Useful Structural Patterns in the Optic Lobe (inferring microstructure from macrostructure)

[TR77]

@AzureJay “Great minds…”, AJ…I’ve been working on Tm cells lately as well.

Also, note that as per my previous post on the preserved arrangement of Tm/TmY cells in the medulla vs lobula, you can see from the last two graphics in your post that the arrangement of the Tm2’s is preserved as well. This is also apparent in the link you provided. So, this example demonstrates that the preservation of arrangement goes for Tm2’s as well as for Tm4’s (and for many others, in my experience).

Outstanding work, AJ!

[Krzysztof_Kruk]

Created a small table to show differences between TmY5 and TmY5a, which are hardest to differentiate for me. The table describes arborizations in different layers of optic lobule’s neuropils (according to FlyBase.org).
The easiest ones are in the Medulla layer - TmY5 starts its arborization “lower” and it’s usually finer (no bleb until M10).

Layer	TmY5	TmY5a
Medulla
M1
M2		fine
M3	fine	fine
M4		fine
M5	fine	fine
M6	fine	fine & bleb
M7	fine
M8	fine	fine & bleb
M9	fine	fine
M10	fine & bleb
Lobula
L1
L2
L3
L4	fine	fine & bleb
L5	fine & bleb	fine & bleb
L6	fine & bleb	fine & bleb
Lobula Plate
LP1	fine & bleb	fine
LP2	fine & bleb
LP3	fine & bleb	fine & bleb
LP4	fine & bleb

[AzureJay]

This is a breakdown of the medulla layers (M1-M10) as described by The Fischbach Paper. This is useful in knowing how these layers are defined, because they also help us, in turn, know how to define some of our neurons.

Note: I use “inner” instead of “proximal” and “outer” instead of “distal” for reading clarity’s sake. Both terms refer to position to the central brain (inner/proximal = against central brain; outer/distal = away from central brain)

Medulla Layer	Defining Characteristics
M1	Extent of L1 Arborization in the outer medulla
M2	Extent of L2 Arborization. Contains outer L4 arborization, T1 arborization. Inner M2 defined by outer expansion of L5 (which fills M1 and M2).
M3	Extends from inner border of L2 arborization to L3 terminal. Contains main medulla arbor of Lawf, most R8 terminals. Border between M2 and M3 marked by Dm3
M4	Slim layer between L3 terminal and inner L1 arbor. Contains inner L4 arbors, and long R8 terminals.
M5	Inner L1 arborization. Contains inner L5 terminal.
M6	Between L1 arborization and serpentine layer. Contains terminal enlargements of R7. Can be subdivided into M6a and M6b depending on R7 terminal.
M7	Serpentine layer. Axons of many medulla tangential (Mt) neurons.
M8	Stratified arbors of some Medulla intrinsic (Mi) neurons, ex Mi3-Mi7, Tm and TmY neurons, and C2’s lower arbor (M8-M10).
M9	Houses several types of Proximal Medulla (Pm) cells.
M10	Height of the bushes of T4 neurons.

[TR77]

@Krzysztof_Kruk @AzureJay Nice work, KK and AJ!

On moving forward, with Tm/TmY cells…

My primary goal is, of course, to proofread and complete the cells of the optic lobe. As such, my primary interest in cell type classifications is to aid in that goal (i.e. if you know what type of cell it is, then you have a pretty good idea of what the cell “ought” to look like, when complete). At some point though, assigning cells to further and further subdivided categories becomes a purely academic exercise, from the standpoint of proofreading. As an example, consider the distinction between the Tm1 cell, and its Tm1a variant. In my experience, the examples of Tm1a that I have found, have always been in place of the Tm1 cell, in a particular medulla column, rather than being an additional cell in the column. From the standpoint of proofreading, and useful structural patterns to aid in proofreading, it makes no difference to me whether it is a typical Tm1, or some slight variation of the Tm1 theme…it is sufficient for proofreading purposes to simply know that every medulla column has a Tm1-ish cell, with a proximal medulla arbor somewhere near the distal end of the Mi1 axonal arbor, and an axon extending somewhat superficially into the lobula, etc.

