Current projects

Anatomical circuit mapping

We are interested in both the physical and the functional connections that underlie brain function, so we use both anatomical and optogenetic approaches for circuit mapping.  Anatomical approaches include imaging the architectures of individual neurons within a brain region of interest.  This allows us to view these brain regions as collections of discernible cell types, which is a prerequisite for explaining how those cells form functional circuits.

A single tectal periventricular neuron labelled the BGUG technique
(adapted from Scott et al., 2009).

Cerebellar projection neurons.  See caption for more explanation.

Tracings of five cerebellar projection neurons to the tectum, and a spatially registered map of topography between the structures (adapted from Heap et al, 2013).


We also use targeted photoconversions to resolve patterns of projections between different brain regions.  By illuminating Kaede-expressing neurons with violet light, we can cause Kaede to change from a green to a red fluorophore.  If we target this photoconversion to particular cells and wait for the red Kaede to diffuse into the axons, we can identify these cells’ targets throughout the brain. 

Kaede photoconversions

tectal neuropil from a larva in which both the thalamus and retina are labelled with Kaede

These are images of the tectal neuropil from a larva in which both the thalamus and retina are labelled with Kaede. We have photoconverted Kaede in thalamic, but not retinal, projection neurons, allowing us to determine which layers of the tectal neuropil receive inputs from each.


Sensory processing and integration

We are interested in addressing how the brain perceives and integrates information about the outside world, and we have expanded our experiments to cover as many sensory modalities as possible.  Most of these projects have been based around calcium imaging in large populations of neurons (the whole brain in some cases), followed by further anatomical or functional analyses of the neurons involved.

In the visual system, we have characterised the tuning of neurons in the tectum to visual stimuli with different levels of saliency (Thompson et al, 2016).  We have also used calcium imaging, anatomical mapping, ablation experiments, and behavioural analyses to describe a role for the thalamus in detecting changes in luminance, and have described how this information contributes to the likelihood and the direction of escape behaviours in larval zebrafish presented with a threatening visual stimulus (Heap et al, 2018).

The tectum’s responses to predator-like stimuli depend on input from the thalamus, while thalamic input is not involved in the tectum’s responses to other visual stimuli (left).  The animal’s behavioural responses are also dependent on this thalamic activity.  They are less likely to startle, and reverse the direction of their startles, when thalamo-tectal communication is interrupted on the relevant side of the brain (right). Adapted from Heap et al, 2018 


We have recently published the first whole-brain analysis of auditory processing in larval zebrafish, indicating that it is rudimentary in terms of frequency discrimination and tonotopy in larvae.  These auditory responses are, however, useful to characterise, as this now puts us in a position to combine auditory stimuli with stimuli from other modalities in studies of sensory integration.

To study the perception of water flow, we have designed a microfluidics device capable of delivering controlled flow over the trunk of larvae while they remain stationary for calcium imaging.  This permits us to observe the neurons in the brain that respond to water flow, their response properties when the direction or strength of the flow changes, and their locations within the brain.  This will allow us to build models of how flow is perceived in the early stages of this pathway, and how emergent properties, such as representations of accumulated flow over time, arise later in the sensory processing cascade.

A microfluidics device for stimulating the lateral line system during whole-brain calcium imaging (top left). Neural responses to a simple stimulus, alternating reverse flow followed by forward flow, fall into eight categories with distinct response properties (bottom left), and these different types of responsive neurons map back to characteristic locations within the brain (right).  Unpublished data from Gilles Vanwalleghem.

Our work in the vestibular system is aimed at addressing one of the fundamental challenges to studying this modality: vestibular stimuli necessarily move the subject, and this makes most approaches for studying activity (electrophysiology, fMRI, calcium imaging) difficult or impossible.  Using the physics technique of optical trapping, we have applied forces to the otoliths (ear stones) of immobilised zebrafish larvae, tricking them into thinking that they are moving when they are stationary.  This permits us to perform calcium imaging to reveal the cells and circuits involved in vestibular processing.  The optical physics and behaviour for this approach are described in (Favre-Bulle et al, 2017) and our description of the brain-wide vestibular processing is (Favre-Bulle et al, 2018).

