Article Figures & Data
Figures
- Figure 1.
Mapping areas of mouse parahippocampal cortex. A, Tangential section through middle layers of flatmounted postrhinal cortex stained with an antibody against M2 muscarinic acetylcholine receptor (magenta). Inset shows low-magnification image of M2 expression in L4 of V1 and surrounding cortex. Green and red spots indicate AAV2/1.hSyn.EGFP (green) and AAV2/1.hSyn.tdT (red) injection sites into nasal/central and peripheral/upper visual field representations, respectively. B, C, Anterogradely labeled clusters of axon terminals labeled after AAV injections into V1 (A, inset) show retinotopic maps in LM, LI, P, POR, and PORa. Note that to show the projections in higher visual areas, the image was overexposed, which made the injection sites artificially large and gives the false impression that viral tracers spilled across the lateral V1/LM border. B, inset, A diagram of visuotopic maps obtained from two representative experiments, whose V1 injections sites and projections are marked by solid and open green and red contours. Blue arrows denote the progression from lower/nasal to upper/peripheral visual field representations in different areas, derived from five different experiments. The overlay reveals that projection maps in POR and PORa are contained within the M2+ cortex (C). D, Expression of PVtdT in postrhinal cortex shows a foot-shaped pattern similar to M2 expression in C. M2 and PVtdT expression is extremely sparse in P, ECT, and TEp. E, F, Coronal section from Chmr2-tdT mouse (approximate location indicated by the green line in A) shows sharp transitions of laminar M2 pattern at the POR/ECT and ECT/ENTm borders (E); after counterstaining with NeuroTrace 425/455, the transitions coincide with a loss of the granular appearance of L4 and the distinctions with the larger cells in the more sparsely populated L5 of ECT were no longer apparent (F). Axes: A–D: A, anterior; L, lateral; E–F: D, dorsal; L, lateral.
- Figure 2.
Patchy M2 expression in higher cortical areas. A, B, Tangential section through occipital cortex of Chrm2-tdT mouse shows patchy M2 expression throughout the thickness of L1 (A) and uniform expression in L4 (B) of V1, LM, LI, POR, PORa, AUD, and RSP. Scale and orientation in B are the same as in A. C, Higher-magnification image of LM, LI, POR, and PORa shows that M2+ patches are elongated parallel to the major axes (green double-headed arrows) of LM, LI, and POR. D, Mean ± SEM ratio of major/minor axial lengths of M2+ patches along the major axis of the containing area in V1, LM, LI, and POR (mean, 63.2 patches/mouse analyzed; Tukey's test). E, Mean ± SEM spatial density of M2+ patches across areas (same mice and patches as in D; Tukey's test). F, Mean ± SEM M2+ patches per square degree of visual space, determined by multiplying spatial densities with magnification factor of respective area (Garrett et al., 2014; same mice and patches as in D; Tukey's test; *p < 0.05; **p < 0.01; ***p < 0.001).
- Figure 3.
Axonal projections of visual cortex → amygdala pathways. A–J, In situ ex vivo images of callosal connections traced by retrograde transport of bisbenzimide from the posterior third of the right (contralateral) hemisphere. The arrows point to bright spots that represent the BDA injection sites in V1, LM, P, LI, POR, AL, RL, PM, AM, and A of the left (ipsilateral) hemisphere. A′–J′, Dark-field images of coronal sections showing anterogradely BDA-labeled axonal projections (yellow) from V1, ventral (LM, P, LI, POR), and dorsal (AL, RL, PM, AM, A) stream areas terminating in different nuclei of the amygdala. Note that the white structures represent unlabeled, highly reflective myelinated fibers of the internal capsule (int). Scale bars: A–J, 1 mm; A′–J′, 0.5 mm.
- Figure 4.
Relative strength of anterogradely BDA-labeled projections of mouse visual cortex to 31 subcortical targets. Nomenclature is after that in Dong (2008) and Franklin and Paxinos (2008). A–E, Average ± SEM optical density of projections from V1 and the ventral stream areas LM, LI, P, and POR in percentage of the summed density of all subcortical projections labeled from each area. F–J, Projections from the dorsal stream areas AL, RL, AM, PM, and A. Colors indicate functionally broadly similar groups of subcortical targets (Jones, 2007). B–E, insets, Relative strength of anterogradely BDA-labeled projections from ventral stream areas LM, LI, P, and POR to LA, BLA, BMA, CeL, and CeM. Percentage (mean ± SEM) of total optical density of labeled terminal axonal branches (black bars). Values are background-subtracted percentage (mean ± SEM) of optical density relative to the injection site (gray bars). K, Comparison of projection strengths from different visual areas to LA. Two-sample t test: *p <0.001; **p < 0.0001. (–), undetectable; CeL, centrolateral amygdala; CeM, centromedial amygdala).
