Potential connectivity patterns within the corticostriatal projection.

A rich and overlapping cortical innervation of the striatum (lines) permits different connectivity patterns: Left, synapses are formed promiscuously. In this case convergence is high as each SPN receives broad cortical input, from many cortical origins (colored bars on top). Middle and right, only a fraction of the potential connections are actually formed and, in these cases, convergence is lower. The selected presynaptic neurons are either scattered (middle) or topographically (right) positioned in the cortex. Increasing the selectivity of the corticostriatal connections while keeping their broad origin (middle) could generate maximal heterogeneity in the patterns of input for SPNs sharing the same striatal volume, due to a large number of combinations of cortical origins. In the case of a topographic innervation (right), neighboring SPNs form synapses with presynaptic neurons located in the same cortical region, reducing the local heterogeneity.

A novel slice preparation for investigating the spatial organization of the somatosensory projections to the DLS neurons.

A. A somatosensory corticostriatal slice in which axons from L5 neurons were labeled with biocytin iontophoresis. Black star, the electrode position in L5a. Right, focus on a region in the striatum.

B. Right, Example dendritic arbors of L5 cortical cells labeled with biocytin in the slice.

C. Montage of a corticostriatal slice (left) and layout of the experiment (right). A SPN was recorded in the dorsolateral striatum while cortical neurons were photostimulated with LSPS. The grid of LSPS (blue) was positioned on the barrel cortex. GPe, globus pallidus, external segment; IC, internal capsule; hipp., hippocampus.

D. Top, examples of synaptic input maps for individual SPNs showing one or four clusters of input (SPN 1 and 2, respectively). In the color maps, each pixel of color indicates the peak amplitude of EPSCs detected within a 50 ms window after the stimulus onset. The different cortical layers are represented by solid white vertical lines on the left side of the map. The green boxes at the top of the maps, the clusters of sites in the connectivity maps collapsed in the vertical axis whose stimulation evokes EPSCs. Bottom, EPSC traces evoked at the sites indicated by letters in the maps above. Two repetitions are superimposed (black and orange). Vertical dashed lines, the stimulus onsets (2 ms stimulus).

E. LSPS evoked excitation of two L5a pyramidal neurons recorded in current-clamp mode. LSPS was a 8 x 8 (left) or 24 x 8 (right) 50 μm spacing grid. Cortical neurons are positioned at the center (white triangles). Traces with an action potential are in red. Bottom, Average number of action potentials evoked at every site of the LSPS grid.

F. Overlay of the sites in the barrel cortex (green polygons) where stimulations evoked EPSCs in SPNs (green symbols; n =101 cells, N = 54 mice). The red dashed vertical line is the reference (Refhor in the Methods) used for aligning slices across experiments horizontally.

G. Contribution of each cortical layer to the SPN innervations. The 16 rows of the grid correspond to different cortical layers (Layer 2/3: 1-6; L4: 7-9; L5a: 10; L5b: 11-13; L6: 14-16). The horizontal white band is L5a.

H. Amplitude of SPN EPSCs as a function of their laminar origin. Median (red) and 25-75th percentiles (boxes). * Kruskal-Wallis p = 0.000125, Dunn-Šidák posthoc tests, p = 0.0001 and 0.03415.

I. Bottom, Positions of SPNs in the striatum for each position of connected sites on the horizontal axis of the LSPS grid. Top, maximal distance between SPNs (dSPN) for every connected site position in the barrel cortex on the horizontal axis, binned every 150 µm, in other words the width of the projection zone of one cortical column within the striatum

J. Schematic of the slice with the size of the cortical and striatal regions that are connected by projections to SPNs (pink) and the size of the striatal domain with functional projections from a single cortical column (shades of green).

Sparse functional projections from the barrel cortex to individual SPNs in the DLS.

A. Example slice with 3 recordings. The hatched horizontal band is L5a, the polygons are the stimulation sites evoking EPSCs, the solid boxes on top are the connectivity clusters and the open rectangle the input field in the collapsed connectivity map. The circles at the bottom mark the positions of SPNs on the horizontal axis. Recordings are color coded.

B. Fraction of SPNs receiving inputs from 1 to 6 clusters of projections in the barrel cortex. In inset, median (red) and 25-75th percentiles (box). Clusters are defined as the ensemble of contiguous sites in the connectivity map collapsed in the vertical axis, whose stimulation evokes EPSCs (see examples in A, boxes at the top). n = 101 cells, N = 54 mice.

C. Median (red) and 25-75th percentiles (box) of cluster width and spacing.

D. Left, Fraction of cells receiving inputs from 1 to 8 cortical columns in the barrel area. In inset, median (red) and 25-75th percentiles (box). Right, Number of cortical columns innervating individual SPNs.

Symbols are cells, lines are slices. Light gray symbols, slices with ≥ 2 recordings. Dark symbols, one recording per slice.

E. Fraction of SPNs with input fields from 0.075 to 1.6 mm.

F. Top to bottom, For every input field width, the number of clusters, cluster width (expressed in cortical column and μm) and proportion of connectivity gap, in input fields collapsed along the vertical axis. Gray symbols, cells with ≥ 2 connectivity clusters. Other symbols, cells with 1 cluster.

G. Bottom, width of input fields in the barrel cortex for different positions of SPNs in the striatum. SPNs with an input field the size of a single or of several cortical columns are shown (open and solid symbols, respectively). Top, The average width of input fields as a function of SPN positions binned every 250 µm.

H. Schematized average and principal connectivity patterns. In pie charts, the SPN proportions.

Individual SPNs have unique cortical innervation

A. Four example slices with 2 to 4 recordings. Same as in Fig. 2A.

B. Left, fraction of pairs with overlap in their input field. Right, width of overlap as a function of the horizontal distance separating cells. n = 70 pairs.

