Delivery of small molecules to zebrafish embryos and larva is a simple and straightforward process. Because the compounds are directly absorbed through the skin and the embryos are swimming in the media, dosing is extremely precise and repeatable. The appropriate dosage, the duration of exposure, and the embryonic stage of exposure are variables the investigator should consider. The ability to treat large numbers of animals allows for rapid evaluation of these parameters and is a major advantage of using zebrafish. In order to better visualize the effects of treatments on the fluorescently labeled neurons of transgenic lines, embryos can be raised in fish water containing 30 mg PTU/L fish water (see Note 16 ). This will prevent the development of pigment that can obscure GFP in deeper structures.

For chemical treatment, embryos are arrayed into multi-well culture plates, 6-well to 96-well as appropriate for the experimental design. The number of embryos per unit volume depends on the duration of treatment and the likelihood that the treatment will have adverse effects on water quality. For chemical screening, we routinely treat three embryos for 24 h, in 100 μL solution, in 96-well plates. After arraying, allow the embryos to develop to the desired developmental stage (staging can be found at http://​zfin.​org/​zf_​info/​zfbook/​stages/​index.​html), remove most of the fish water and then replace it with the experimental solution. Control embryos should be from the same mating pair (siblings), treated with carrier. To end treatment, pipette off the solution, rinse once with fish water, and replace with fresh fish water.

Specific subsets of neurons can be directly visualized in the developing brain using transgenic lines that express fluorescent marker proteins in given neurons (see a 3D brain map at Zebrafishbrain.org). For visualizing the effects of chemical treatments, the embryos/larva are first immobilized in tricaine. The embryos or larva are pipetted onto a depression slide, most of the fish water is removed, and then replaced with a 2× tricaine solution (see Note 17 ). Under a stereomicroscope, the anesthetized embryos/larva are carefully transferred to 3 % methylcellulose in fish water in a new depression slide (see Notes 18 and 19 ). The embryos should be gently oriented using an eyelash brush. It may take some practice to avoid injuring them. Once this is mastered it is useful to line up ten experimental embryos on top and ten control embryos on the bottom with their heads facing each other. In this way, control and experimental embryos can be directly compared and any differences in morphology noted. Once lined up and oriented appropriately, fluorescent neurons are visualized by epi-fluorescence microscopy. A 40× objective will be needed if examining fine neuronal processes; however, 10× or 20× objectives are adequate for visualizing neuronal cell bodies and major processes. At this resolution the numbers of cells are easily counted and the appearance of new neurons is detected by quickly scanning the entire animal. Changes in cell numbers or appearance should be quantified at this time (see Fig. 3) (see Note 20 ).

Lining the fish up in groups of ten experimental and ten controls helps in keeping track of the embryos when under the microscope. Remove the embryos/larva very carefully from the methylcellulose by gently sucking them into a glass pipette. Release them back into their respective wells slowly. The methylcellulose containing the embryos/larva is the consistency of toothpaste, but as long as the fish are extruded slowly with no sheering forces, they should be unharmed. The methylcellulose will gradually dissipate into the fish water, completely freeing the embryos/larva. The fish are then raised to adulthood for behavioral testing.

The groups of experimental and control fish should be raised keeping the density of fish per tank consistent amongst the treatment groups. Wherever possible, all conditions should be kept constant between the treatment groups in order to eliminate confounding variables. If raising fish in 10 gallon aquaria, perforated plastic dividers can be used to separate groups. As soon as the fish reach maturity (between 3 and 4 months of age) they can be tested in the shoaling assay.

Zebrafish are a highly social species that prefer to shoal with other members of their own species. Several labs have reported variations of the shoaling assay [3, 31]. The following protocol describes our own variation of the shoaling assay based on Wright and Krause, created according to the specific needs and resources of our lab [32].

In order to test sociality in individual zebrafish, we set up a clean 10 gallon (38 L) test tank filled with 8 L of fresh fish water. The sides of the tanks are covered with cardboard, preventing the fish from seeing the researchers (see Notes 21 and 22 ). The tank is marked on each side with tape, 10 cm from the edge of the tank.

