BrainVISAの結果をMNE-Cで使う方法

MNE-Cで電流源推定... 1

導体モデルの作成... 1

脳磁図データファイルとその共分散ファイルのコピー... 1

mne_analyzeで座標合わせ... 1

電流源推定... 1

mne_analyzeで結果表示... 1

皮質メッシュで電流源推定... 1

MNE-Cではvolume dataであるT1.mgz、brain.mgz、polygon mesh dataであるlh.white、rh.whiteがあればとりあえずdSPMで電流源の推定結果を表示することは可能です。

BrainVISAでは

~/default_acuisition/subject.nii.gz、(/mri/T1.mgz)

~/default_acquisition/default_analysis/segmentation/brain_subject.nii.gz、(/mri/brain.mgz　但しT1.mgzデータのbrain.mgz部分)

~/default_acuisition/default_analysis/segmentation/mesh/subject_Lwhite.gi、(/surf/lh.white)

~/default_acquisition/default_analysis/segmentation/mesh/subject_Rwhite.gi、(/surf/rh.white)

がそれに相当します。

NIfTIファイルからmgzファイルへの変換はFreeSurferのmri_convet、

GIfTIファイルからsurfファイルへの変換はFreeSurferのmris_convert

で行います。が、座標系の変換はしてくれませんので書き換える必要があります。

MATLAB R2017b以降でNIfTIファイルの読み書きが可能になっています。またGIfTIはGIfTI toolbox for MATLABを使うことでファイルの読み書きが可能です。

MNE MATLAB toolboxというのがあるのはあるのですが、基本的にFIFFファイル読み書き用のtoolです。FreeSurferのsurfファイルは読み書きできるものの、MGZファイルは扱えないようです。

座標変換がない場合

水平断のDICOM画像をNIfTI化し、BrainVISAで処理したNIfTIデータとFreeSurferで処理したMGZデータを比較します。

MATLAB R2018aのコードです

尚、MGZファイルはFreeSurferのmri_convertコマンドでNIfTIファイル化しています。

clear;

close all;

a.T1.data=niftiread('c:\test\a\mri\T1.nii.gz');

a.brain.data=niftiread('c:\test\a\mri\brain.nii.gz');

% a.T1.info=niftiinfo('c:\test\a\mri\brain.nii.gz'); % 読み込みエラーになる

a.brain.info=niftiinfo('c:\test\a\mri\brain.nii.gz');

b.T1.data=niftiread('c:\test\b\mri\T1.nii.gz');

b.brain.info=niftiinfo('c:\test\b\mri\brain.nii.gz');

figure('color',[1,1,1]);

colormap(gray(256));

for n=1:2

if n==1

Z=a.T1.data;

scale=a.brain.info.PixelDimensions;

else

Z=b.T1.data;

scale=b.brain.info.PixelDimensions;

end

subplot(2,3,n*3-2);

imagesc(squeeze(Z(size(Z,1)/2,:,:)));

daspect([scale(2),scale(3),scale(1)]);

subplot(2,3,n*3-1);

imagesc(squeeze(Z(:,size(Z,2)/2,:)));

daspect([scale(1),scale(3),scale(2)]);

subplot(2,3,n*3);

imagesc(squeeze(Z(:,:,size(Z,3)/2)));

daspect([scale(1),scale(2),scale(3)]);

end

set(findobj(gcf,'type','axes'),'clim',[0,255]);

上段がBrainVISA、下段がFreeSurferです。三次元配列の第2・3軸、第1・3軸、第1・2軸の順となっています。

BrinVISAでは座標系は右左、後前、下上の順になっています。

FreeSurferでは座標系は右左、上下、後前の順になっています。

これらの画像をもとに脳を切り出してpolygon mesh化したデータを比較します。

尚surfファイルはmris_convertコマンドで予めGIfTIファイルに変換しておきます。

a.meshL=gifti('c:\test\a\surf\Lwhite.gii');

a.meshR=gifti('c:\test\a\surf\Rwhite.gii');

% 何故かmris_convert lh.white Lwhite.giiはできず

b.meshL=gifti('c:\test\b\surf\lh.white.gii');

b.meshR=gifti('c:\test\b\surf\rh.white.gii');

figure('color',[1,1,1]);

for n=1:2

if n==1

p1=a.meshL.vertices;

t1=a.meshL.faces;

p2=a.meshR.vertices;

t2=a.meshR.faces;

else

p1=b.meshL.vertices;

t1=b.meshL.faces;

p2=b.meshR.vertices;

t2=b.meshR.faces;

end

subplot(2,3,n*3-2);

meshL=trimesh(t1,p1(:,1),p1(:,2),p1(:,3),'LineStyle','none','facecolor',[0,0,1],'facealpha',0.1);hold on;

meshR=trimesh(t2,p2(:,1),p2(:,2),p2(:,3),'LineStyle','none','facecolor',[0,1,0],'facealpha',0.1);

view([0,0]);daspect([1,1,1]);axis tight;grid on;

subplot(2,3,n*3-1);

copyobj(meshL,gca);hold on;

copyobj(meshR,gca);

view([90,0]);daspect([1,1,1]);axis tight;grid on;

subplot(2,3,n*3);

copyobj(meshL,gca);hold on;

copyobj(meshR,gca);

