The Central Nervous System: Meaning and Origin

I recently discussed the origin of the nervous system(s), focusing on the earliest neuron:  a cell with an axon that carries an action potential.  Now I want to extend that discussion to the Central Nervous System (CNS), a subject of several recent blog posts (e.g. at Neurophilosophy and NeuroDojo) and a "News Focus" article at Science Magazine.^[19] 

It may seem a little strange to be combining "meaning" and "origin" in one discussion:  one is a semantic issue, the other a matter of research.  However, the fact is that research is normally done within a paradigm, and part of what makes up a paradigm is the system of definitions, either tacit or explicit, that allow us to precisely define the meaning of a specific research question.

In my previous post on the subject, I suggested that the original purpose of nerve cells was long distance/high speed communication of chemical state, the early form of what we might call emotions in higher mammals (and perhaps other vertebrates).  The modern descendants of this system in humans and other vertebrates are the autonomic and limbic nervous systems, which carry a large variety of chemical messages throughout the body, at sub-second speed.  Some of these messages pretty much duplicate those of hormones distributed more widely but slowly by the bloodstream, while others are probably restricted to specific areas of the brain that require the information.  In addition, there may be some populations of neurons that carry "region-specific" chemical messages:  where the same chemical "means" something different depending on which area of the brain it's released in.  The dopamine and serotonin systems almost certainly fall into this category (at least in vertebrates).

The CNS in vertebrates is specifically defined as the brain and spinal cord, which are protected by meninges, a series of membranes that separate the cerebrospinal fluid from the bloodstream, and the bony structures of the skull and vertebrae.  This definition has been extended by analogy to such invertebrates as insects, which also have brains and systems of nerve cords.

The CNS and the peripheral nervous system (PNS) together make up the entirety of the vertebrate nervous system.  However, the PNS is divided by function into two very separate systems: the somatic nervous system (SNS) and the autonomic nervous system (ANS).  The SNS extends the CNS to the "voluntary" muscles and external senses, allowing the body to act as a unit in response to external stimuli.  The ANS makes contact with the internal sensory system which monitors such things as the levels of carbon dioxide, oxygen and sugar in the blood, arterial pressure and the chemical composition of the stomach and gut content, and communicates to such organs as the salivatory glands, heart, gut, etc.

Both parts of the PNS actually communicate between the CNS and their target senses and organs.  However, the control of "voluntary" muscles is largely a different system from that of the viscera, even within the brain and spinal cord.  Even though they travel through the spinal cord and are centralized in the brain, these systems are pretty much independent, although they are hooked together at several points.

The visceral parts of the body are mostly connected to the ANS, which communicates (via the spinal cord) with the limbic system in the brain.  This system is often thought of as being concerned with emotions, and it seems to include neurons that produce a much larger number of types of neurotransmitters and neurohormones than the parts associated with control of "voluntary" muscles and overall movement of the body.

Most of the vertebrate brain seems to be dedicated to control of the "voluntary" muscles, which is hardly surprising, since this function requires a lot more precision in location than is needed for communicating chemical information.  It's hardly surprising, therefore, that many more neurons, and a much larger number of connections, is needed for it.

The fact that these systems seem to be fairly independent doesn't mean that they don't communicate in several important ways.  For instance, the limbic system extends axons into the rest of the brain, where it delivers a variety of neurohormones involved with emotional state, allowing all the "voluntary" and "conscious" calculations to use the current emotional state as input.

Nevertheless, the difference between these systems is extremely important, probably much more so than the classical distinction between the CNS and the PNS.

When we look for the origins of the CNS then, we need to begin with a precise definition of what we're looking for.  As mentioned above, the "classical" definition is restricted to vertebrates, and the extension to insects and other invertebrates was by analogy.  Basically, the distinction was anatomical:  nerve cords containing lots of axons and neuropil were considered to constitute CNS, while distributed systems of nerves and ganglia were not.  However, most ganglia contain neuropil, so the distinction is somewhat fuzzy.

