Genetic and neural mechanisms underlying emerging social behavior in zebrafish
Our goal is to understand emerging collective behaviors of groups, such as schooling and shoaling in fish.
Our approach is to dissect basic sensorimotor transformations in the zebrafish, which we believe play a
fundamental role in explaining emerging social interactions. We have identified two simple and well described
reflexive behaviors: 1) the optomotor reflex (OMR), where fish swim along with whole field motion stimuli and
2) object evoked re-orienting responses (OER) where fish turn away or towards moving objects, depending
on the object’s size and movement. We have shown in preliminary modeling studies that an implementation
of these two simple “motor primitives” in virtual agents can explain a significant fraction of the emerging social
behaviors in adult fish. A compelling advantage of focusing our studies on these two simple reflexes is that
they are robustly expressed in 7 day old larvae, which facilitates a detailed and quantitative behavioral
analysis of the related visuomotor transformation, as well as a dissection of their underlying neural circuitry.
A critical element in our proposal is the generation of mutant zebrafish that we have shown to display subtle
but distinctive social behavioral phenotypes at the adult stage. We found that, even in the larval stage, and
prior to onset of robust schooling and shoaling behaviors, these mutants already reveal behavioral
phenotypes in the context of the OMR and OER, and that these phenotypical deviations are predictive of the
later emerging differences in schooling and shoaling in adults. One of our central goals is the dissection of
the specific changes in neural circuitry in the mutants that are responsible for these altered behavioral
phenotypes. Some such changes in neural phenotype may manifest at the level of global brain structures, but
many are likely to disrupt micro-circuits - either at the level of cellular identities or synaptic connectivity - that
underlie both simple behavior in the embryo and more complex behaviors in the adult. Notably, we already
have generated realistic circuit models that form specific hypotheses about the neural networks underlying
the OMR and OER in wild-type animals, and these models are readily adjusted to identify and constrain the
specific latent variables that are changed in the mutant animals. Such adjusted models serve as ideal priors
and specific hypotheses to be tested in brain wide functional imaging experiments. Lastly, the identification
of detailed neural phenotypes in mutant animals in terms of anatomical location, neuronal cell fate and
synaptic specificity will facilitate linkage of these anatomical and physiological changes to specific cell fates
and molecular pathways. Our parallel ongoing efforts in describing and modelling brain wide neural circuits in
zebrafish (within the framework of the U19 Team-Research BRAIN Circuits program) will allow us to narrow
down which of all these observable neural phenotypes in the mutants are responsible and causally related to
the specific neural changes that underlie the changes in behavior.