view([0,90]);daspect([1,1,1]);axis tight;grid on;

end

上がBrainVISA、下がFreeSurferです。青が左半球、緑が右半球です。左が正面、中央が左側面、右が上から見たところです。

BrainVISAは中点が128くらい、FreeSurferは中点は0くらいになっています。

MNE-CはFreeSurfer用の座標でないとうまく電流源推定できません。

freeviewでT1.ni.gzとLwhite.giiを開いてみました。

mne_analyzeでLwhite.giiとRwhite.giiをsurfaceファイルに変換して開いたものです。

座標変換した場合

NIfTIとGIfTIと別々に座標変換します。まずNIfTIから書き換えます。三次元配列(data)とヘッダー部分(info)を書き換えます。

% 三次元配列の書き換え

d=size(a.T1.data);

data=repmat(a.T1.data(1,1,1),d(1),d(3),d(2));

for m=1:2

data0=a.T1.data;

if m==2

data0(a.brain.data<1)=0;

end

for n=1:d(3)

data(:,n,:)=data0(:,:,d(3)-n+1);

end

if m==1

a.T1.data0=data;

else

a.brain.data0=data;

end

% infoの書き換え

info=a.brain.info;

info.ImageSize=[d(1),d(3),d(2)];

info.PixelDimensions=info.PixelDimensions([1,3,2]);

d=info.ImageSize.*info.PixelDimensions;

d=d([1,3,2]);

info.SpaceUnits='Millimeter';

info.TimeUnits='Second';

T=info.Transform.T;

T(:,[2,3])=T(:,[3,2]);

T(3,2)=-T(3,2);

T(4,2)=-T(4,2)+T(3,2);

T(:,1:3)=-T(:,1:3);

T(1,1)=-T(1,1);

T(4,1)=d(1)/2;

info.Transform.T=T;

info.raw.dim([2,3])=info.raw.dim([3,2]);

info.raw.pixdim([3,4])=info.raw.pixdim([3,4]);

info.raw.qoffset_x=T(4,1);

info.raw.qoffset_y=T(4,2);

info.raw.qoffset_z=T(4,3);

info.raw.srow_x=T(:,1)';

info.raw.srow_y=T(:,2)';

info.raw.srow_z=T(:,3)';

figure('color',[1,1,1]);

colormap(gray(256));

for n=1:2

if n==1

% Z=a.T1.data;

% scale=a.brain.info.PixelDimensions;

Z=a.T1.data0;

scale=info.PixelDimensions;

else

Z=b.T1.data;

scale=b.brain.info.PixelDimensions;

end

subplot(2,3,n*3-2);

imagesc(squeeze(Z(size(Z,1)/2,:,:)));

daspect([scale(2),scale(3),scale(1)]);

subplot(2,3,n*3-1);

imagesc(squeeze(Z(:,size(Z,2)/2,:)));

daspect([scale(1),scale(3),scale(2)]);

subplot(2,3,n*3);

imagesc(squeeze(Z(:,:,size(Z,3)/2)));

daspect([scale(1),scale(2),scale(3)]);

end

set(findobj(gcf,'type','axes'),'clim',[0,255]);

上段はBrainVISA、下段はFreeSurferです。座標が揃いました。

MNE-CはmriのT1.mgz、brain.mgzというファイルを使います。既存のファイルのファイル名を変更し、座標変換後のファイルをT1.nii.gz、brain.nii.gzというファイル名にしておきます。

% ファイル保存

copyfile c:\test\a\mri\T1.nii.gz c:\test\a\mri\orig_T1.nii.gz

copyfile c:\test\a\mri\brain.nii.gz c:\test\a\mri\orig_brain.nii.gz

% delete c:\test\a\mri\T1.nii.gz % 上書きされるので不要

% delete c:\test\a\mri\brain.nii.gz

niftiwrite(a.T1.data0,'c:\test\a\mri\T1',info,'compressed',true);

niftiwrite(a.brain.data0,'c:\test\a\mri\brain',info,'compressed',true);

次はGIfTIです。

d=a.brain.info.ImageSize.*a.brain.info.PixelDimensions;

X0=[eye(3,4);-d(1)/2,-d(2)/2,-d(3)/2,1];

X1=[-1,0,0,0;

0,-1,0,0;

0,0,-1,0;

0,0,0,1];

X1=X0*X1*[1,0,0,0;0,1,0,0;0,0,1,0;0,0,0,1];

figure('color',[1,1,1]);

for n=1:2

if n==1

p=a.meshL.vertices;

p(:,4)=1;

p=p*X1;

p1=p(:,1:3);

t1=a.meshL.faces;

p=a.meshR.vertices;

p(:,4)=1;

p=p*X1;

p2=p(:,1:3);

t2=a.meshR.faces;

else

p1=b.meshL.vertices;

t1=b.meshL.faces;

p2=b.meshR.vertices;

t2=b.meshR.faces;

end

subplot(2,3,n*3-2);

meshL=trimesh(t1,p1(:,1),p1(:,2),p1(:,3),'LineStyle','none','facecolor',[0,0,1],'facealpha',0.1);hold on;

meshR=trimesh(t2,p2(:,1),p2(:,2),p2(:,3),'LineStyle','none','facecolor',[0,1,0],'facealpha',0.1);

view([0,0]);daspect([1,1,1]);

