Highly amplified, multiplexable smFISH in the medial amygdala.

A. Schematic of smFISH readout systems. Panel created with BioRender.com/r99q760. B. Representative images from Parvalbumin smFISH comparing amplification methods. Circles indicate examples of extracellular regions with non-specific signals. C. Quantification of smFISH signal intensity from different amplification methods, n = 500 signals (50 from 10 cells) for each condition. Signal intensity is background subtracted (mean of 50 background intensities subtracted from transcript signal). p > 0.0001, unpaired T-test. D. Schematic of final hamFISH readout system incorporating bridge-readout probes and highly amplified v1 design. E. Representative images known MeA marker gene transcripts in the MeA, detected using the hamFISH scheme shown in D. F. Proportion of bits that give overlapping single molecule localized signals in a sequential smFISH experiment where all 32 bits were used to develop a different gene.

The MeA is comprised of distinct molecular cell types that occupy specific spatial locations.

A. Two-dimensional t-distributed stochastic neighbor embedding (t-SNE) plots based on differentially expression of 31 genes for n = 163,925 cells from 12 animals, colored by cluster. Those cells positive for either Slc32a1 (inhibitory) or Slc17a6 (excitatory) are highlighted in the right two panels. B. Marker gene expression distributions within each cluster, represented by violin plots for cells shown in A. C. Heatmap representation of the 30 identified clusters from one animal (male). Note gene order differences between plots to aid with visualization of clusters. D. The spatial position of every cell from each cluster from one animal (male) across the MeA. E. t-SNE plots based on spatial features of each clusters. MeA clusters (left) have been separated into inhibitory (middle) and excitatory (right) classes. Color scheme the same as indicated in B, separable clusters are indicated with boxes.

Specific MeA cell types are associated with projection targets and are activated differentially to social stimuli.

A. Schematic of CTB-injection scheme along with a representative MeA image from triple-injected animals. Note that injection sites of the 3 different CTB-conjugated fluorophores were swapped around in different animals, but for the purpose of this figure all subsequent data adhere to the color scheme green = MPOA-projecting cells, red = BNST-projecting cells, blue = VMHvl- projecting cells. B. Representative position of CTB-labelled projection neurons in a single animal injected in all 3 target regions. C. t-SNE plots. Shaded cells in top row represent either inhibitory or excitatory cells (left panel is t-SNE without shading). Bottom row shading indicates cells labelled with CTB injected into either the BNST, MPOA or VMHvl. n = 263,474 cells from 13 animals, colored by cluster. D. Comparison of the proportion of all cells that are excitatory vs projection neurons labelled by CTB that are excitatory. For VMHvl p < 0.0001, paired T-test. E. Comparison of the proportion of all cells that are cluster X vs projection neurons labelled by CTB that are cluster X. Paired T-tests. F. Schematic of activity mapping workflow. Panel created with BioRender.com/j05e655. G. c-fos-positive cells per cluster as a proportion of total c-fos-positive cells. Enrichment indicated by * (female vs male), † (male vs pup), and ‡ (female vs pup), unpaired T-tests. H. k-means clustering (k=3) of c-fos expression pattern in control, male-exposed and female- exposed mice. Selected clusters are shown for clarity.

Amplification scheme v1 reduces background signals and increases hybridisation efficiency versus amplification scheme v0.

A. Comparison of background signal abundance between v0 and v1 amplification schemes. n = 5 background areas (each 10 um2), p < 0.01, unpaired T-test. Note that left lobe (containing the MeA) was dissected for cell type profiling so is not present in these sections. Green = CTB-Alexafluor 488 conjugate, Red = CTB-Alexafluor 555 conjugate, Blue = CTB-Alexafluor 647 conjugate, Grey = DAPI. B. Probes were designed to target two different regions of Pvalb mRNA (region 1 and region 2) which were each developed with a different fluorophore. . C. Representative examples of Pvalb smFISH signal from region 1 and region 2 using the v0 and v1 probe designs. D. Quantification of overlap of signal from each probe design. Each data points represents the overlap of real Pvalb+ signal vs background noise. n = 3 cells/background areas per sample, n = 3 samples. Comparison of signal overlap between v0 and v1, p = 0.047, unpaired T-test.

Highly amplified smFISH v2 further increases signal, making it comparable to other highly amplified methods of transcript detection.

A. Left; representative images of Pvalb smFISH signal developed using our different amplification methods. Right; quantification of signal intensity comparison between v1 and v2, p < 0.0001, unpaired T-test. B. Left; representative images of Somatostatin (Sst) smFISH using with probes that were compatible with either highly amplified v2 smFISH or rolling circle amplification. Right; quantification of fluorescence intensity from each method. n = 100 each, comprising 10 ROIs from 10 cells. p = 0.23, unpaired T- test.

smFISH signal using all 32 orthogonal amplifier sets.

Each bit was used to detect a different gene. All images are taken from MeA tissue.

MeA clusters map onto previously publish scRNA-seq data.

