How To 3D Print Your Brain In A Few Simple Steps - TUTORIAL 2022

[@JesseIrwinArt]

I followed your tutorial and shared the resulting print on reddit, and now tons of people are asking how I did it so I’m just linking them to your video

[@IZLabs]

Haha I was wondering where all these people were coming from! Thanks for the referral and glad it was useful for you. I took a look and your brain came out well lol!

[@lucaswiseman1259]

So we meet again

[@Blckjack18]

I am using data from a clinical study. Amazing detail and an amazing tutorial so far.

[@IZLabs]

Awesome! Thanks for the kind words!

[@AndrewSink]

Great tutorial! I did something similar back in 2014, and it was a massive pain. This looks so much easier in comparison! Thanks for sharing!

[@IZLabs]

That’s awesome! Yeah I tried to find the least painful way of doing this haha

[@slurmking]

Your tutorial was the best one, thank you!!!

[@IZLabs]

Thank you! I’m glad you found it useful😁

[@daneverhart7649]

DUDE! You must 3D print a Jello brain mold! Imagine at your next party, serving your brain laced with fruits, sweets, maybe nuts, and of course "happy" ingredients. Lmao! Also 3D print your skull, as a serving dish for the brain. Definitely.

[@IZLabs]

Oooh I love this idea! I'm going to give it a shot! I'll post a video about it if I get it to work

[@hurbsnarebergowitz4083]

Thank you! I participate in the LE TBI study at UWMC in Seattle and I get MRI scans... so soon maybe i can print my brain

[@IZLabs]

That’s awesome! I hope it works out for you!

[@wornwoodroaming]

that's awesome! great job

[@IZLabs]

Thanks! Glad you enjoyed it!

[@Darrylx444]

Could you make your skull instead of your brain, using the same scans? That would be very cool.

[@IZLabs]

Theoretically yes! A CT scan would be better suited to picking out the bone, but it’s possible.

[@LiquidRetro]

Great video, my MRI scan has many views, and I'm unsure which to pick to do this. The ones I have picked so far seem to be low resolution for the side and rear, high res on the top.

[@IZLabs]

Glad you liked it! This seems to be an issue many people face. They don’t always do a 3D high resolution scan. I mentioned a potential fix in one of the other recent comments, but not sure if anyone has tested it yet. Sorry I can’t be more help!

[@Joebus100]

Thank you for this, exactly what I was looking for! Unfortunately when I press apply no 3d model appears in the viewer.

[@Me-kt3gh]

Have you clicked "Show 3D Model"? If so and it still isn't showing, do you notice any error messages?

[@Joebus100]

@@Me-kt3gh Hi Me. I did click "Show 3D model". What's strange is I started completely fresh today and ran through it and it worked! I'm not sure what is different this time but I was missing something the first time. Before I noticed the brain wasn't glowing green(mask) like in this vid and when I tried selecting stuff things got weird. Thanks again for the help :)

[@IZLabs]

Glad you found it helpful! That may happen if you don’t have a large enough selection in your masking step. Are you seeing green areas highlighted on each of the 2D views? -edit, just saw the previous comments too haha. Yep I think the lack of green highlighting was the issue!

[@rodolfofrancisco]

They could do an AI tool to automate the conversion from MRI to 3d model. That would be awesome. Thanks for the video!

[@IZLabs]

That would be cool to see for sure! Glad you enjoyed the video!

[@jswallet]

This is a great tutorial. My 3d model in Slicer isn't as defined as yours. Is that because of the actual file sizes of the MRI or is it just settings in the software? I'm wondering if the images on my CD from the radiologist aren't hi-res enough.

[@IZLabs]

It’s likely the scan resolution. You can see if there are other files in your scan package as sometimes they do lower res scans first to locate and then high res. Good luck!

[@spartenz14]

This is the most comprehensive tutorial I have found online for this, thank you!!! Quick question though, whenever I pull my files into Slicer, the back of my brain is always less detailed than the top front. Do you have any suggestions to fix this? All the tissue folds basically combine into one semi-flat semi-bumpy area. Scrubbing through the files in the DICOM viewer shows that the rear area of my scans looks identical to the rest. So I don't know why I'm having this issue. If I change the threshold it's either too flat on the back, or if I adjust the threshold so the back looks good, the front top becomes way too sharp and the folds aren't even close to one another and it looks like veins instead of tissue folds. Is there any way to apply a gradient to the threshold? Any advice is appreciated!!!

