-post was glued to the skull along the midline. The head was fixed to a stereotaxic platform via the head post, and a small craniotomy (~3 mm diameter) was made with microscissors on the lateral side of the left hemisphere targeting the auditory cortex location. A circular cover glass (3 mm diameter) was placed over the exposed brain and sealed dental cement. Post-surgery, mice recovered in a temperature-controlled environment held at 37 °C.

Simultaneous imaging of AC and IC

Following a 1-h recovery period from anesthesia, mice were secured in a swaddling apparatus made from a modified 15 mL conical centrifuge tube, with the upper portion removed to allow access to the headbar and visualization of the craniotomy. Pups were head-fixed and maintained at 37 °C using a feedback-controlled heating pad (TC-1000; CWE). Pups remained largely immobile throughout imaging, with occasional limb or tail twitches observed.

For widefield epifluorescence imaging of AC and IC regions (Fig. 3), data were acquired at 10 Hz using a Hamamatsu ORCA-Flash4.0 LT digital CMOS camera mounted on a Zeiss Axio Zoom.V16 stereo zoom microscope. For most preparations, viral-mediated fluorescence signals in the IC were much stronger than subplate signals in the AC. In these cases, a neutral density filter with optical density (OD) between 0.2 and 1 was positioned between the midbrain and objective to prevent clipping of IC signals. Control experiments performed in Supplementary Fig. 7 were imaged at 2 or 5 Hz to enhance sensitivity. A 6 × 6 mm field of view (FOV) was imaged at 11× zoom. Images were captured at 512 × 512 pixel resolution (16-bit depth), using 2 × 2 binning. Each imaging session lasted 10 min. Imaging this region also allowed analysis of visual cortex and midbrain, as they were contained within the FOV.

Imaging of cortex through cranial windows

Following a 1-h recovery period from anesthesia, mice were head-fixed beneath a custom widefield microscope and maintained at 37 °C using a feedback-controlled heating pad (Rodent Warmer X1, Stoelting). The cortical surface was illuminated with a 470 nm LED (M470L3, Thorlabs) through a 4× air objective (UPlanSApo 4×/0.16, Olympus) mounted at ~45° from vertical. Excitation light was directed via a long-pass dichroic mirror (MD499-FITC, Thorlabs), and emitted light was filtered (AT535/40 m, Chroma), focused with a tube lens (AC508-150A, Thorlabs), and captured using a pco.edge 4.2 sCMOS camera (Excelitas Technologies). The camera and sound stimuli were synchronized using an external trigger (USB-6259, National Instruments). Imaging was performed at 5 or 10 Hz, with 330 × 330 pixel resolution that encompassed the entire cranial window (~3 mm²).

Two-photon calcium imaging was performed using a Thorlabs Bergamo microscope equipped with a resonant-galvo scanner and a Coherent Discovery laser tuned to 920 nm. Images were acquired at ~2 or ~4 Hz (averaging 16 frames or 8 frames, respectively) with a resolution of 512 × 512 pixels using a 16× Nikon water-immersion objective (0.8 NA), yielding a ~1 × 1 mm² field of view at 0.6× zoom. A low zoom setting was used to expand the scan field for large-scale topographic mapping, resulting in vignetting at the corners due to overfilling the objective’s field of view. Imaging was performed using ThorImageLS 4.1 software. During recordings, animals were head-fixed in a natural supine position using a custom 3D printed headbar and clamp system. The microscope body was rotated 50–60° relative to vertical to match the cortical surface

L1 imaging was conducted 40–60 µm below the pial surface using average laser powers of 20–40 mW, while SPNs were imaged at depths of 700–800 µm using up to average laser power of 250 mW. Despite the higher power used at depth, the large ~1 mm² scan area significantly reduced local power density. No histological evidence of laser-induced damage, tissue burning, or sustained calcium loading was observed following any imaging session.

Bilateral cochlear ablation and enucleation

To eliminate peripheral sensory input, mice underwent bilateral enucleation and cochlear ablation under deep anesthesia following an initial imaging session. For bilateral enucleation, eyeballs were surgically excised from the orbit, followed by the application of a topical hemostatic agent and wound closure. Bilateral cochlear ablation was performed via a retroauricular approach. Following a post-auricular incision and removal of the bulla, the cochlea was mechanically disrupted using a fine surgical pick and aspirated. Following both procedures, the surgical sites were bonded with surgical adhesive, and subjects were monitored closely for recovery, receiving post-operative analgesia to ensure comfort and stability. We performed post-ablation imaging sessions ~45 min after removal from anesthesia.

Analysis

Two-photon calcium imaging preprocessing

Calcium imaging data were preprocessed using a custom Python pipeline to extract normalized fluorescence signals and align neuronal responses to stimulus onset. Raw fluorescence traces were obtained from ROIs automatically segmented by Suite2p⁸⁵.

Neuropil contamination was corrected by subtracting 70% of the surrounding neuropil signal from each ROI. To compute ΔF/F, a rolling percentile baseline was calculated for each ROI using a 100-frame window stepped every 50 frames. Within each window, the 10th percentile of fluorescence values was used as an estimate of the local baseline, and interpolation was used to match the baseline length to the original trace. ΔF/F was then calculated as:

Δ F / F = \frac{F - F_{0}}{F_{0} + ε}

where F₀ is the baseline trace, and ε = 20 is a small constant added for numerical stability, as subplate signals were often near the lower limit of the detectable range. No additional corrections were made for photobleaching.

To extract trial-aligned responses, ΔF/F traces were segmented into epochs surrounding each stimulus presentation. Trial onsets were obtained from a stimulus timing file recorded during acquisition. For each neuron, response epochs were extracted spanning a baseline period (4 frames) and post-stimulus period (12 frames) for each frequency, attenuation, and repetition. This resulted in a five-dimensional array of aligned responses with shape (neurons × time × frequency × attenuation × repetitions). This structure was used for subsequent response quantification, tuning analysis, and statistical testing. All data processing was performed using custom scripts in Python using NumPy, SciPy, and Mat73, and saved in MATLAB-compatible format for downstream analysis.

Correlation maps

For simultaneous imaging of AC and IC, correlation maps were generated by using the signal of the IC as a seed for correlations across the rest of the image. Briefly, a circular ROI (radius = 25 pixels) was placed over the IC, and the signal was computed as the mean of all pixels within that ROI. Then, for every pixel, a Pearson correlation coefficient was calculated, and the resulting matrix plotted as a heat map, with red colors indicating high correlation and blue colors indicating low correlation (Fig. 3e).

These correlation maps guided selection of AC and VC ROIs for analysis in Fig. 3f. Subsequent ROI analysis were performed with a zero-lag Pearson correlation coefficient, as this lag yielded the highest correlation values across all n = 10 mice. The AC ROI was consistently placed over a ~1.5 mm circular region in the lateral portion of the cortex that exhibits strong correlation with the IC seed across all animals²². The VC ROI was positioned medial and slightly posterior to this region. To assess the statistical significance of observed correlations, a shuffle control was performed by generating a null distribution for each ROI. Specifically, the AC and VC traces were circularly permuted across their full duration to preserve their amplitude distribution while disrupting its temporal alignment with IC activity. For each shuffled iteration, correlation values were computed between the shuffled traces and the IC seed, and the resulting values for each ROI were collected to form the null distribution. The 95th percentile of each null distribution was then used to assess whether observed correlation values exceeded chance-level expectations (Fig. 3f).

A similar approach was used to generate correlation maps for both widefield and two-photon datasets. For each dataset, seed ROIs were defined, and pixelwise correlation coefficients were computed to produce spatial maps of functional connectivity.

For widefield auditory datasets, the cortical surface was segmented using k-means clustering (n = 5 clusters). In all animals, this identified a consistent region of interest (~1.5 mm diameter) corresponding to the auditory cortex. Seed points were then positioned at standardized spatial offsets relative to the cluster centroid. Specifically, three seeds were arranged in a diagonal orientation: a primary seed was placed at the centroid’s AP position but shifted ~150 μm dorsally. Relative to this primary seed, a second seed was shifted 145 μm dorsally and 290 μm anteriorly, while a third seed was placed at a posterior offset (145 μm ventrally and 290 μm posteriorly). These seeds correspond to future middle, high, and low frequency responsive areas of A1^22,86. This standardized analysis was applied to both Cplx3-CreER; lsl-GCaMP6s and Thy1-GCamP6s mice.

