Diagnose tensor orientation problems - Revision history

Jadrian Miles at 15:11, 16 July 2009

2009-07-16T15:11:23Z

← Older revision		Revision as of 15:11, 16 July 2009
Line 1:		Line 1:
	Sometimes when [[Diffusion Processing Pipeline\|processing]] [[diffusion MRI]] data, the DWIs come to us ~~flipped or~~ in a different coordinate system than the b-vectors. This issue manifests itself in orientation problems with the fit tensors, which can be very subtle and hard to detect.		Sometimes when [[Diffusion Processing Pipeline\|processing]] [[diffusion MRI]] data, the DWIs come to us in a different coordinate system than the b vectors. This issue manifests itself in orientation problems with the fit tensors, which can be very subtle and hard to detect.

	There are two types of mis-orientations:		There are two types of mis-orientations:
	* Permutations, where ''e.g.'' X has been exchanged with Y. These are easy to detect using an RGB-mapped visualization of the tensor field, or even by examining the DWIs if you know the b-vectors.		* Permutations, where ''e.g.'' the X dimension has been exchanged with Y. These are easy to detect using an RGB-mapped visualization of the tensor field, or even by examining the DWIs if you know the b vectors.
	* Flips, where the images have been mirrored in ''e.g.'' X. These are easy to detect using a tensor ellipse visualization or by looking at a tractogram, a fair bit harder to spot by looking at a field of the off-diagonal elements of the tensors, and impossible to spot in an RGB visualization of the tensor field due to the mirror symmetry of the colormap. Since the ~~tensors have three-way mirror symmetry (that is,~~ orientation but not direction matters), only non-axis-aligned tensors display any irregularity when flipped in one dimension.		* Flips, where the images have been mirrored in ''e.g.'' X. These are easy to detect using a tensor ellipse visualization or by looking at a tractogram, a fair bit harder to spot by looking at a field of the off-diagonal elements of the tensors, and impossible to spot in an RGB visualization of the tensor field due to the mirror symmetry of the colormap. Since the orientation but not the direction of each tensor matters, only non-axis-aligned tensors display any irregularity when flipped in one dimension.

	Since generating streamtubes takes a long time, it is ideal to detect orientation problems just by looking at the tensors, before generating the tubes. However, the effects of incorrect orientation are far more apparent in the tube visualization than in the tensor map.		Since generating streamtubes takes a long time, it is ideal to detect orientation problems just by looking at the tensors, before generating the tubes. However, the effects of incorrect orientation are far more apparent in the tube visualization than in the tensor map. On the first pass through a new dataset, it's recommended that you [[run the diffusion processing pipeline at low resolution]] so that you don't have to wait so long to detect problems.

	== ~~HOWTO Detect~~ Tensor Orientation Problems ==		== Detecting Tensor Orientation Problems ==

	=== ... Before Generating Tubes ===		=== ... Before Generating Tubes ===
Line 38:		Line 38:
	Email "Brown 3T Cohen Study brain tubes (at last!)" contains screenshots from $G/data/diffusion/brown3t/cohen_hiv_study.2007.02.07/patient024/tubes.2008.07.27, a healthy brain with no orientation problems. I believe $G/data/diffusion/brown3t/cohen_hiv_study_registered.2007.02.07/patient116/tubes.2008.11.06 has an X flip. Should generate screenshots of RGB maps and streamtubes for examples.		Email "Brown 3T Cohen Study brain tubes (at last!)" contains screenshots from $G/data/diffusion/brown3t/cohen_hiv_study.2007.02.07/patient024/tubes.2008.07.27, a healthy brain with no orientation problems. I believe $G/data/diffusion/brown3t/cohen_hiv_study_registered.2007.02.07/patient116/tubes.2008.11.06 has an X flip. Should generate screenshots of RGB maps and streamtubes for examples.
	-->		-->

			== Choosing which Coordinate System to Trust ==
			Since the source of an orientation problem is a disagreement between the coordinate systems of the b-vectors and the DWIs, one must choose which coordinate system to treat as the "true" one when fixing the problem. Due to conflicting "standards" for medical image formats, sometimes the DWIs use a left-handed coordinate system or otherwise store images in a funky way; if this is the case, you probably want to trust the coordinate system of the b vectors. On the other hand, if the DWIs seem to be stored correctly, you probably want to trust their coordinate system.

