Harvard University COVID-19 updates

Department News

Feeling the Heat – How Does the Zebrafish Brain Process Temperature? [Engert- and Schier Labs]

Feeling the Heat – How Does the Zebrafish Brain Process Temperature? [Engert- and Schier Labs]

The Engert and Schier labs address in a recent paper in Neuron (PDF) how larval zebrafish senses temperature and transforms the sensory representation to generate behavior with the ultimate goal of heat avoidance.

The sense of temperature gives us important information on how to interact with the world. In order to maintain a comfortable temperature, we change our clothing with the seasons (with exceptions – such as the last author of this paper) and seek shelter from extreme heat or cold.

For animals, such as zebrafish, that cannot regulate their own body-temperature, behavioral strategies for seeking out preferred environmental temperatures are especially important. These strategies allow the animal to avoid areas in which the temperature is either too hot or too cold. To avoid noxious heat the animal first needs to detect environmental temperature, then neural circuits in the brain need to extract relevant information (temperature level and change) and ultimately generate the appropriate behavior (swim straight or turn).

In a recent paper in Neuron (PDF), Martin Haesemeyer, Alex Schier, Florian Engert and colleagues uncovered how these processes are implemented in the brain of larval zebrafish. They used brain wide 2-photon calcium imaging to record neuronal activity during heat stimulation while simultaneously recording behavioral output. This approach made it possible to identify neurons throughout the brain that are specifically related to temperature changes as well as behavioral elements.

Temperature sensing cells were identified in the trigeminal ganglia, a pair of sensory ganglia of the trigeminal nerve that are thought to be responsible for encoding a variety of somatosensory modalities in all vertebrate animals. In the zebrafish two types of heat sensitive cells were identified in these ganglia: “ON cells” that were excited by increases in temperature and “OFF cells” that were inhibited upon temperature rise. Both cell types tracked the stimulus with slow dynamics and thus essentially represented the current temperature.

Strikingly, heat responses diversified in the hindbrain area that receives information from the trigeminal ganglia. Here, five types of temperature responsive cells were identified, including two cell types that were most sensitive to an increase in heat or a decrease in temperature. Thus, cells and circuits in the hindbrain are computing a derivative of the temperature stimulus that informs the larval zebrafish about the direction and magnitude of temperature change (warmer or colder). Importantly, these new cell types contained information that was required to explain larval zebrafish behavior with a simple linear model. Haesemeyer and colleagues use this finding to argue that larval zebrafish compute and use such information about temperature change to guide behavioral output as predicted by a previous behavioral study by the same authors (Haesemeyer et al., Cell Systems, 2015).

Haesemeyer et al. next developed a dynamic circuit model for heat perception in larval zebrafish that includes connectivity between cell types as well as how cells temporally transform their inputs. The strength of such models in neuroscience is that they allow researchers to form strong, testable predictions about circuit architecture. Here, for example, the model predicts mixed excitatory and inhibitory influences of heat responsive cell types on pre-motor cells in the hindbrain that ultimately drive heat avoidance. By combining functional imaging with labeling specific neurotransmitters Haesemeyer et al. could confirm that the different hindbrain cell types indeed contain excitatory and inhibitory neurons.

In summary, Haesemeyer and colleagues identified a hindbrain region that implements a small computational motif that calculates a derivative of its input to generate behavior. Modeling predicts that this is achieved by a combination of adapting and non-adapting neurons in this region. Further studies guided by this model, including electrophysiology and connectomics, can reveal the full circuit implementation of this computational motif. Since the computation of the direction of change of a sensory stimulus allows for a simple form of prediction – if I know how something changes I know its value in the near future – it can be expected that this brain circuit motif is widespread throughout the animal kingdom.

by Martin Haesemeyer


Florian Engert faculty profile, Engert lab website

Alex Schier faculty profile, Schier lab website



(l to r) Florian Engert, Drew Robson, Martin Haesemeyer, Jennifer Li, and Alex Schier

(l to r) Florian Engert, Drew Robson, Martin Haesemeyer, Jennifer Li, and Alex Schier