For visual cortex datasets, functional connectivity was mapped using a standardized five-point seed constellation. This configuration was referenced to a fixed coordinate at the posterolateral boundary of the primary visual area as defined by visualization of retinal waves within this region. Individual seed locations were defined by the following offsets (AP:X/DV:Y, with negative values indicating anterior/dorsal directions) from this reference point: Seed 1: −551 μm/−145 μm, Seed 2: −1015 μm/−580 μm, Seed 3: −493 μm /−725 μm, and Seed 4: −58 μm/−522 μm. During analysis, this constellation was manually aligned to the cortical surface for each preparation until the resulting pixel-wise correlation maps resolved the characteristic “teardrop” functional architecture of the neonatal V1²³. This alignment procedure ensured that topographic sampling remained consistent across surgical preparations, accounting for minor variations in cranial window placement.

For visualization of multiple correlation maps, we used a “max merge” approach in which, at each pixel, the maximum correlation value across all maps was retained while lower values were discarded. This allowed for clearer display of regions showing strong correlations in at least one condition, without inflating background values as would occur in an additive merge. Unless otherwise stated, correlation maps shown in figures were normalized by rescaling values across the image to span the range from 0 to 1.

Correlation as a function of distance

To examine the spatial structure of spontaneous activity in two-photon imaging datasets, we computed pairwise Pearson correlation coefficients between individual pixels and analyzed how correlation strength varied with distance. Image stacks were acquired at 2 Hz and downsampled to 100 × 100 pixels to spatially bin the data. A fluorescence trace for each downsampled pixel was extracted and reshaped into a two-dimensional array (time × pixel), and the first 1200 frames (10 min) were used for analysis. A pairwise Euclidean distance matrix was calculated for all pixel pairs, scaled to the physical dimensions of the imaging field. Pearson correlation coefficients were computed between all pixel pairs, yielding a symmetric correlation matrix. To restrict analysis to a central region of the field of view and avoid edge artifacts, a circular mask (radius = 50 pixels) was applied, and only pixel pairs within this mask were considered. To examine how correlation decayed with distance, pairwise correlations were grouped into 100 µm-wide distance bins. For each bin, the 95th percentile correlation coefficient was computed from all pixel pairs whose spatial separation fell within the bin; subsequent analyses normalized these correlation values to the 100 μm bin.

Sound-evoked two-photon imaging

Auditory stimuli consisted of sinusoidal amplitude-modulated tones (500 ms or 1 s duration, 10 Hz modulation frequency) with a 10 ms cosine ramp applied at onset and offset to minimize spectral artifacts. Tones were presented at frequencies ranging from 4 to 64 kHz and at sound pressure levels (SPLs) of 50–70 dB SPL, covering the typical hearing range of mice. Stimuli were generated using custom MATLAB scripts and delivered via a Tucker-Davis Technologies RX6 multifunction processor. Sounds were played through an electrostatic speaker (ES1) and driver (ED1; Tucker-Davis Technologies) positioned ~10 cm from the contralateral (right) ear. Output levels for each tone frequency were calibrated to a 70 dB SPL reference using a Brüel & Kjær ultrasonic microphone (Type 4939) and preamplifier, ensuring flat spectral output across the tested range.

To assess whether Cplx3⁺ SPNs exhibit sound-driven responses prior to ear canal opening, we presented sound stimuli consisting of a 1-s pure tone (8 kHz, 90 dB SPL) or broadband white noise (90 dB SPL), delivered 50 times with a 30-s interstimulus interval. Image stacks were acquired at ~4 Hz. Fluorescence traces were extracted from neuronal regions of interest (ROIs) automatically detected by suite2p⁸⁵, and ΔF/F was calculated as described above. A one-way repeated-measures ANOVA was performed for each neuron, comparing baseline activity (frames 0–3) to two test windows: an onset window (frames 6–9) and an offset window (frames 10–13). These windows were empirically selected based on the response kinetics observed in positive-control hearing animals (P13–14). To control for multiple comparisons (i.e. multiple neurons within an imaging session), p-values were corrected using the Benjamini–Hochberg false discovery rate (FDR) procedure with a threshold of q < 0.2. In Supplementary Fig. 8, we present nominal p-value distributions for all neurons examined with and without neuropil subtraction and epsilon correction.

To assess the spatial organization of frequency tuning in SPNs, mice were presented with a series of tone pips (1-s duration) across a range of frequencies (4–64 kHz with ½ octave spacing) and sound levels (50–70 dB SPL). Image stacks were acquired at ~4 Hz. For each neuron, the best frequency (BF) was defined as the tone frequency that elicited the maximum average ΔF/F response across all sound levels, while the characteristic frequency (CF) was defined as the frequency that elicited the strongest response at the lowest sound level. BF and CF distributions were computed across neurons within each animal and across animals.

To visualize spatial organization, CF maps were generated by assigning each neuron a color corresponding to its CF and plotting its spatial location. To account for the inherent biological variability in auditory cortex morphology and cranial window location across animals, we employed a functional registration protocol. For each animal, the field of view was aligned to a common coordinate system by calculating the spatial centroid of all neurons with low-frequency CFs (<8 kHz). Each map was then registered via rigid translation to align these functional centroids. To assess tonotopic gradients, the absolute distance of each neuron from the low-frequency centroid was calculated and compared across neurons with different CFs. Local map structure was quantified by computing the standard deviation of CFs within 100 μm circular neighborhoods for each neuron, providing a measure of local CF dispersion. In addition, the spatial spread of tone-evoked activity was quantified for each frequency by calculating the standard deviation of each responsive neuron’s location in the anterior-posterior direction. To compare subplate maps to those of L4 neurons, we performed parallel two-photon imaging in age-matched Thy1-GCaMP6s mice at a distance 300 μm from the pial surface. All analyses, including CF assignment, spatial registration, local dispersion, and response spread, were applied identically to both datasets.

Quantification and statistical analysis

Statistical analyses were performed in Python or MATLAB. Data are presented as mean ± standard deviation (SD) unless otherwise indicated. For group comparisons of normally distributed data, unpaired or paired two-tailed t-tests were used as appropriate. For non-normally distributed data, Mann–Whitney U tests were applied. Normality was assessed using the Shapiro–Wilk test. For within-subject designs with multiple conditions, repeated measures or mixed-design ANOVA was used, followed by post hoc t-tests when appropriate. For multiple group comparisons, p-values were corrected for false discovery rate (FDR) using the Benjamini–Hochberg procedure (FDR = 0.05).

Asterisks in figures denote Benjamini–Hochberg corrected p-values unless otherwise noted: p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***). Individual data points represent biological replicates (e.g., mice) unless otherwise specified in the figure legends. Both male and female mice were used; however, sex was not included as a factor in statistical analyses as the study was not powered to detect sex-specific differences. Linear mixed-effects models were implemented using the statsmodels package in Python, with mouse ID treated as a random factor. Effect sizes (e.g., η², β coefficients) are reported where relevant. Sample sizes and number of replicates are listed in the figure legends. No statistical methods were used to predetermine sample sizes, but values were consistent with those used in the field.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(1.7MB, pdf)}

41467_2026_77454_MOESM2_ESM.pdf^{(5.8KB, pdf)}

Description of Additional Supplementary Files

Supplementary Video 1^{(28.8MB, mp4)}

Supplementary Video 2^{(10.1MB, mp4)}

Supplementary Video 3^{(27.3MB, mp4)}

Supplementary Video 4^{(26.3MB, mp4)}

Supplementary Video 5^{(25.5MB, mp4)}

Supplementary Video 6^{(12.7MB, mp4)}

Reporting Summary^{(107.7KB, pdf)}

Transparent Peer Review File^{(334KB, pdf)}

Source data

Source Data^{(408.1KB, xlsx)}

Acknowledgements

We thank the members of the Kanold laboratory for valuable discussions on the project. We also thank the Johns Hopkins Transgenic Core Laboratory, led by Chip Hawkins, for their support in generating the Cplx3-CreER mouse line.

