Thread by @armanafzadeh

A belated #SICB2020-inspired thread — if you ever use #XROMM, watch #XROMM talks, or read #XROMM papers, the following unsolicited FYI is for you. Please r/t for visibility! @mechsNmorph @elbrainerd @jd_lc @KonowLab @JohnRHutchinson

***disclaimer, I don’t claim to be an expert on any of this, but I’ve made my fair share of mistakes over the past three years, and hopefully learned from some of them 🤞🏽

#XROMM studies like to include graphs with a whole lot of joint angles. Let’s talk about where those numbers come from, because it’s not as simple as you might think — and that actually matters.

We pretty much all measure angles using Joint Coordinate Systems (JCSs), a method we stole from the human biomechanics folks (thanks, human biomechanics folks) — here’s the oft-cited explanation from Grood & Suntay (1983). Everything’s super clear now, right? Didn’t think so.

Alright, let’s back up. When we want to measure the rotations between two bones, first we give each bone its own Anatomical Coordinate System (ACS) with X, Y, and Z axes.

Then we grab pieces of those ACSs to construct a JCS: the Z axis of the proximal bone, the X axis of the distal bone, and a dynamic Y axis that remains orthogonal to both X and Z (but X + Z can/do break orthogonality!). This is all that complicated Grood & Suntay figure shows 🎉

What we actually MEASURE are the rotations about those three axes that are necessary to get from a chosen zero (aka reference) joint pose to any other joint pose, in a ZYX rotation order. This is what you’re seeing when you’re shown those red, blue, and green kinematic curves.

Here’s a handy dandy chart to make that all a little more concrete, with a simple knee system as an example. Check out Kambic et al. 2014 in JEB to walk through some other hindlimb joints’ construction for practice.

Fun fact: you can actually visualize this 3-rotation system with a 4-bar linkage, if you’re hip with the times and that’s what floats your boat (cc @aarolsen, linkage king).

Figure from Grood & Suntay (1983)

(@jd_lc I think you should 3-D print one 🤓)

“So? Like, I guess that’s nice and all, but I like biology and this is very much NOT biology, so why should I care?” ***BECAUSE THE FACT THAT WE MEASURE THIS WAY IS REALLY IMPORTANT FOR OUR INTERPRETATIONS***

It matters bc the angles we measure are called “Euler angles,” and they’re REALLY FREAKING WEIRD. Bc — bear with me — 3-D rotations actually mathematically live in 4-D, and we’re trying to smush them down into 3-D. That’s a lot like trying to smush a globe into a flat map.

When we do that smushing, the joint angles we end up measuring are NOT INDEPENDENT. That means common graphical representations like ternary diagrams aren’t actually valid, and neither are the common stats that we’d all like to be able to run.

It also ends up meaning that HOW WE SET UP JCSs IS A BIG DEAL — certain setups can make us think our joints are rotating way more than they actually are. I won’t get into that here, but I’ll leave you with the vague analogy that it’s like how Greenland looks way too big on maps.

If you’ve made it this far, I have lots more thoughts and feelings about Euler angles and JCSs and would genuinely love to chat about them any time — my DMs are open, or email armita AT brown DOT edu✌🏽

A follow-up for the most FAQ — “where can I find #XROMM help and resources?” @elbrainerd has been leading a great charge at @EEB_Brown to get these updated and organized. To my knowledge, here’s where everything currently lives:

#XROMM resources (cont.):
XMALab instructions: bitbucket.org/xromm/xmalab/w…
XROMM Maya tool instructions: bitbucket.org/xromm/xromm_ma…
XROMM workflow help: wiki.brown.edu/confluence/dis…

N.B., that last link is likely to change soon — check xromm.org for links to the latest info!