Figure 1

Schematics of experimental process and behavioral assay paradigms. (A) Experimental flow. Embryos obtained on 0 days-post-fertilization (dpf) through natural spawning. Healthy larvae were plated onto a pair of 48-well plates on 3–4 dpf. Assays were performed on 5 dpf unless otherwise stated. (B) Baseline assays. Larvae were videorecorded without any exogenous stimuli for about 12 h between 9:30 a.m. and 10:30 p.m. in light or darkness. (C) Dark–light repeat assays. Fish were observed in a changing illumination regimen i.e., 30-min acclimation + 4 × [7.5-min dark + 7.5-min light] + 25-min dark. Some assays do not include the final [25-min dark] component since they were performed before the experimental protocol was established. (D) Statistical analysis pipeline post-experiment. GAM modeling produces main effects of the predictors (i.e., genotype, illumination, developmental stages). When the main effect of the predictor is statistically significant, pairwise post-hoc analyses were conducted to identify the significant difference temporally (time points) and in the comparison group. The outcome of the statistical significance in the time window was summarized into proportions of significance over the entire range of time. Subsequently, the proportion was assessed for statistical significance to obtain the difference between different assay regimens (i.e., [7.5 + 2-min] assay vs. [7.5 + 7.5-min] assay). Inferences for the proportions were not made for baseline assays because there was only one assay regimen for the baseline assay (continuous recording in a dark or lit condition). GAM: generalized additive model.