Author contributions

T.A.B. and M.C. contributed equally to this work. T.A.B., M.C., and P.O.K. conceived the project and designed the experiments. T.A.B. and M.C. developed the methodology and performed formal analysis. T.A.B., M.C., R.J., M.C.O., J.M., Y.C., L.C., and D.G. performed the experiments. T.A.B. and M.C. developed software for data analysis. T.A.B., M.C., and P.O.K. supervised the research. T.A.B. and P.O.K. acquired funding. T.A.B., M.C., and P.O.K. wrote the original draft. T.A.B., M.C., P.O.K., M.C.O., J.M., and Y.C. reviewed and edited the manuscript.

Peer review

Peer review information

Nature Communications thanks and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by the National Institutes of Health (R01DC009607 to P.O.K.) and an NIH NRSA F32 fellowship (F32DC019842 to T.A.B.).

Data availability

Two-photon and widefield calcium imaging ROI traces and other data generated in this study have been deposited in the Johns Hopkins Research Data Repository (https://doi.org/10.7281/T1GMWOHB). Source data are provided with this paper.

Code availability

All custom code used for data analysis has been deposited at the Johns Hopkins Research Data Repository (https://doi.org/10.7281/T1GMWOHB) and is publicly available.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Travis A. Babola, Minzi Chang.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77454-8.