axis tight;grid on;

subplot(2,3,n*3-1);

copyobj(meshL,gca);hold on;

copyobj(meshR,gca);

view([90,0]);daspect([1,1,1]);

axis tight;grid on;

subplot(2,3,n*3);

copyobj(meshL,gca);hold on;

copyobj(meshR,gca);

view([0,90]);daspect([1,1,1]);

axis tight;grid on;

end

上段はBrainVISA、下段はFreeSurferです。これで座標が合いました。

MNE-Cはsurfのlh.whiteとrh.whiteというファイルを使います。既存のファイルのファイル名を変更し、座標変換後のファイルをLwhite.gii、Rwhite.giiというファイル名にしておきます。またMNE-Cはpolygonメッシュ面に裏と表があり、このままだと裏返ってしまうので三角メッシュの頂点の順番を入れ替えておきます。

% ファイル保存

copyfile c:\test\a\surf\Lwhite.gii c:\test\a\surf\orig_Lwhite.gii;

copyfile c:\test\a\surf\Rwhite.gii c:\test\a\surf\orig_Rwhite.gii;

for n=1:2

if n==1

g=a.meshL;

else

g=a.meshR;

end

gg=[];

gg.mat=eye(4);

gg.faces=g.faces;

p=g.vertices;

p(:,4)=1;

p=p*X1;

gg.vertices=p(:,[1,2,3]);

gg.faces=g.faces(:,[1,3,2]);

gg=gifti(gg);

if n==1

save(gg,'c:\test\a\surf\Lwhite.gii','Base64Binary');

else

save(gg,'c:\test\a\surf\Rwhite.gii','Base64Binary');

end

freeviewでT1.nii.gzとLwhite.giiを開いてみました。

mne_analyzeでLwhite.giiとRwhite.giiをsurfaceファイルに変換して開いてみました。

上記のようであれば、MNE-Cが使えます。

visa_convets.m

BrainVISAで出力されたフォルダを指定することで上記の処理を行うMATLABの関数Mファイル、visa_converts.mというのを作成しました。

addpath(GIfTI toolbox for MATLABのフォルダ)

visa_convert

とするだけでt1mriフォルダの隣にBVフォルダを作成します。BVフォルダの中には座標変換後の~/mri/T1.nii.gz、~/mri/brain.nii.gz、~/surf/Lwhite.gii、~/surf/Rwhite.giiが含まれます。head.gii、Lhemi.gii、Rhemi.gii、Lwhite_inflated.gii、Rwhite_inflated.giiなどもあれば、それらも座標変換して保存されます。

実際のファイルは以下の通りです。MATLAB R2017b以降で使用可能です。

function visa_converts

folder=uigetdir;

x=strfind(folder,'\');

name=folder((x(end)+1):end);

[d,flag]=process1(folder,name);

process2(folder,name,d,flag);

function [d,flag]=process1(folder,name)

T1name=[folder,'\t1mri\default_acquisition\',name,'.nii.gz'];

T1folder=[folder,'\BV\mri'];

if exist(T1folder,'dir')~=7;mkdir(T1folder);end

brainname=[folder,'\t1mri\default_acquisition\default_analysis\segmentation\brain_',name,'.nii.gz'];

T1=niftiread(T1name);

brain=niftiread(brainname);

T2=T1;

T2(brain<1)=0;% brainはゼロ以上と仮定

info=niftiinfo(brainname);

info.Transform.T

d=size(T1);

T0=repmat(T1(1,1,1),[d(1),d(3),d(2)]);

T3=T0;

flag=1;

if all(info.ImageSize==[256,256,256])

if all(info.PixelDimensions==[1,1,1])

if all(all(info.Transform.T(1:3,1:3)==[-1,0,0;0,0,-1;0,1,0])==[1,1,1])

flag=0;

end

if flag==1

for z=1:d(3)

T0(:,z,:)=T1(:,:,d(3)-z+1);

T3(:,z,:)=T2(:,:,d(3)-z+1);

end

info.ImageSize=[d(1),d(3),d(2)];

info.PixelDimensions=info.PixelDimensions([1,3,2]);

d=info.ImageSize.*info.PixelDimensions;

d=d([1,3,2]);

info.SpaceUnits='Millimeter';

info.TimeUnits='Second';

T=info.Transform.T;

T(:,[2,3])=T(:,[3,2]);

T(3,2)=-T(3,2);

T(4,2)=-T(4,2)+T(3,2);

T(:,1:3)=-T(:,1:3);

T(1,1)=-T(1,1);

T(4,1)=d(1)/2;

info.Transform.T=T;

info.raw.dim([2,3])=info.raw.dim([3,2]);

info.raw.pixdim([3,4])=info.raw.pixdim([3,4]);

info.raw.qoffset_x=T(4,1);

info.raw.qoffset_y=T(4,2);

info.raw.qoffset_z=T(4,3);

info.raw.srow_x=T(:,1)';

info.raw.srow_y=T(:,2)';

info.raw.srow_z=T(:,3)';

% info.raw

else

d=[256,256,256];

disp('FreeSurfer format');

T0=T1;

