Fig. 4: CICADA: Cell-type identification by Ca²⁺ coupled activity through drug activity.

a Overview of CICADA showing the intracellular mechanisms (inset) and the agonists used to target the specified receptors (top). b Representative ΔF/F Ca²⁺ traces for each of the CICADA ligands. Each trace is a different cell. SP, 1 µM; OXT, 1 µM; GRP, 300 nM; Oxotremorine M (Oxo.) 50 µM; Taltirelin, 3 µM; NMB, 100 nM; CCK, 200 nM; NKB, 500 nM. c t-SNE plot of excitatory superficial clusters (Ex) identified by CICADA. Each dot represents a CICADA-identified cell, color-coded for the appropriate cluster with the cluster regions color coded by shading. Cluster numbers (e.g. Ex2) are followed by the principal receptor found in that population of cells (Ex2_TACR1). d Normalized Ca²⁺ traces of each CICADA ligand for Ex1-7. Each cell’s ΔF/F trace was normalized to a maximum of 1 and minimum of 0. Data are shown as median ± interquartile range. Ex1_TACR3/1, 57 cells; Ex2_TACR1, 75 cells; Ex3_GRPR/NMBR, 57 cells; Ex4_CCKBR, 21 cells; Ex5_TRHR, 65 cells; Ex6_CHRM3, 49 cells; Ex7_OXTR, 74 cells (cells are pooled from 4 mice). e Radar chart showing the response to CICADA ligands within each population. The spoke length is the population average of normalized response to CICADA ligands ((X-X_min)/X_max). f Violin plot showing the depth of each population relative to the surface of the dorsal grey matter. g Relative abundance of neurons making up the excitatory populations identified by CICADA. h Cardinal tuning of Ex1-7. For vector angles: 1-way ANOVA (F (7, 1541) = 3.983), P = 0.0003. For vector magnitudes: 1-way ANOVA (F (7, 1541) = 3.289), P = 0.0018. Post hoc testing was performed comparing each population’s vector angle (#Q < 0.05) and magnitude (*Q < .05) to the average of all recorded cells. Dotted circle is 1 unit vector in diameter. i Percent of cells within each cluster which show cardinal (MHC) polymodal. Multiple logistic regression model was used to test how predictive each cluster was for MHC modality, ***P = .0001 (P values are calculated from an F test). Odds ratio (OR) shown for significant results. Full logistic regression model shown in source data. As described in methods: all tests were 2 tailed and any post hoc comparisons were corrected for multiple comparisons by False Discovery Rate correction.