References

1.Penfield, W. & Boldrey, E. Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain60, 389–443 (1937). [Google Scholar]
2.Kalatsky, V. A. & Stryker, M. P. New paradigm for optical imaging: temporally encoded maps of intrinsic signal. Neuron38, 529–545, 10.1016/s0896-6273(03)00286-1 (2003). [DOI] [PubMed] [Google Scholar]
3.Schuett, S., Bonhoeffer, T. & Hubener, M. Mapping retinotopic structure in mouse visual cortex with optical imaging. J. Neurosci.22, 6549–6559, 10.1523/JNEUROSCI.22-15-06549.2002 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Petersen, C. C. The functional organization of the barrel cortex. Neuron56, 339–355, 10.1016/j.neuron.2007.09.017 (2007). [DOI] [PubMed] [Google Scholar]
5.Woolsey, T. A. & Van der Loos, H. The structural organization of layer IV in the somatosensory region (SI) of mouse cerebral cortex. The description of a cortical field composed of discrete cytoarchitectonic units. Brain Res.17, 205–242, 10.1016/0006-8993(70)90079-x (1970). [DOI] [PubMed] [Google Scholar]
6.Stiebler, I., Neulist, R., Fichtel, I. & Ehret, G. The auditory cortex of the house mouse: left-right differences, tonotopic organization and quantitative analysis of frequency representation. J. Comp. Physiol. A181, 559–571, 10.1007/s003590050140 (1997). [DOI] [PubMed] [Google Scholar]
7.Issa, J. B. et al. Multiscale optical Ca²⁺ imaging of tonal organization in mouse auditory cortex. Neuron83, 944–959, 10.1016/j.neuron.2014.07.009 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Bandyopadhyay, S., Shamma, S. A. & Kanold, P. O. Dichotomy of functional organization in the mouse auditory cortex. Nat. Neurosci.13, 361–368, 10.1038/nn.2490 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Kanold, P. O., Nelken, I. & Polley, D. B. Local versus global scales of organization in auditory cortex. Trends Neurosci.37, 502–510, 10.1016/j.tins.2014.06.003 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Patel, G. H., Kaplan, D. M. & Snyder, L. H. Topographic organization in the brain: searching for general principles. Trends Cogn. Sci.18, 351–363, 10.1016/j.tics.2014.03.008 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Kanold, P. O. & Luhmann, H. J. The subplate and early cortical circuits. Annu. Rev. Neurosci.33, 23–48, 10.1146/annurev-neuro-060909-153244 (2010). [DOI] [PubMed] [Google Scholar]
12.Ghosh, A. & Shatz, C. J. Pathfinding and target selection by developing geniculocortical axons. J. Neurosci.12, 39–55, 10.1523/JNEUROSCI.12-01-00039.1992 (1992). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Kostovic, I. & Rakic, P. Developmental history of the transient subplate zone in the visual and somatosensory cortex of the macaque monkey and human brain. J. Comp. Neurol.297, 441–470, 10.1002/cne.902970309 (1990). [DOI] [PubMed] [Google Scholar]
14.De Carlos, J. A. & O’Leary, D. D. Growth and targeting of subplate axons and establishment of major cortical pathways. J. Neurosci.12, 1194–1211, 10.1523/JNEUROSCI.12-04-01194.1992 (1992). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Molnar, Z., Adams, R., Goffinet, A. M. & Blakemore, C. The role of the first postmitotic cortical cells in the development of thalamocortical innervation in the reeler mouse. J. Neurosci.18, 5746–5765, 10.1523/JNEUROSCI.18-15-05746.1998 (1998). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Barkat, T. R., Polley, D. B. & Hensch, T. K. A critical period for auditory thalamocortical connectivity. Nat. Neurosci.14, 1189–1194, 10.1038/nn.2882 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Molnar, Z., Kurotani, T., Higashi, S., Yamamoto, N. & Toyama, K. Development of functional thalamocortical synapses studied with current source-density analysis in whole forebrain slices in the rat. Brain Res. Bull.60, 355–371, 10.1016/s0361-9230(03)00061-3 (2003). [DOI] [PubMed] [Google Scholar]
18.Mrsic-Flogel, T. D. et al. Altered map of visual space in the superior colliculus of mice lacking early retinal waves. J. Neurosci.25, 6921–6928, 10.1523/JNEUROSCI.1555-05.2005 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Xu, H. P. et al. An instructive role for patterned spontaneous retinal activity in mouse visual map development. Neuron70, 1115–1127, 10.1016/j.neuron.2011.04.028 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Kersbergen, C. J., Babola, T. A., Rock, J. & Bergles, D. E. Developmental spontaneous activity promotes formation of sensory domains, frequency tuning and proper gain in central auditory circuits. Cell Rep.41, 111649. 10.1016/j.celrep.2022.111649 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Penn, A. A., Riquelme, P. A., Feller, M. B. & Shatz, C. J. Competition in retinogeniculate patterning driven by spontaneous activity. Science279, 2108–2112, 10.1126/science.279.5359.2108 (1998). [DOI] [PubMed] [Google Scholar]
22.Babola, T. A. et al. Homeostatic control of spontaneous activity in the developing auditory system. Neuron99, 511–524 e515, 10.1016/j.neuron.2018.07.004 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Ackman, J. B., Burbridge, T. J. & Crair, M. C. Retinal waves coordinate patterned activity throughout the developing visual system. Nature490, 219–225, 10.1038/nature11529 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Molnar, Z., Luhmann, H. J. & Kanold, P. O. Transient cortical circuits match spontaneous and sensory-driven activity during development. Science370. 10.1126/science.abb2153. (2020). [DOI] [PMC free article] [PubMed]
25.Ghosh, A. & Shatz, C. J. Involvement of subplate neurons in the formation of ocular dominance columns. Science255, 1441–1443, 10.1126/science.1542795 (1992). [DOI] [PubMed] [Google Scholar]
26.Ghosh, A. & Shatz, C. J. A role for subplate neurons in the patterning of connections from thalamus to neocortex. Development117, 1031–1047, 10.1242/dev.117.3.1031 (1993). [DOI] [PubMed] [Google Scholar]
27.Ghosh, A. & Shatz, C. J. Segregation of geniculocortical afferents during the critical period: a role for subplate neurons. J. Neurosci.14, 3862–3880, 10.1523/JNEUROSCI.14-06-03862.1994 (1994). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Kanold, P. O., Kara, P., Reid, R. C. & Shatz, C. J. Role of subplate neurons in functional maturation of visual cortical columns. Science301, 521–525, 10.1126/science.1084152 (2003). [DOI] [PubMed] [Google Scholar]
29.Kanold, P. O. & Shatz, C. J. Subplate neurons regulate maturation of cortical inhibition and outcome of ocular dominance plasticity. Neuron51, 627–638, 10.1016/j.neuron.2006.07.008 (2006). [DOI] [PubMed] [Google Scholar]
30.Tolner, E. A., Sheikh, A., Yukin, A. Y., Kaila, K. & Kanold, P. O. Subplate neurons promote spindle bursts and thalamocortical patterning in the neonatal rat somatosensory cortex. J. Neurosci.32, 692–702, 10.1523/JNEUROSCI.1538-11.2012 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Wess, J. M., Isaiah, A., Watkins, P. V. & Kanold, P. O. Subplate neurons are the first cortical neurons to respond to sensory stimuli. Proc. Natl. Acad. Sci. USA114, 12602–12607, 10.1073/pnas.1710793114 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Viswanathan, S., Sheikh, A., Looger, L. L. & Kanold, P. O. Molecularly defined subplate neurons project both to thalamocortical recipient layers and thalamus. Cereb. Cortex27, 4759–4768, 10.1093/cercor/bhw271 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Chang, M., Nehs, S., Xu, Z. & Kanold, P. O. Distinct distribution of subplate neuron subtypes between the sensory cortices during the early postnatal period. J. Comp. Neurol.532, e25594. 10.1002/cne.25594 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Hoerder-Suabedissen, A. et al. Expression profiling of mouse subplate reveals a dynamic gene network and disease association with autism and schizophrenia. Proc. Natl. Acad. Sci. USA110, 3555–3560, 10.1073/pnas.1218510110 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Wang, W. Z. et al. Subplate in the developing cortex of mouse and human. J. Anat.217, 368–380, 10.1111/j.1469-7580.2010.01274.x (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Hoerder-Suabedissen, A. & Molnar, Z. Molecular diversity of early-born subplate neurons. Cereb. Cortex23, 1473–1483, 10.1093/cercor/bhs137 (2013). [DOI] [PubMed] [Google Scholar]
37.Hoerder-Suabedissen, A. et al. Temporal origin of mouse claustrum and development of its cortical projections. Cereb. Cortex33, 3944–3959, 10.1093/cercor/bhac318 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Grimstvedt, J. S. et al. A multifaceted architectural framework of the mouse claustrum complex. J. Comp. Neurol.531, 1772–1795, 10.1002/cne.25539 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Zolnik, T. A. et al. Layer 6b is driven by intracortical long-range projection neurons. Cell Rep.30, 3492–3505 e3495, 10.1016/j.celrep.2020.02.044 (2020). [DOI] [PubMed] [Google Scholar]
40.Hoerder-Suabedissen, A. & Molnar, Z. Morphology of mouse subplate cells with identified projection targets changes with age. J. Comp. Neurol.520, 174–185, 10.1002/cne.22725 (2012). [DOI] [PubMed] [Google Scholar]
41.Liao, C. C. & Lee, L. J. Evidence for structural and functional changes of subplate neurons in developing rat barrel cortex. Brain Struct. Funct.217, 275–292, 10.1007/s00429-011-0354-5 (2012). [DOI] [PubMed] [Google Scholar]
42.Zhao, C., Kao, J. P. & Kanold, P. O. Functional excitatory microcircuits in neonatal cortex connect thalamus and layer 4. J. Neurosci.29, 15479–15488, 10.1523/JNEUROSCI.4471-09.2009 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Ehret, G. Development of absolute auditory thresholds in the house mouse (Mus musculus). J. Am. Audiol. Soc.1, 179–184 (1976). [PubMed] [Google Scholar]
44.Kandler, K. & Friauf, E. Development of electrical membrane properties and discharge characteristics of superior olivary complex neurons in fetal and postnatal rats. Eur. J. Neurosci.7, 1773–1790, 10.1111/j.1460-9568.1995.tb00697.x (1995). [DOI] [PubMed] [Google Scholar]
