Fig. 4: Top–down inputs are CaMKII-expressing, while bottom–up inputs are mostly GAD1-expressing.

a Representative image in the CeA, PIR, AI, and CP demonstrating neuronal subtypes of input neurons (red) identified by CaMKII or GAD1 FISH probes (green) (scale bar: 100 μm). b The percentages of CaMKII- and GAD1-expressing inputs in 44 nested subregions (front) and their anatomical regions (background). In all these input subregions—except for the ORB, NLOT, NDB, VP, MD, PF, PSTN, STN, and PAG—there were more inputs that expressed CaMKII than those that expressed GAD1. Additionally, only the hypothalamus had more inputs expressing GAD1 than those expressing CaMKII. c Nearly half of the inputs expressed CaMKII (49.09 ± 1.05%), while less than one-third of the inputs expressed GAD1 (28.89 ± 0.80%). d The number of CaMKII-expressing inputs was markedly higher than that of GAD1-expressing inputs (F (2, 6) = 115.8, p < 0.0001). e Nearly 50% of CaMKII-expressing inputs were located in the CP (12.81 ± 0.35%), SS (11.87 ± 0.37%), CeA (9.96 ± 0.21%), PIR (9.43 ± 0.20%), and AI (5.99 ± 0.19%). f While the major brain regions of GAD1-expressing inputs included the CeA (13.52 ± 0.24%), SS (11.02 ± 1.19%), CP (9.92 ± 0.69%), and PIR (6.66 ± 0.63%). Furthermore, these expression data suggest that the CeA, SS, and PIR may play both excitatory and inhibitory roles in CeA-CRF-related behaviors. g The line refers to the mean number of CaMKII- and GAD1-expressing inputs and the shadow refers to the corresponding s.e.m. from anterior-to-posterior, throughout the brain are shown. The distribution patterns of inputs for these two major neuronal types were relatively similar, but there were more inputs expressing CaMKII in the first half of this spatial distribution than those expressing GAD1, whereas in the second half the numbers of these two populations were similar. h The percentage of CaMKII-expressing neurons in the top–down group was significantly larger than that in the bottom–up group (F [2, 6] = 7.867, p = 0.0210). i The percentage of GAD1-expressing neurons in the top–down group was markedly smaller than that in the other two groups (F [2, 6] = 6.458, p = 0.0319). These results indicated that the distribution of CaMKII- and GAD1-expressing inputs were also concentrated, consistent with the previous results. In addition, the top–down inputs in the cortical regions may convey mostly excitatory information, while bottom–up inputs were mostly inhibitory. Data are mean ± s.e.m., N = 3, one-way ANOVA with Tukey correction.