A diagram (left) of our optical setup, where an infrared optical trap is applied to the utricular (vestibular) otolith while we preform whole-brain volumetric calcium imaging.  The centre panels show a larva before and during the application of the optical trap, demonstrating the postural change that the trap elicits.  This tail bend (and eye movements not shown here) scale with the strength of the laser used, and therefore with the perceived magnitude of the fictive vestibular stimulation.  The full story can be read in (Favre-Bulle et al, 2018).

We observe three distinct functional clusters of neurons responding to vestibular stimulation (mean responses shown on the left), and these are spread across characteristic brain regions (dorsal view shown in the centre, with sagittal views on the right). The full story can be read in (Favre-Bulle et al, 2018).

As we further refine our techniques for stimulating a range of sensory modalities and get baseline data for how each modality is processed in the brain’s circuits, we are shifting our attention to important biological question that are made accessible by these techniques.  These include how information from different modalities influence one another and are integrated, how habituation and attention are manifested in this circuitry, and how there circuits change in models of human sensory disorders.

Functional circuit mapping with optogenetics

The anatomical and imaging approaches described above still don’t concretely demonstrate how information passes through the brain circuits.  To address this, we need to test the functional relationships among the cells that we are studying.  For this, we turn to optogenetics.   By driving activity specifically in one brain region that we are interested in, and performing calcium imaging of the neurons in another region that may be connected, we can see whether and in what way one region’s activity affects cells in the downstream region.  Given the complex and interconnected nature of the nervous system, this does not necessarily mean that the stimulated cells and the responsive cells are synaptic partners (anatomical mapping is needed for that), but it does establish a functional relationship between one region and another.

Optogenetic circuit mapping

Image showing optogenetic circuit mapping.  See text below for explanation.

Optogenetic stimulation using sculpted laser light and Channelrhodopsin2 (blue circle, left) can be used to drive activity in a small patch of the cerebellum.  The resulting activation in tectal neurons is shown in green.  The right panel shows a raster plot of individual tectal cells that respond to this cerebellar stimulation and the average response of this population to three consecutive trials. (Unpublished data from Lucy Heap.)

Zebrafish models of Autism Spectrum Disorder

Psychiatric disorders necessarily have their origins in the dysfunction or loss of neural circuitry.  Our lab has recently begun modelling Autism Spectrum Disorder (ASD) by mutating genes linked to ASD in humans.  The strength of the zebrafish model system is that it permits access to the functioning brain, thus allowing analyses of developmental dynamics, population-level neural coding, and synapse dynamics in vivo. 

We are currently in the process of using CRISPRs to mutate a number of zebrafish homologs to ASD-linked human genes.  These mutants will be used to gauge the anatomical, developmental, and physiological mechanisms underlying ASD etiology, at the cellular level. 

Whole-brain calcium imaging of neural activity

The mechanisms by which signals from the outside world are perceived, processed, and integrated are largely mysterious.  We hope to reveal some of these processes with whole-brain imaging of neural activity while our animals are experiencing sensory stimuli.  This involves expression of GCaMP6 (a genetically encoded calcium indicator developed at HHMI Janelia Research Campus) throughout the brain, combined with selective plane illumination microscopy (SPIM).  You can learn about our home-made SPIM microscopes and bioinformatics approaches on our Methods and Machines page.

To investigate a range of sensory modalities, we have designed microscopes that provide open space around the larva.  This permits us to surround the larva with LCD screens for visual stimuli, auditory speakers, and even microfluidics devices for generating water flow across the animal’s body.  We then look at activity across tens of thousands of neurons while these stimuli are presented alone or in combination.  The result is a map of what types of stimuli elicit what types of neural responses, where the responding neurons are, and what their relationships with each other are.  All of this moves us a little closer to understanding how these senses are encoded and integrated by the brain.

Responses of segmented neurons in the tectum to a visual stimulus, as revealed by SPIM.
This image is from (Thompson et al, 2016)