45.Scott, L. L., Mathews, P. J. & Golding, N. L. Posthearing developmental refinement of temporal processing in principal neurons of the medial superior olive. J. Neurosci.25, 7887–7895, 10.1523/JNEUROSCI.1016-05.2005 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Rose, H. J. & Metherate, R. Auditory thalamocortical transmission is reliable and temporally precise. J. Neurophysiol.94, 2019–2030, 10.1152/jn.00860.2004 (2005). [DOI] [PubMed] [Google Scholar]
47.Katz, B. & Miledi, R. The effect of temperature on the synaptic delay at the neuromuscular junction. J. Physiol.181, 656–670, 10.1113/jphysiol.1965.sp007790 (1965). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Markram, H., Lubke, J., Frotscher, M., Roth, A. & Sakmann, B. Physiology and anatomy of synaptic connections between thick tufted pyramidal neurones in the developing rat neocortex. J. Physiol.500, 409–440, 10.1113/jphysiol.1997.sp022031 (1997). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Cruikshank, S. J., Rose, H. J. & Metherate, R. Auditory thalamocortical synaptic transmission in vitro. J. Neurophysiol.87, 361–384, 10.1152/jn.00549.2001 (2002). [DOI] [PubMed] [Google Scholar]
50.Tritsch, N. X. et al. Calcium action potentials in hair cells pattern auditory neuron activity before hearing onset. Nat. Neurosci.13, 1050–1052, 10.1038/nn.2604 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Tritsch, N. X., Yi, E., Gale, J. E., Glowatzki, E. & Bergles, D. E. The origin of spontaneous activity in the developing auditory system. Nature450, 50–55, 10.1038/nature06233 (2007). [DOI] [PubMed] [Google Scholar]
52.Sonntag, M., Englitz, B., Kopp-Scheinpflug, C. & Rubsamen, R. Early postnatal development of spontaneous and acoustically evoked discharge activity of principal cells of the medial nucleus of the trapezoid body: an in vivo study in mice. J. Neurosci.29, 9510–9520, 10.1523/JNEUROSCI.1377-09.2009 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Clancy, B. & Cauller, L. J. Widespread projections from subgriseal neurons (layer VII) to layer I in adult rat cortex. J. Comp. Neurol.407, 275–286 (1999). [PubMed] [Google Scholar]
54.Xue, B., Meng, X., Xu, Y., Kao, J. P. Y. & Kanold, P. O. Transient coupling between infragranular and subplate layers to layer 1 neurons before ear opening and throughout the critical period depends on peripheral activity. J. Neurosci.42, 1702–1718, 10.1523/JNEUROSCI.2505-20.2021 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Hanganu, I. L., Ben-Ari, Y. & Khazipov, R. Retinal waves trigger spindle bursts in the neonatal rat visual cortex. J Neurosci.26, 6728–6736, 10.1523/JNEUROSCI.0752-06.2006 (2006). [DOI] [PMC free article] [PubMed]
56.Blankenship, A. G. & Feller, M. B. Mechanisms underlying spontaneous patterned activity in developing neural circuits. Nat Rev Neurosci.11, 18–29, 10.1038/nrn2759 (2010). [DOI] [PMC free article] [PubMed]
57.Mehra, M., Mukesh, A. & Bandyopadhyay, S. Earliest experience of a relatively rare sound but not a frequent sound causes long-term changes in the adult auditory cortex. J. Neurosci.42, 1454–1476, 10.1523/JNEUROSCI.0431-21.2021 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Mukherjee, D., Meng, X., Kao, J. P. Y. & Kanold, P. O. Impaired hearing and altered subplate circuits during the first and second postnatal weeks of otoferlin-deficient mice. Cereb. Cortex32, 2816–2830, 10.1093/cercor/bhab383 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Makarov, R. et al. Bone conducted responses in the neonatal rat auditory cortex. Sci. Rep.11, 16777. 10.1038/s41598-021-96188-9 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Zhang, Y. et al. Fast and sensitive GCaMP calcium indicators for imaging neural populations. Nature615, 884–891, 10.1038/s41586-023-05828-9 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Herrmann, K., Antonini, A. & Shatz, C. J. Ultrastructural evidence for synaptic interactions between thalamocortical axons and subplate neurons. Eur. J. Neurosci.6, 1729–1742, 10.1111/j.1460-9568.1994.tb00565.x (1994). [DOI] [PubMed] [Google Scholar]
62.Lopez-Bendito, G. & Molnar, Z. Thalamocortical development: how are we going to get there? Nat. Rev. Neurosci.4, 276–289, 10.1038/nrn1075 (2003). [DOI] [PubMed] [Google Scholar]
63.Hoerder-Suabedissen, A. et al. Novel markers reveal subpopulations of subplate neurons in the murine cerebral cortex. Cereb. Cortex19, 1738–1750, 10.1093/cercor/bhn195 (2009). [DOI] [PubMed] [Google Scholar]
64.Osheroff, H. & Hatten, M. E. Gene expression profiling of preplate neurons destined for the subplate: genes involved in transcription, axon extension, neurotransmitter regulation, steroid hormone signaling, and neuronal survival. Cereb. Cortex19, i126–i134, 10.1093/cercor/bhp034 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Hanganu, I. L., Staiger, J. F., Ben-Ari, Y. & Khazipov, R. Cholinergic modulation of spindle bursts in the neonatal rat visual cortex in vivo. J. Neurosci.27, 5694–5705, 10.1523/JNEUROSCI.5233-06.2007 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Hanganu, I. L., Kilb, W. & Luhmann, H. J. Functional synaptic projections onto subplate neurons in neonatal rat somatosensory cortex. J. Neurosci.22, 7165–7176, 10.1523/JNEUROSCI.22-16-07165.2002 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Friauf, E., McConnell, S. K. & Shatz, C. J. Functional synaptic circuits in the subplate during fetal and early postnatal development of cat visual cortex. J. Neurosci.10, 2601–2613, 10.1523/JNEUROSCI.10-08-02601.1990 (1990). [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Chang, M. et al. Specific functional connectivity of molecular subtypes of subplate and layer 6b neurons. J. Neurosci.45. 10.1523/JNEUROSCI.2094-24.2025. (2025). [DOI] [PMC free article] [PubMed]
69.Meng, X., Mukherjee, D., Kao, J. P. Y. & Kanold, P. O. Early peripheral activity alters nascent subplate circuits in the auditory cortex. Sci. Adv.7. 10.1126/sciadv.abc9155. (2021). [DOI] [PMC free article] [PubMed]
70.Anton-Bolanos, N. et al. Prenatal activity from thalamic neurons governs the emergence of functional cortical maps in mice. Science364, 987–990, 10.1126/science.aav7617 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Scala, F. et al. Layer 4 of mouse neocortex differs in cell types and circuit organization between sensory areas. Nat. Commun.10, 4174. 10.1038/s41467-019-12058-z (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Schuman, B., Dellal, S., Pronneke, A., Machold, R. & Rudy, B. Neocortical layer 1: an elegant solution to top-down and bottom-up integration. Annu. Rev. Neurosci.44, 221–252, 10.1146/annurev-neuro-100520-012117 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Takesian, A. E., Bogart, L. J., Lichtman, J. W. & Hensch, T. K. Inhibitory circuit gating of auditory critical-period plasticity. Nat. Neurosci.21, 218–227, 10.1038/s41593-017-0064-2 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Ved, N., Curran, A., Ashcroft, F. M. & Sparrow, D. B. Tamoxifen administration in pregnant mice can be deleterious to both mother and embryo. Lab. Anim.53, 630–633, 10.1177/0023677219856918 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Sun, M. R., Steward, A. C., Sweet, E. A., Martin, A. A. & Lipinski, R. J. Developmental malformations resulting from high-dose maternal tamoxifen exposure in the mouse. PLoS ONE16, e0256299, 10.1371/journal.pone.0256299 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Luskin, M. B. & Shatz, C. J. Studies of the earliest generated cells of the cat’s visual cortex: cogeneration of subplate and marginal zones. J. Neurosci.5, 1062–1075, 10.1523/JNEUROSCI.05-04-01062.1985 (1985). [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Chun, J. J. & Shatz, C. J. Interstitial cells of the adult neocortical white matter are the remnant of the early generated subplate neuron population. J. Comp. Neurol.282, 555–569, 10.1002/cne.902820407 (1989). [DOI] [PubMed] [Google Scholar]
78.Valverde, F., Lopez-Mascaraque, L., Santacana, M. & De Carlos, J. A. Persistence of early-generated neurons in the rodent subplate: assessment of cell death in neocortex during the early postnatal period. J. Neurosci.15, 5014–5024, 10.1523/JNEUROSCI.15-07-05014.1995 (1995). [DOI] [PMC free article] [PubMed] [Google Scholar]
79.Zolnik, T. A. et al. Layer 6b controls brain state via apical dendrites and the higher-order thalamocortical system. Neuron112, 805–820, 10.1016/j.neuron.2023.11.021 (2024). [DOI] [PubMed] [Google Scholar]
80.Garcia, M. M. et al. Noncanonical short-latency auditory pathway directly activates deep cortical layers. Nat. Commun.16, 5911. 10.1038/s41467-025-61020-9 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Babola, T. A., Donovan, N., Darcy, S. S., Spjut, C. D. & Kanold, P. O. Limiting hearing loss in transgenic mouse models. eNeuro12. 10.1523/ENEURO.0465-24.2025. (2025). [DOI] [PMC free article] [PubMed]
82.Kim, S. J. et al. A consensus definition for deep layer 6 excitatory neurons in mouse somatosensory, visual, and motor cortex. Cell Rep.44, 116167. 10.1016/j.celrep.2025.116167 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
83.Chang, M. & Kawai, H. D. A characterization of laminar architecture in mouse primary auditory cortex. Brain Struct. Funct.223, 4187–4209, 10.1007/s00429-018-1744-8 (2018). [DOI] [PubMed] [Google Scholar]
84.Xu, Y. K. T. et al. Brain-wide mapping of oligodendrocyte organization, oligodendrogenesis, and myelin injury. Cell189, 1535–1554, 10.1016/j.cell.2026.01.025 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Pachitariu, M. et al. Suite2p: beyond 10,000 neurons with standard two-photon microscopy. Preprint at bioRxiv, 10.1101/061507. (2017). [DOI]
86.Romero, S. et al. Cellular and widefield imaging of sound frequency organization in primary and higher order fields of the mouse auditory cortex. Cereb. Cortex30, 1603–1622, 10.1093/cercor/bhz190 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(1.7MB, pdf)}

