Manual to 3D-print brain

To print your brain scan as a plaster brain, read on! What you need is a T1 weighted anatomical scan of your brain, preferably acquired on a 3T scanner. I used a simple T1-FFE sequence, which comes pretty standard on a Philips MR.

I) If you have a DICOM file: Transfer this to nifti, e.g. using the DICOM importer in SPM (for Matlab). If you have a PAR/REC file: Transfer into nifti format (e.g. using r2aGUI, or the dcm2nii found in mricron).

II) Read nifti with SPM, in Matlab

a) Use Segment option to tear apart white from gray matter. Batch Editor interface; SPM > Spatial > Segment; default values are OK, but maybe a bit too coarse for the perfectionist as it uses probability maps and smoothes the image for a bit. Use the c1 output file for a nice gray matter volume and the c2 output file for a white matter volume. If you’d like to integrate the two, go to the ImCalc feature of SPM and add the two together.

b) check with MRIcroGL if you’ve got a good volume matrix (not a mesh yet!)

III) Read nifti in ITK-snap (although ITK-Gray seems to be doing a better job at this)

a) Open grayscale image, select .nii file (nifti)

b) Click on the ‘Snake ROI tool’ in the Main Toolbox window (top left); click ‘Segment 3D’; choose segmentation on the basis of ‘intensity regions’, click on ‘preprocess image’ and play with the details to make sure your brain appears in one clear colour. Press ‘okay’, and ‘next’.

c) Step 2 involves placing bubbles in your brain which will be used as point from where brain and no-brain volumes will be rendered. I found placing the bubbles outside the brain yield the easiest .stl files to read in Meshlab. Place the bubbles, press the play button, and press stop when the whole image has been calculated. Press ‘finish’. NB: unfortunately this method is coarse, and will not get all the detail from your SPM model.

d) Now go to Segmentation > Export as surface mesh, export Label 1, save as a STL Mesh file.

IV) Read .stl in Meshlab, you now need to:

a) Simplify your model, do this by selecting areas you don’t want. This can be cumbersome, haven’t found a good solution to skip this step yet. Make sure your brain is one whole piece and that there is a decent connection between your R and L hemisphere.

b) Waterproof your model: normals, piosson reconstruction, see this video (I’m not entirely sure if I did this right; I’m a neuro guy, not a 3D modeller).

V) Check your model with MiniMagics to make sure your model will be printable by an average 3d printer. If there is any error, go back to Meshlab and (try to) improve your model.

VI) You’ll now have a .stl file which will be accepted by the majority of 3D-printers. (If no printer is near, you can even have it printed and shipped to you, see Shapeways) NB: make sure that you have one (1) object and not many by checking it in 3D-print software (e.g. ZPrint for a ZCorps 3D-printer); my model had multiple objects in the center (anatomical name: the ventricles) which I could not remove myself, but a handy man at the 3D-printing lab could using Rhinoceros 4.0. Load the .stl in the programme, select the outer mesh, then select the inverse and press delete.

VII) Press print! I have printed mine in plaster, which looks quite OK in my opinion, but you can also print them in any hard material. I’ve seen options to print it in gold, even…

Additionally, what you can do, is to make a negative food-safe mold from your brain. With that negative mold, you can make your brain in chocolate… See Make your own molds for an overview what products you can use to create your own moulds – I’ve used mould-making stuff from sillicreations, and it worked quite well. Pay attention to the shifting of your brain in the drying mold; it will have the tendency to move through the medium.

Additionally II: I’ve started working with FreeSurfer, which outputs the surface area of left and right hemisphere automatically. Take a look here to see how those meshes can be read into Blender – 3D software that will also save a printer-compatible .stl file. Please note: I haven’t tried this!

Sources, which give a manual of some kind, but not one completely clear:

http://www.instructables.com/id/Edible-Chocolate-Brain-from-MRI-Scan/

http://i.imgur.com/lP2Lp.jpg

http://www.polycount.com/forum/showthread.php?t=97924

Originally, I posted this on reddit, look for iloveoxytocin's post.

Let me know if you’ve printed your brain and what you think of it..!

Previous Post Next Post