			You can use the Matlab commands <tt>readMriImage()</tt>, <tt>stackToVideo()</tt>, and <tt>viewSlice()</tt> to examine the source DWIs. The standard coordinate system is right-handed and based on the perspective of a doctor standing upright and facing a standing patient. Positive X goes to the doctor's right (patient's left), positive Y goes to the posterior (back) side of the patient, and positive Y goes to the superior side of the patient. In a Cartesian view of a single X-Y slice, then, the anterior (front) of the patient should be facing downward, with the patient's left side on the right side of the slice; this is the same perspective the doctor would have looking straight down on the top of the patient's head. As Z increases, the slice moves up the body (superior) and therefore closer to the observer.

			If inspection of the DWIs indicates a direction flip of the sort that you observed in the tensors, it makes sense to flip the images. If the DWIs seem correct, it might be better to flip the b vectors.

	== Fixing Problems ==		== Fixing Problems ==
Line 44:		Line 51:
	# Transform the MRIimage DWIs before fitting the tensors		# Transform the MRIimage DWIs before fitting the tensors
	# Transform the tensors in place before computing tensor measures and the tractogram		# Transform the tensors in place before computing tensor measures and the tractogram
	There are pros and cons to each approach. In the short term, when you've discovered an orientation problem and are trying to fix it, the last option has the lowest cost. In the long term, when you already know the orientation problem and want to fix it automatically for each dataset coming from a particular collaborator, the first ~~option has the lowest~~ cost. Therefore it's advised that you transform the tensors in place when ''checking'' a fix, but once you've confirmed the orientation problem and how to fix it, it's advised that you code up a change to the b vectors.		There are pros and cons to each approach. In the short term, when you've discovered an orientation problem and are trying to fix it, the last option has the lowest cost. In the long term, when you already know the orientation problem and want to fix it automatically for each dataset coming from a particular collaborator, the first two options have lower cost. Therefore it's advised that you transform the tensors in place when ''checking'' a fix, but once you've confirmed the orientation problem and how to fix it, it's advised that you code up a change to the b vectors or the DWIs.

	=== ... by Transforming the b Vectors ===		=== ... by Transforming the b Vectors ===
	By transforming the b vectors before converting the original DWIs from the source format to MRIimage, you avoid transforming the images and instead change the coordinate frame for the b vectors to correspond to the coordinate frame for the images. This operation ~~is faster than transforming the images and also avoids making the images look weird when inspected visually. It~~ makes the assumption that the coordinate frame for the images is the "true" one and that there is some error in the coordinate frame for the b vectors. There may be a genuine problem with the images, though (for example, if they use a left-handed coordinate system); in which case transforming the images is the correct thing to do.		By transforming the b vectors before converting the original DWIs from the source format to MRIimage, you avoid transforming the images and instead change the coordinate frame for the b vectors to correspond to the coordinate frame for the images. This operation makes the assumption that the coordinate frame for the images is the "true" one and that there is some error in the coordinate frame for the b vectors. There may be a genuine problem with the images, though (for example, if they use a left-handed coordinate system); in which case transforming the images is the correct thing to do. Transforming the b vectors is also faster than transforming the images.

	<code>/map/gfx0/common0/diffusion/Interface/bin/data/mri_io.py</code> has handy tools for I/O involving bvecs files. Though no command-line tool has yet been created to transform b vectors, here's an example that flips the y component (in Python):		<code>/map/gfx0/common0/diffusion/Interface/bin/data/mri_io.py</code> has handy tools for I/O involving bvecs files. Though no command-line tool has yet been created to transform b vectors, here's an example that flips the y component (in Python):
Line 64:		Line 71:

	=== ... by Transforming the DWIs ===		=== ... by Transforming the DWIs ===
	As discussed above, transforming the DWIs is more computationally expensive than transforming the b vectors~~, and may also make the images look weird when visually inspected~~. Due to conflicting "standards" for medical imaging coordinate systems, though, sometimes it actually does make sense to transform the DWIs instead. Transforming the DWIs makes the assumption that the coordinate frame for the b vectors is the "true" one, and that there is some error in the coordinate frame for the DWIs.		As discussed above, transforming the DWIs is more computationally expensive than transforming the b vectors. Due to conflicting "standards" for medical imaging coordinate systems, though, sometimes it actually does make sense to transform the DWIs instead. Transforming the DWIs makes the assumption that the coordinate frame for the b vectors is the "true" one, and that there is some error in the coordinate frame for the DWIs.