41467_2026_77454_MOESM2_ESM.pdf^{(5.8KB, pdf)}

Description of Additional Supplementary Files

Supplementary Video 1^{(28.8MB, mp4)}

Supplementary Video 2^{(10.1MB, mp4)}

Supplementary Video 3^{(27.3MB, mp4)}

Supplementary Video 4^{(26.3MB, mp4)}

Supplementary Video 5^{(25.5MB, mp4)}

Supplementary Video 6^{(12.7MB, mp4)}

Reporting Summary^{(107.7KB, pdf)}

Transparent Peer Review File^{(334KB, pdf)}

Source Data^{(408.1KB, xlsx)}

Data Availability Statement

All custom code used for data analysis has been deposited at the Johns Hopkins Research Data Repository (https://doi.org/10.7281/T1GMWOHB) and is publicly available.

[CR1] 1.Penfield, W. & Boldrey, E. Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain60, 389–443 (1937). [Google Scholar]

[CR2] 2.Kalatsky, V. A. & Stryker, M. P. New paradigm for optical imaging: temporally encoded maps of intrinsic signal. Neuron38, 529–545, 10.1016/s0896-6273(03)00286-1 (2003). [DOI] [PubMed] [Google Scholar]

[CR3] 3.Schuett, S., Bonhoeffer, T. & Hubener, M. Mapping retinotopic structure in mouse visual cortex with optical imaging. J. Neurosci.22, 6549–6559, 10.1523/JNEUROSCI.22-15-06549.2002 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Petersen, C. C. The functional organization of the barrel cortex. Neuron56, 339–355, 10.1016/j.neuron.2007.09.017 (2007). [DOI] [PubMed] [Google Scholar]

[CR5] 5.Woolsey, T. A. & Van der Loos, H. The structural organization of layer IV in the somatosensory region (SI) of mouse cerebral cortex. The description of a cortical field composed of discrete cytoarchitectonic units. Brain Res.17, 205–242, 10.1016/0006-8993(70)90079-x (1970). [DOI] [PubMed] [Google Scholar]

[CR6] 6.Stiebler, I., Neulist, R., Fichtel, I. & Ehret, G. The auditory cortex of the house mouse: left-right differences, tonotopic organization and quantitative analysis of frequency representation. J. Comp. Physiol. A181, 559–571, 10.1007/s003590050140 (1997). [DOI] [PubMed] [Google Scholar]

[CR7] 7.Issa, J. B. et al. Multiscale optical Ca²⁺ imaging of tonal organization in mouse auditory cortex. Neuron83, 944–959, 10.1016/j.neuron.2014.07.009 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Bandyopadhyay, S., Shamma, S. A. & Kanold, P. O. Dichotomy of functional organization in the mouse auditory cortex. Nat. Neurosci.13, 361–368, 10.1038/nn.2490 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Kanold, P. O., Nelken, I. & Polley, D. B. Local versus global scales of organization in auditory cortex. Trends Neurosci.37, 502–510, 10.1016/j.tins.2014.06.003 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Patel, G. H., Kaplan, D. M. & Snyder, L. H. Topographic organization in the brain: searching for general principles. Trends Cogn. Sci.18, 351–363, 10.1016/j.tics.2014.03.008 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Kanold, P. O. & Luhmann, H. J. The subplate and early cortical circuits. Annu. Rev. Neurosci.33, 23–48, 10.1146/annurev-neuro-060909-153244 (2010). [DOI] [PubMed] [Google Scholar]

[CR12] 12.Ghosh, A. & Shatz, C. J. Pathfinding and target selection by developing geniculocortical axons. J. Neurosci.12, 39–55, 10.1523/JNEUROSCI.12-01-00039.1992 (1992). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Kostovic, I. & Rakic, P. Developmental history of the transient subplate zone in the visual and somatosensory cortex of the macaque monkey and human brain. J. Comp. Neurol.297, 441–470, 10.1002/cne.902970309 (1990). [DOI] [PubMed] [Google Scholar]

[CR14] 14.De Carlos, J. A. & O’Leary, D. D. Growth and targeting of subplate axons and establishment of major cortical pathways. J. Neurosci.12, 1194–1211, 10.1523/JNEUROSCI.12-04-01194.1992 (1992). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Molnar, Z., Adams, R., Goffinet, A. M. & Blakemore, C. The role of the first postmitotic cortical cells in the development of thalamocortical innervation in the reeler mouse. J. Neurosci.18, 5746–5765, 10.1523/JNEUROSCI.18-15-05746.1998 (1998). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Barkat, T. R., Polley, D. B. & Hensch, T. K. A critical period for auditory thalamocortical connectivity. Nat. Neurosci.14, 1189–1194, 10.1038/nn.2882 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Molnar, Z., Kurotani, T., Higashi, S., Yamamoto, N. & Toyama, K. Development of functional thalamocortical synapses studied with current source-density analysis in whole forebrain slices in the rat. Brain Res. Bull.60, 355–371, 10.1016/s0361-9230(03)00061-3 (2003). [DOI] [PubMed] [Google Scholar]

[CR18] 18.Mrsic-Flogel, T. D. et al. Altered map of visual space in the superior colliculus of mice lacking early retinal waves. J. Neurosci.25, 6921–6928, 10.1523/JNEUROSCI.1555-05.2005 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Xu, H. P. et al. An instructive role for patterned spontaneous retinal activity in mouse visual map development. Neuron70, 1115–1127, 10.1016/j.neuron.2011.04.028 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Kersbergen, C. J., Babola, T. A., Rock, J. & Bergles, D. E. Developmental spontaneous activity promotes formation of sensory domains, frequency tuning and proper gain in central auditory circuits. Cell Rep.41, 111649. 10.1016/j.celrep.2022.111649 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Penn, A. A., Riquelme, P. A., Feller, M. B. & Shatz, C. J. Competition in retinogeniculate patterning driven by spontaneous activity. Science279, 2108–2112, 10.1126/science.279.5359.2108 (1998). [DOI] [PubMed] [Google Scholar]

[CR22] 22.Babola, T. A. et al. Homeostatic control of spontaneous activity in the developing auditory system. Neuron99, 511–524 e515, 10.1016/j.neuron.2018.07.004 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Ackman, J. B., Burbridge, T. J. & Crair, M. C. Retinal waves coordinate patterned activity throughout the developing visual system. Nature490, 219–225, 10.1038/nature11529 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Molnar, Z., Luhmann, H. J. & Kanold, P. O. Transient cortical circuits match spontaneous and sensory-driven activity during development. Science370. 10.1126/science.abb2153. (2020). [DOI] [PMC free article] [PubMed]

[CR25] 25.Ghosh, A. & Shatz, C. J. Involvement of subplate neurons in the formation of ocular dominance columns. Science255, 1441–1443, 10.1126/science.1542795 (1992). [DOI] [PubMed] [Google Scholar]

[CR26] 26.Ghosh, A. & Shatz, C. J. A role for subplate neurons in the patterning of connections from thalamus to neocortex. Development117, 1031–1047, 10.1242/dev.117.3.1031 (1993). [DOI] [PubMed] [Google Scholar]

[CR27] 27.Ghosh, A. & Shatz, C. J. Segregation of geniculocortical afferents during the critical period: a role for subplate neurons. J. Neurosci.14, 3862–3880, 10.1523/JNEUROSCI.14-06-03862.1994 (1994). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Kanold, P. O., Kara, P., Reid, R. C. & Shatz, C. J. Role of subplate neurons in functional maturation of visual cortical columns. Science301, 521–525, 10.1126/science.1084152 (2003). [DOI] [PubMed] [Google Scholar]

[CR29] 29.Kanold, P. O. & Shatz, C. J. Subplate neurons regulate maturation of cortical inhibition and outcome of ocular dominance plasticity. Neuron51, 627–638, 10.1016/j.neuron.2006.07.008 (2006). [DOI] [PubMed] [Google Scholar]

[CR30] 30.Tolner, E. A., Sheikh, A., Yukin, A. Y., Kaila, K. & Kanold, P. O. Subplate neurons promote spindle bursts and thalamocortical patterning in the neonatal rat somatosensory cortex. J. Neurosci.32, 692–702, 10.1523/JNEUROSCI.1538-11.2012 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Wess, J. M., Isaiah, A., Watkins, P. V. & Kanold, P. O. Subplate neurons are the first cortical neurons to respond to sensory stimuli. Proc. Natl. Acad. Sci. USA114, 12602–12607, 10.1073/pnas.1710793114 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Viswanathan, S., Sheikh, A., Looger, L. L. & Kanold, P. O. Molecularly defined subplate neurons project both to thalamocortical recipient layers and thalamus. Cereb. Cortex27, 4759–4768, 10.1093/cercor/bhw271 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Chang, M., Nehs, S., Xu, Z. & Kanold, P. O. Distinct distribution of subplate neuron subtypes between the sensory cortices during the early postnatal period. J. Comp. Neurol.532, e25594. 10.1002/cne.25594 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Hoerder-Suabedissen, A. et al. Expression profiling of mouse subplate reveals a dynamic gene network and disease association with autism and schizophrenia. Proc. Natl. Acad. Sci. USA110, 3555–3560, 10.1073/pnas.1218510110 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Wang, W. Z. et al. Subplate in the developing cortex of mouse and human. J. Anat.217, 368–380, 10.1111/j.1469-7580.2010.01274.x (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Hoerder-Suabedissen, A. & Molnar, Z. Molecular diversity of early-born subplate neurons. Cereb. Cortex23, 1473–1483, 10.1093/cercor/bhs137 (2013). [DOI] [PubMed] [Google Scholar]