A. Heatmap of clustered scRNA-seq data from Chen et al. B. Top: Heatmap showing the correlation in gene expression between scRNA-seq and sequential hamFISH clusters. Values <0.4 were excluded from analysis. Bottom: Hierarchical clustering comparing average gene expression in scRNA-seq (grey) and sequential hamFISH (black) datasets.

MeA cell types are consistent between animals and sexes.

A. Heatmap representation of the 30 identified clusters from one animal (female). Note gene order differences between plots to aid with visualization of clusters. B. Comparison of number of cells per cluster as a proportion of total MeA cells. There is no significant difference between each sex. Female n = 10, Male n = 11. (C) The spatial position of every cell from each cluster from the same animal as A.

Characterization of cluster i3.

A. A representative i3 cell with smFISH signal from Ndnf, Slc32a1 and Chrna7. B. Of 55 Ndnf+/Slc32a1+ neurons from 3 animals, 41 were Chrna7+.

The input-output logic of MeANdnf+ neurons

A. Visualization of MeANdnf+ outputs. B. Projection targets of MeANdnf+ neurons (n=4 animals). Normalized to number of starting cells per animal (n=4 animals). BNST = bed nuclei of the stria terminalis, COAa = cortical amygdalar nucleus, anterior part, COApl = cortical amygdalar nucleus, posterior part, lateral zone, COApm = cortical amygdalar nucleus, posterior part, medial zone, ENT = Entorhinal cortex, AOB = accessory olfactory bulb. C. Representative images of output areas with Synaptophysin+ neurons. D. Visualization of MeANdnf+ inputs. E. Inputs to MeANdnf+ (n=3 animals) and MeASst+ (n=4 animals) neurons. Significance from 2-way ANOVA comparing Ndnf to Sst neurons. BMA = basomedial amygdalar nucleus, HPF = hippocampal formation, OLF = olfactory areas, PALc = central pallidum, PALm = medial pallidum, PALv = ventral pallidum, sAMY = Striatum-like amygdala nuclei, STRd = dorsal striatum, MEZ = hypothalamic medial zone. F. Representative images of output areas with rabies+ neurons.

Projection cells in the MeA were labelled with CTB injections in downstream areas.

A. Cholera toxin subunit b injection sites. Note that left lobe (containing the MeA) was dissected for cell type profiling so is not present in these sections. Green = CTB-Alexafluor 488 conjugate, Red = CTB-Alexafluor 555 conjugate, Blue = CTB-Alexafluor 647 conjugate, Grey = DAPI. B. MeA neurons from triple-injected BNST animal.

MeA projection neurons are made up of specific neuronal classes.

A. Quantification of projecting cells along the anterior-posterior axis. For each target region, the number of CTB-positive cells at each bregma position was normalized to the total number of CTB-positive cells in that animal, with 1 being the mean. n = 8, 11, 10 for BNST, MPOA, VMHvl respectively. For VMHvl, p < 0.0001, F=26.03, one way ANOVA with a significant decrease at bregma -1.7mm (p < 0.0001, Tukey post-hoc pairwise comparison). For BNST and MPOA p=0.97, F=0.43 and p=0.20, F=1.30 respectively, one way ANOVA. B. Quantification of the total number of CTB-labelled neurons. p < 0.01 for BNST vs both MPOA and VMH, unpaired T-test. BNST = 621 ± 83 cells, MPOA = 341 ± 55 cells, VMHvl = 329 ± 55 cells. C. Representative images of single (white arrow), double (grey arrow) and triple (yellow arrow) CTB-labelled MeA neurons. D. Quantification of single/double/triple CTB-labelled cells. Of BNST-projecting neurons, 86±6% were single-labelled, 6±3% co-labelled with MPOA- injected CTB, 7±5% co-labelled with VMHvl-injected CTB, of MPOA-projecting neurons, 71±12% were single-labelled, 13±9% co-labelled with BNST-injected CTB, 13±11% co-labelled with VMHvl-injected CTB, of VMHvl-projecting neurons, 83±4% were single-labelled, 8±6% co-labelled with MPOA-injected CTB, 7±3% co-labelled with BNST injected CTB. E. CTB-positive cells per cluster as a proportion of total CTB-positive cells. Enrichment indicated by † (BNST vs MPOA), * (BNST vs VMHvl) and ‡ MPOA vs VMHvl), unpaired T-tests. F. Same as E, but comparing single with double-labelled cell. Clusters are reordered for each so that proportion of each cluster in the single-projecting category is presented as highest to lowest, left to right, to emphasise differences. Differences between single and double-projecting neurons for enriched clusters are indicated in the category that has significantly more enrichment (unpaired T-test). G. Summary model of projections from MeA cell types.

Specific MeA cell types are associated with projection targets and are activated differentially in response to social stimuli.

A. Heatmap of c-fos expression pattern in control, male-exposed, pup-exposed and female-exposed mice. Selected clusters are those with significant responses compared with non-exposed controls. Scale bar represents proportion of cells in cluster positive for c-fos. B. c-fos positive cells per cluster as a proportion of total c-fos positive cells (n = 6 (control),10 (female), 8 (pup), 10 (male)). Enrichment indicated by shading (vs non exposure controls), * (female vs male), † (male vs pup), and ‡ (female vs pup), unpaired T-tests.