[@IZLabs]

Hi there! Glad you've enjoyed the tutorial! I had similar issues isolating the cerebellum haha. The thresholding + smoothing approach struggles a bit here for sure. Couple things you can try: first, you can create multiple segmentations and segment masks through the segment editor and have one which better matches the cerebellum structures. Then you just have to combine them in your model. Secondly, you could use the manual painting tool in the segment editor to manually pick which regions you want to keep. This would be a little time consuming, but since it's only part of the brain it's doable. Lastly, I would play with your smoothing settings. If I remember correctly, the smoothing is what really took away the detail from my scan. Truth be told, a combination of all three methods (or at least the last 2) may be your best option. Hope that helps! Let me know when you get it figured out!

[@IvyToller]

About how long did the design (pre-print) process take? Hard to tell since you had to edit and clean up for YT (not complaining, just trying to understand a reasonable expectation)

[@IZLabs]

Umm it’s hard to remember for sure, but I’d estimate somewhere in the 2-4hr range. That was outside of the initial time to learn the different software packages

[@bit_mic171]

Great video, super helpful! I'm also struggling with lack of resolution in the resulting model. I can't see any individual folds the way that you have in your resulting model. I've played with the thresholding a lot and it hasn't helped. What size do you consider to be high res? The 'series' I am using is 512x512 size, with 176 as the 'count' column.

[@IZLabs]

Glad you found it helpful! It's less about the 'size' and more about the inter-layer spacing. I don't see a place for 'count', but looking in the Volumes window, under Volume Information, the model I used has an image spacing of between 0.95-1mm, i.e. a new scan layer taken ~1mm apart. On some of my less 'high-res' scans, the spacing can go as high as 6mm. My image dimensions are only 240x240 in the version I used. It seems many people are having issues with the lack of layer resolution in their scans. Maybe check out the process I linked below which supposedly allows for interpolation and smoothing over the higher layer separation, though I haven't used it personally. Good luck! Let me know how things work out for you. discourse.slicer.org/t/i-cannot-remove-ridges-in-3d-segmentation/4792/4

[@bit_mic171]

@@IZLabs Wow I really can't express how grateful I am for this reply, thank you! I will try out it out and share back here. Btw, I see the 'count' column when I go to the 'Add DICOM data' section. Looking at volume information, here are the stats. Image Dimensions: 512x512x176 Image Spacing: 0.466mm x 0.466m x 1mm So that's odd, it seems my resolution may be matching yours yet I still get a smoothed out model at the end, missing the folds. It also seems that the 'count' column from the dicom section is the same as the last dimension under image dimensions.

[@IZLabs]

I now see the count you’re talking about. I may be misunderstanding you (my apologies if so), but if the actual images look ok to you in the three planes but only lose detail during the smoothing step, that may be the issue? Make sure the smoothing size you are using isn’t too high. And of course, always happy to help 😁; also shameless plug, if it’s been useful, please consider subscribing as it helps the channel!

[@bit_mic171]

​@@IZLabsjust liked and subscribed!:) I tried reducing the smoothing but that didn't help. Could you share the image dimensions and spacing dimensions for your high res model? I saw it on the volume tab. Just want to get a sense of how similar they are.

[@IZLabs]

@bit_mic171 I appreciate the sub! That's a shame the smoothing didn't help. If it's still struggling, perhaps try following the steps using the included MRIHead sample data that comes with 3D slicer. That's what I used to make the tutorial video (to not share my brain with the world haha), so I know it should work. If that doesn't work, then perhaps it's the version of Slicer you're using? I've been on 4.11.2 just because I never got around to updating. My parameters for the actual scan of myself that I used are as follows: Image Dimensions: 240 x 240 x 158 Image Spacing: 0.95 x 0.95 x 1.00

[@4posZa]

Hi, How do I get in touch with you? My son has had a corpus callosotomy performed in 2011. It now appears on the MRI that 2 connections have remained resulting in sever epilepsy again. The neuro surgeon suggested a 3D printed version to inspect each part to ascertain if further surgery will be beneficial... I know - tough one...

[@IZLabs]

Hi there, you can reach out via email at: [e-mail]. Full disclosure, I'm not a medical professional and can't provide medical advice, but I'm happy to see if I can help or point you in the right direction.

[@luluknowslp]

thankyou!

[@IZLabs]

Glad you liked it!

[@josefdjarf7622]

Great video! Thank you👌

[@IZLabs]

Glad you liked it!

[@Iamranasameer]

How to check vein that is going above trigeminal nerve into 3d to see contact is there or not??

[@IZLabs]

This may be tough without a contrast agent being used. But if there was contrast used with your scan you may be able to see it. I'd probably look at an anatomy reference and then look at the layers in 2D to get close. Then you can convert just a small section to 3D and carve away to find what you're looking for.