[CR37] 37.Hoerder-Suabedissen, A. et al. Temporal origin of mouse claustrum and development of its cortical projections. Cereb. Cortex33, 3944–3959, 10.1093/cercor/bhac318 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Grimstvedt, J. S. et al. A multifaceted architectural framework of the mouse claustrum complex. J. Comp. Neurol.531, 1772–1795, 10.1002/cne.25539 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Zolnik, T. A. et al. Layer 6b is driven by intracortical long-range projection neurons. Cell Rep.30, 3492–3505 e3495, 10.1016/j.celrep.2020.02.044 (2020). [DOI] [PubMed] [Google Scholar]

[CR40] 40.Hoerder-Suabedissen, A. & Molnar, Z. Morphology of mouse subplate cells with identified projection targets changes with age. J. Comp. Neurol.520, 174–185, 10.1002/cne.22725 (2012). [DOI] [PubMed] [Google Scholar]

[CR41] 41.Liao, C. C. & Lee, L. J. Evidence for structural and functional changes of subplate neurons in developing rat barrel cortex. Brain Struct. Funct.217, 275–292, 10.1007/s00429-011-0354-5 (2012). [DOI] [PubMed] [Google Scholar]

[CR42] 42.Zhao, C., Kao, J. P. & Kanold, P. O. Functional excitatory microcircuits in neonatal cortex connect thalamus and layer 4. J. Neurosci.29, 15479–15488, 10.1523/JNEUROSCI.4471-09.2009 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Ehret, G. Development of absolute auditory thresholds in the house mouse (Mus musculus). J. Am. Audiol. Soc.1, 179–184 (1976). [PubMed] [Google Scholar]

[CR44] 44.Kandler, K. & Friauf, E. Development of electrical membrane properties and discharge characteristics of superior olivary complex neurons in fetal and postnatal rats. Eur. J. Neurosci.7, 1773–1790, 10.1111/j.1460-9568.1995.tb00697.x (1995). [DOI] [PubMed] [Google Scholar]

[CR45] 45.Scott, L. L., Mathews, P. J. & Golding, N. L. Posthearing developmental refinement of temporal processing in principal neurons of the medial superior olive. J. Neurosci.25, 7887–7895, 10.1523/JNEUROSCI.1016-05.2005 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Rose, H. J. & Metherate, R. Auditory thalamocortical transmission is reliable and temporally precise. J. Neurophysiol.94, 2019–2030, 10.1152/jn.00860.2004 (2005). [DOI] [PubMed] [Google Scholar]

[CR47] 47.Katz, B. & Miledi, R. The effect of temperature on the synaptic delay at the neuromuscular junction. J. Physiol.181, 656–670, 10.1113/jphysiol.1965.sp007790 (1965). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Markram, H., Lubke, J., Frotscher, M., Roth, A. & Sakmann, B. Physiology and anatomy of synaptic connections between thick tufted pyramidal neurones in the developing rat neocortex. J. Physiol.500, 409–440, 10.1113/jphysiol.1997.sp022031 (1997). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Cruikshank, S. J., Rose, H. J. & Metherate, R. Auditory thalamocortical synaptic transmission in vitro. J. Neurophysiol.87, 361–384, 10.1152/jn.00549.2001 (2002). [DOI] [PubMed] [Google Scholar]

[CR50] 50.Tritsch, N. X. et al. Calcium action potentials in hair cells pattern auditory neuron activity before hearing onset. Nat. Neurosci.13, 1050–1052, 10.1038/nn.2604 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Tritsch, N. X., Yi, E., Gale, J. E., Glowatzki, E. & Bergles, D. E. The origin of spontaneous activity in the developing auditory system. Nature450, 50–55, 10.1038/nature06233 (2007). [DOI] [PubMed] [Google Scholar]

[CR52] 52.Sonntag, M., Englitz, B., Kopp-Scheinpflug, C. & Rubsamen, R. Early postnatal development of spontaneous and acoustically evoked discharge activity of principal cells of the medial nucleus of the trapezoid body: an in vivo study in mice. J. Neurosci.29, 9510–9520, 10.1523/JNEUROSCI.1377-09.2009 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Clancy, B. & Cauller, L. J. Widespread projections from subgriseal neurons (layer VII) to layer I in adult rat cortex. J. Comp. Neurol.407, 275–286 (1999). [PubMed] [Google Scholar]

[CR54] 54.Xue, B., Meng, X., Xu, Y., Kao, J. P. Y. & Kanold, P. O. Transient coupling between infragranular and subplate layers to layer 1 neurons before ear opening and throughout the critical period depends on peripheral activity. J. Neurosci.42, 1702–1718, 10.1523/JNEUROSCI.2505-20.2021 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Hanganu, I. L., Ben-Ari, Y. & Khazipov, R. Retinal waves trigger spindle bursts in the neonatal rat visual cortex. J Neurosci.26, 6728–6736, 10.1523/JNEUROSCI.0752-06.2006 (2006). [DOI] [PMC free article] [PubMed]

[CR56] 56.Blankenship, A. G. & Feller, M. B. Mechanisms underlying spontaneous patterned activity in developing neural circuits. Nat Rev Neurosci.11, 18–29, 10.1038/nrn2759 (2010). [DOI] [PMC free article] [PubMed]

[CR57] 57.Mehra, M., Mukesh, A. & Bandyopadhyay, S. Earliest experience of a relatively rare sound but not a frequent sound causes long-term changes in the adult auditory cortex. J. Neurosci.42, 1454–1476, 10.1523/JNEUROSCI.0431-21.2021 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Mukherjee, D., Meng, X., Kao, J. P. Y. & Kanold, P. O. Impaired hearing and altered subplate circuits during the first and second postnatal weeks of otoferlin-deficient mice. Cereb. Cortex32, 2816–2830, 10.1093/cercor/bhab383 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Makarov, R. et al. Bone conducted responses in the neonatal rat auditory cortex. Sci. Rep.11, 16777. 10.1038/s41598-021-96188-9 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Zhang, Y. et al. Fast and sensitive GCaMP calcium indicators for imaging neural populations. Nature615, 884–891, 10.1038/s41586-023-05828-9 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Herrmann, K., Antonini, A. & Shatz, C. J. Ultrastructural evidence for synaptic interactions between thalamocortical axons and subplate neurons. Eur. J. Neurosci.6, 1729–1742, 10.1111/j.1460-9568.1994.tb00565.x (1994). [DOI] [PubMed] [Google Scholar]

[CR62] 62.Lopez-Bendito, G. & Molnar, Z. Thalamocortical development: how are we going to get there? Nat. Rev. Neurosci.4, 276–289, 10.1038/nrn1075 (2003). [DOI] [PubMed] [Google Scholar]

[CR63] 63.Hoerder-Suabedissen, A. et al. Novel markers reveal subpopulations of subplate neurons in the murine cerebral cortex. Cereb. Cortex19, 1738–1750, 10.1093/cercor/bhn195 (2009). [DOI] [PubMed] [Google Scholar]

[CR64] 64.Osheroff, H. & Hatten, M. E. Gene expression profiling of preplate neurons destined for the subplate: genes involved in transcription, axon extension, neurotransmitter regulation, steroid hormone signaling, and neuronal survival. Cereb. Cortex19, i126–i134, 10.1093/cercor/bhp034 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] 65.Hanganu, I. L., Staiger, J. F., Ben-Ari, Y. & Khazipov, R. Cholinergic modulation of spindle bursts in the neonatal rat visual cortex in vivo. J. Neurosci.27, 5694–5705, 10.1523/JNEUROSCI.5233-06.2007 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Hanganu, I. L., Kilb, W. & Luhmann, H. J. Functional synaptic projections onto subplate neurons in neonatal rat somatosensory cortex. J. Neurosci.22, 7165–7176, 10.1523/JNEUROSCI.22-16-07165.2002 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] 67.Friauf, E., McConnell, S. K. & Shatz, C. J. Functional synaptic circuits in the subplate during fetal and early postnatal development of cat visual cortex. J. Neurosci.10, 2601–2613, 10.1523/JNEUROSCI.10-08-02601.1990 (1990). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR68] 68.Chang, M. et al. Specific functional connectivity of molecular subtypes of subplate and layer 6b neurons. J. Neurosci.45. 10.1523/JNEUROSCI.2094-24.2025. (2025). [DOI] [PMC free article] [PubMed]

