Skip to main content
. Author manuscript; available in PMC: 2019 Feb 18.
Published in final edited form as: Cell Rep. 2019 Jan 2;26(1):159–167.e6. doi: 10.1016/j.celrep.2018.12.040

Figure 3. Statistical Analysis of Input Biases to VTA Neurons.

Figure 3.

(A) Linear regression between the percentage of total rabies-labeled inputs from each of the 22 input sites for each animal (n = 76) and the starter cell COM, the mouse Cre driver line, or both the COM and the Cre driver line. (B) Similar regressions were performed like in (A), but inputs were run against the COM, output site, or both. (C) Regressions were run against the COM, the percentage of starter cells immunostaining for TH, or both. (D and E) A linear regression model was trained on a subset of the 76 brains and tested on the remainder of the dataset. An r2 value was calculated over 500 randomized training/testing set divisions and compared to control sets where the link between COM and inputs was scrambled. The obtained r2 value for the GPe (D) was 0.38 and was 0.33 for the NAcCore (E). Each graph shows 1, 2, and 3 SD from the mean in the positive direction. See Table S1 for related data.