This is work in progress, in collaboration with Alexis Prevost and Laetitia Pontani. After a long personal reflection on the conceptual foundations of theoretical and computational neuroscience, I have come to the realization that most models in systems neuroscience are not in fact biological models, in that function is only obtained by imposing abstract constructs such as “decoders” or “ideal observers” of neural activity, which have no biological basis. Thus, I have concluded that to make decisive progress on modeling perception, one should consider sensorimotor systems, in which behavior is explicitly modeled as the result of physiological processes. I have also come to the sad conclusion that, at this stage, this appears nearly impossible for any mammalian brain, and even extremely challenging for small nervous systems such as C. Elegans.
I recently got interested in Paramecium. Paramecium is a large unicellular eukaryote organism which swims in fresh water using its cilia. When it hits an obstacle, mechanosensitive channels open, depolarize the membrane and trigger a calcium-based action potential. In turn, the action potential triggers a reversal of the swimming direction, followed by a change of direction. This is called the “avoiding reaction”, illustrated in the figure below (A) from Jennings (1906, an amazing book on the behavior of microorganisms; I have collected a number of papers and books he wrote on the subject). Panel B shows the action potential with the ciliary reorientation (from Eckert & Naitoh, 1970).
It displays some of the basic physiological features found in nervous systems of multicellular organisms (except obviously for synapses, although it does have GABA receptors), and for this reason it has been called “a swimming neuron”. Paramecium provides a unique opportunity to study and model an entire sensorimotor system, linking the biophysics of sensory transduction to ecologically relevant behaviors, such as navigation and food foraging in complex and crowded environments. Paramecium is also sensitive to chemicals, light, gravity, water currents, temperature. For example, to locate food (it eats bacteria), it uses the avoiding reaction in the following way: when chemical concentration decreases, it spikes, which triggers a change in direction, so that by trial and error, it manages to find the source.
With my collaborators, we have started developing experiments to characterize Paramecium physiology and behavior, and to develop integrative models (see some preliminary videos featuring electrophysiology and automatic tracking). We have first developed a simple device to immobilize Paramecium to perform electrophysiological recordings (1).
We are currently developing a quantitative model of the action potential, models of chemotaxis, and detailed quantitative descriptions of trajectories. I am also currently writing a detailed review of Paramecium from a neuroscience (and theoretical neuroscience) perspective – it turns out that there is a very dense literature on the subject, mainly from the 1970s (see the peeriodical I am editing on the subject). We have obtained some seed funding from the CNRS and Sorbonne Université to develop this line of research.
Let me know if you are interested in collaborating on this project. We are interested in technical developments as well as in the analysis of data (we are generating videos and electrophysiological data), or even normative approaches to modeling. Or if you are a biologist (preferably in Paris) and want to help with e.g. genetics or other approaches, you are welcome!
Relevant publications (chronological order):