When it comes to the actual beginnings of the CNS, we need to be more careful.  Even cnidarians have neuropil,^[18] while the acoels, arguably the most basal-branching bilaterians, have a single "ganglion" in their larval form.^[16] What are probably the most basal of the eumetazoans, the ctenophores, have a nervous system "composed of a subepidermal nerve net, with neurons concentrated around the mouth and apical organ; there are also nervous elements in the mesoglea, with high concentrations in the tentacles, and near the base of the endoderm (Fig. 1d; [ref])."^[17]


Figure 1: Ctenophore body plan and development.  Click on image to see full illustration with caption in article. (From Ref 17.)


Figure 2: Cross-sections of the body wall and brain of Neochildia juvenile.  Muscle fibers are labeled "msf".  Click on image to see original illustration and caption in article.  (From Ref 16 Figure 3.)

What I would propose is to start by excluding the ANS and limbic system from consideration.  While parts of these systems are associated with the CNS, IMO this was purely opportunistic:  a specialized system developed for controlling the muscles, and these nerves moved in too, taking advantage of the protection and supplies of food and oxygen....

Here we need to consider the original system of the common ancestor of the surviving Eumetazoans.  The prevailing view is that the primitive system was a diffuse nerve net, however I've just recently argued that the diffuse nets in the Cnidarians and Ctenophores probably represent specialized adaptations for their lifestyle, and cannot be assumed to represent the ancestral pattern.  If we consider the various small larval forms of the acoels and sponges, lineages that somewhat bracket the forms we're looking at, we find a creature small enough to coordinate its movements with a single bundle of neuropil.

Given this, we can probably suppose a beginning with a single bundle of neuropil, with the muscles extending processes into it to collect control information.  This is consistent with the larval acoel Neochildia fusca, while the adult has a more diffuse structure: 

The size of the brain in juvenile Neochildia is considerable.  Other flatworm taxa possess a ganglion-like brain consisting of a central neuropile surrounded by a cortex of cell bodies.  In adults of Neochildia and other acoel species such a brain has been reported absent.  Instead, neuronal somata and neurites, together with muscle fibers and gland cell processes form a loose reticulated mass at the anterior pole of the animal ([refs]).  However, we find that in late embryos and juveniles of Neochildia, the brain is more compact.  A 3-4 cell-diameter-thick layer of neurons forms a cortex surrounding a central neuropile that is relatively free of cell bodies.  In addition, an "orthogonal" array of muscle fibers penetrates the brain (Fig.  3).^[16]

In cnidarians we find a highly sophisticated system for synchronizing muscle activity over long distances.^{[11] [12] [18]}  Most of this sophistication (IMO) is necessary to protect from the poisoning or damage of large areas of the bell during conflicts, however the need for synchronization would be present even without the need for such distributed systems.

In this regard, I want to especially mention a recent paper:  Cnidarians and the evolutionary origin of the nervous system by (Hiroshi Watanabe, Toshitaka Fujisawa, and Thomas W. Holstein).  Here we have a summary of recent research into the localized expression in Cnidarians of genes which, in bilaterians, control the differential development of motor nerves and sensory systems, along with an analysis of the localized development of different types of nerve cell.  This paper provides a useful and very recent survey of what's known regarding the "CNS" of Cnidarians, and the likelihood that it is homologous to that of the Bilaterians: 

Anatomically and physiologically elaborated neural structures observed in cnidaria clearly reflect a considerable degree of regionalization of the cnidarian nervous system.  Although these data have prompted several authors to propose that the nerve ring is the cnidarian CNS ([ref]) that may be even homologous to the bilaterian CNS ([ref]), it still remains unclear whether these regionalized (centralized) neural networks in Cnidaria represent a homologous feature to the CNS of Bilateria.  There is no reason why anatomical features of nerve tracts observed in cnidarians should necessarily be homologous to the CNS of bilaterians.  Even if one accounts for a great degree of physiological complexity observed in the regionalized nervous system and eye-bearing sensory complex (Rhopalia) of cnidarians, an increasing number of examples indicating the evolutionary convergence in both nervous and sensory systems ([ref]) make it difficult to simply compare cnidarian and bilaterian neural structures and to place the evolutionary origin of CNS as commonly discussed ([ref]).  The idea that the rudimentary neural centralization observed in some cnidarians is an antecedent characteristic of the eumetazoan nervous system should be examined at least by comparative molecular analysis.^[11]

I should point out that this review was written based on the assumption that the distributed nerve net in Cnidarians (and ctenophores) was primitive.  If, as I have proposed, it is secondary, the possibilities of homology between the concentrations of nerve cells in specific regions and similar concentrations in bilaterians are much more plausible. 

