Introduction

The general goal of the laboratory is the comprehensive identification and examination of neural circuits controlling behavior using the larval zebrafish as a model system. To that end, we have established and quantified a series of visually induced behaviors and analyzed the individual resulting motor components. Using these assays in combination with various calcium indicators and two-photon microscopy we have monitored neuronal activity throughout the fish brain in an awake and intact preparation. An extended goal is the study of how changes or variations in the behavior are reflected in changes in the underlying neuronal activity. To that end, we have developed several quantitative learning assays and tools for in vivo monitoring of neural activity in freely swimming larvae.  

Background

Neuroscientists have long been working to understand how biological structures can produce the complex behaviors that are generated by the nervous system. However, even the basic operational principles governing a brain’s interconnected network of cells have remained painfully elusive. My laboratory is working on a scientific strategy focused on building a complete, multi-level picture of simple neural circuits that will advance our basic understanding of brain function and offers a complete view into the neuronal activity underlying a series of relatively complex behaviors.

Generally the question of how the brain, or neural circuits in particular, function ought to be reduced to the question of what the brain or particular parts of the brain are doing. This is best addresses by rigorous and quantitative behavioral assays that allow us to relate a particular set of input stimuli – in our case visual signals that get translated by the retina into action potentials in ganglion cells – to a set of motor actions that are controlled by an array of spikes in motor neurons – the output of the system. Both data-sets can be assessed to a first order by behavioral experiments. The second question then has to be how the neurons and synapses within the circuit actually perform this computation. Important insights into this problem can be gained by recording activity in every neuron of the fish’s brain in the context of a specific behavior. This would be a truly daunting task in any mammalian preparation but the small and transparent brain of the larval zebrafish allowed us to develop and extend the technology that makes it possible to acquire these datasets in fish expressing genetically encoded indicators in all neurons¹. In parallel we have tried to simplify the question by the examination of individual modules in the brain. Here, we have started to make progress by finding specific nuclei in the fish brain - the optic tectum which processes inputs directly from the retina² on the one hand and the reticular spinal formation, a set of identified neurons that exclusively controls behavior^3,4, on the other hand - that provide stepping stones on the long journey to a complete understanding of the brain and its role in producing behavior.

Current and future projects – Neuronal circuits controlling innate behaviors.

The opto-motor response (OMR)

When presented with visual stimuli containing whole field motion, fish will turn and swim to follow the moving patterns. In order to study the neuronal networks controlling this particular behavior we implemented a method to label all neurons projecting into the spinal cord – the reticular spinal formation - with a dextran-conjugated calcium indicator dye. These neurons forms a bottle neck of circa 300 identified cells that have exclusive control over tail motion and therefore the majority of all fish behaviors. Data describing the response properties of these neurons and their role in transducing visual information into motor output is described in Orger et al. ¹¹. Using a combination of quantitative behavioral analysis, in-vivo two photon imaging and targeted ablations we found that a surprisingly small set of identified neurons – not more than twelve cells on each side of the hindbrain – are responsible for right- or left- turning respectively. These spinal projection neurons link sensory processing in the brain to motor output in the spinal cord and, therefore, provide an excellent starting point for studying the 
complete sensorimotor transformations underlying behavior.

Figure 1: (Left panel) Behavioral scheme with OMR stimulus.  (RIght panel) Functional properties of hindbrain neurons to various moving gratings.  See Orger et al., 2008

Having identified which neurons control particular motor patterns, we can now ask how their activity is decoded in the spinal cord to produce the associated behavior and, importantly, start to investigate the upstream circuitry that leads to the selective activation of these descending control neurons. To that end we have generated a series of transgenic fish lines that express genetically encoded calcium indicators in all neurons.   Head-fixed fish that expresses G-CaMP2 in all neurons views OMR inducing stimuli while activity in its neurons is recorded with a two-photon microscope. This technique will allow us to map out the neuronal activity related to this particular behavior in the complete fish brain with sub-cellular resolution.  So far we find that surprisingly small sets of neurons show selective response properties to stimuli that reliably induce forward swimming or turning in the fish and, remarkably, all neurons upstream of the retina that respond to whole field motion stimuli show very distinct directional preference. Complementary to the functional imaging we are using extra- and intracellular electrophysiology recordings for a detailed characterization of the calcium signals and as an important probe to examine the precise temporal patterns of evoked action potentials and synaptic inputs.

