JingyiGF / HERBS

MIT License
18 stars 15 forks source link

wrong probe insertion coordinates #38

Open KacperKon opened 1 year ago

KacperKon commented 1 year ago

Hello! In a couple of experiments I observed that HERBS places the probe insertion point not on the top of the brain, but significantly deeper. I guess it's related to the fact that in all these cases the brain is not physically continous (there's some empty space between the two structures and it's not described with any anatomical label). I'm attaching a picture of one probe reconstruction to show what I mean - the probe obviously was inserted through cortex, but according to HERBS the insertion point is deeper, in superior colliculus. Do you have some idea on how this kind of error could be corrected?

herbs_probe_example
KacperKon commented 1 year ago

Hello, did you manage to confirm if that's indeed a systematic bug? Or do you suspect that it was a random error or I'm doing something wrong while fitting the probe tract?

JingyiGF commented 1 year ago

Hi, sorry for the late reply. I was busy with other stuff and forgot to come back for this. So sorry for this and thank you so much for leaving the second message.

I am wondering where is the most top point(entering point) you chose? Could you please show me an example?

In my memory, I extrapolate along the probe line from the top point to outer surface until there is no brain parts exist. So I would think the top point you chose in the example is placed below that "black crack". If this is not true, I need to do some further inspection on it.

I wonder what will happen if you place the top point above that "black crack"?

Last week was Easter Holiday and this week I have no functional computer to test anything. I will get my computer back next Monday. So will fix bugs afterwards. Please do let me know what you find, many many thanks :-)

KacperKon commented 1 year ago

Hi, now I'm sorry for the late reply :) Yes indeed, it works now - thank you!

KacperKon commented 7 months ago

Hi! I've reopened the issue because unfortunately, I now have a few animals in which there are no visible entry points in the slices. That means that I cannot mark any points in the upper part of the brain (above the 'black crack'). Do you think it would be hard to change the code, so HERBS would extrapolate the probe more and find the real top of the brain? Or maybe you have some suggestions on how I could post-process the HERBS output to fix that?