Another consideration has to do with using cilia for swimming.  In general, only the ctenophores among larger animals still use this method, but it was almost certainly the ancestral method for at least normal swimming, with muscles reserved for special situations such as capturing food (e.g. in ctenophores) and escape from predators or natural hazards such as produces by wave action.  The need for nerves to coordinate ciliary action would become increasingly important as size increases, and, unlike muscles, ciliated cells don't naturally have the features needed to coordinate their activity over large areas.  I have suggested that the action potential began as a feature of muscle fibers, to coordinate their activity, and the need for the sort of rapid control over ciliary swimming would probably not have arisen in a small creature, where a slower, more diffuse, system depending on the sort of calcium waves seen in astrocytes might well have been fast enough.

Once larger creatures using coordinated muscle action had evolved, and neurons had evolved to carry the very fast signals needed for this coordination, a somewhat diffuse neural net would probably have been safer than a concentrated cord that would have been vulnerable to a crippling attack.  These creatures probably didn't have the vast redundancies found in Cnidarians, but some redundancy would have been useful, along with spreading out the various nerves so point attacks would have had small effect.

The concentration of nerves into centralized cords, containing neuropil, is generally defined as the beginning of a CNS.  By this definition, many cnidarians posses a CNS, and even the ctenophores have the rudiments of one, although this is mostly a concentration of axons right under the bands of cilia.  This fact, actually, allows us a new approach into the reasons for this concentration:  energy.

Both the transmission of action potentials along axons, and the more analog calculations that take place in the neuropil, require a lot of energy.  There were, basically, two sources of energy for this function in early animals:  burning sugar and the anaerobic process of glycolysis.  Both require a good supply of glucose, or something that can be converted to it (such as fructose).  Fatty acids can also be burned, as can amino acids, but this is unlikely to have been used by the early nerve cells.  In fact, even glycolysis is seldom or never used by most nerve cells,^[20] suggesting that it simply doesn't provide enough energy, or perhaps its effect on the internal conditions of the cell (e.g. pH) make it unsuitable.  (Nerve cells can also burn acetoacetate, which is produced by the partial breakdown of triglycerides.  How much this mattered in these early times is hard to fathom.)

In general, then, the source of fuel was glucose, which could be provided without a high-capacity blood flow, because large concentration gradients could be provided.  However, we can assume that extensions of the enteric system were close enough to the surface that sugar could be "pumped" across the enteric membrane into the interior where it would diffuse towards the surface.  That is where the oxygen was, and that is where the original ciliated swimming cells would have been located.  In fact, when we look at the ctenophores, we see that the nerve cords underlie the ciliated cells, which allows them access to oxygen from the moving water (stirred up by the cilia), and also allows supplies of fuel (sugar or whatever) to be provided for both swimming and nervous activity.

This logic is also applicable to the more distributed nerve nets that probably predated the CNS.  Without the ciliary activity, oxygen could be depleted by too many nerve cells or axons in the same place.  Thus, a distributed system of axons and ganglia makes the most sense, until a more centralized and active system of oxygen transport was evolved.

But the fact that this network was distributed doesn't mean it didn't serve the same functions as the CNS:  overall coordinated control of bodily activity that impinges on the external environment:  ciliary activity for swimming and rapid changes to the body's shape for capture of food and escape swimming.  Again, our perceptions of the activity of "diffuse" nervous systems has, IMO, been distorted by looking at the cnidarians, because their massively redundant cross-connected networks have been perceived as primitive, rather than a secondary adaptation to their lifestyle.

I would suggest that the original system for communicating chemical state, ancestral to the ANS and the limbic system, was also sufficient for controlling the various regions of ciliated cells that did the swimming.  A few disparate signal chemicals would have been sufficient to communicate desired speed, direction, and twisting/turning, and a distributive network for each dimension of this 6- to 10-dimensional vector would have required no more cells than the modern ANS.  Each ciliated cell could have computed for itself the required response to the vector, using separate receptors for each chemical signal, and independently calculating its response based on its knowledge of its position on the body.