T3=T2;

end

niftiwrite(T0,[T1folder,'\T1'],info,'compressed',true);

niftiwrite(T3,[T1folder,'\brain'],info,'compressed',true);

function process2(folder,name,d,flag)

folder2=[folder,'\BV\surf'];

if exist(folder2,'dir')~=7;mkdir(folder2);end

folder=[folder,'\t1mri\default_acquisition\default_analysis\segmentation\mesh\'];

str{1}='_head.gii';

str{2}='_Lhemi.gii';

str{3}='_Rhemi.gii';

str{4}='_Lwhite.gii';

str{5}='_Rwhite.gii';

str{6}='_Lwhite_inflated.gii';

str{7}='_Rwhite_inflated.gii';

X0=[eye(3,4);-d(1)/2,-d(2)/2,-d(3)/2,1];

X1=[-1,0,0,0;

0,-1,0,0;

0,0,-1,0;

0,0,0,1];

if flag==1

X1=X0*X1*[1,0,0,0;0,1,0,0;0,0,1,0;0,0,0,1];

else

X1=X0*X1*eye(4,4);

end

x1=[1,2,3];

for n=1:7

st=str{n};

loadname=[folder,name,st];

if exist(loadname,'file')~=2;continue;end

savename=[folder2,'\',st(2:end)];

savename=strrep(savename,'white_','');

g=gifti(loadname);

P=g.vertices;

P(:,4)=1;

P=P*X1;

gg=[];

gg.mat=eye(4);

gg.faces=g.faces(:,[1,3,2]);

gg.vertices=P(:,x1);

gg=gifti(gg);

save(gg,savename,'Base64Binary');

end

実行後にt1mriの隣にBVというフォルダが作成されます。このフォルダ名をsubject名に変更してコピーすればMNE-Cで使用可能となります。

MNE-Cで電流源推定

導体モデルの作成

以下のようなファイルを作成し、$MNE_ROOT/mne_preparebv.cshというファイル名で保存しました。

# make COR.fif from T1.mgz and brain.mgz

cd mri

mri_convert T1.nii.gz T1.mgz

mri_convert brain.nii.gz brain.mgz

cd ..

mne_setup_mri

# make 7mm space nodes on cortices from lh.white and rh.white

cd surf

mris_convert Lwhite.gii lh.white # using white mesh

mris_convert Rwhite.gii rh.white #using white mesh

# mris_convert Lhemi.gii lh.pial # using pial mesh

# mris_convert Rhemi.gii rh.pial # using pial mesh

cd ..

mne_setup_source_space # --surface white # default

# mne_setup_source_space --surface pial # using pial mesh

# make 4-layer trigon mesh

mne_watershed_bem --atlas

# make dense scalp surface

# mkheadsurf -subjid $SUBJECT -srcvol T1.mgz -hemi lh

# mkheadsurf -subjid $SUBJECT -srcvol T1.mgz -hemi rh

# cd $SUBJECTS_DIR/$SUBJECT/bem

# mne_surf2bem --surf ../surf/lh.seghead --id 4 --check --fif ${SUBJECT}-head-dense.fif

# mv ${SUBJECT}-head.fif ${SUBJECT}-head-sparse.fif

# cp ${SUBJECT}-head-dense.fif ${SUBJECT}-head.fif

# convert into inner.tri

cd $SUBJECTS_DIR/$SUBJECT/bem/watershed

mne_convert_surface --surf ${SUBJECT}_inner_skull_surface --triout ../inner_skull.tri

# forward model

cd $SUBJECTS_DIR

mne_setup_forward_model --homog --noswap --ico4

用意するファイルは

の2つのフォルダの

mriフォルダに

surfフォルダに

です。名前は変更しないでください。

terminalを開いて以下のようにします。尚subjects_dirは/home/centos/Desktop、subject名はaです。

bash: export: `on/home': not a valid identifier

bash: whchvirtualenvwrapper.sh: command not found

bash: source: filename argument required

source: usage: source filename [arguments]

[centos@localhost Desktop]$ csh

-------- freesurfer-Linux-centos6_x86_64-stable-pub-v6.0.0-2beb96c --------

Setting up environment for FreeSurfer/FS-FAST (and FSL)

FREESURFER_HOME /usr/local/freesurfer

FSFAST_HOME /usr/local/freesurfer/fsfast

FSF_OUTPUT_FORMAT nii.gz

SUBJECTS_DIR /home/centos/Desktop

MNI_DIR /usr/local/freesurfer/mni

MNE software location set to: /usr/local/MNE-2.7.0-3106-Linux-x86_64

MATLAB software not available

/usr/local/MNE-2.7.0-3106-Linux-x86_64/bin added to PATH

/usr/local/MNE-2.7.0-3106-Linux-x86_64/lib added to LD_LIBRARY_PATH

/usr/local/MNE-2.7.0-3106-Linux-x86_64/share/app-defaults/%N added to XUSERFILESEARCHPATH

Note: Remember to set SUBJECTS_DIR and SUBJECT environment variables correctly.

Note: FreeSurfer environment is needed to access tkmedit from mne_analyze.

