# problem in BEM computation inner skull

**URL:** <https://mne.discourse.group/t/problem-in-bem-computation-inner-skull/11489>\
**Category:** Support & Discussions\
**Tags:** bem\
**Created:** [September 14, 2025, 9:45pm UTC](https://mne.discourse.group/t/problem-in-bem-computation-inner-skull/11489 "2025-09-14T21:45:50Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![chiaram](https://avatars.discourse-cdn.com/v4/letter/c/35a633/32.png) [@chiaram](https://mne.discourse.group/u/chiaram)\
**Post date:** [September 14, 2025, 9:45pm UTC](https://mne.discourse.group/t/problem-in-bem-computation-inner-skull/11489/1 "2025-09-14T21:45:50Z")

</div>

Dear all,

I am having trouble with the function `mne.bem.make_watershed_bem`. I have preprocessed the T1w data with FastSurfer and I am now computing the BEM. However the inner skull seems to have problems (specifically it is outside the outer skull and has weird shapes on the bottom). I have tried the follwing:

- increased the `preflood`: it is slightly better until `preflood`=45, then gets worse (outer skulls extremely wide. I also tried with `atlas` and `gcaatlas` but it does not improve. Below, you can see a picture of the preflood=45 result.

 ![image](https://global.discourse-cdn.com/free1/uploads/mne/original/2X/e/edf768cd7bfdf1aeb2aa303788930d3868c0ca8d.jpeg)

- If I follow [these tutorials](https://mne.tools/stable/auto_tutorials/forward/80_fix_bem_in_blender.html#tut-fix-meshes), I manage to fix it until the inner skull surface does not exceed the outer skull one. But still the lines don’t follow well the brain even if they are inside (see below)

What should I try next? Manually running in FreeSurfer the mri\_watershed command? are there any specific parameters that could be helpful for my case?

---

<div class="post-metadata">

**Author:** ![bonokat](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/bonokat/32/1901_2.png) [@bonokat](https://mne.discourse.group/u/bonokat)\
**Post date:** [September 17, 2025, 7:42am UTC](https://mne.discourse.group/t/problem-in-bem-computation-inner-skull/11489/2 "2025-09-17T07:42:36Z")

</div>

I believe I solved this by adjusting `--preflood` height (`-p`) when generating watershed BEM in freesurfer:

```python
mne watershed_bem -s subject_name -d SUBJECTS_DIR -p 15 --overwrite

```

I don’t remember, if I used -p 15, but it’s worth playing with this parameter. Hopefully it’ll help!