Neural canal ridges: the director’s cut

Here’s a grab-bag of follow-up stuff related to our new paper on neural canal ridges in dinos (Atterholt et al. 2024, see the previous post and sidebar page).

Neural canal ridges, or bony spinal cord supports?

I got into the habit of calling the inwardly-projecting bony prominences in the neural canals of sauropods and other critters “neural canal ridges” partly because I was thinking about them for literally years before I knew what they were, and I had to call them something, and partly because “neural canal ridges” is a reasonably accurate descriptive term that does not imply a specific function. NCRs became part of my internal lexicon.

Later on, thanks first to David Wake, and later to Skutschas & Baleeva (2012), we discovered that extant fishes and salamanders have bony spinal cord supports, and we think that’s the best explanation for why NCRs show up in so many dinos. “Bony spinal cord supports” is not function-neutral, it takes a stand. Since the whole point of our paper is not only to describe these things in dry terms, but to also take a stand on their associated soft tissues, it would be more coherent to cowboy up and call them “bony spinal cord supports” instead of “neural canal ridges”, and that’s exactly what Jessie Atterholt did in the tables and figure captions of the new paper. Also, sometimes the bony spinal cord supports are not ridges, but shelves or planks or spikes — check out that tuna vertebra up top, and the salamander verts in Fig. 1 of the new paper — so “neural canal ridges” doesn’t even accurately describe them all the time. If I call them NCRs in my blogging, it’s out of habit, and because — so far — that does accurately describe the appearance of the bony spinal cord supports in dinos.

Denticulate ligaments: sometimes double, sometimes absent

Here’s something that turned up late in our research on this project. Elvan et al. (2020) is a nice paper on the denticulate ligaments in developing humans (it is of course tragic when fetuses are miscarried or stillborn, but what we learn from them can help keep others alive). One of the curious things they mention, and figure, is that the denticulate ligaments that suspend the spinal cord inside the dura mater are occasionally doubled on one side, and occasionally absent.

This shouldn’t be super surprising. Variation exists in part because developmental programs are messy. “Asymptomatic anatomical variation”, “pathological variation”, “congenital anomaly” (“birth defect”), and “fatal malformation” are points on a spectrum — and all of us are somewhere on that spectrum. “Normal” human anatomy is normal in the statistical sense, in that the majority of folks end up in the big middle, but that middle encompasses a lot of variation, and there are long tails in lots of directions for almost every body part and body system, and things can sometimes be pretty non-standard under the hood without causing noticeable symptoms.

In particular, if there’s a developmental program for building structure X — whether structure X is a hair follicle, a muscle, nerve, or blood vessel, a finger or toe, a gill arch, a vertebra and its associated body segment, or an entire limb — then inevitably there will be counting errors from time to time, omissions or duplications, and embryos, fetuses, or offspring produced with fewer or more of structure X than is typical. At the small end of the scale we might not even notice, and at the large end of the scale the variation might not be viable.

ANYWAY, finding the Elvan et al. paper was an “Aha!” moment for me. Back in 2018 when I’d been photographing tuna vertebrae in the OMNH collections, I found some that had not one but two inward-pointing bony spikes on each side. I figured these were just a fancier system of bony spinal cord supports, probably indicating doubled denticulate ligaments. I didn’t know for sure that the latter existed, so in assembling figures for the paper we went with the tuna vertebra that most closely resembled the salmon vertebra figured by Skutschas & Baleeva (2012). Later on, the Elvan et al. paper confirmed for us that doubled denticulate ligaments sometimes occur, at least in humans, so it’s plausible that they happen in fish, too, and maybe regularly given that I found the quad-spike setup in multiple tuna vertebrae. But that seemed like a lot of extra yap and figures to make a rather minor point, which is why you’re hearing about this in a blog post instead of in the paper.

I assume that these spikes and whatever attaches to them were described back in the 1800s in some obscure paper, probably published in Germany or Great Britain, but if so I’ve not yet tracked down that hypothetical publication. Even if said publication exists, I’m sure it’s illustrated with a hand-drawn diagram. It occurs to me that someone could go to a fish market, buy a chunk of tuna with the bone in, do a little careful dissecting, get some hi-res color photos, and have everything they’d need to publish a nice little paper, either describing these spikes and their soft-tissue correlates for the first time, or redescribing them and providing the first good color photos. Realistically I’m unlikely to get around to that, so if you want it, go nuts.