[centos@localhost ~/Desktop]$ setenv SUBJECT a

[centos@localhost ~/Desktop]$ cd $SUBJECT

[centos@localhost a]$ $MNE_ROOT/mne_preparebv.csh

これをEnterすると、以下のような表示が出て・・・結構計算に時間がかかります。

MNE software location set to: /usr/local/MNE-2.7.0-3106-Linux-x86_64

MATLAB software not available

/usr/local/MNE-2.7.0-3106-Linux-x86_64/bin already in PATH

/usr/local/MNE-2.7.0-3106-Linux-x86_64/lib added to LD_LIBRARY_PATH

/usr/local/MNE-2.7.0-3106-Linux-x86_64/share/app-defaults/%N added to XUSERFILESEARCHPATH

Note: Remember to set SUBJECTS_DIR and SUBJECT environment variables correctly.

Note: FreeSurfer environment is needed to access tkmedit from mne_analyze.

mri_convert.bin T1.nii.gz T1.mgz

$Id: mri_convert.c,v 1.226 2016/02/26 16:15:24 mreuter Exp $

reading from T1.nii.gz...

WARNING: MRIsetRas2VoxFromMatrix(): voxels sizes are inconsistent

(0.9375,0.9375) (2,0.9375) (0.9375,2)

This is probably due to shear in the vox2ras matrix

Input Vox2RAS ------

-0.93750 -0.00000 -0.00000 120.00000;

-0.00000 -0.00000 2.00000 -84.00000;

-0.00000 -0.93750 -0.00000 120.00000;

0.00000 0.00000 0.00000 0.00000;

TR=1000.00, TE=0.00, TI=0.00, flip angle=0.00

i_ras = (-1, -0, -0)

j_ras = (-0, -0, -1)

k_ras = (-0, 1, -0)

writing to T1.mgz...

mri_convert.bin brain.nii.gz brain.mgz

$Id: mri_convert.c,v 1.226 2016/02/26 16:15:24 mreuter Exp $

reading from brain.nii.gz...

WARNING: MRIsetRas2VoxFromMatrix(): voxels sizes are inconsistent

(0.9375,0.9375) (2,0.9375) (0.9375,2)

This is probably due to shear in the vox2ras matrix

Input Vox2RAS ------

-0.93750 -0.00000 -0.00000 120.00000;

-0.00000 -0.00000 2.00000 -84.00000;

-0.00000 -0.93750 -0.00000 120.00000;

0.00000 0.00000 0.00000 0.00000;

TR=1000.00, TE=0.00, TI=0.00, flip angle=0.00

i_ras = (-1, -0, -0)

j_ras = (-0, -0, -1)

k_ras = (-0, 1, -0)

writing to brain.mgz...

-----------------------------------------------------------------------------------------------

Setting up T1...

Creating /home/centos/Desktop/a/mri/T1-neuromag/sets/COR.fif...

mne_make_cor_set version 1.5 compiled at Dec 21 2009 19:47:29

mgh/mgz input file : /home/centos/Desktop/a/mri/T1.mgz

output file : COR.fif

Include data in the output file

Reading /home/centos/Desktop/a/mri/T1.mgz...[done]

Adding Talairach transforms...[failed]

Could not load Talairach transforms : Transform file not present in these MRI data

Talairach transforms will not be included into the output file

All 256 slices composed.

Wrote COR.fif

Created /home/centos/Desktop/a/mri/T1-neuromag/sets/COR.fif

-----------------------------------------------------------------------------------------------

Setting up brain...

Creating /home/centos/Desktop/a/mri/brain-neuromag/sets/COR.fif...

mne_make_cor_set version 1.5 compiled at Dec 21 2009 19:47:29

mgh/mgz input file : /home/centos/Desktop/a/mri/brain.mgz

output file : COR.fif

Include data in the output file

Reading /home/centos/Desktop/a/mri/brain.mgz...[done]

Adding Talairach transforms...[failed]

Could not load Talairach transforms : Transform file not present in these MRI data

Talairach transforms will not be included into the output file

All 256 slices composed.

Wrote COR.fif

Created /home/centos/Desktop/a/mri/brain-neuromag/sets/COR.fif

-----------------------------------------------------------------------------------------------

Complete.

Saving lh.white as a surface

Saving rh.white as a surface

Setting up the source space with the following parameters:

SUBJECTS_DIR = /home/centos/Desktop

Subject = a

Surface = white

Grid spacing = 7 mm

>>> 1. Creating the source space file /home/centos/Desktop/a/bem/a-7-src.fif...

mne_make_source_space version 2.5 compiled at Dec 21 2009 19:46:50

Loading /home/centos/Desktop/a/surf/lh.white...

Triangle file : created by centos on Mon Sep 3 21:30:12 2018 nvert = 70458 ntri = 140912

Triangle and vertex normals and neighboring triangles...[done]

Vertex neighbors...[done]

Distances between neighboring vertices...[422736 distances done]

Decimating...[done]

loaded lh.white 2430/70458 selected to source space (approx. spacing = 7 mm)

Loading /home/centos/Desktop/a/surf/rh.white...

Triangle file : created by centos on Mon Sep 3 21:30:13 2018 nvert = 69911 ntri = 139818

Triangle and vertex normals and neighboring triangles...[done]

Vertex neighbors...[done]

Distances between neighboring vertices...[419454 distances done]

Decimating...[done]

loaded rh.white 2384/69911 selected to source space (approx. spacing = 7 mm)

Wrote /home/centos/Desktop/a/bem/a-7-src.fif

You are now one step closer to computing the gain matrix.

>>> 2. Creating ASCII versions for checking with MRI lab...

Reading from /home/centos/Desktop/a/bem/a-7-src.fif...

