Final Report Summary - CORTEX SIMPLEX (Function and computation in three-layer cortex)
We are, in most ways, our cerebral cortex: its circuits serve to shape our perception of the world, store our memories, plan our behavior. Cerebral cortex is found only among vertebrates, and among them only in mammals (such as rats cats and primates) and in non-avian reptiles such as lizards, turtles and crocodiles. Cortex does not exist in frogs, salamanders or in fish. (Birds, which are the descendants of dinosaurs, reptiles themselves, have lost their reptilian cortical architecture and thus do not technically have a cortex.)
Mammals and reptiles originate from a common ancestor that lived some 320 million years ago. We know that this ancestor had a small cortex with three layers, because such a structure is found today in some cortices (such as the hippocampus) of mammals and in all cortices of modern reptiles. The most plausible interpretation for this observation is that a three-layered cortex existed in their common (“amniote”) ancestor. But since cortex does not exist in fish or amphibians, the evolution of cortex probably happened in that amniote ancestor and not earlier. By comparing the cortex of today’s reptiles to that of today’s mammals, we can look for similarities—potential ancestral traits—and differences—results of their parallel evolutions, and thus reconstruct the main features of cortical evolution. By studying reptilian cortex, we also probably study something close to the ancestral cortex, that in which cortical operations first got established. In all likelihood, this ancestral cortex is probably simpler than mammalian neocortex. Chances are, therefore, that understanding its computational operations will be easier.
It is with this reasoning in mind that we initiated a new research program with the help of ERC funding. This approach is already bearing fruit. In the past few years for example, we discovered REM (rapid-eye-movement) sleep and slow-wave sleep in a reptile. This discovery suggests that REM is not the result of convergent evolution among avians and mammals, as previously thought, but rather, the probable result of common descent from their common amniote ancestor. This result was published in Science two years ago (and received world-wide press attention) (Shein-Idelson et al., 2016).
Using state-of-the-art molecular techniques, we also attempted to answer a set of long-standing questions about vertebrate brain and cortical evolution. Using single-cell transcriptomics approaches, we could derive molecular maps of reptilian cortex and discover that reptiles have neurons that correspond to those found in the subdivisions the mammalian hippocampus, a structure involved in episodic memory and spatial navigation. By staining these neurons in the reptilian brain, their precise location could be determined allowing us to draw unambiguous one-to-one correspondence between known mammalian structures, their reptilian equivalent and possibly, their common ancestor as well. With this molecular approach, some of the neuronal types that populate our neocortex can be traced back to cells in the reptilian-mammalian ancestor. This work (Tosches et al., 2018) is now in press at Science.
Using a multi-techniques approach to record neuronal activity in isolated slabs of reptilian cortex (Shein-Idelson et al., Nature Methods 2017), we have recently discovered reliable spike-sequence patterns among populations of excitatory and inhibitory neurons in cortex, indicating the existence of highly reliable connection lines among cortical neurons, and of reliable and input-specific cortical dynamics. This work, now being written up, has important and general implications for cortical dynamics, computation and cortical codes.
In conclusion, by centering our approach on a simple ancestral-like cortex, we have been able to address questions of general significance for brain function, and already shed light on key issues about cortical and behavioral evolution among vertebrates.
Mammals and reptiles originate from a common ancestor that lived some 320 million years ago. We know that this ancestor had a small cortex with three layers, because such a structure is found today in some cortices (such as the hippocampus) of mammals and in all cortices of modern reptiles. The most plausible interpretation for this observation is that a three-layered cortex existed in their common (“amniote”) ancestor. But since cortex does not exist in fish or amphibians, the evolution of cortex probably happened in that amniote ancestor and not earlier. By comparing the cortex of today’s reptiles to that of today’s mammals, we can look for similarities—potential ancestral traits—and differences—results of their parallel evolutions, and thus reconstruct the main features of cortical evolution. By studying reptilian cortex, we also probably study something close to the ancestral cortex, that in which cortical operations first got established. In all likelihood, this ancestral cortex is probably simpler than mammalian neocortex. Chances are, therefore, that understanding its computational operations will be easier.
It is with this reasoning in mind that we initiated a new research program with the help of ERC funding. This approach is already bearing fruit. In the past few years for example, we discovered REM (rapid-eye-movement) sleep and slow-wave sleep in a reptile. This discovery suggests that REM is not the result of convergent evolution among avians and mammals, as previously thought, but rather, the probable result of common descent from their common amniote ancestor. This result was published in Science two years ago (and received world-wide press attention) (Shein-Idelson et al., 2016).
Using state-of-the-art molecular techniques, we also attempted to answer a set of long-standing questions about vertebrate brain and cortical evolution. Using single-cell transcriptomics approaches, we could derive molecular maps of reptilian cortex and discover that reptiles have neurons that correspond to those found in the subdivisions the mammalian hippocampus, a structure involved in episodic memory and spatial navigation. By staining these neurons in the reptilian brain, their precise location could be determined allowing us to draw unambiguous one-to-one correspondence between known mammalian structures, their reptilian equivalent and possibly, their common ancestor as well. With this molecular approach, some of the neuronal types that populate our neocortex can be traced back to cells in the reptilian-mammalian ancestor. This work (Tosches et al., 2018) is now in press at Science.
Using a multi-techniques approach to record neuronal activity in isolated slabs of reptilian cortex (Shein-Idelson et al., Nature Methods 2017), we have recently discovered reliable spike-sequence patterns among populations of excitatory and inhibitory neurons in cortex, indicating the existence of highly reliable connection lines among cortical neurons, and of reliable and input-specific cortical dynamics. This work, now being written up, has important and general implications for cortical dynamics, computation and cortical codes.
In conclusion, by centering our approach on a simple ancestral-like cortex, we have been able to address questions of general significance for brain function, and already shed light on key issues about cortical and behavioral evolution among vertebrates.