	<code>mritransp</code> can perform dimension permutations and flips on DWIs. Here's an example that flips the y dimension in a collection of DWIs (in tcsh):		<code>mritransp</code> can perform dimension permutations and flips on DWIs. Here's an example that flips the y dimension in a collection of DWIs (in tcsh):
Line 81:		Line 88:
	<pre>		<pre>
	cd voxel_models; mv tensors original_tensors		cd voxel_models; mv tensors original_tensors
	mritensormult -i voxel_models/original_tensors -o voxel_models/tensors -mat -1 0 0 0 1 0 0 0 1		mritensormult -i voxel_models/original_tensors -o voxel_models/tensors -mat 1 0 0 0 -1 0 0 0 1
	</pre>		</pre>

Jadrian Miles: /* Fixing Problems */

2009-07-15T20:38:08Z

Fixing Problems

← Older revision		Revision as of 20:38, 15 July 2009
Line 47:		Line 47:

	=== ... by Transforming the b Vectors ===		=== ... by Transforming the b Vectors ===
			By transforming the b vectors before converting the original DWIs from the source format to MRIimage, you avoid transforming the images and instead change the coordinate frame for the b vectors to correspond to the coordinate frame for the images. This operation is faster than transforming the images and also avoids making the images look weird when inspected visually. It makes the assumption that the coordinate frame for the images is the "true" one and that there is some error in the coordinate frame for the b vectors. There may be a genuine problem with the images, though (for example, if they use a left-handed coordinate system); in which case transforming the images is the correct thing to do.

	<code>/map/gfx0/common0/diffusion/Interface/bin/data/mri_io.py</code> has handy tools for I/O involving bvecs files. Though no command-line tool has yet been created to transform b vectors, here's an example that flips the y component (in Python):		<code>/map/gfx0/common0/diffusion/Interface/bin/data/mri_io.py</code> has handy tools for I/O involving bvecs files. Though no command-line tool has yet been created to transform b vectors, here's an example that flips the y component (in Python):
	<pre>		<pre>
Line 62:		Line 64:

	=== ... by Transforming the DWIs ===		=== ... by Transforming the DWIs ===
			As discussed above, transforming the DWIs is more computationally expensive than transforming the b vectors, and may also make the images look weird when visually inspected. Due to conflicting "standards" for medical imaging coordinate systems, though, sometimes it actually does make sense to transform the DWIs instead. Transforming the DWIs makes the assumption that the coordinate frame for the b vectors is the "true" one, and that there is some error in the coordinate frame for the DWIs.

	<code>mritransp</code> can perform dimension permutations and flips on DWIs. Here's an example that flips the y dimension in a collection of DWIs (in tcsh):		<code>mritransp</code> can perform dimension permutations and flips on DWIs. Here's an example that flips the y dimension in a collection of DWIs (in tcsh):
	<pre>		<pre>
Line 72:		Line 76:

	=== ... by Transforming the Tensors ===		=== ... by Transforming the Tensors ===
			Transforming the tensors in a DTI is equivalent to transforming the b vectors, and so makes the same assumption that the coordinate frame of the DWIs is the "true" one and that there was some error with the coordinate frame of the b vectors. It has the downside of being more computationally expensive than either of the previous techniques, though if you've already computed the tensors, it is more immediately expedient.

	<code>mritensormult</code> can perform general transformations on all the tensors in a DTI, leaving the tensors in the same positions but changing their values in place. Here's an example that flips the y dimension in a DTI:		<code>mritensormult</code> can perform general transformations on all the tensors in a DTI, leaving the tensors in the same positions but changing their values in place. Here's an example that flips the y dimension in a DTI:
	<pre>		<pre>

Jadrian Miles: /* Fixing Problems */

2009-07-15T20:24:42Z

Fixing Problems

← Older revision		Revision as of 20:24, 15 July 2009
Line 40:		Line 40:

	== Fixing Problems ==		== Fixing Problems ==
	There are ~~two~~ ways to correct tensor orientation problems: ~~you can either transform each tensor~~ in place, or you ~~can transform all~~ the ~~DWIs~~ and ~~then re-fit~~ the ~~tensors~~. '''~~Transforming the DWIs~~''' ~~is cheaper than~~ '''~~transforming the tensors~~''', ~~which is cheaper than~~ '''~~fitting the tensors~~'''~~. Therefore there are pros and cons to each approach.~~		There are three ways to correct tensor orientation problems, each of which comes at a different point in the pipeline:
			# Transform the b vectors before converting the DWIs to MRIimage
			# Transform the MRIimage DWIs before fitting the tensors
			# Transform the tensors in place before computing tensor measures and the tractogram
			There are pros and cons to each approach. In the short term, when you've discovered an orientation problem and are trying to fix it, the last option has the lowest cost. In the long term, when you already know the orientation problem and want to fix it automatically for each dataset coming from a particular collaborator, the first option has the lowest cost. Therefore it's advised that you transform the tensors in place when ''checking'' a fix, but once you've confirmed the orientation problem and how to fix it, it's advised that you code up a change to the b vectors.

