Fig. S2

Additional Details of MultiMAP and Compatibility with In Situ Hybridization, Related to Figure 2

(A) For general user reference: commands used for volume registration in CMTK (STAR Methods).

(B) 16 z-planes extracted from tissue volumes, from a single example fish. Live GCaMP (green) and fixed GCaMP after registration (red); yellow indicates overlap. Each plane corresponds to z-plane in live volume that was imaged during behavior.

(C) 16 z-planes extracted from tissue volumes, in same example fish as panel b. Live GCaMP (green) and Z-brain atlas Tg(elavl3:H2B-RFP) reference after registration (red).

(D) Protocol for annotation of live-recorded cells with neurochemical and anatomical information, with example for TH+ cells in the locus coeruleus (LC). To distinguish between antibody labeling of cell bodies versus axonal and dendritic projections, cell bodies were manually identified. For each ROI identified from live-imaged z-plane, the ROI is assigned to a given neuromodulatory region if 75% of the pixels within that ROI overlap with the antibody cell body label, and 100% of the pixels within that ROI overlap with the anatomical region label(s). In the example image, TH+ LC neurons are labeled in red, whereas TH+ cells that do not overlap with the LC anatomical mask are labeled in blue. Scale bars: 100 μm.

(E) Higher magnification images of example brains, showing cellular-resolution of registration approach, even in densely packed tissue. Scale bars: 30 μm.

(F) Demonstration of registration method compatibility with fluorescent in situ hybridization (fISH). Here we show registration in the hypothalamus, even after the harsh treatment of fISH. We used the hybridization chain reaction method (STAR Methods), which allows for multiplexed molecular labeling with orthogonal fluorophores. Detection of RNA over protein is particularly useful for neuropeptides such as hypocretin/orexin (hcrt, used here), where the protein is primarily localized to axonal terminals. Scale bars: 50 μm and 5 μm (inset).