Read 2 source spaces from /home/centos/Desktop/a/bem/a-7-src.fif with a total of 4814 source locations

Selection triangulation information is not available.

Source spaces are now in MRI (surface RAS) coordinates.

Wrote /home/centos/Desktop/a/bem/a-7-lh.dip

Wrote /home/centos/Desktop/a/bem/a-7-rh.dip

Wrote /home/centos/Desktop/a/bem/a-7-lh.pnt

Wrote /home/centos/Desktop/a/bem/a-7-rh.pnt

Wrote 2430 vertices into /home/centos/Desktop/a/bem/a-7-lh.w

Wrote 2384 vertices into /home/centos/Desktop/a/bem/a-7-rh.w

Temporary files removed.

Finished.

Running mri_watershed for BEM segmentation with the following parameters

SUBJECTS_DIR = /home/centos/Desktop

Subject = a

Result dir = /home/centos/Desktop/a/bem/watershed

Temporary files removed.

Mode: Atlas analysis

Mode: use surfaceRAS to save surface vertex positions

Mode: Saving out BEM surfaces

*********************************************************

The input file is /home/centos/Desktop/a/mri/T1.mgz

The output file is /home/centos/Desktop/a/bem/watershed/ws

conforming input...

MRIchangeType: Building histogram

*************************WATERSHED**************************

Sorting...

first estimation of the COG coord: x=127 y=134 z=118 r=75

first estimation of the main basin volume: 1812151 voxels

global maximum in x=97, y=96, z=135, Imax=255

CSF=23, WM_intensity=169, WM_VARIANCE=4

WM_MIN=158, WM_HALF_MIN=164, WM_HALF_MAX=173, WM_MAX=178

preflooding height equal to 25 percent

done.

Analyze...

main basin size= 1769363 voxels, voxel volume =1.000

= 1769363 mmm3 = 1769.363 cm3

done.

PostAnalyze...

***** 0 basin(s) merged in 1 iteration(s)

***** 0 voxel(s) added to the main basin

done.

****************TEMPLATE DEFORMATION****************

second estimation of the COG coord: x=127,y=120, z=113, r=9749 iterations

^^^^^^^^ couldn't find WM with original limits - expanding ^^^^^^

GLOBAL CSF_MIN=0, CSF_intensity=10, CSF_MAX=86 , nb = 39914

CSF_MAX TRANSITION GM_MIN GM

GLOBAL

before analyzing : 86, 90, 93, 135

after analyzing : 86, 92, 93, 102

mri_strip_skull: done peeling brain

highly tesselated surface with 10242 vertices

matching...69 iterations

*********************VALIDATION*********************

curvature mean = -0.013, std = 0.011

curvature mean = 71.011, std = 7.652

Rigid alignment...

scanning 32.00 degree nbhd, min sse = 1.95 at ( 0.00, 0.00, 0.00)

scanning 16.00 degree nbhd, min sse = 1.88 at ( 0.00, 0.00, 4.00)

scanning 8.00 degree nbhd, min sse = 1.80 at ( 0.00, 0.00, -2.00)

scanning 4.00 degree nbhd, min sse = 1.79 at (-1.00, -1.00, 0.00)

done

before rotation: sse = 1.95, sigma = 2.73

after rotation: sse = 1.79, sigma = 2.57

Localization of inacurate regions: Erosion-Dilation steps

the sse mean is 1.86, its var is 2.29

before Erosion-Dilatation 0.10% of inacurate vertices

after Erosion-Dilatation 0.00% of inacurate vertices

Validation of the shape of the surface done.

Scaling of atlas fields onto current surface fields

********FINAL ITERATIVE TEMPLATE DEFORMATION********

Compute Local values csf/gray

Fine Segmentation...

この辺で止まって・・・

Fine Segmentation...80 iterations

mri_strip_skull: done peeling brain

Brain Size = 1812930 voxels, voxel volume = 1.000 mm3

= 1812930 mmm3 = 1812.930 cm3

outer skin surface matching...139 iterationsINFO: MRImask() using MRImaskDifferentGeometry()

******************************

Saving /home/centos/Desktop/a/bem/watershed/ws

non-standard value for type (4, usually 0) in volume structure

non-standard value for height (86, usually 256) in volume structure

non-standard value for thick (0.9375, usually 1) in volume structure

non-standard value for ps (0.9375, usually 1) in volume structure

done

mri_watershed utimesec 729.152152

mri_watershed stimesec 0.745886

mri_watershed ru_maxrss 380088

mri_watershed ru_ixrss 0

mri_watershed ru_idrss 0

mri_watershed ru_isrss 0

mri_watershed ru_minflt 98815

mri_watershed ru_majflt 41

mri_watershed ru_nswap 0

mri_watershed ru_inblock 16736

mri_watershed ru_oublock 15208

mri_watershed ru_msgsnd 0

mri_watershed ru_msgrcv 0

mri_watershed ru_nsignals 0

mri_watershed ru_nvcsw 1320

mri_watershed ru_nivcsw 6828

mri_watershed done

mne_surf2bem version 1.6 compiled at Dec 21 2009 19:47:09

input file # 1 : /home/centos/Desktop/a/bem/watershed/a_outer_skin_surface / id = 4 / sigma N/A

output file : a-head.fif

Triangle file : created by centos on Mon Sep 3 21:47:21 2018 nvert = 10242 ntri = 20480

Triangle and vertex normals and neighboring triangles...[done]

Vertex neighbors...[done]

Distances between neighboring vertices...[61440 distances done]

/home/centos/Desktop/a/bem/watershed/a_outer_skin_surface read. id = 4

Topology checks skipped.

a-head.fif written.

