Cortical Reconstruction with FreeSurfer

surfer.nmr.mgh.harvard.eduwww.nmr.mgh.harvard.edu/~greve/surferfest3.ppt

Douglas N. Greve

MGH-NMR CenterHarvard University

greve@nmr.mgh.harvard.edu

Acknowledgements• Bruce Fischl• Anders Dale• Marty Sereno• Doug Greve• Xiao Han• Nick Schmansky• Florent Segonne • Yasunari Tosa • CMA (Nikos Makris, Dave Kennedy)• Jenni Pacheco, Brian Quinn, Rahul

Desikan, Ron Killiany, …

Outline• What you can get from FreeSurfer• Surface Reconstruction Theory• Surface Reconstruction with FreeSurfer• Demo• Tutorial• Stuff not covered:

• Mapping functional data to the surface• Surface-based group analysis

Administratia• surfer.nmr.mgh.harvard.edu/fswiki • Platforms: Linux and Mac• Registration• Versions – stable vs dev• FreeSurfer mailing list• Important! How to report a bug!

• Version• Command-line• Error description• subjid/scripts/reconall.log• freesurfer@nmr.mgh.harvard.edu

What FreeSurfer Can Do for You

• “Models” of cortical surfaces

• Morphological measures of cortex

• Surfaced-based registration to atlas

• Automatic surface labeling/segmentation

• Automatic volume labeling/segmentation

• Surface-based group analysis

• Visualization in the volume and surface

Surface Model

• Mesh (“Finite Element”)

• Vertex = point of 6 triangles

• XYZ at each vertex

• Triangles/Faces ~ 150,000

• Area, Distance

• Curvature, Thickness

• Moveable

Surfaces: White and Pial

Inflation

Cortical Morphometry

Thickness

Sulcal Curvature

Gyrus

“Mean” CorticalSurface+

-Pial

White/Gray

Sulcus

Inter-Subject Registration

Inter-Subject AveragingS

ubje

ct 1

Sub

ject

2

Native Spherical Spherical

Surface-to- Surface

Surface-to- Surface

GLM

Demographics

mris_glm

Automatic Surface Labeling

• Based on folding• Fine-tuned to subject

Automatic Volume Labeling

• Subcortical Segmentation

• Caudate• Putamen• Pallidum• Hippocampus• Amygdala• Nucleus Acumbens• Ventricles • White Matter• Cerebellum• …

Visualization

Surface Reconstruction Theory• Find white/gray surface

• Find pial surface

• “Find” = create mesh– Vertices, neighbors, triangles, coordinates

– Accurately follows boundaries between tissue types

– “Topologically Correct”

• closed surface, no donut holes

• no self-intersections

– Subcortical Segmentation along the way

Find “Subcortical Mass”

• All White Matter

• All Subcortical Structures

• Ventricles

• Excludes brain stem and cerebellum

• Hemispheres separated

• Connected (no islands)

• Many Stages … More Later …

Tessellation and Topology Fixing

• Mosaic of triangles (“tessellation”)

• Donut holes, handles

• Automatic topology fixer

orig surface surf/lh.orig surf/rh.orig

White Matter Surface• Nudge orig surface

• Follow T1 intensity gradients

• Smoothness constraint

• Vertex Identity Stays

Pial Surface

• Nudge white surface

• Follow T1 intensity gradients

• Vertex Identity Stays

Details on Getting the Subcortical Mass

• Hardest part

• Some FreeSurfer background

• All Stages of Processing

• Volumetric Processing

FreeSurfer Directory Tree

bert

bem label morph mri scripts surf tiff

orig T1 brain aseg wm filled

Subject ID

mksubjdirs SubjectId

Each data set has its own unique SubjectId (eg, bert)

mksubjdirs bert

Exercise A:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/Conversion

Individual StagesVolumetric Processing Stages (subjid/mri):1. Motion Cor, Avg, Conform (orig.mgz)2. Talairach transform computation3. Non-uniform inorm (nu.mgz)4. Intensity Normalization 1 (T1.mgz)5. Skull Strip (brain.mgz)

6. EM Register (linear volumetric registration) 7. CA Intensity Normalization 8. CA Non-linear Volumetric Registration 9. CA Label (Volumetric Labeling) (aseg.mgz)

10. Intensity Normalization 2 (T1.mgz)11. White matter segmentation (wm.mgz)12. Edit WM With ASeg13. Fill and cut (filled.mgz)

Surface Processing Stages (subjid/surf):14. Tessellate (?h.orig)15. Smooth1 (?h.smoothwm)16. Inflate1 (?h.inflated)17. QSphere (?h.qsqhere)18. Automatic Topology Fixer (?h.orig)19. Euler Number20. Smooth221. Inflate222. Final Surfs (?h.white,?h.pial)23. Cortical Ribbon Mask

24. Spherical Morph25. Spherical Registration 26. Spherical Registration27. Map average curvature to subject28. Cortical Parcellation (Labeling)29. Cortical Parcellation Statistics30. Cortical Parcellation mapped to ASeg

recon-all -help Note: ?h.orig means lh.orig or rh.orig

Green = Manual Intervention?

Motion Correction and Averaging

001.mgz

002.mgz

+ rawavg.mgz

orig

001.mgz 002.mgz

mri

rawavg.mgzDoes not change native resolution.