- Figure 5.
Retrograde tracing in PVtdT mice of subcortical and cortical source neurons projecting to POR. A, Tangential section through L2/3/4 of flatmounted PVtdT-expressing (red) cortex, shows DY injection site (false colored blue) in POR. B, Overexposed black/white image (notice the artificially large injection site gives the false impression that DY spilled beyond the borders of POR) showing DY-labeled neurons (white dots) in LA and multiple cortical areas segmented according to Gămănuţ et al. (2018). C, Overlay of DY-labeled neurons with PVtdT expression. D–G, Coronal sections through thalamus and amygdala. D, E, Representative sections through the dLGN of two different mice (17134 and 17185) showing that DY-labeled neurons are present in the shell (arrowheads) but not in the core. F, G, Dense labeling in the LPLR and LPMR (F), and LA (G).
- Figure 6.
Axonal projections of lateral amygdala → POR pathway. A, Tangential section through L1 of flatmounted POR in Chrm2-tdT mouse showing nonuniform pattern of M2+ patches (magenta) and M2– interpatches (black). B, Patchy axonal projections to L1 labeled by anterograde tracing with AAV2/1.hSyn.EGFP from the LA. Coronal section of injection site in LA (B, inset). C, Overlay of A and B shows that LA → POR projection to L1 terminates preferentially in M2– interpatches. D, Heat map of M2 expression in POR (A). E, Partitioning of heat map (D) into six equal-area quantiles (different shades of gray) for the delineation of M2+ patches and M2– interpatches. F, Normalized average ± SEM (N = 5 mice) EGFP fluorescence intensity in each of the six quantiles shows that LA → POR inputs are preferentially associated with M2– interpatches (R = −0.86, p < 10−8, Pearson correlation).
- Figure 7.
Axonal projections of dLGN → POR and LP → POR pathways. A, C, Immunostained M2+ patches in L1 of flatmounted tangential section through POR and PORa. B, dLGN → POR axonal projections traced anterogradely with AAV2/1.hSyn.EGFP. B, inset, Coronal section of injection site spreading across core and shell of dLGN. White contours indicate borders between fourth and fifth intensity quantiles (of six) of M2 expression (A). Overlay of contours shows overlapping patterns of M2+ patches and dLGN → POR projections (B). D, AAV2/1.hSyn.EGFP-labeled LP → POR projections to M2+ patches (C) in L1 of POR. D, inset, Coronal section of injection site in LP. E, Laminar distribution of anterogradely labeled LA/BLA → POR, dLGN → POR, and LP → POR projections. F, Mean ± SEM patch: interpatch EGFP intensity ratio of dLGN → POR, LP → POR, dLGN → LM, LP → LM, dLGN → LI, LP → LI, LA/BLA → POR, and V1 → POR axonal labeling in L1 of the projection target. **p < 0.01; *p < 0.05; ns, not significant (one-sample t test).
- Figure 8.