Citing the Deep Magic

I’m gonna geek out for a sec on the developmental underpinnings of the denticulate ligaments and the vertebrae they’re associated with. And to do that, we have to orient ourselves to the various bits sticking out of the spinal cord and how they relate to the vertebral column.

Here’s a chunk of sauropod tail in left lateral view (modified from Wedel et al. 2021: fig. 2a) — specifically, a 3D-printed section of Haplocanthosaurus tail that Alton Dooley put together for the “Tiny Titan” exhibit at the Western Science Center a few years ago, seen in medial view in the second image down in this post. The laterally-facing bony loop formed by the central and zygapophyseal articulations of two adjacent vertebrae is the intervertebral foramen, and it’s through the intervertebral foramina that the spinal nerves leave the neural canal (blood vessels enter and leave through these openings, too). Assuming that sauropods were built like reptiles rather than mammals, and lacked epidural fat, a horizontal section through this bit of tail on the black line indicated by the Xs might look something like this:

Anterior is toward the top now. There’s a lot going on in this image, so let’s take it one piece at a time. The neural arch pedicles are the paired black-and-white pillars on either side of the spinal cord, defining the lateral walls of the neural canal. (The section in the photo also went through the caudal ribs but I was too lazy to draw those.) The meninges — the dura, arachnoid, and pia mater, and the subarachnoid space — are by now old friends; this diagram is showing us the same structures as this one from the previous post, just in horizontal section rather than transverse. Bundles of spinal nerve roots come together to form the spinal nerves, which exit the neural canal at the intervertebral foramina between adjacent neural arch pedicles. The various meninges form little sideways-projecting meningeal sleeves over the first little section of each spinal nerve; imagine making 3-layer coveralls for a centipede and you’ll have a good mental model of the whole meningeal system of the spinal cord (for real geekery, past the ends of the meningeal sleeves the nerves are jacketed in a different connective tissue called epineureum). The denticulate ligaments attach the spinal cord to the dura mater (or even through the dura mater) level with the neural arch pedicles of the vertebrae, so if you’re looking at a section of the cord in dorsal or ventral view you’ll see bundles of spinal nerve roots (at the intervertebral foramina) alternating cranio-caudally with denticulate ligaments (in between intervertebral foramina). You can check that with the dorsal-view photos of human spinal cords above and in this image in the previous post.

(Note for any confused med students who might be reading this: anatomical position for humans is upright, so horizontal and transverse sections are synonymous. Most other animals carry their bodies horizontally, so a horizontal section through a sauropod would be similar to a coronal or frontal section through a human vertebral column. Also, humans do have epidural fat, unlike this sauropod, and our denticulate ligaments do not go through the dura mater to attach to bone. So don’t use these sauropod diagrams to study for your human anatomy courses! Instead, a great learning exercise would be to redraw this diagram so it was accurate for a human. If you do that, feel free to drop me a line in the comments and we can talk about your results. Standing offer, good forever.)

At the bottom of the image I labeled segmental muscles and intermuscular septum. You’ve seen these before, although you may not have known it: they make the zig-zag patterns in the meat of fishes, where we call the segmental muscles myomeres (“muscle parts”) and each intermuscular septum a myoseptum, plural myosepta (“muscle partition”).

Each myomere is associated with a particular spinal level — a paired set of spinal nerves, like the C7 or T10 spinal nerves in a human — and each myoseptum is associated with a particular vertebra, like, er, C7 or T10 in a human (or a sauropod, although we’d call it D10 for dorsal 10 in a sauropod; sauropod dorsals all have big ribs that were mobile at some point, so there’s no need to separate them into thoracic [dorsals with mobile ribs] and lumbar [dorsals without mobile ribs]). Put a pin in that thought for a moment, we need to wrap up something fishy.