With the above in mind, I would personally like to organzie the Transmedullary cells into two broad categories - those that are coaxial with the medulla columns, and those that are interstitial to the medulla columns. The figure below illustrates what I mean by this distinction. In my experience, Tm1 and Tm2 can always be found within, and coaxial with, a medulla column. The Tm3 cell, on the other hand, can always be found in the interstitial space between medulla columns. I would then like to come up with a count of all coaxial Tm/TmY cells, in a typical medulla column, and a count of all interstitial Tm/TmY cells, at a typical interstitial location…and then compare those totals to the Fischbach classification scheme. Why? Because Fischbach acknowledges in the paper that, for example, Tm8 and Tm22 may actually just be slight variations of the same cell type, and my experience suggests that Tm5 may also be yet another variation of that same cell type. There are other examples, but this paragraph provides the basic idea of what I am looking for.

I am not asking the community to do anything here, I will pursue this avenue of exploration myself (although anyone is more than welcome to help, if they like)…just providing an update on where I am at, regarding the question of identifying Tm/TmY cells.

Cheers, all.

[Nseraf]

How to tell apart ‘normal’ cell Somas/CBs from Glia ones?

Iit’s not 100%, but very often the glia ones are going to be nucleus and surrounding area all ‘pitch’ black, unlike the non-glia neuron CBs which are ‘white’ (or shades of grey), like in the picture, the 2 glia cells are much blacker than the other cells.

The other way to know is they’re usually larger than the average cell, with the exceptions of the really large non-glia cells’ CBs.
FlyWire the link contains 3 black CB glia cells and one large CB, the 3 are not reconstructed so they’re probs mergery.

[Krzysztof_Kruk]

Yes, and often the branches themselves are usually at least a little bit darker for glia cells.
Maybe the matter, from which the cells are built is denser and keeps more osmium (or whatever was used for staining), that the neurons.

[TR77]

Hi all,

Something has come up on my end, so I will be away (afk) for a while. Keep up the good work, all!

Cheers.

[annkri]

have anyone testet/noticed if the widefield amacrine neurons make layers thight layers like the dm 4 does? Have a feeling that if it does that might help finding the missing clusters

[Krzysztof_Kruk]

Didn’t test, but that’s a good idea.

[Nseraf]

Idk abt a specific cell type but in the CBs layer in 182648, 80934, 5572

and the one ‘above’ I have noticed that the stems/backbones going from the CBs to the dendrites/axons etc all cluster around like this;

which isnt 100% that the CBs will be in one place but they look like this:

So, I’m thinking this: Say that u have either a CB w/o backbone, or vice versa backbone w/o a CB, scroll a bit up or down and find the other backones in the ‘bunch’ (or bucket? following the DMs flowers analogy lol) and if u have say 10 CBs/Backbones and 8 or 9 of them have a match in the boucket/bunch then it becomes a whole lot easier to find the missing backbone/CB from 1 or 2 than 1000k (cells in the brain lol).

Disclaimer: ive not actually tried to find a backone’s CB yet but I keep seeing layers of ‘endless’ CBs and the CBs seem to alway seem or tend to go to clusters of backbones. May not end up helping - this observation- as the black spill area might just cover up/destroy the pathways from cb to backbone, but maybe if some of the other cbs’ pathways in the bunch are restructurable/ or restructured then deductiveness can bring results on the missing stuff and while it may not work for every case it might help us more than not trying at all.

[annkri]

I have been testing out trying to find any pattern on the Lamina wide field, from my observations so far each bundle have 4 different cells together in the same knot. Also this is a type of cell you never get finished with. most of my exsamples with half of them already is beeing complete but missing half the cell. FlyWire

[Krzysztof_Kruk]

Lately I’ve been working with many C cells and I’ve seen many already completed cells, that were missing their most outer branch. Id’s say, about 1 in 5 of the completed cells didn’t have it.

