Martin’s paper “Convergent Temperature Representations in Artificial and Biological Neural Networks” was published in Neuron!
Martin is a postdoc in the Engert and Schier labs and has been interested in how temperature is processed by the zebrafish larval brain. In this exciting paper Martin trained artificial neural networks to navigate a heat gradient. When comparing these results to known biological neural networks (building on work in his previous paper), there were striking similarities of representations and computation of heat sensation. The artificial neural network predicted previously undescribed neurons which led to the identification of a new temperature-responsive cell type in the zebrafish cerebellum. This work is an excellent example of how artificial networks can be used to uncover biological insight.
Read the Harvard Gazette’s write up of the work here