			=== ... by Transforming the b Vectors ===
			<code>/map/gfx0/common0/diffusion/Interface/bin/data/mri_io.py</code> has handy tools for I/O involving bvecs files. Though no command-line tool has yet been created to transform b vectors, here's an example that flips the y component (in Python):
			<pre>
			from mri_io import *
			bvecfname = findUpward('.', 'bvecs')
			fid = open(bvecfname, 'r')
			vecs = parseBvecs(fid.read())
			fid.close()
			for i in range(len(vecs)):
			vecs[i][1] = -1.0 * vecs[i][1]
			fid = open('bvecs.yflipped', 'w')
			writeBvecs(vecs, fid)
			fclose(fid)
			</pre>

	=== ... by Transforming the DWIs ===		=== ... by Transforming the DWIs ===
	<code>~~$G/bin/~~mritransp</code> can perform dimension permutations and flips on DWIs.		<code>mritransp</code> can perform dimension permutations and flips on DWIs. Here's an example that flips the y dimension in a collection of DWIs (in tcsh):
	* '''Long-Term Advantage:''' the ~~next time you get~~ a ~~dataset from the same protocol, you can fix the orientation problem cheaply at the DWI stage and then the tensors that you fit to the~~ DWIs ~~will be correct.~~		<pre>
	*; {{infobox\|Total processing (~~long term~~) <~~nowiki~~>~~=</nowiki> transform DWIs + fit tensors}}~~		mkdir -p dwi/flipped
	* '''Immediate Disadvantage:''' when you discover the orientation problem for a single dataset, you have already fit the tensors, so fixing it by modifying the DWIs means you will have to re-~~fit the tensors, which takes time.~~		foreach n ( `/bin/ls dwi/original` )
	*; {{infobox\|Total processing (~~short term~~) ~~<nowiki>=~~</~~nowiki~~> ~~transform DWIs + fit tensors}}~~		mkdir dwi/flipped/$n
			mritransp -i dwi/original/$n -o dwi/flipped/$n -flipy
			end
			</pre>

	=== ... by Transforming the Tensors ===		=== ... by Transforming the Tensors ===
	<code>~~$G/bin/~~mritensormult</code> can perform general transformations on all the tensors in a DTI, leaving the tensors in the same positions but changing their values in place.		<code>mritensormult</code> can perform general transformations on all the tensors in a DTI, leaving the tensors in the same positions but changing their values in place. Here's an example that flips the y dimension in a DTI:
	* '''Long-Term Disadvantage:''' the ~~next time you get~~ a ~~dataset from the same protocol, the same orientation problem will probably be present, and so you will have to correct it, and transforming the tensors is expensive.~~		<pre>
	*; {{infobox\|Total processing (long term) <~~nowiki~~>~~=</nowiki> fit~~ tensors ~~+ transform tensors}}~~		cd voxel_models; mv tensors original_tensors
	* '''Immediate Advantage:''' when you discover the orientation problem for a single dataset, you don't have to re-~~fit the~~ tensors~~, which means that you can quickly move on to the next step in the [[Diffusion Processing Pipeline\|pipeline]] without waiting.~~		mritensormult -i voxel_models/original_tensors -o voxel_models/tensors -mat -1 0 0 0 1 0 0 0 1
	*; {{infobox\|Total processing (short term) <nowiki>=</~~nowiki~~> ~~transform tensors}}~~		</pre>

	== Automatic Diagnosis ==		== Automatic Diagnosis ==

Jadrian Miles: /* ... After Generating Tubes */

2009-07-15T17:28:07Z

... After Generating Tubes

← Older revision		Revision as of 17:28, 15 July 2009
Line 30:		Line 30:
	## Look at a front view of the tubes through the callosum and make sure they complete a U the goes all the way to the superior cortex. Since the CC is almost always axis-aligned, it rarely visibly breaks.		## Look at a front view of the tubes through the callosum and make sure they complete a U the goes all the way to the superior cortex. Since the CC is almost always axis-aligned, it rarely visibly breaks.
	##* If they're broken, you have a flip in either the superior-inferior direction or the left-right direction.		##* If they're broken, you have a flip in either the superior-inferior direction or the left-right direction.
	## Look at a top view of tubes and make sure that forceps major and minor complete a pair of U shapes that go anterior and posterior from the callosum.		## Look at a top view of tubes and make sure that [[Brain anatomy#Commissural Tracts\|forceps major and minor]] complete a pair of U shapes that go anterior and posterior from the callosum.
	##* If they're broken, you have a flip in either the anterior-posterior direction or the left-right direction.		##* If they're broken, you have a flip in either the anterior-posterior direction or the left-right direction.
			[[Image:A-P flip.png\|thumb\|500px\|center\|A tractogram with an anterior-posterior flip. Note the absence of the forceps major and minor.]]
	<!--		<!--