Two smaller 2.5 L stimulus tanks are placed at either end of the test tank. Water level in the stimulus tanks is maintained at an equivalent level to the water in the test tank. One stimulus tank contains a shoal of ten fish with wild-type phenotype of mixed sizes, ages, and sex (see Note 23 ). The other stimulus tank contains an equivalent amount of water, but no fish. The stimulus tanks are visually blocked from the test tank by white blinders placed between the sides of the stimulus tanks and the test tank (see Fig. 4).

Before the test begins, test fish are isolated in small individual tanks for 1 h. At the start of each test, individual fish are transferred as smoothly as possible via a net to the test tank, where they are allowed to acclimatize for 5 min. After 5 min, the time the fish spent past the marking tape on either end of the tank is recorded with a stopwatch for 2 min (using a different stopwatch for each side). If the fish passes the marker, it is considered to be within shoaling distance from the stimulus tank, as zebrafish with the wild type phenotype typically form shoals of 12–14 cm [31]. A fish should be considered to be within shoaling range when its entire body is within the marked-off area.

After 2 min, the blinders are carefully removed, revealing both stimulus tanks (see Note 24 ). The time the fish spends within shoaling distance near either tank is recorded for another 2 min. After 2 min, the blinds are carefully replaced, and the time the fish spends within shoaling distance of either tank is recorded again. The blinds are removed again after another 2 min, and again, the time the fish spends near each stimulus is recorded.

Sides of the stimulus tanks (fish vs. no fish) should be varied randomly during and between trials. The water should be thoroughly mixed by stirring with a net before each trial, preventing confounding variables arising as a result of fear pheromones released by the fish. Fish that freeze for longer than 2 min should be removed from the assay (see Notes 25 and 26 ).

Experimental variables for this procedure are the amount of time the test fish spends within shoaling distance of the stimulus shoal, as well as the amount of time the fish spends furthest away from the shoal. Time the fish spends between the two regions of the tank is also recorded. The age, genetic background, and sex of the fish should be recorded as well.

As in the human population, significant variation in sociality exists within a fish population. Some fish show a high degree of sociality, spending almost all of their time near the stimulus shoal. At the other extreme, some fish show extremely low sociality, spending most of their time on the side furthest from the shoal. We saw this subpopulation of fish that did not appear to display sociality in all the treatment groups we testing, indicating that it is a natural variation within the zebrafish population in our lab (see Figs. 5 and 6). Most fish will show some variation between these two extremes.

Behavioral data can be analyzed via a Mann–Whitney U-test or via one-way ANOVA, followed by post-hoc Tukey Kramer test or Games Howell test of significance.

If fish are behaving unusually compared to the rest of the population, then their results should be removed from the assay. For example, if fish freeze for the full 5 min of the acclimatization period and show no signs of responding to the shoal, they should be removed from the study. Fish that freeze for a significant amount of time either near or away from the stimulus shoal should also be removed from the study, as fish may be responding fearfully to the new tank environment or external stimuli rather than to the test shoal. If fish show any signs of injury or illness, they should not be included in the test.

Zebrafish display anxiety in response to novelty and upon isolation from other fish [5]. Fear responses in the zebrafish are complex, but almost always involve huddling near the bottom of the tank. Therefore, in order to quickly and simply measure anxiety in our fish, we use an assay measuring the amount of time zebrafish spent near the bottom half of a small tank based on Cachat et al. [33]. A test tank is prepared (Fig. 7), consisting of an empty 1 L tank with a narrow black line about the tank circumference at 400 mL (the half-way mark).

Prior to the test, individual fish are isolated in 1 L tanks containing 800 mL of water for 1 h. After 1 h, the fish is gently poured along with its water from an isolation tank to the test tank. The amount of time it takes for the fish to swim above the 400 mL mark is measured using a stopwatch. Fish that freeze for longer than 5 min should be removed from the assay (see Notes 27 and 28 ).

The main experimental variable for this procedure is the amount of time the test fish spends beneath the halfway mark on the test tank. The age, genetic background, and sex of the fish are also recorded.