[CR69] 69.Meng, X., Mukherjee, D., Kao, J. P. Y. & Kanold, P. O. Early peripheral activity alters nascent subplate circuits in the auditory cortex. Sci. Adv.7. 10.1126/sciadv.abc9155. (2021). [DOI] [PMC free article] [PubMed]

[CR70] 70.Anton-Bolanos, N. et al. Prenatal activity from thalamic neurons governs the emergence of functional cortical maps in mice. Science364, 987–990, 10.1126/science.aav7617 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR71] 71.Scala, F. et al. Layer 4 of mouse neocortex differs in cell types and circuit organization between sensory areas. Nat. Commun.10, 4174. 10.1038/s41467-019-12058-z (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR72] 72.Schuman, B., Dellal, S., Pronneke, A., Machold, R. & Rudy, B. Neocortical layer 1: an elegant solution to top-down and bottom-up integration. Annu. Rev. Neurosci.44, 221–252, 10.1146/annurev-neuro-100520-012117 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR73] 73.Takesian, A. E., Bogart, L. J., Lichtman, J. W. & Hensch, T. K. Inhibitory circuit gating of auditory critical-period plasticity. Nat. Neurosci.21, 218–227, 10.1038/s41593-017-0064-2 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR74] 74.Ved, N., Curran, A., Ashcroft, F. M. & Sparrow, D. B. Tamoxifen administration in pregnant mice can be deleterious to both mother and embryo. Lab. Anim.53, 630–633, 10.1177/0023677219856918 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR75] 75.Sun, M. R., Steward, A. C., Sweet, E. A., Martin, A. A. & Lipinski, R. J. Developmental malformations resulting from high-dose maternal tamoxifen exposure in the mouse. PLoS ONE16, e0256299, 10.1371/journal.pone.0256299 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR76] 76.Luskin, M. B. & Shatz, C. J. Studies of the earliest generated cells of the cat’s visual cortex: cogeneration of subplate and marginal zones. J. Neurosci.5, 1062–1075, 10.1523/JNEUROSCI.05-04-01062.1985 (1985). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR77] 77.Chun, J. J. & Shatz, C. J. Interstitial cells of the adult neocortical white matter are the remnant of the early generated subplate neuron population. J. Comp. Neurol.282, 555–569, 10.1002/cne.902820407 (1989). [DOI] [PubMed] [Google Scholar]

[CR78] 78.Valverde, F., Lopez-Mascaraque, L., Santacana, M. & De Carlos, J. A. Persistence of early-generated neurons in the rodent subplate: assessment of cell death in neocortex during the early postnatal period. J. Neurosci.15, 5014–5024, 10.1523/JNEUROSCI.15-07-05014.1995 (1995). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR79] 79.Zolnik, T. A. et al. Layer 6b controls brain state via apical dendrites and the higher-order thalamocortical system. Neuron112, 805–820, 10.1016/j.neuron.2023.11.021 (2024). [DOI] [PubMed] [Google Scholar]

[CR80] 80.Garcia, M. M. et al. Noncanonical short-latency auditory pathway directly activates deep cortical layers. Nat. Commun.16, 5911. 10.1038/s41467-025-61020-9 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR81] 81.Babola, T. A., Donovan, N., Darcy, S. S., Spjut, C. D. & Kanold, P. O. Limiting hearing loss in transgenic mouse models. eNeuro12. 10.1523/ENEURO.0465-24.2025. (2025). [DOI] [PMC free article] [PubMed]

[CR82] 82.Kim, S. J. et al. A consensus definition for deep layer 6 excitatory neurons in mouse somatosensory, visual, and motor cortex. Cell Rep.44, 116167. 10.1016/j.celrep.2025.116167 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR83] 83.Chang, M. & Kawai, H. D. A characterization of laminar architecture in mouse primary auditory cortex. Brain Struct. Funct.223, 4187–4209, 10.1007/s00429-018-1744-8 (2018). [DOI] [PubMed] [Google Scholar]

[CR84] 84.Xu, Y. K. T. et al. Brain-wide mapping of oligodendrocyte organization, oligodendrogenesis, and myelin injury. Cell189, 1535–1554, 10.1016/j.cell.2026.01.025 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR85] 85.Pachitariu, M. et al. Suite2p: beyond 10,000 neurons with standard two-photon microscopy. Preprint at bioRxiv, 10.1101/061507. (2017). [DOI]

[CR86] 86.Romero, S. et al. Cellular and widefield imaging of sound frequency organization in primary and higher order fields of the mouse auditory cortex. Cereb. Cortex30, 1603–1622, 10.1093/cercor/bhz190 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

A topographic protomap in mouse subplate neurons precedes the onset of sensory experience

Travis A Babola

Minzi Chang

Rachel Jun

Marina Cardoso de Oliveira

Jessica E Messier

Yunru Chen

Lillian Choi

Dave R Gutierrez

Patrick O Kanold

Abstract

Introduction

Results

Cplx3-CreER mouse line enables specific labeling of SPNs

Fig. 1. Generation and validation of subplate neuron-specific recombination in Cplx3-CreER mice.

Cplx3+ SPNs mature early and receive functional thalamic input before and after ear canal opening

Fig. 2. Cplx3+ SPNs receive functional thalamic input.

Spontaneous activity in Cplx3+ SPNs reflects cochlear-driven input before ear canal opening

Fig. 3. Peripherally generated cochlear spontaneous activity propagates to SPNs.

Early subplate activity reveals a diffuse protomap of tonotopy in auditory cortex

Fig. 4. Correlation structure of subplate spontaneous activity maps future auditory organization in neonatal mice.

Divergent SPN-layer 1 projections mask the subplate protomap

Fig. 5. SPN somata exhibit a tighter correlational structure than their L1 projections.

Retinal waves entrain Cplx3+ SPNs in visual cortex

Fig. 6. Retinal waves traverse SPNs in the primary visual cortex.

Cplx3+ SPNs do not exhibit responses to sounds prior to ear canal opening

Fig. 7. Cplx3+ SPNs do not exhibit sound-evoked responses prior to ear canal opening.

Cplx3+ SPNs exhibit tonotopic organization after ear canal opening

Fig. 8. SPNs exhibit coarse frequency organization in the auditory cortex.

Discussion

Methods

Contact for reagent and resource sharing

Experimental model and subject details

Transgenic animal models

Terminology

Histological preparation and analysis

Definition of subplate/L6b and cell count analysis

Fluorescence intensity analysis

Brain clearing

Electrophysiology

Brain slice preparation and recording conditions

Intrinsic properties

Electrical thalamic stimulation and analysis

Optogenetic thalamic stimulation and analysis

In vivo calcium imaging

Intracranial virus injection

Cranial window implantation

Simultaneous imaging of AC and IC

Imaging of cortex through cranial windows

Bilateral cochlear ablation and enucleation

Analysis

Two-photon calcium imaging preprocessing

Correlation maps

Correlation as a function of distance

Sound-evoked two-photon imaging

Quantification and statistical analysis

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Funding

Data availability

Code availability

Competing interests

Footnotes

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Cplx3⁺ SPNs mature early and receive functional thalamic input before and after ear canal opening

Fig. 2. Cplx3⁺ SPNs receive functional thalamic input.

Spontaneous activity in Cplx3⁺ SPNs reflects cochlear-driven input before ear canal opening

Retinal waves entrain Cplx3⁺ SPNs in visual cortex

Cplx3⁺ SPNs do not exhibit responses to sounds prior to ear canal opening

Fig. 7. Cplx3⁺ SPNs do not exhibit sound-evoked responses prior to ear canal opening.

Cplx3⁺ SPNs exhibit tonotopic organization after ear canal opening