Created /home/centos/Desktop/a/bem/a-head.fif

Complete.

mne_convert_surface version 1.10 compiled at Dec 21 2009 19:46:42

surf input file : a_inner_skull_surface

ASCII tri output file : ../inner_skull.tri

Triangle file : created by centos on Mon Sep 3 21:37:58 2018 nvert = 10242 ntri = 20480

Read a_inner_skull_surface (10242 vertices 20480 triangles)

Wrote ../inner_skull.tri (10242 vertices 20480 triangles)

Used counterclockwise vertex ordering.

Coordinates were written in millimeters

Setting up the BEM with the following parameters:

SUBJECTS_DIR = /home/centos/Desktop

Subject = a

Inner skull = /home/centos/Desktop/a/bem/inner_skull.tri (20480 triangles)

brain conductivity = 0.3 S/m

Resulting BEM = /home/centos/Desktop/a/bem/a-20480-bem.fif

>> 1. Creating the BEM geometry file...

mne_surf2bem version 1.6 compiled at Dec 21 2009 19:47:09

input file # 1 : /home/centos/Desktop/a/bem/inner_skull.tri / id = 1 / sigma = 0.3 S/m

output file : /home/centos/Desktop/a/bem/a-20480-bem.fif

Loaded a surface from /home/centos/Desktop/a/bem/inner_skull.tri with 10242 nodes and 20480 triangles.

Node normals were not included in the source file.

Triangle and vertex normals and neighboring triangles...[done]

Vertex neighbors...[done]

Distances between neighboring vertices...[61440 distances done]

/home/centos/Desktop/a/bem/inner_skull.tri read. id = 1

inner skull CM is 0.50 -14.99 7.61 mm

Surfaces passed the basic topology checks.

/home/centos/Desktop/a/bem/a-20480-bem.fif written.

>> 2. Creating ascii pnt files for MRIlab...

mne_list_bem version 1.9 compiled at Dec 21 2009 19:47:02

input file : /home/centos/Desktop/a/bem/a-20480-bem.fif

output file : /home/centos/Desktop/a/bem/a-inner_skull-20480.pnt

surface id : 1

output in millimeters.

output the vertex coordinates only.

done.

mne_list_bem version 1.9 compiled at Dec 21 2009 19:47:02

input file : /home/centos/Desktop/a/bem/a-20480-bem.fif

output file : /home/centos/Desktop/a/bem/a-inner_skull-20480.surf

surface id : 1

output in FreeSurfer format.

done.

>> 3. Calculating BEM geometry data (this takes several minutes)...

mne_prepare_bem_model version 1.2 compiled at Dec 21 2009 19:46:09

BEM surface file : /home/centos/Desktop/a/bem/a-20480-bem.fif

Solution file : /home/centos/Desktop/a/bem/a-20480-bem-sol.fif

Approximation method : linear collocation

Loading surfaces...

Triangle normals and neighboring triangles...[done]

Vertex neighbors...[done]

Distances between neighboring vertices...[61440 distances done]

Homogeneous model surface loaded.

Computing the linear collocation solution...

Matrix coefficients...

inner skull (10242) -> inner skull (10242) ...

何回か止まりながら処理が進んでいきます。

Computing the linear collocation solution...

Matrix coefficients...

inner skull (10242) -> inner skull (10242) ... [done]

Inverting the coefficient matrix...

LU factorization...

Compute inverse...

この辺り遅いです。なかなか進みませんが、あと一息です。

Computing the linear collocation solution...

Matrix coefficients...

inner skull (10242) -> inner skull (10242) ... [done]

Inverting the coefficient matrix...

LU factorization...

Compute inverse...

Solution ready.

Saving...

Saved the result to /home/centos/Desktop/a/bem/a-20480-bem-sol.fif

BEM geometry computations complete.

The model /home/centos/Desktop/a/bem/a-20480-bem-sol.fif is now ready for use

Complete.

[centos@localhost a]$

計算終了です。bemフォルダが作成されます。

bemフォルダの中身です。

mriフォルダの中身です。

surfフォルダの中身です。

あとは通常の解析手順となります。

脳磁図データファイルとその共分散ファイルのコピー

共分散ファイルは通常mne_browse_rawで作成しますが、加算波形の場合は横着してhns_megで作成することも可能です。

元波形にかける雑音処理や周波数フィルタ別に共分散行列を計算する必要があります。

mne_analyzeで座標合わせ

Save defaultを押すと頭座標とMRI座標間の変換行列ファイル*-trans.fifが作成されます。

Save MRI setを押すと~/mri/T1-neuromag/sets/COR-xx-日付-時間.fifファイルが作成されます。

電流源推定

あと２つMNE-Cのコマンドを実行するだけです。

1つ目。順問題。ちょっと時間がかかります。

[centos@localhost a]$ mne_do_forward_solution --meas SEF.fif --megonly

mne_forward_solution version 2.9 compiled at Dec 21 2009 19:47:25

Source space : /home/centos/Desktop/a/bem/a-7-src.fif

MRI -> head transform source : ./SEF-trans.fif

Measurement data : SEF.fif

BEM model : /home/centos/Desktop/a/bem/a-20480-bem.fif

Accurate field computations

Do computations in head coordinates.

