If two nerve cells of the same type were to be exchanged with each other, both might no longer be able to fulfill their tasks. Studies on the retina of mice show that previous assumptions about nerve cells need to be reconsidered: Tiny details in the anatomical characteristics can lead to significant changes in the cells’ response behavior. A cell that does its job well and processes signals in a meaningful way at one specific location within the neuronal network could be completely useless at another location of the network because its anatomical properties would not allow meaningful signal behavior there. This was investigated by a research team from the Vienna University of Technology and Harvard Medical School. A publication on the subject has now appeared in the journal "Science Advances".
Cells in the mouse retina
"The light sensitivity of some cell types of a mouse retina is different in the upper part of the visual field than in the lower part," says Paul Werginz, first author of the study. "In the lower area there is prey, above is the sky, which it is brighter". Evolution has thus led to the fact that the neuronal networks in different areas of the mouse retina react with different sensitivities, although the cells are of the same type.
Years ago, Paul Werginz studied the response behavior of nerve cells on a theoretical level at the TU Wien (Vienna) and developed computational models for this purpose. Funded by an FWF Schrödinger Fellowship he then went to Harvard Medical School for two years, where he was able to analyze the nerve cells individually under the microscope in numerous experiments. Now he has returned to TU Wien and is presenting his results together with the Harvard team.
"Until now, when simulating nerve networks on the computer, people usually assumed that neurons of the same type had basically identical properties," says Paul Werginz. When we learn something, the strength of the connection between the cells changes. The way in which the activity of one cell influences the activity of the next cell is modified - this is also the basic idea behind artificial neural networks, which are used in software development for creating artificial intelligence. However, it was generally assumed that the cells do not significantly change their anatomical and physiological properties in the process and that all cells have the same possibilities to convert signals from neighboring cells into new signals.
But as it turns out, this view is not sufficient to explain the behavior of biological nerve networks. "In the initial portion of the axon of the nerve cell, where the signals from the neighboring cells are integrated and where it is decided whether the cell fires or not, there are important anatomical differences between the individual cells," says Paul Werginz. "We have investigated many different parameters, such as the length of the section known as the Axon Initial Segment, and we have found that such anatomical parameters play a crucial role in determining the behavior of the nerve cell.” This means that not every cell is capable of every kind of signal processing. The anatomical details of the cell must be adapted to its task.
"The cell can only be useful for data processing if it delivers a signal in certain situations and not in others," says Vineeth Raghuram, co-first author of the paper. "But this is only possible if the anatomical parameters of the cell, such as the length of the Axon Initial Segment, have exactly the right values - and these values must match the environment of the nerve cell exactly; they may be different in the upper field of vision of the mouse than in the lower one.". If these parameters were different, the cell would not deliver a signal that can be interpreted by the brain in a meaningful way.
"This means that even nerve cells of the same cell type behave differently due to their different anatomical details - exactly as their task in their local environment requires," explains Werginz. If two cells were to be exchanged, both might be useless just because their anatomical features are slightly different.
Small differences with great significance
How these subtle differences develop and whether they are genetically predetermined has not yet been fully clarified. But it is obvious that this is not just ordinary neuroplasticity as it underlies normal learning processes. "If we want to understand the behavior of nerve networks, we must in any case take into account that these subtle differences can completely change the signaling behavior of a single cell," emphasizes Paul Werginz. "I can well imagine that this could also play an important role in the investigation of neuronal diseases.”
P. Werginz, V. Raghuram, S.Fried: Tailoring of the axon initial segment shapes the conversion of synaptic inputs into spiking output in OFF-α T retinal ganglion cells, Science Advances Vol. 6, no. 37., opens an external URL in a new window
Dr. Paul Werginz
Institute for Analysis und Scientific Computing
+43 1 58801 10166
Dr. Florian Aigner
PR und Marketing
Technische Universität Wien