Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2011 Oct 6;72(1):111-23.
doi: 10.1016/j.neuron.2011.07.029.

Long-range Neuronal Circuits Underlying the Interaction Between Sensory and Motor Cortex

Affiliations
Free PMC article

Long-range Neuronal Circuits Underlying the Interaction Between Sensory and Motor Cortex

Tianyi Mao et al. Neuron. .
Free PMC article

Abstract

In the rodent vibrissal system, active sensation and sensorimotor integration are mediated in part by connections between barrel cortex and vibrissal motor cortex. Little is known about how these structures interact at the level of neurons. We used Channelrhodopsin-2 (ChR2) expression, combined with anterograde and retrograde labeling, to map connections between barrel cortex and pyramidal neurons in mouse motor cortex. Barrel cortex axons preferentially targeted upper layer (L2/3, L5A) neurons in motor cortex; input to neurons projecting back to barrel cortex was particularly strong. Barrel cortex input to deeper layers (L5B, L6) of motor cortex, including neurons projecting to the brainstem, was weak, despite pronounced geometric overlap of dendrites with axons from barrel cortex. Neurons in different layers received barrel cortex input within stereotyped dendritic domains. The cortico-cortical neurons in superficial layers of motor cortex thus couple motor and sensory signals and might mediate sensorimotor integration and motor learning.