Here’s an example of such cell:

So, whenever you see a cell, that has its CB outside LP or between LP and Medulla, and looks like the one in the picture, look for extensions at the “top” (outer) part. There should always be a thick twisted branch, that extends to a thin straight bulbous branch, that goes through the Lamina.

Here’s the same cell with the missing branch added:

It’s probably the most common error, I have seen (with any type of cells).

[Nseraf]

ive probs done that mistake, yeah. But they easily confuse me with a C type of cell, which has very similar shape except for the axon bit.

[Krzysztof_Kruk]

Yep, you did (had one of your neurons finished like that). Around the same time I had a cell completed by myself, which also didn’t have the branch . That’s why I like working on neurons from the same area, because then I know exactly, what to look for.
The C types can be mistaken with Ts and Ys, becuase all of them have their CBs in this area, however both T and Y have some forking, while Cs are single lines in shape of a squared letter C (hence the name, probably)

[TR77]

Hi all…just dropping in for a minute, to say hello…it looks like a lot has happened. You all are doing incredible work!

I would like to share an anecdotal analogy that occurred to me recently, regarding cell type classification…

Suppose that you are on a trip, and you wake up one morning in an unfamiliar location. The buildings outside distinctly resemble traditional Chinese architecture, and you wonder if you are actually in China. How could you tell, if you are in China or not? By the look of the buildings? No, of course not. Why? Because you may be in Chinatown, in San Francisco…a lot of the buildings there look pretty much the same. Or vice versa, if you were in downtown Shanghai (in China), most of the buildings there do not resemble traditional Chinese architecture. So again, how could you tell, if you are in China or not?

…?

By checking your location, relative to other known locations on earth (preferably using GPS, these days, lol)…that, is how you could tell if you are in China or not.

There is a potentially useful analogy in this anecdote, that lends itself well as an objective criteria for classifying cells in the optic lobe. As a familiar example, it would be entirely possible for a degenerate L1 cell to resemble an L2 cell. In such a case, subjective classification of the cell, based on appearance, would be inadequate for identification. It is the relative position and arrangement of the cell, relative to other known cells (within the repeated structural units of the optic lobe), that would determine that it is actually the L1 cell of its respective column (particularly if there is already a prototypical L2 cell present in the column).

Nothing revelatory, of course…just a useful anecdotal analogy that occurred to me. Keep up the good work, all!

Cheers.

[AzureJay]

I caught sight of some structural similarities today I wanted to shared.

I’ve talked before about the structure of Tm2 which, of all the Tm cells, has a unique shape to its upper arbor in the M2-M3 layers.

Today I realized that Mi8 and Mi9 cells also have very similar upper arbors to Tm2 - and in fact, they also interface in the M2-M3 layers.

Form usually follows function, and the fact that these are all M2-M3 layers probably indicates that these three cell types have some sort of similar interaction in the optic lobe!

Here’s a link of the three cells displayed above in the same Flywire view for comparison. Notice that the Tm2’s second arbor also is in the same layers as the final arbors of the Mi8/Mi9!

[jones]

@TR77 @Nseraf @AzureJay @annkri @KrzysztofKruk This has been an extremely informative thread. If I adapted this into a post for the FlyWire blog and credited you for all the hard work here, does that sound good to all of you? Would any of you like to help draft the post in general since you’ve been working on this stuff inside and out? Just let me know!

[Nseraf]

heyyyy, sounds good to me, and I wouldnt mind helping (you) write a draft but Im in vacas w gf in South East Asia until the 27th so I wont be able to work on it until 29th or so (jet lag etc). If this can wait until then i’ll be more than glad to help, if not, is ok.

[jones]

No worries, there’s no time pressure and I’d be waiting to hear back from everyone before getting started anyway. Thanks for the offer however! Duly noted!

[annkri]

no problem for me