C. Same as in B for the fraction of vertical alignment between SPN connectivity maps.

D. Same as in B for the fraction of overlap between SPN connectivity maps.

Topographic organization of the functional vibrissal innervation of SPNs in the DLS

A. Overlay of the sites in the barrel cortex where stimulations evoked EPSCs in SPNs. The colors, yellow to blue, indicate the position of the SPNs (gray circles) along the horizontal axis in the striatum (axis at the bottom, 0 is Refhor in the Methods; n = 101, N = 54). Blue shades are for lateral SPNs.

B. Position of the connectivity center of mass (CM) as a function of the SPN position in dorsal striatum. Zero on the x and y axis is the position of the vertical dashed line shown in A.

C. Ratio of the largest EPSC over the smallest EPSC for each recording. Median (red) and 25-75th percentiles (box).

D. Position of the synaptic input CM as a function of the SPN position in dorsal striatum. In light gray, the connectivity CM.

E. EPSC sum obtained in the strongest connectivity cluster relative to others sum. Cells with ≥ 2 clusters, n = 55.

D1 and D2 SPNs have similar patterns of innervation from the barrel cortex

A. Overlay of the sites in the barrel cortex where stimulations evoked EPSCs in D1 (blue) or D2 SPNs (red). The shades, light to dark, indicate the position of the SPNs (bottom circles) along the medio-lateral axis in the dorsal striatum (axis at the bottom). D1 cells, n = 47, N = 36; D2 cells, n = 54, N = 37.

B. Position of the connectivity center of mass (CM) as a function of the SPN position in the dorsal striatum (D1, red; D2, blue). R = 0.70 (D1, n = 47, N = 36); R = 0.57 (D2, n = 54, N = 37).

C. Left, D2 SPN input field width as a function of the D1 SPN input field width, in the same slice. The black symbol and lines are the median value and 25-75th percentiles. Large symbols indicate n = 2 to 3 pairs. n = 42 pairs. Right, medians (thick lines) and 25-75th percentiles (boxes) of D1 and D2 input field widths, across all cells. n = 101.

D. Same as in C for the number of cortical columns.

E. Same as in C for the width of connectivity clusters.

F. Contribution of each cortical layer to the SPN innervations. The horizontal hatched band is L5a.

G. EPSC sum for D1 and D2 SPNs. Outliers were not shown for clarity (1 D1, 2 D2, 1.2-1.5 nA)

H. EPSC sum obtained from the strongest cluster relative to others. Cells with ≥ 2 clusters. D1, n = 24; D2, n = 30.

I. Top, D1 and D2 SPN EPSC sum as a function of laminar origin. For each layer, only cells with inputs are included. Outliers are not shown for clarity (3 D1, 2 D2, 350-700 pA). * indicates a significant difference (p = 0.005 and p = 0.008; Mann-Whitney). Bottom, fraction of cells with input.

Connectivity pattern within the somatosensory corticostriatal projection to SPNs.

Left, Our results support the model in which each SPN integrates limited and heterogeneous rather than exhaustive inputs transmitted by the barrel cortex, intermediate between the middle and right panels in Fig. 1. Right, Each SPN representation of the whisker cortical columns complements the representations of its neighbors.

A. Left, the corticostriatal slice generated from the Allen mouse brain reference atlas (CutNII custom-angle slice visualization tool; G. Csucs). bf, barrel field. str, striatum. VPL, ventral posterolateral thalamic nucleus. VPM, ventral posteromedial thalamic nucleus. Center, the section angles illustrated on horizontal and longitudinal sections of the brain. Right, the section angle with respect to the orientations of the arcs and rows in the barrel field of the left hemisphere (Allen mouse brain atlas).

B. A simulated somatosensory corticostriatal slice generated using data from the Allen mouse connectivity atlas (https://connectivity.brain-map.org/projection/experiment/cortical_mp/293728197). eGFP was expressed in L2/3 and L5 pyramidal cells principally, in 3-4 arcs of the D and E rows of whisker columns (whole brain visualization in the green box at the bottom left; the dashed lines indicate the slice orientation). Acquisitions of serial two-photon tomography were stacked in Fiji and the reconstituted brain was sectioned using the angles to produce the acute somatosensory corticostriatal slices (BigDataViewer plugging). The red box identifies the “experimental slice” with intact projections from the barrel cortex to the striatum. Fluorescence in the striatum (blue) and in the corpus callosum (cc, pink) was quantified on consecutive max projections (300 µm thick) obtained in the anterior-posterior axis (A-P). Note that slices, being from the opposite hemisphere to our experiments, have the reverse orientation in the medio-lateral axis compared to Fig. 2A,C.

A. Vertical (left) and horizontal (right) excitation profile of L5 pyramidal cells recorded in current clamp, in the juveniles and adolescents combined (20 and 25 mW stimulation, respectively). Evoked action potentials (APs) were summed along each column (left) or line (right) of the LSPS grid (50 µm spacing). Solid line, median. In gray, 25-75th percentiles. n = 30, N = 14.

B. Total number of APs evoked in the stimulation grid for L5 (left) and L2/3 (right) pyramidal cells in juvenile (juv, P22-30) and adolescent (ado, P31-41) mice. A higher stimulation intensity was used for adolescent mice (25 mW instead of the standard -std -intensity, 20 mW), so that the total number of APs matched between the two age groups. L5, juv, n = 22, N = 9; ado std, n = 22, N = 13; ado high: n = 8, N = 5; L2/3, juv: n = 16, N=3; ado std and high n = 10, N = 2; * p < 0.001, Mann-Whitney.