Larval zebrafish as a model for studying individual variability in translational neuroscience research

Abstract

The larval zebrafish is a popular model for translational research into neurological and psychiatric disorders due to its conserved vertebrate brain structures, ease of genetic and experimental manipulation and small size and scalability to large numbers. The possibility of obtaining in vivo whole-brain cellular resolution neural data is contributing important advances into our understanding of neural circuit function and their relation to behavior. Here we argue that the larval zebrafish is ideally poised to push our understanding of how neural circuit function relates to behavior to the next level by including considerations of individual differences. Understanding variability across individuals is particularly relevant for tackling the variable presentations that neuropsychiatric conditions frequently show, and it is equally elemental if we are to achieve personalized medicine in the future. We provide a blueprint for investigating variability by covering examples from humans and other model organisms as well as existing examples from larval zebrafish. We highlight recent studies where variability may be hiding in plain sight and suggest how future studies can take advantage of existing paradigms for further exploring individual variability. We conclude with an outlook on how the field can harness the unique strengths of the zebrafish model to advance this important impending translational question.

Keywords: in vivo imaging; individual variability; neural circuit; translational neuroscience; zebrafish.

FIGURE 1

(A) Top: illustration of two “ideal” normal distributions with the same mean (0) and different variances (blue = 0.5, orange = 1.5). Bottom: two examples of different sample sizes drawn from the distributions illustrated at the top. A small (n = 15) sample size may fail to reveal the differences in variances and even erroneously indicate differences in the mean, when in actuality both samples are drawn from distributions with the same mean but different variances. A larger sample size (n = 100) more accurately represents both the mean and variance of the population the samples were drawn from. (B) Top: illustration of a normal distribution (blue) and a distribution that results from a combination of two distributions with different means (orange). When only a small portion of the sample (here 10%) comes from a distribution with a different mean, this is difficult to detect in the average. Bottom: a small sample size may not reveal any samples that deviate from the majority, while a larger sample size can make outliers more obvious and detectable. 10% of a sample can appear like a small component that may get discarded as outliers, but it is estimated that 5–10% of the human population is left-handed (Knecht et al., 2000). This percentage is high enough that most people know at least one left-handed person, and discarding these outliers would be akin to neglecting the existence of left- handed individuals.