Supplementary Figure 3 : Discriminability and cell-type promiscuity of the pERK indicator.

From: Whole-brain activity mapping onto a zebrafish brain atlas

Supplementary Figure 3

(a) The Reciever Operator Characteristic plots for 4 different fish, comparing the ability of pERK to discriminate between true and false positives, where a ROI with >30 s of detected Ca2+ activity is considered to be a true positive and an active cell. Strong leftward deviation from the diagonal line indicates good discriminability in all four fish analyzed. (b) The area under the ROC curve (AUC) yields a number between 0 and 1, representing the probability that the pERK will be higher in a randomly chosen active neuron than in a randomly chosen silent neuron. By plotting the AUC as a function of the threshold of required activity observed during Ca2+ imaging to classify a cell as ‘active’, we see that the AUC increases as we increase this stringency on activity. We observe values of nearly chance (0.5) for very low activity thresholds (<1 s), while after 10 and 30 s thresholds we see values of ~0.7 and ~0.8, and we observe perfect discrimination in Fish 2 at the 41-s threshold. This indicates that the discriminability of the pERK indicator increases with increasing levels of activity. (c) Examples of Tg(Vglut2a:GFP)64 positive (arrows) and negative (arrowheads) cells exhibiting high pERK levels in the Habenula (Hab) and Telencephalon (Tel). (d) Examples of a Tg(Glyt2:GFP)53 positive (arrow) and negative (arrowhead) cell exhibiting high pERK levels in the hindbrain.