Conform

rawavg.mgz

orig.mgz

mri

rawavg.mgz

Changes to 256^3, 1mm^3All volumes will be conformed.

orig.mgz

Talairach Transform

• Computes 12 DOF transform matrix

• Does NOT resample

• MNI305 template

• Used to help find structures (eg, CC)

• Can also be used to localize functional activation

• mri/transforms/talairach.xfm

mri

transforms

talairach.xfm

Intensity Normalization

• Removes B1 bias field

• NU (MNI) nu.mgz

• Presegmentation (T1.mgz)– All WM = 110 intensity– Pre- and Post-Skull Strip

T1 Volume

Skull Strip

• Removes all non-brain– Skull, Eyes, Neck, Dura

• brain.mgz

Orig Volume Brain Volume

Automatic Volume Labeling

• Used to fill in subcortical structures for creating subcortical mass• Usefull in its own right• aseg.mgz

ASeg Volume

White Matter Segmentation• Separates white matter from everything else

• “Fills in” subcortical structures

• Cerebellum removed, brain stem still there

• Somewhat redundant with aseg

• wm.mgz

WM Volume

Fill and Cut (Subcortical Mass)• Fills in any voids• Removes any islands• Removes brain stem• Separates hemispheres (each hemi has different value) • filled.mgz – “Subcortical Mass”

WM Volume Filled Volume

Pressing the Buttons

• Creating a FreeSurfer subject directory• Adding your data• The SUBJECTS_DIR environment variable• Automatic reconstruction• Actual Reconstruction and Troubleshooting• Downloading and Installing• Getting Help• Tutorial Overview

FreeSurfer Directory Tree

bert

bem label morph mri scripts surf tiff label

orig T1 brain wm aseg

Subject ID

mksubjdirs SubjectId

Each data set has its own unique SubjectId (eg, bert)

mksubjdirs bert

Exercise A:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/Conversion

SUBJECTS_DIR Environment Variable

bert

$SUBJECTS_DIR

fred jenny margaret …

Subject ID

Add Your Data• mri_convert yourdicom.dcm bert/mri/orig/001.mgz• mri_convert yourdicom.dcm bert/mri/orig/002.mgz• …

bert

bem label morph mri scripts surf tiff label

orig

001.mgz 002.mgz

Exercise A:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/Conversion

MGZ – Compressed MGH Format

Automatic Reconstruction

Come back in 48 hours …

Check your results – do the white and pial surfaces follow the boundaries?

-- Not normally how we proceed …

recon-all –s bert –autorecon-all

Partial Voluming –WM classified as Non-WMin thin strands in Temporal Lobe

Exercise F:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/FixingDefects

Eye Socket classified as WM.Skull Strip Failure.

Exercise F:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/FixingDefects

Exercise F:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/FixingDefects

“Hypo-Intensities”White Matter Lesions

Blood VesselWhite/Gray OK, butPial Inaccurate

Exercise G:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/PialEdits

Exercise C:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/ControlPoints

Intensity Normalization Failure. All WM in T1 volume (T1.mgz) should be 110.Can fix by adding “Control Points”. Beware partial voluming!

Exercise D:https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/SkullStripTutorial

Skull Strip Failure

Individual StagesVolumetric Processing Stages (subjid/mri):1. Motion Correct and Average (orig.mgz)2. Talairach transform computation3. Non-uniform inorm (nu.mgz)4. Intensity Normalization 1 (T1.mgz)5. Skull Strip (brain.mgz)

6. EM Register (linear volumetric registration) 7. CA Intensity Normalization 8. CA Non-linear Volumetric Registration 9. CA Label (Volumetric Labeling) (aseg.mgz)

10. Intensity Normalization 2 (T1.mgz)11. White matter segmentation (wm.mgz)12. Edit WM With ASeg13. Fill and cut (filled.mgz)

Surface Processing Stages (subjid/surf):14. Tessellate (?h.orig)15. Smooth1 (?h.smoothwm)16. Inflate1 (?h.inflated)17. QSphere (?h.qsqhere)18. Automatic Topology Fixer (?h.orig)19. Euler Number20. Smooth221. Inflate222. Final Surfs (?h.white,?h.pial)23. Cortical Ribbon Mask

24. Spherical Morph25. Spherical Registration 26. Spherical Registration27. Map average curvature to subject28. Cortical Parcellation (Labeling)29. Cortical Parcellation Statistics30. Cortical Parcellation mapped to ASeg

recon-all -help Note: ?h.orig means lh.orig or rh.orig

Green = Manual Intervention?

Actual Workflow

1. recon-all –autorecon1 (Stages 1-5)2. Check talairach transform, skull strip, normalization (?)3. recon-all –autorecon2 (Stages 6-23)4. Check surfaces

1. Add control points: recon-all –autorecon2-cp (Stages 10-23)

2. Edit wm.mgz: recon-all –autorecon2-wm (Stages 13-23)3. Edit brain.mgz: recon-all –autorecon2-pial (Stage 23)

5. recon-all –autorecon3 (Stages 24-30)

Note: all stages can be run individually

Tutorial Exerciseshttps://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/ExercisesYou don’t actually have run anything!

• Exercise A: mksubjdir and convert a DICOM volume into mgz format • Exercise B: SKIP (check talairach registration)• Exercise C: Using control points for intensity normalization • Exercise D: Fixing skull stripping • Exercise E: View orig, T1, brain, and wm volumes with tkmedit• Exercise F: Recognizing and fixing inaccuracies in the white matter surfaces • Exercise G: Correcting Pial Surfaces

Administratia• surfer.nmr.mgh.harvard.edu/fswiki • Platforms: Linux and Mac• Registration• Versions – stable vs dev• FreeSurfer mailing list• Important! How to report a bug!

• Version• Command-line• Error description• subjid/scripts/reconall.log

• Stuff not covered:• Mapping functional data to the surface• Surface-based group analysis

Study Questions1. What are the two types of surfaces?

2. How are surfaces modeled?

3. What is meant by a topologically correct surface?

4. What is the purpose of the SUBJECTS_DIR?