Modular distribution of retrogradely labeled apical dendrites of POR → ENTm and POR → amygdala projecting neurons in L1 of POR. A–C, Tangential section through flatmounted POR showing patchy expression of M2 immunostaining in L1 (A) and nonuniform branching pattern of spine-covered apical dendrites (false-colored green, denoted by white arrows) in L1 of POR → ENTm projecting neurons labeled by retrograde tracing with rAAV2-Retro/CAG-Cre in Ai9 mice (B); overlay of A and B shows that dendritic branches are preferentially located in M2– interpatches (C). D, Tracings of retrogradely labeled dendrites (green) in L1 of POR overlaid onto contour plot of M2 expression. Shades of gray represent six quantiles of fluorescence intensity. E, F, Tangential section through flatmounted POR showing patchy expression of M2 immunostaining in L1 (E) and nonuniform branching pattern of apical dendrites (false-colored green) of POR → LA/BLA projecting neurons labeled by retrograde tracing with rAAV2-Retro/CAG-Cre (F). White arrows indicate dendrites branching in the plane of the pial surface. Red arrow points to a thick vertically ascending dendritic trunk at the branchpoint of terminal tuft at the L1/2 border. Because the section is not perfectly parallel to L1 and cuts into L2 it shows a retrogradely labeled cell body at the L1/2 border (arrowhead). G, Overlay of E and F shows that dendritic branches (green) and cell bodies at the L1/2 border are preferentially aligned with M2– interpatches. H, Length of dendritic branches of POR → LA/BLA and POR → ENTm projecting neurons in M2+ patches and M2– interpatches of L1 of POR. The fraction (mean ± SEM) of total dendritic lengths for both types of neurons shows a strong (POR → LA/BLA: R = −0.93, p < 10−10, N = 4 mice; POR → ENTm: R = −0.99, p < 10−8, N = 4 mice, Pearson correlation) preference for M2– interpatches. I, J, Laminar distribution (mean ± SEM) of retrogradely labeled cells in POR, PORa, LM, LI, and P projecting to ENTm (I) and LA/BLA (J), respectively. Same axes in I and J.
- Figure 9.
Modular distribution of retrogradely labeled cell bodies of POR → LA/BLA and POR → ENTm projecting neurons in L2/3 of POR. A, B, Tangential section through flatmounted POR showing nonuniform distribution of somata (false-colored green) of POR → ENTm projecting L2/3 neurons labeled by retrograde tracing with rAAV2-Retro/CAG-Cre in Ai 9 mice (A) overlaid with patchy expression of M2 immunostaining in L1 (B). C, Overlay of A and B shows that cell bodies and basal dendrites are preferentially aligned with M2– interpatches. D, Normalized cell counts (mean ± SEM) of POR → ENTm in M2 quantiles (shades of gray) of POR. The fraction of total cell bodies shows a strong preference for M2– interpatches. E–G, Tangential section through flatmounted POR showing nonuniform distribution of somata (false-colored green) of POR → LA/BLA projecting L2/3 neurons labeled by retrograde tracing with rAAV2-Retro/CAG-Cre in Ai 9 mice (E) overlaid with patchy expression of M2 immunostaining in L1 (F). G, Overlay of E and F shows that cell bodies are preferentially aligned with M2– interpatches. H, Normalized cell counts (mean ± SEM) of POR → LA/BLA in M2 quantiles of POR. The fraction of total cell bodies shows a preference for M2– interpatches.
- Figure 10.
Anterogradely labeled POR inputs to CA1 of hippocampus. A, Tangential section through L4 of flatmounted cortex stained with an antibody against M2 (magenta). Green spot indicates injection site of AAV2/1.hSyn.EGFP in POR. B, Tangential section through flatmounted HPF. For flatmounting, the hippocampus was separated from the external capsule (ec) along the alveus (alv) and flipped backward (inset) so that the ventral side faced the pial surface and the dorsal pole (cross hairs) came to lie at the posterior edge of the section through the flatmounted left cortical hemisphere (inset). POR → CA1 axons (green) are most densely labeled at the dorsal pole and gradually fade away toward the ventral pole. DG, Dentrate gyrus; slm, stratum lacunosum moleculare; so, stratum oriens; sp, stratum pyramidale; sr, stratum radiale; Sub, subiculum; dHPF, dorsal HPF; vHPF, ventral HPF.
- Figure 11.
Circuit diagram of the modular thalamo-amygdalo-entorhinal network with POR. Modules in POR are depicted as M2+ patches (dark red) and M2– interpatches (light red) in L1, which indicate clusters or putative columns across the layers below. The M2-defined modularity is supported by patch- and interpatch-specific inputs and outputs. Inputs from the dLGN shell and the lateral posterior nucleus (LP or pulvinar) go to M2+ patches. Inputs from V1 to POR are equally dense in M2+ patches and M2– interpatches. Inputs from the amygdala (Amy) and, hypothetically, the ENTm (dotted line) terminate in M2– interpatches. M2– interpatches also contain the dendrites and cell bodies (red) of POR neurons projecting to Amy and ENTm. The projection targets of patch neurons (black) are unknown. POR projections to the HPF are biased toward the dorsal pole (dHPF) and are undetectable in the ventral part (vHPF). It is unknown whether POR → HPF neurons are preferentially associated M2+ patches or M2– interpatches or reside in both modules.
