Project Timespan
September 2019
Inspiration
Everyone has those “off” days during which one feels that their brain has been switched into the “resting, don’t bother me” setting. On the flip side, there are also those days during which one’s brain is just “on.” I call these days a “small-brain day” and “big-brain day”, respectfully.
One of the upsides of being in the medical field is that there are plenty of opportunities to participate as “healthy controls” in studies, in which the research group will compensate you for your time and participation. In this case, my big-brain moment came when I signed up to be a healthy control for a multi-center study investigating chronic fatigue syndrome.
Over the course of two days, I took part in radiographic and physiologic analysis involving MRI, PET/CT, and biking fitness tests. At the end of my participation in the study, in addition to receiving a healthy amount of monetary compensation for my time, I was able to obtain MRI scans of my head!
Project Overview
There is an amazing tutorial by user 3d_printed_brain on Instructables of segmentation of one’s brain from an MRI scan. I basically followed this tutorial, step by step, so in lieu of redundancy, I will refer you to there for the details.
MRI Files
I spent about an hour in the MRI scanner and an hour in the PET/CT scan in consecutive days, and by the end of the process I was all scanned-out. I did receive a nice CD containing the various MRI sequences of my brain. Unfortunately, I was not able to get CT scans for myself, or I would have been able to segment out the bony structures as well.
The MRI sequences I was given included FSPGR Bravo, SWI, MAG, QSM, T2 Flair, and SWI. Of these, I chose the axial FSPGR Bravo sequence as it provided the best contrast and resolution.
Axial FSPGR Bravo Sequence
Coronal FSPGR Bravo Sequence
Axial T2 Flair sequence
FreeSurfer
As I was running the program on a Windows computer, I had to download FreeSurfer for VirtualBox and follow the command prompt instructions as presented in the tutorial. It is really cool how automated the process is – as someone who spends a fair bit of time segmenting out 3D geometries of structures such as the tibial cartilage and menisci, I know how time-consuming this process can be from a manual perspective. Nevertheless, the automated process took a long time. In the tutorial, user 3d_printed_brain described the program running for up to two whole days, and I corroborate his claim as it took my computer something like 25 hours or so to complete the program.
MeshLab
After exporting the .stl files of the right and left hemispheres from FreeSurfer, I refined the mesh within MeshLab. MeshLab is an excellent program for smoothing and parametrizing 3D geometries. I use it frequently in research projects in medical school, and it is completely free and open-source.
The images below are the sample outputs of this process – my cerebrum along the axial, coronal, and sagittal views, along with a sagittal view of the left cerebral hemisphere.
3D Printed Brain
I printed two versions of my cerebrum, one at 50% scale and one at 100% scale. For the former, I used the school’s printer, a MakerBot Replicator 5th generation, to print a version in white PLA. For the full-scale brain, I was able to go the dedicated 3D printing facility at Cornell to print out the brain for free, as it was part of a research study. The printer they used, a FDM Stratasys uPrint, is much, much better than the one I used. I requested this print to be in ivory color, and it turned out wonderfully!
Lessons Learned & Improvements
I am quite satisfied in the end results of the 3D printing, though there are a few areas of improvement. For one, whether this is vanity or insanity, I really wanted to print out my cerebrum with a metal 3D printer, though the cost of this process using third-party services was too exorbitant for me to seriously entertain.
There are some neat 3D printing materials available, including models of organs that are composed of transparent material such that one can visualize the underlying vessels. One of my early thoughts was to 3D print the brain and the calvarium in separate pieces, the latter being composed of transparent material. However, the process of extracting out the brain was difficult enough.
Finally, this model of the brain lacks the brain stem and cerebellum. The brain stem, in particular, would have made for an excellent stand for display, and there is a tutorial for extraction of these subcortical regions, but I didn’t quite figure it out.
Reflection
I find it a strange feeling to hold the 1:1 version of my cerebrum in my hands. The brain is such a fascinating organ, and I am grateful that mine is able to churn out some big-brain moments, once in a while, at least.