[@Artech1923]

hello is there a tool that merge 3 different series? because when i open one series axial is good but others bad and in the other coronal is good and in the other sagittal is good i want to merge good ones to use in segment editor to make 3 model

[@IZLabs]

Hey, this is a pretty common question. I’ve provided some links to potential workarounds in the other comments, but long story short you might just be out of luck as none of them have reported back with success

[@jacoposchenetti2612]

What if we wanted to print also the inside structures in order to create an assemblable model?

[@IZLabs]

That's a great question. The MRI does contain that data, so you'd just have to be careful with how you are segmenting the files to make sure the internal structures are included in your selection. I would imagine they will be as they share similar densities. I think your best choice would be to individually isolate each part you want to print separately in 3D slicer using the paintbrush tool and then export each as its own model and go through the process of cleaning up each segment using the smoothing tools. That should allow you to relatively easily capture the anatomic structures with more fidelity. I haven't tested it myself, but I imagine that would be a reasonable way of going about this process. What is the use case you're wanting to break the model up for? I'd be curious to see how it comes out! Let me know if you have other questions.

[@samiteee6857]

What about from CD? How do you get it on the computer to begin with?

[@IZLabs]

You can get the CD from your doctor and then just need to find a CD reader to upload it to your computer.

[@samiteee6857]

@@IZLabs any cd reader? That’s great! So have the 3D slicer to access it ready and with the cd reader, I should be good to go? I thought I’d need access to RadiAnt or a medical processing software. I just have a laptop but I’ll look for a cd reader on Amazon or something.

[@iiaurel9847]

So I followed the steps but was a little concerned when I saw my brain in 3d editor looking like a literal sponge, no folds in the brain just a sponge, do you have any fixes for that or is it the MRI issue?

[@IZLabs]

Without looking at it, I can’t say for sure, but it sounds like it may be a low resolution MRI. If you have multiple MRI files try another and look for the one that’s the clearest. Usually they do a couple low res scans to align and then high resolution one for diagnostics

[@iiaurel9847]

@@IZLabs Thank you, for a fast response I'll look into that, thank you

[@iiaurel9847]

@@IZLabs so I tested all the models and all of them get "spongy" tho I have noticed every scan when used with swiss skull stripper turns really blocky (by this I mean the outlines get square shaped) except one scan, tho I still can't get the folds to show

[@IZLabs]

I took another look at my data and think I’ve found the blockiness you’re talking about. That’s what happens with lower resolution models (mine labeled as isoReg DWI HR SENSE has the blocks you describe). For my a file called T1W_3D_TFE SENSE had the highest resolution in all three axes (thus the 3D). If none of your models are showing up with higher resolution it’s possible they didn’t perform a high res 3D scan, but might be worth taking another look! Hope that helps, but do let me know

[@iiaurel9847]

@@IZLabs Sadly I don't have one named 3D tho thank you for looking, I might've just goten low sens MRI thank you for trying and spending your time, have an amazing day!

[@danseiler27]

Thats a great video! I just followed your instruction, but wen it comes to the "Make Solid" step in Meshmixer, than I just get a transparent brain and whenI accept it, its just gone. Do you have any idea what I did wrong?

[@IZLabs]

Glad you enjoyed it! As for the error you're running into... It sounds almost like the make solid step is somehow eliminating your geometry as a whole. Is the part you're processing really thin or otherwise weirdly shaped? Are you using the default settings on the make solid tool? If that's the only step causing issues, you can skip it and things may still work. Let me know, I'm happy to dive in a bit deeper to help if needed.

[@IZLabs]

Also check the object browser (View -> Show Objects Browser) on Mac, as the make solid step creates a new body, which may be hidden for you.

[@michaelweber385]

I ran into this issue too. I was able to solve it by loading the model in Windows 3D Builder. When you import the model, there should be a dialogue box in the bottom right corner that will ask if you want to repair the model. For some reason, this repair method worked for me. After repairing the model, save and import back into meshmixer. You should be able to make the model solid now.

[@IZLabs]

Thanks for sharing!

[@auzzierocks]

I have a different series for each angle (axial, sag, coronal) and combining them into one image in 3D slicer has been a nightmare

[@legeat2555]

Have you been able to do it? Have the same case.

[@IZLabs]

Yeah there are a couple comments on ideas for handling this below, but I haven’t heard anyone become successful yet. Good luck!

[@IanV-2024]

My MRI images are in JPG, I never downloaded a DICOM file and I doubt the records are easily accessible from being years ago, and can't be stuffed contacting them for it. I could spend time learning how to add that into 3D Slicer to be usable, or convert it to DICOM, but I just learnt that the set of images is missing the Coronal axis. Unless you tell me otherwise, this seems important for the process, and I have stopped.

