One or two fma:Thalamus?

[Alan Ruttenberg]

In debugging why the Right thalamus isn't inferred to be (always) part

of some Brain in my setup, I realize there might be some ambiguity in

the FMA about the thalamus.

Specifically, I think the answer is that there are two fma:Thalamus,

the left and right, which are subclasses of Thalamus. Most of their

parts that are asserted as part of the thalamus (which inherit to the

right and left) are bilateralized as one would expect.

The only exception is the second partition into allothalamus and

isothalamus. If I understand things correctly those they could be bilaterally

differentiated as well, but are not at the moment.

The thalamus is asserted to be regional part of the diencephalon.

The Thalamic complex has parts Left thalamus, Right Thalamus, and the

Interthalamic adhesion. This looks like an exhaustive partition. It

isn't part of anything, leaving the Interthalamic adhesion an orphan.

I think it might be clearer to have the Thalamic complex be part of

the diencephalon. Then it would be clear that next level of component

below diencephalon is the thalamic complex. Not asserting diencephalon

having part thalamus might also cause less possibility for confusion

between thalamus and thalamic complex. The left and right would be

inferred to be part of the diencephalon because the they are part of

the thalamic complex.

There is a general issue when one has the pattern

<non-bilateral> (e.g. diencephalon)

 has_part <superclass of bilateral> (e.g. thalamus) 

<bilateral> 

  <left> (e.g. left thalamus)

  <right> (e.g. right thalamus)

which is that we don't know (from the axioms) whether the

<non-bilateral> has both or only one of the <left> and <right>.

The all-some axiom is: ALL <non-bilateral> has_part some SOME

<bilateral>.

This is satisfied by it having either a <left> or a <right>. (think of

tool that is manufactered in left and right hand versions - some parts

are in common to both, and some are specialized to right and left, but

there would be only one of left or right).

Even if we have the inverses - that ALL <left> is part of SOME

<non-bilateral>, and ALL <right> is part of SOME <non-bilateral> we

still have a situation in which having a single <left> or <right> would

satisfy the axioms.

I am thinking that what I should do is just recognize the pattern and add axioms

<non-bilateral>

has_part 

  <left> 

  <right>

-Alan

[mej...@comcast.net]

You're right, the thalamic complex is the proper part of the diencephalon. I have intended to do this when I created that class but for some reason I got sidetracked.

 I do agree with the axiom you propose but I'm still trying to come up with examples where the non-bilateral structure may indeed only have as part one lateralized structure. I doubt it but I'll get back to you on that.

Thanks again.
Onard