It's with muscles that things become more complex.  Muscular activity in many places must be coordinated across the entire body, which requires some sort of "centralized" headquarters, with its calculations being distributed to the various muscles involved.  As mentioned above, for small animals, all that's needed a simple bundle of neuropil with the muscles directly extending processes into it to acquire their directions.  As the animal gets larger, extensions of the nerve cells that end up calculating the required activity for each group of muscle fibers could reach out to the muscles themselves, and their reach could be extended by branching systems or redundant networks of nerve cells that act as simple relays:  passing along the actions potentials just as received without calculation:  perhaps even using direct electrical connections to transfer the action potential from one cell to another.

As the centralized, whole-body activity was supplemented by local responses to local stimuli, the need for local calculations could be met by local ganglia:  concentrations of nerve cells and neuropil that could calculate the appropriate response to a local stimulus in view of the current directions coming from the central headquarters.  Thus, we can see that the SNS may well have developed after the first rudimentary CNS.  This also allows us to refine our definition of the CNS:  It consists of the neurons and associated cells that act as the "central headquarters" for calculating the whole-body activity in response to external stimuli and internal conditions.  The SNS, by contrast, can be redefined as the system of local ganglia that calculate local response(s) to local stimuli, integrating them into the whole-body directions.  This definition is independent of whether these systems are concentrated into "nerve cords", which is probably more a function of the need for oxygen.

This can also allow us to estimate the time that the original CNS arose:  when the systems of muscles became complex enough to require coordinated activity, while the size was too large to allow each muscle fiber to extend its own process into the central bundle of neuropil that functioned as a "brain".  In fact, as we saw above, many adult acoels probably fit this definition, and despite the fact that they have no "nerve cords", and the "neuronal somata and neurites, together with muscle fibers and gland cell processes [that] form a loose reticulated mass at the anterior pole of the animal"^[16] would not be defined as a CNS according to the traditional definition, I think we should allow it according to the definition I have proposed.  Given that such a system is present in both cnidarians and ctenophores, I would propose that the origin of the CNS should be placed prior to the first divergence of these groups.

Hiroshi Watanabe, Toshitaka Fujisawa, and Thomas W. Holstein (2009). Cnidarians and the evolutionary origin of the nervous system Development, Growth & Differentiation, 51 (3), 167-183 DOI: 10.1111/j.1440-169X.2009.01103.x

Links: (I've only included the links called out in this leader.) Not all of these are called out in the text.  Use the back key if you came via a footnote.


1.  Anteroposterior Patterning in Hemichordates and the Origins of the Chordate Nervous System

2.  Centralization of the Deuterostome Nervous System Predates Chordates paywall

3.  Defining a neuron:  neuronal ELAV proteins paywall

4.  Brain Size:  A Global or Induced Cost of Learning?

5.  Gene Duplication, Co-Option and Recruitment during the Origin of the Vertebrate Brain from the Invertebrate Chordate Brain paywall

6.  Rapid functional diversification in the structurally conserved ELAV family of neuronal RNA binding proteins
(Requires free registration)

7.  The interplay of nuclear mRNP assembly, mRNA surveillance and export paywall

8.  Post-transcriptional operons and regulons co-ordinating gene expression (Galley copy:  for the published article behind the paywall, go here.)

9.  Regulation of mRNA stability in mammalian cells paywall (2001=old)

10.  The evolution of nervous system centralization

11.  Cnidarians and the evolutionary origin of the nervous system

12.  Anatomy and development of the nervous system of Nematostella vectensis, an anthozoan cnidarian paywall

13.  The neuroendocrine system of invertebrates:  a developmental and evolutionary perspective

14.  Development of the insect stomatogastric nervous system

15.  Larval and adult brains paywall

16.  Embryonic development in the primitive bilaterian Neochildia fusca:  normal morphogenesis and isolation of POU genes Brn-1 and Brn-3

17.  Developmental expression of homeobox genes in the ctenophore Mnemiopsis leidyi

18.  Bilateral symmetric organization of neural elements in the visual system of a coelenterate, Tripedalia cystophora (Cubozoa) paywall

19.  On the Origin of The Nervous System paywall

20.  Energy Substrates for Neurons During Neural Activity:  A Critical Review of the Astrocyte-Neuron Lactate Shuttle Hypothesis