[@IZLabs]

I would bet the records are still around! They make a point of keeping them. You will need DICOM to make this process work easily I’m afraid. I’m not sure of the process at your hospital, but it shouldn’t be too crazy an ask for them to supply the files if you call the patient services number and ask for patient records request

[@anterprites]

Nice tutorial, but I don't like how you did 3D slicing part: Making cylindrical voids adds more paths for printer, you should just change infill amount. Splitting onto 3 parts is kinda pointless, just make it whole thing. Or splice it by brain halfs. Or make it smaller.

[@IZLabs]

Glad you liked most of it. I agree that the voids do add extra complexity, but there was a reason for the madness haha. If you check out my other video on it you can see, but basically the voids were for me to add lights and other electronic components inside the brain. Splitting in 3 parts was done cause I didn’t want to have to use support material and needed a large flat surface to get good bed adhesion. Thanks for your input though!

[@anterprites]

@@IZLabs Thanks for answering :D I used before FreeSurfer that automatically extract the brain in high quality (same example model as you did). Also 3D printed it. The problem I have is something called anisotropic voxel resolution images that are only high quality in one direction, split by multiple scans. Which no software can correctly extract so far. Very annoying.

[@IZLabs]

Ahhh yeah I’ve seen that issue crop up for a lot of people, very frustrating indeed. One method I’ve seen (but never used personally) is described in this blog post. Maybe worth a shot? Let me know if that works! (discourse.slicer.org/t/i-cannot-remove-ridges-in-3d-segmentation/4792/4)

[@ShopTherapy623]

Is there a way to do this but for the bones instead????

[@IZLabs]

Haha it’s definitely possible! Usually done with a CT scan instead of an MRI though as it can better pick up the bone. You’d follow pretty similar steps

[@ShopTherapy623]

@@IZLabs but how would you separate the bones on the CT scan?

[@IZLabs]

Similar process, you use the segmentation tools in 3DSlicer to select by density for the bone and then you can manually select the region of interest. There are also some plugins which can automate the process

[@jimbone007]

When I am looking at my mri images on the 3 different planes only 1 is high res and the other 2 are not. I can click on a different image and then a different image out of the 3 planes is high res. In the example you show all 3 planes are high res. Do all 3 need to be high res or is there a way for me to take the high res of each image and combine them so I have all 3 planes high res?

[@jimbone007]

When I use the software on the mri disc I have It shows all planes in high res. I’m So confused as to how to get them to show as high res on all 3 planes in 3D slicer

[@IZLabs]

Hi there, sorry you're having trouble! It sounds like you had a series of scans done, but may not be viewing the high res 3D scan in 3D scanner. Usually they will do a high res sequence in each plane to try to localize before doing a single full res scan. On the left hand side of the 3D slicer tab do you see the list of multiple scans (under the DICOM, loaded data pane)? Try clicking through a couple different scans, and see if you see one labeled 3D or something similar. When I had mine done, I have about 25 different scans which I can view, and only 2-3 were the full resolution in all three planes. Let me know if that doesn't help and I can try and dig a little deeper for you.

[@jimbone007]

Thanks for the reply! I did check on the left were it says loaded data. I’m on 3D slicer 5.2.2 so it looks a little different but I check all 11 images and they all only have 1 plane as high res. Each of them will have a different plane that is high res which makes me wonder if I can combine them somehow to get all 3 planes to be high res. Also I’ve had 3 head mri and I’m trying the data on all 3. The 11 images is just for 1 of the 3. Thanks for the help btw. I’d really like to get this to work. Such a cool thing print and have!

[@IZLabs]

Ok that would have been the easy fix, but it sounds like you’ve encountered an issue many have where they don’t do a full 3 plane high res scan. It’s quite difficult to align the three scans to form one 3D image due to small movements between scans. One technique is shown in the link below (though I haven’t personally tried it) which may work. But I hope maybe one of your other two scans does contain a combined high res scan as that’s definitely the easiest! Do let me know how it turns out and if I can be of any more help! (discourse.slicer.org/t/i-cannot-remove-ridges-in-3d-segmentation/4792/4)

[@jimbone007]

That post is interesting. I’ll give it a try and hopefully it doesn’t change dimensions since I’d prefer it to be as accurate possible. Either way I’ll reply with my results and thank you for all your help!

[@AlvaroALorite]

Bro, i don't mean to be a party pooper, but you printed your brain with -45 IQ points...you cut away a lot of gray matter hahahaha (that segmentation was too aggresive)

[@IZLabs]

Haha I was going for a clean print and a nice visual. All in the eye of the beholder…😂

[@Y3SS1N]

Me ending up with an empty .STL file😢

[@IZLabs]

Oh no! Usually happens if the selection isn't completed. Did you make sure the 2D images had sections highlighted in green?