Figures

Figure 1
Figure 1. Mapping Output from Somatosensory Cortex (Barrel Cortex, vS1) and Vibrissal Motor Cortex (vM1)
(A–C) Viral injections in vS1 and projections to vM1 and other targets. (A) Schematic, injection in vS1 and projection to vM1. (B) Representative images of injections in vS1 and projections to vM1 and other targets. (B1) AAV-tdTomato injected into vS1 (asterisks) and projection to vM1 (arrowhead). Dashed lines correspond to the sections containing the injection site in vS1 (inj) and the projection site in vM1 (proj). (B2) Coronal section through the vS1 injection site (asterisk). Also shown are projections to second somatosensory cortex (S2), thalamus (Th), and ectorhinal/perirhinal cortex (Ect), and fibers passing through the internal capsule (ic). (B3) Coronal section through vM1. (B4) Confocal image of vS1 axons in vM1, overlaid with a bright field image of the brain slice. (C) Fraction of vS1 output to various brain areas, rank-ordered by strength (quantified based on fluorescence; three separate experiments). (D and E) Topographic projections from vS1 to vM1 and S2. (D) Schematic, pairs of injections in vS1and projections to vM1. (E1) AAV-eGFP and AAV-tdTomato injected in nearby parts of vS1 (green, centered on barrel C2; red, centered on barrel C5) (additional examples in Figure S3). The fluorescence image is from a section of flattened cortex, overlaid on a brightfield image showing the cytochrome oxydase stained section to highlight barrels (see Experimental Procedures). (E2) Locations of the projections to vM1 and S2 (dashed boxes). (E3) Contrast-enhanced image showing projections to vM1 (arrowheads indicate split projection). (E4) Contrast-enhanced image showing projections to S2. (F–H) Projections from vM1 to vS1 and other targets. (F) Schematic, injection in vM1 and projection to vS1. (G1) AAV-eGFP injected into vM1 (asterisks) and projection to vS1 (arrowhead). Dashed lines correspond to the sections containing the injection site in vM1 (inj) and the projection site in vS1 (proj). (G2) Coronal section through the injection site (asterisk) and projection to contralateral vM1. (G3) Coronal section showing vS1, S2, and dorsal lateral striatum (Str). (G4) Confocal image showing vM1 axons in vS1. (H) Fraction of vM1 output to various brain areas, rank-ordered by strength (quantified based on fluorescence; two separate experiments). Str, striatum; S2, secondary somatosensory cortex; vM1, primary motor, and some area of FrA; Th, thalamus, including PO, Rt, and VPM in (C) and PO, VA,VL in (H); SC, superior colliculus; Ect, ectorhinal/perirhinal cortex; cvS1, contralateral vS1; ZI, zona incerta; MS1, medial primary sensory cortex, medial to vS1 (also see Figure S1); APT, anterior pretectal nucleus; cEct, contralateral side of ectorhinal/perirhinal cortex; cMS1, contralateral side of medial primary sensory cortex; LPtA, lateral parietal association cortex (see Figure S1); OC, orbital cortex; Re/Rh, both ipsilateral and contralateral sides of reuniens thalamic nucleus and rhomboid thalamic nucleus; DP, infralimbic cortex/dorsal peduncular cortex. S1+S2, primary sensory cortex and also including some of secondary sensory cortex; FrA, frontal associate cortex, also might include some intra-vM1 axons; cvM1, contralateral side of primary motor, and contralateral side of FrA; cStr, contralateral side of striatum; RSA, retrosplenial agranular cortex; cOC, contralateral orbital cortex; cCl, contralateral side of claustrum; *one animal’s data not shown, either because projections could not be clearly segregated from injection site (for MS1 in C) or not visible (for DP in C), or slightly damaged (for Ect in H). Unless explicitly stated, the region of interest is on the same side as the viral infection. See also Figures S1–S4, Table S1, and Movie S1.
Figure 2
Figure 2. Relationship between Laminar Location and Projection Targets in vM1
(A) Bright field image of a vM1 brain slice. The vertical lines demarcate the recording locations. Horizontal lines indicate layer boundaries. (B) A representative retrograde labeling experiment. Fluorescent microbeads were injected in vS1 and imaged in vM1. (C) Fluorescent microbeads were injected in brain areas that are targets of vM1 projections. vS1, black; zona inserta, ZI, magenta; superior colliculus, SC, gray; brain-stem, BS, green; posterior thalamic nucleus, PO, blue; three separate experiments for each target brain region; 15 experiments total. Bead-positive cells were mapped and their density plotted against the relative cortical depth (see Experimental Procedures). Error bars, SEM. See also Figures S4 and S5 and Table S1.
Figure 3
Figure 3. Subcellular Channelrhodopsin-2-Assisted Circuit Mapping (sCRACM)
(A) A vM1 slice overlaid with the photostimulation grid (12 × 26, 50 μm spacing) and the reconstructed dendrites of two sequentially recorded cells (L5A, magenta; L5B, blue) in the same column of the same slice. White fluorescence, vS1 axons. (B) Excitatory postsynaptic currents (EPSCsCRACM) recorded from the L5A cell (magenta in A), evoked by photostimulation on a grid (black traces are reproduced at higher magnification in C). EPSCsCRACM are caused by local depolarization of ChR2-positive axons, triggered by blue light. (C) EPSCsCRACM were reproducible across repetitions (three repetitions; photostimulus locations as for black traces in B). Blue ticks indicate the photostimuli. The blue and gray ticks demarcate the window for calculating the response plotted in sCRACM input maps. (D) sCRACM input maps. Left panel, L5A cell in (A). Right panel, L5B cell in (A). The pixel value is proportional to the strength of input from ChR2-positive axons to particular locations of the dendritic arbor. The triangles indicate the soma locations. Two maps were obtained under the same stimulation and recording conditions. See also Figures S4 and S5.
Figure 4
Figure 4. Input from vS1 to vM1 as a Function of Layer
(A) Representative dendritic arbors in vM1, sorted by depth in the cortex. (B) sCRACM input maps in different layers. The maps were thresholded to show pixels with significant signal (Experimental Procedures). Note differences in color scales. (C–E) Comparison of input using L5A as the reference. A neuron in L2/3, L5B, or L6 was recorded in the same brain slice as a neuron in L5A. Input (summed pixels with significant signal) for L2/3 (C) (n = 19 pairs of cells, 11 mice), L5B (D) (n = 31 pairs of cells, 15 mice), or L6 (E) (n = 21 pairs of cells, 8 mice) neurons is plotted against input to L5A neurons. Statistics, signed-rank test. (F) Summary of pair wise comparisons. The histogram corresponds to the slopes of the regression lines in (C)–(E). See also Figure S6 and Table S1.
Figure 5
Figure 5. Laminar Input to vM1 Neurons from vS1
(A) Averaged sCRACM input maps aligned by pia. Triangles, soma locations (number of cells in each group are noted in each panel; total n = 77 cells, 35 mice). (B) Averaged sCRACM input maps aligned by soma locations. Right, averaged dendritic length density as a function of depth in the cortex (dendritic reconstructions were performed for a subset of cells; L2/3, n = 7; L5A, n = 19; L5B, n = 15; L6, n = 5). The integral of the dendritic density was normalized to 1. See also Figure S6 and Table S1.
Figure 6
Figure 6. Low Input to Brainstem-Projecting L5B Pyramidal Neurons from vS1
(A) Input to upper layer neurons (L2/3 and L5A) compared to L5B neurons (left, actual values; right, normalized) (n = 36 pairs of cells, 17 mice). Statistics, signed-rank test. The circle in the right panel indicates the mean (0.18). (B) Same experiments as (A), except without antagonist cocktails (CPP, TTX, and 4-AP) (n = 23 pairs of cells, 4 mice). Statistics, signed-rank test. The circle in the right panel indicates the mean (0.26). (C and D) Input to upper layer neurons (L2/3 and L5A) compared to brainstem-projecting L5B neurons. (C) Individual maps of a L5A cell (left) and a bead-positive L5B cell (middle). Triangles indicate the soma locations. Both maps were obtained under identical experimental conditions in the same brain slice. An overlay of DIC and red fluorescence shows a bead-positive L5B cell (right). (D) Input to upper layer neurons (L2/3 and L5A) compared to bead-positive L5B neurons. (left, actual values; right, normalized) (n = 8 pairs of cells, 5 mice). Statistics, signed-rank test. The circle in the right panel indicates the mean (0.28). See also Figures S5–S8 and Table S1.
Figure 7
Figure 7. vS1-Projecting Neurons in vM1 Receive Strong Input from vS1
(A) Schematic of the experiment. AAV-ChR2-venus (green) and fluorescent microbeads (red) were coinjected in vS1. Recordings were made in vM1 (dashed line). (B) Representative images of vS1 axons (left), bead-labeled cells (vS1-projecting cells; middle), and the overlay (right) in vM1. (C) Example sCRACM input map recorded in a retrograde bead-positive cell (triangle, soma location). (D) Average sCRACM maps for bead-positive L2/3 neurons (left) aligned on the pia (n = 7 cells, 6 mice) and for bead-positive L5A neurons (right) aligned on the pia (n = 8 cells, 6 mice). (E) Input to bead-positive and nearby (distance <50 μm) bead-negative cells (left, actual values; right, normalized) (n = 15 pairs of cells, 12 mice). Statistics, signed-rank test. The circle in the right panel indicates the mean (0.40). See also Figures S6 and S9 and Table S1.
Figure 8
Figure 8. The Long-Range Circuits Connecting vS1 and vM1
Red, projections from vS1 to vM1. Blue, projections from vM1 to vS1. Gray, three strongest intracortical projections. Line thickness is proportional to the strength of connection strength. Targets of vM1 projections: zona inserta (ZI, magenta), superior colliculus (SC, cyan), braisnstem (BS, green), and posterior thalamic nucleus (PO, blue).

Similar articles

See all similar articles

Cited by 169 articles

See all "Cited by" articles

Publication types

MeSH terms

Feedback