Supplementary Figure 6: ArchT expression in VIP BLA interneurons.

a, Representative example image of ArchT-GFP expression in VIP interneurons in the BLA of VIP-cre mice. Scale bar, 20 µm. b, Quantification of co-localization of ArchT-GFP expression with VIP detected by immunohistochemistry (N=3 mice). c, Representative patch-clamp recording of an ArchT-GFP expressing VIP BLA interneuron. Top, suppression of spontaneous action potential generation by 4.5 s yellow light. Scale bars, 20 mV, 500 ms. Middle, ArchT activation with yellow light diminishes action potentials evoked by depolarizing current steps (-50 pA, 0 pA, and +50 pA current injections while holding the cell at -60 mV). Scale bars, 20 mV, 200 ms. Bottom, confocal image of the same ArchT-GFP+ cell filled with biocytin during whole-cell recordings to confirm VIP expression. Scale bar, 10 µm. d, Spontaneous action potentials are reliably inhibited by application of yellow light (Wilcoxon matched-pairs signed rank test, two-sided, P=0.001, n=11 cells). e, ArchT activation decreases excitability of VIP BLA interneurons. Spike rate was normalized to the maximum frequency in baseline condition for each cell. Sigmoidal curve fitting reveals a significant shift to the right of input-output curves with ArchT activation (15.7 pA shift; I_half baseline 99.9 pA, I_half light 115.4 pA; paired t-test, two-sided, t₍₁₁₎=2.356, P=0.0381; n=12 cells) and decreased gain (slope baseline 43.5%/pA, light 36.6%/pA; paired t-test, two-sided, t₍₁₁₎=2.740, P=0.0192; n=12 cells) without affecting maximum output. f, Representative example traces from a VIP BLA interneuron expressing ArchT-tdTomato, demonstrating reliable spike suppression with yellow, but not blue light. Further, only yellow but not blue light activates ArchT at a holding potential of -60 mV, leading to membrane potential hyperpolarization. Scale bars, 20 mV, 500 ms. g, Yellow light significantly decreases spike probability in VIP BLA interneurons, while blue light has no effect (Friedman test, F=40.71, P<0.0001; Dunn’s multiple comparisons test, C yellow vs C yellow +light, P<0.0001, C yellow +light vs C blue +light, P=0.0002; n=15 cells). Note that blue light used for nVoke in vivo imaging experiments was further of shorter wavelength and lower intensity (448 nm, 0.4-0.7 mW) compared to slice electrophysiology to exclude unwanted cross-excitation of ArchT. h, Yellow light (yellow line; 590 nm, 12 mW at objective, 20 s) does not affect Ca²⁺ fluorescence in VIP BLA interneurons expressing GCaMP6 in vivo (n=95 cells from N=3 mice, trace represents mean and s.e.m.). Scale bars, 0.5% ∆F/F, 10 s. i, Yellow light induces a decrease in Ca²⁺ fluorescence in VIP BLA co-expressing GCaMP6 and ArchT-tdTomato (n=80 from N=3 mice). Scale bars, 0.5% ∆F/F, 10 s. j, Average amplitude during yellow light application (20 s) is significantly different between GCaMP6-only controls and VIP interneurons expressing GCaMP6 with ArchT (Mann-Whitney U test, two-sided, P<0.0001; control, n=95; ArchT, n=80). k, Schematic illustrating reconstructed implant sites of GRIN lenses within the BLA for VIP nVoke experiments shown in Fig. 5 matched to a mouse brain atlas (gray lines, GCaMP6 in VIP, N=4 mice; yellow lines, GCaMP6 and ArchT in VIP, N=3). LA, lateral amygdala; BA, basal amygdala; CEA, central amygdala. Confocal images and example traces are representative for N=3 mice (a), n=12 cells (c), n=15 (f). Circles in b show individual data points, connected circles in d and g represent individual paired data points, horizontal lines additionally indicate the mean. Box-and-whisker plots show median values and 25^th/75^th percentiles with 10^th to 90^th percentile whiskers, dots additionally indicate the mean. All other data is presented as mean and s.e.m. * P<0.05, *** P<0.001. All details of statistical analysis are listed in Supplementary Table 3.