The YFP/CFP ratio showed correlation with the relatively large movement under this recording condition (data not shown). AVE and AVB coimaged showed out-of-phase profiles and negative correlation ( Figure 1F). The YFP/CFP value for AVE and AVB recording in each sample was normalized by mean and SD. Pearson’s correlation coefficient was determined by R. Under this recording condition, backward motion was hyperstimulated compared to standard

culturing conditions. For correlation analyses of the averaged YFP/CFP ratio change during transitions of directions, YFP/CFP ratios before and after ABT-263 concentration directional change were collected and normalized against the YFP/CFP value immediately before the directional change. Traces from nine AVA/AVE and 15 AVB recordings were used for correlation analysis in Figures 1D and 1E. For correlation analysis between AVE and AVB activity, seven AVE/AVB recordings were analyzed to obtain the data shown in Figure 1F. To compare the interneuron calcium

signals between wild-type and innexin see more mutant animals, we compared the averaged YFP/CFP ratio instead of ΔR/R. YFP/CFP ratio for each sample during 5 min was presented by raster plots. The averaged YFP/CFP value over 5 min of recording for each sample was considered a single data point and presented as scatter plots (Figure 6; Figures S3A, S3B, and S7). This is because neurons analyzed in this study showed relatively high-frequency activation, and we rarely observed the decline of the calcium level to the basal value. In this case, measuring ΔR/R probably leads to an inaccurate measurement of neuronal activity. Imaging of motoneurons was carried out with a protocol modified from the AVA and AVE single-neuron imaging method (Figure 2, Figure 4 and Figure 8; Figure S1D). We dropped 20 μl M9 buffer onto a 2% dried agarose pad, and ∼20 adult animals were placed in the liquid as spacers. Ten last-larval stage (L4) hpIs171 animals were placed in the buffer, covered by a coverslip, and imaged with a 63× objective. Neurons were identified

by their stereotypic anatomical organization. Most data presented in Figure 4 and Figure 8 were obtained by manually recentering the moving animals during the recording and scoring the forward, backward, and kinking motion manually based on the direction of the body-bend propagation. During later parts below of the study, we utilized an in-house-developed automated tracking software to recenter animals, which allowed the automated analysis of the directional movement, as well as correlation between calcium transients with directions and velocity (Figures 2A and 2B, bottom). Samples that show sustained forward or backward movement (Figure 4 and Figure 8), instead of frequent directional change (Figure S1D), were quantified for the mean calcium level in continuous directional movement (Figure 4 and Figure 8). Locomotion direction and calcium transients showed similar correlation pattern in both data sets.