Learning-Induced Changes in Neuronal Network Activity in the Behaving Larval Zebra-Fish

Florian Engert, Ph.D.

Harvard University, Cambridge, MA

Grant Program:

David Mahoney Neuroimaging Program

Funded in:

June 2006, for 1 years

Funding Amount:


Lay Summary

Learning-Induced Changes in Neuronal Network Activity in the Behaving Larval Zebra-Fish

Lay Results:
The brain collects information from the world and uses it to select appropriate behaviors. The neural command is sent to the muscles via a group of cells which project into the spinal cord.  The translucent brain of zebrafish allows us to look at the activity of all these neurons, so we can ask how this command is encoded.  Do all neurons participate in every behavior, or are individual actions controlled by dedicated subsets?  We find that small groups of neurons in the fish control different behaviors, revealing that the brain’s control system is composed of surprisingly organized pathways, connecting sensory stimuli to behavioral responses.

Scientific Results:
We have developed novel techniques for real-time behavioral analysis, recording of visually-evoked activity by 2-photon imaging and targeted cell ablations, to address a long-standing question about the organization of the brain’s motor command, using zebrafish as a model system. We have mapped visual responses in the entire population of motor control neurons, revealing distinct groups of identified control neurons that are necessary for generating specific behaviors.  This result is consistent with hypothesized models of motor control, but has not been previously demonstrated in a vertebrate system.


Learning-Induced Changes in Neuronal Network Activity in the Behaving Larval Zebra-Fish

How acquired memories are stored and retrieved in functional neuronal networks in living animals is a fundamental question of great interest that is still largely unresolved. The current proposal offers an approach to gain insight into this question with two complementary strategies. The first is to develop and optimize a learning assay for larval zebrafish in which the animals learn to associate two distinct but intrinsically neutral visual stimulus patterns with danger and safety, respectively. This is achieved by pairing the presence of the fish in a visually distinct area of the tank with the delivery of mild electroshocks. In several pilot experiments, we show that the fish quickly learn to avoid these specific areas and seek the locations which are marked with complementary visual cues. This behavior is robust, lasts for many hours, and is dynamic such that the cues can be used after training to manipulate the fish into specific locations of the tank.

The second proposed strategy is to develop an in-vivo calcium imaging assay in those larvae to allow the monitoring of activity in large populations of neurons in response to these specific visual stimuli. The small size of the larval zebrafish brain in combination with the almost perfect translucence of the preparation permits functional in-vivo imaging of any chosen part of the central nervous system. The widespread use of the zebrafish as a genetic model system further allows to target specific population of neurons for the expression of genetically encoded calcium indicators, which offer an attractive alternative to synthetic dyes for the purpose of functional imaging studies.

Initially, the well established methods of single cell electroporation of dextran dyes as well as the injection of calcium indicators in the form of AM-esthers will be used in our studies, but a spectrum of recently designed genetically encoded indicators will also be used to target and analyze neuronal response properties in parallel. The function and reliability of synthetic as well as genetically encoded calcium indicators is analyzed by simultaneous in vivo patch clamp recordings of the labeled neurons.

The general aim of these experiments is to seek and describe specific neuronal populations that respond differentially to correlates of acquired information in the vertebrate brain.



The storage of new information in the vertebrate brain requires the rearrangement of synaptic connections and consequently will result in different patterns of neuronal activity in response to sensory inputs. We hypothesize that these changes in neuronal activity in response to specifically learned stimuli can be localized and quantified with calcium imaging methods in the intact zebrafish brain.

Aim 1. generate a robust assay for visual learning in the larval zebrafish

Aim 2. improve and adapt functional two-photon laser scanning microscopy in the living zebrafish to analyse learning induced changes in neuronal activity

Learning assay: We are developing a robust learning assay that allows us to teach fish from early developmental (pf7 – seven days post fertilization) to juvenile stages to swim away from specific visual cues. In order to induce this aversion, freely swimming fish receive mild electroshocks in synchrony with the delivery of previously neutral visual stimuli.

Imaging: calcium indicators like Oregon-Green-Bapta are delivered either to individual neurons by single cell electroporation of its dextran form or, alternatively, neurons are bulk labeled by injection of the AM-esther into the brain region of choice. Calcium signals are then used to report neuronal activity in different areas of the fish brain. A particularly interesting example is the identified reticulospinal neurons, MeLc and MeLr, which are found in the nucleus of the medial longitudinal fasciculus of the tegmentum. These neurons extend dendrites into the ipsilateral tectum and project axons into the spinal cord and are thus placed in a key position to integrate relevant sensory information and send it on to the spinal cord.

Selected Publications

Orger M.B., Kampff A.R., Severi K.E., Bollmann J.H., and Engert F. Control of visually guided behavior by distinct populations of spinal projection neurons.  Nat Neurosci. 2008 Mar;11(3):327-33 .