Innate attraction and repulsion

We found that larval fish, when presented with small moving objects, will track these visual stimuli or turn away from them depending primarily on the object’s size and contrast. The central module of the set-ups used to quantify these (and several other) behaviors in larval zebrafish is described in Figure 2. A freely swimming fish is monitored by a high-speed camera, its body position and gaze direction extracted online and visual stimuli are presented through a LPD projector onto a 360 degree screen in an online-feedback loop. This “head-fixed” stimulation gives the experimenter absolute control over the visual input and evoked (as well as spontaneous) motor-output can be recorded with millisecond precision and high spatial resolution. A dataset acquired by this method is shown in the right panel. Larval zebrafish will track virtual prey-like stimuli projected onto the screen as long as their size doesn’t exceed 3 degrees,  approximately the size of paramecia or artemia in a natural hunting situation. Larger stimuli elicit a reliable turning away from the stimulus and projection pattern akin to whole field motion (20 degrees) elicit optomotor-like behavior. These are important findings that tell us which parameters in visual stimulus space get mapped onto distinct output patterns. Experiments to study the role of different brain regions in these transformations are currently under way using 2-photon microscopy and transgenic fish-lines that express genetically encoded calcium indicators in all neurons. However, a first examination of the processing of these stimuli in the optic tectum - one of the first processing stations after the retina - was undertaken using bulk labeling by bolus injection of a calcium indicator conjugated to AM-esther. By combining two-photon imaging, binocular motion stimuli and artificial re-wiring of retinofugal projections we could show that this nucleus serves as a module to extract direction selective information from prey like stimuli ¹². This is particularly interesting since it is strongly reminiscent of higher cortical function in mammalian preparations while in lower vertebrates the retina has been thought to be responsible for most of the extraction of direction selective information.

Figure 2: left panel - a 360° X 90° virtual environment controlled by a custom DirectX (Microsoft ©) graphics engine, coupled with a custom real-time CCD-based object tracking system, which allows the online association of zebrafish behaviour and stimulus presentation. Right panel – responses of a zebrafish larva to moving spots of varying sizes. Spots appear at frame zero straight ahead of the fish and are moved at constant speed for the next 100 frames either to the left or the right. Fish respond with whole body tracking motion to spots of 1 degree size and with distinct turns away from stimuli with a diameter of 5 degrees or larger.

Somatosensory stimuli, targeted stimulation and escape

A brief touch to the head or body of a zebrafish larva will elicit a rapid and pronounced escape response. In order to examine the role of the primary sensory neurons in this behavior we have generated transgenic fish lines expressing Channelrhodopsin2 (ChR2) and an archaeal light-driven chloride pump (NpHR) ¹⁶ under various promotors which enables us to selectively excite or inhibit specific neurons involved in the behavior under scrutiny. A first study employing this technique probed the role of trigeminal and Rohon-Beard neurons in the context of escape behaviors in the larval zebrafish. Here we could show that single, ChR2 evoked action potentials in individual sensory neurons are sufficient to trigger complex escape reflexes in the animal ¹³. This is a first demonstration of the feasibility of this novel technique in the zebrafish preparation and reveals a degree of efficiency in sensory coding that is normally found only in the central processing components of neural circuits.

Figure 3: (Left panel) Channelrhodopsin expression and bahvioral stimulation.  (Right panel) Channelrhodopsin elicits single spikes.  See Douglass et al., 2008

Current and future projects – correlates of learning and memory

Space avoidance

In the past two years we have developed and optimized 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 ¹⁵. This is achieved by pairing the presence of the fish in a visually distinct area of the tank with the delivery of mild electroshocks. Adult fish learn quickly to avoid these specific areas and seek the locations which are marked with visual cues encoding safety. This behavior is robust, lasts for many hours and is dynamic such that the cues can be used after training to chase the fish into specific locations of the tank. We furthermore found that in larvae operant conditioning constitutes a distinctly stronger assay than classical conditioning which needs to be further strengthened by repetitive training sessions. In future studies we will use functional imaging in combination with these assays to seek and describe specific neuronal populations that respond differentially to correlates of acquired information in the brain. These experiments should give insight into the general mechanisms that the vertebrate brain uses to store and retrieve information and will provide valuable clues about the neuronal underpinnings of learning and memory.

Adaptive motor learning

Fish with their heads firmly embedded in agarose will still attempt to swim forward and turn to follow OMR inducing whole field motion patterns. Using high speed cameras we can then analyze the tail motion induced by these visual stimuli, extract the expected motion of the fish from these parameters and feed back the correct visual motion onto the screen. This allows us to put the fish into a virtual reality environment in which it basically uses its tail like a joystick to navigate the virtual world. We have used this set-up in a series of proof of principle experiments that demonstrate that zebrafish larva can – very much like fruit flies in the Heisenberg fly-simulator ¹⁷ - navigate a virtual environment and correct for experimentally induced gain changes in the feedback loop in less than a minute. These forms of motor adaptation constitute a robust learning in the zebrafish larva and open the possibility to use optophysiology in these head-fixed preparations to study the underlying changes in the brain before, during and after learning occurs.