You maybe looking at the mild zig-zaggy-ness of the myomeres in that first salmon diagram, and the target-like concentric circles in the photo of the salmon steaks up above, and thinking something doesn’t add up. And you’re right — the surface zig-zaggy-ness of the myomeres is not their full extent, they have anterior and posterior cones arranged concentrically, presumably to allow each myomere to exert force over more of the vertebral column. And that’s why fish comes apart in such interesting ways when you eat it, especially if it’s cooked.

Anyway, back to the segmental muscles and intermuscular septa in the sauropod — and in yourself, for that matter. It’s not immediately obvious that amniotes are built on the same myomere/myoseptum infrastructure as sharks and salmon, because our development involves a lot of splitting and recombining and stretching of muscles across multiple spinal levels. But if you go deep enough, we all have some single-segment muscles that bridge adjacent body segments — intercostal muscles between our ribs, and interspinales, intertransversarii, and rotatores breves between adjacent vertebrae.

Now here’s the part that I think is awesome, what this whole section has been building toward: the myomeres and myosepta were there from very early on in development, and the myosepta originally ran from spinal cord to skin. Denticulate ligaments are just what we call the little stretch of myoseptum between the spinal cord and the dura mater, sorta like how we use ‘Foothill Boulevard’ for the stretch of US Route 66 that runs through Claremont and adjacent townships. The pedicles of the neural arches — in fact, the entire left and right halves of each neural arch — form within the myosepta. The light gray boxes around “denticulate ligament”, “neural arch pedicle”, and “intermuscular septum” in my cross-sectional diagram above unite the different portions or aspects of the embryonic myoseptum. I didn’t work all this out myself, mind, I learned it from Skutschas & Baleeva (2012), who demonstrate it all very convincingly with developmental work on larval salamanders.

And that brings us to the weirdness of mammals.

NCRs? No thanks, we’re mammals

I’ve gotten some questions about whether mammals could have NCRs. I doubt it. Not to put too fine a point on it, but as a species we just care more about our own anatomy and that of dogs and cattle and rabbits and rats, than we do about any other critters, and I think if mammals had NCRs they’d have been found and logged by now.

Also, I don’t think we mammals have the capacity to have bony spinal cord supports, because those are the attachment scars of the denticulate ligaments to the inner walls of the neural canals, and our denticulate ligaments don’t work that way. Our denticulate ligaments connect our spinal cords to our dural sacs, but we have epidural fat between the dura and the neural arch pedicles, and apparently when in development the dura pulls away from the neural arch pedicles and epidural fat starts to be laid down in between, whatever embryonic connection existed between the denticulate ligament and the rest of the myoseptum is broken.

I said “I doubt it” rather than a flat “no” because apparently there is very little to no epidural space in the cervical region of most mammals. IF there are mammals in which the dura mater fuses to the periosteum in the cervical region, then maybe the embryonic myoseptal connection could be maintained, the resulting denticulate ligaments could be tied down to bone, and bony spinal cord supports could exist. I wouldn’t rule it out, because if there’s one thing we as a species are even worse about than caring about non-mammals, it’s peering into neural canals.

But we’re working on it.

References

• Atterholt, J., Wedel, M.J., Tykoski, R., Fiorillo, A.R., Holwerda, F., Nalley, T.K., Lepore, T., and Yasmer, J. 2024. Neural canal ridges: a novel osteological correlate of postcranial neuroanatomy in dinosaurs. The Anatomical Record, 1-20. https://doi.org/10.1002/ar.25558
• Brown, M.D., 1996, February. A Six-Legged Rattus (Todentia: Muridae) in Oklahoma. Proceedings of the Oklahoma Academy of Science 76: 91-92.
• Elvan, Ö., Kayan, G. and Aktekin, M., 2020. The anatomical features of denticulate ligament in human fetuses. Surgical and Radiologic Anatomy 42: 969-973.
• Liem, K.F., Bemis, W.E., Walker, W.F., and Grande, L. 2001. Functional Anatomy of the Vertebrates. (3rd ed.). Thomson/Brooks Cole, Belmont, CA.
• Skutschas, P. P., & Baleeva, N. V. (2012). The spinal cord supports of vertebrae in the crown-group salamanders (Caudata, Urodela). Journal of Morphology, 273(9), 1031–1041.

Source: https://svpow.com/2024/09/01/neural-canal-ridges-the-directors-cut/