Free source orientations

Destination for the solution : ./SEF-7-fwd.fif

Reading /home/centos/Desktop/a/bem/a-7-src.fif...

Read 2 source spaces a total of 4814 active source locations

Coordinate transformation: MRI (surface RAS) -> head

0.998619 0.018430 -0.049193 -4.10 mm

-0.011331 0.989965 0.140856 20.99 mm

0.051296 -0.140104 0.988807 36.49 mm

0.000000 0.000000 0.000000 1.00

Read 306 MEG channels from SEF.fif

Coordinate transformation: MEG device -> head

0.997806 -0.066204 -0.000753 -6.05 mm

0.066190 0.997208 0.034567 -4.23 mm

-0.001538 -0.034541 0.999402 42.28 mm

0.000000 0.000000 0.000000 1.00

EEG not requested. EEG channels omitted.

57 coil definitions read

Head coordinate coil definitions created.

Source spaces are now in head coordinates.

Setting up the BEM model using /home/centos/Desktop/a/bem/a-20480-bem-sol.fif...

Loading surfaces...

Triangle normals and neighboring triangles...[done]

Vertex neighbors...[done]

Distances between neighboring vertices...[61440 distances done]

Homogeneous model surface loaded.

Loading the solution matrix...

Loaded linear collocation BEM solution from /home/centos/Desktop/a/bem/a-20480-bem-sol.fif

Employing the head->MRI coordinate transform with the BEM model.

BEM model /home/centos/Desktop/a/bem/a-20480-bem-sol.fif is now set up

Source spaces are in head coordinates.

Checking that the sources are inside the inner skull (will take a few...)

Thank you for waiting.

Setting up compensation data...

No compensation set. Nothing more to do.

Composing the field computation matrix...[done]

2 processors. I will use one thread for each of the 2 source spaces.

Computing MEG at 4814 source locations (free orientations)...done.

writing ./SEF-7-fwd.fif...done

Finished.

[centos@localhost a]$

*-7-fwd.fifというファイルが作成されます。

あとコマンド１つです。

[centos@localhost a]$ mne_do_inverse_operator --fwd SEF-7-fwd.fif

mne_inverse_operator version 2.27 compiled at Dec 21 2009 19:47:03

Compute the MNE inverse operator decomposition

Forward solution : SEF-7-fwd.fif

Sensor noise covariance matrix : ./SEF-cov.fif

Source covariance matrix : identity matrix

Destination for the inverse operator data : ./SEF-7-meg-inv.fif

Include MEG data.

Reading the forward solution....

Read data for 306 MEG channels and 4814 sources

Free source orientations.

The forward computation was performed in head coordinates.

Read 2 source spaces with a total of 4814 source locations

Source spaces are now in head coordinates.

Channel description list matched with the composite forward solution matrix.

Global Cartesian head coordinate system forward matrix will be employed.

No linear projection information in ./SEF-cov.fif.

Projection will not have any effect on selected channels. Projection omitted.

Read a full noise covariance matrix from ./SEF-cov.fif

Picked appropriate channels from the sensor noise covariance matrix.

MEG/EEG correlations omitted.

No regularization applied to the noise-covariance matrix

Decomposing the noise covariance...

Eigenvalue decomposition done.

Verdict: 153 small eigenvalues

done.

Creating the source covariance matrix:

done

Whitening the forward solution...done

Scaling the source covariance...done

Decomposing...

Applying a priori source weighting to the forward solution...done

Transpose...done

SVD...done

largest singular value = 0.591894

scaling factor to adjust the trace = 17.4929

done

Writing the solution to ./SEF-7-meg-inv.fif...done

Attaching the environment to ./SEF-7-meg-inv.fif...done

Inverse operator file ./SEF-7-meg-inv.fif ready. Thank you for waiting.

[centos@localhost a]$

*-7-meg-inv.fifというファイル名が作成されています。

mne_analyzeで結果表示

右正中神経電気刺激で左体性感覚誘発磁場の結果です。脳磁図データ読み込み後load surfaceで皮髄境界whiteを選択します。いい感じで推定されています。

皮髄境界whiteから皮質pialに変えてみます。

mris_convertコマンドを使ってLhemi.giiとRhemi.giiはlh.pialとrh.pial、Lwhite_inflated.giiとRwhite_inflated.giiはlh.inflatedとrh.inflatedに変換しておきます。

pialを選択すると・・・ポリゴンメッシュの頂点数、メッシュ数が異なるため表示できません。

膨張皮質inflatedを選択します。

inflatedはLwhite.giiとRwhite.giiを膨らましてLwhite_inflated.giiとRwhite_inflated.giiで保存してlh.inflatedとrh.inflatedにしたものです。visa_converts.m処理後は_whiteが消えてLinflated.giiとRinflated.giiとなっています。頂点数、三角メッシュ数は一致しるのでちゃんと表示されます・・・が、残念ながら、どこが脳回表面でどこが脳溝面か白・灰色表示には対応していないようです。