Appetitive conditioning

Electrical self stimulation of dopaminergic nuclei in the rat is one of the most powerful appetitive operant conditioning paradigms in the animal kingdom. We plan to extend our investigations of functional plasticity in the fish brain to include the study of this fundamental feature in animal learning. In order to implement a self-stimulation paradigm in the fish we express Channelrhodopsin2 under dopamine specific promoters. This will allow fish to self-stimulate by seeking out a blue laser beam parked in the fish tank. We have shown proof of principle by expressing ChR2 under the islet-1 promoter which drives expression in touch sensitive neurons in the fish head ¹³ and transgenic fish lines expressing ChR2 and NpHR specifically in the dopaminergic system (TH and DA promoters) are currently generated in the laboratory. Functional imaging of neuronal activity in the dopaminergic and several sensory systems before and after learning in more specific assays will then help to understand the functional changes induced during learning. 

In order to examine response properties of these specific neuronal subpopulation in freely swimming larvae we have developed an imaging technique based on the bioluminescence of Aquorin-GFP that does not require excitation light and can therefore operate on a zero background signal ¹⁸. This optical signal can be collected with large angle optics and detected as a one-dimensional temporal signal by a photon multiplier tube. Figure 3 shows data from a first series of experiments in which Aquorin is expressed in the hypocretin system, a small nucleus consisting of less than 20 neurons that are known to regulate the sleep-wake cycle.

The left side shows an in-vivo two-photon image of the transgenic fish and a diagram of the set-up. On the right side bioluminescence photon counts from the same fish are shown in green below traces describing the simultaneously monitored behavior of the fish in black. These data illustrate the feasibility of recording neuronal activity in a small set of identified neurons in a freely behaving animal with high temporal resolution and over many hours.        

   1.   Higashijima,S.I., Masino,M.A., Mandel,G. & Fetcho,J.R. Imaging neuronal activity during zebrafish behavior with a genetically encoded calcium indicator. J. Neurophysiol. 90, 3986-3997 (2003).

   2.   Niell,C.M. & Smith,S.J. Functional imaging reveals rapid development of visual response properties in the zebrafish tectum. Neuron 45, 941-951 (2005).

   3.   Kimmel,C.B., Powell,S.L. & Metcalfe,W.K. Brain neurons which project to the spinal cord in young larvae of the zebrafish. J. Comp Neurol. 205, 112-127 (1982).

   4.   O'Malley,D.M., Kao,Y.H. & Fetcho,J.R. Imaging the functional organization of zebrafish hindbrain segments during escape behaviors. Neuron 17, 1145-1155 (1996).

   5.   Engert,F. & Bonhoeffer,T. Synapse specificity of long-term potentiation breaks down at short distances. Nature 388, 279-284 (1997).

   6.   Engert,F. & Bonhoeffer,T. Dendritic spine changes associated with hippocampal long-term synaptic plasticity. Nature 399, 66-70 (1999).

   7.   Engert,F., Tao,H.W., Zhang,L.I. & Poo,M.M. Moving visual stimuli rapidly induce direction sensitivity of developing tectal neurons. Nature 419, 470-475 (2002).

   8.   Tao,H.W., Zhang,L.I., Engert,F. & Poo,M. Emergence of input specificity of ltp during development of retinotectal connections in vivo. Neuron 31, 569-580 (2001).

   9.   Vislay-Meltzer,R.L., Kampff,A.R. & Engert,F. Spatiotemporal specificity of neuronal activity directs the modification of receptive fields in the developing retinotectal system. Neuron 50, 101-114 (2006).

10.   Bollmann,J.H. & Engert,F. Topopraphic distribution of visually driven dendritic activity in the vertebrate visual system. (2008).

11.   Orger,M.B., Kampff,A.R., Severi,K.E., Bollmann,J.H. & Engert,F. Control of visually guided behavior by distinct populations of spinal projection neurons. Nat. Neurosci. 11, 327-333 (2008).

12.   Ramdya,P. & Engert,F. Binocular Circuit Properties Emerge Following Retinotectal Rewiring. Nature Neuroscience in press, (2008).

13.   Douglass,A.D., Kraves,S., Deisseroth,K., Schier,A.F. & Engert,F. Escape behavior elicited by single, Channelrhodopsin-2-evoked spikes in zebrafish somatosensory neurons. Curr. Biol. (2008).

14.   Naumann,E.A., Kampff,A.R., Prober,D., Schier,A.F. & Engert,F. In vivo detection of neural calcium signals in the freely behaving zebrafish: A bioluminescence assay. in preparation (2008).

15.   Valente,A., Huang,K.H., Portugues,R. & Engert,F. Learning and memory in the developing and adult zebrafish. PLoS. Biol. submitted, (2008).

16.   Zhang,F. et al. Multimodal fast optical interrogation of neural circuitry. Nature 446, 633-639 (2007).

17.   Heisenberg,M., Wolf,R. & Brembs,B. Flexibility in a single behavioral variable of Drosophila. Learn. Mem. 8, 1-10 (2001).

18.   Baubet,V. et al. Chimeric green fluorescent protein-aequorin as bioluminescent Ca2+ reporters at the single-cell level. Proc. Natl. Acad. Sci. U. S. A 97, 7260-7265 (2000).