Jadrian Miles: /* ... After Generating Tubes */

2009-03-17T15:01:37Z

... After Generating Tubes

← Older revision		Revision as of 15:01, 17 March 2009
Line 25:		Line 25:

	=== ... After Generating Tubes ===		=== ... After Generating Tubes ===
	# Generate tubes with <~~code~~>~~$G/bin/~~tubegen</~~code~~> (check a [[diffusion processing pipeline]] Makefile to see how).		# Generate tubes with <tt>[[tubegen]]</tt> (check a [[diffusion processing pipeline]] Makefile to see how).
	# View them with [[Brainapp]].		# View them with [[Brainapp]].
	# Check for orientation problems:		# Check for orientation problems:

Jadrian Miles: /* ... Before Generating Tubes */

2009-03-17T14:53:13Z

... Before Generating Tubes

← Older revision		Revision as of 14:53, 17 March 2009
Line 22:		Line 22:
	##* Again, an incorrect color indicates a permutation.		##* Again, an incorrect color indicates a permutation.

	Lots of other groups seem to use planes of diffusion ellipses to diagnose orientation problems, but we don't currently have a way to generate those from tensors in [[MRIimage]] format. If we did, we could detect dimension flips before ~~fitting tubes~~.		Lots of other groups seem to use planes of diffusion ellipses to diagnose orientation problems, but we don't currently have a way to generate those from tensors in [[MRIimage]] format. If we did, we could detect dimension flips before generating streamtubes.

	=== ... After Generating Tubes ===		=== ... After Generating Tubes ===

Jadrian Miles: /* Fixing Problems */

2009-03-17T14:50:02Z

Fixing Problems

← Older revision		Revision as of 14:50, 17 March 2009
Line 43:		Line 43:
	=== ... by Transforming the DWIs ===		=== ... by Transforming the DWIs ===
	<code>$G/bin/mritransp</code> can perform dimension permutations and flips on DWIs.		<code>$G/bin/mritransp</code> can perform dimension permutations and flips on DWIs.
	* '''Long-Term Advantage:''' the next time you get a dataset from the same protocol, you can fix the orientation problem cheaply at the DWI stage and then the tensors that you fit to the DWIs will be correct. Total processing~~: '''~~transform DWIs~~'''~~ + ~~'''~~fit tensors~~'''.~~		* '''Long-Term Advantage:''' the next time you get a dataset from the same protocol, you can fix the orientation problem cheaply at the DWI stage and then the tensors that you fit to the DWIs will be correct.
	* '''Immediate Disadvantage:''' when you discover the orientation problem for a single dataset, you have already fit the tensors, so fixing it by modifying the DWIs means you will have to re-fit the tensors, which takes time. Total processing~~: '''~~transform DWIs~~'''~~ + ~~'''~~fit tensors~~'''.~~		*; {{infobox\|Total processing (long term) <nowiki>=</nowiki> transform DWIs + fit tensors}}
			* '''Immediate Disadvantage:''' when you discover the orientation problem for a single dataset, you have already fit the tensors, so fixing it by modifying the DWIs means you will have to re-fit the tensors, which takes time.
			*; {{infobox\|Total processing (short term) <nowiki>=</nowiki> transform DWIs + fit tensors}}

	=== ... by Transforming the Tensors ===		=== ... by Transforming the Tensors ===
	<code>$G/bin/mritensormult</code> can perform general transformations on all the tensors in a DTI, leaving the tensors in the same positions but changing their values in place.		<code>$G/bin/mritensormult</code> can perform general transformations on all the tensors in a DTI, leaving the tensors in the same positions but changing their values in place.
	* '''Immediate Advantage:''' when you discover the orientation problem for a single dataset, you don't have to re-fit the tensors, which means that you can quickly move on to the next step in the [[Diffusion Processing Pipeline\|pipeline]] without waiting~~. Total processing: '''transform tensors'''~~.		* '''Long-Term Disadvantage:''' the next time you get a dataset from the same protocol, the same orientation problem will probably be present, and so you will have to correct it, and transforming the tensors is expensive.
	* '''Long-Term Disadvantage:''' the next time you get a dataset from the same protocol, the same orientation problem will probably be present, and so you will have to correct it, and transforming the tensors is expensive. Total processing~~: '''fit tensors''' + '''~~transform tensors~~'''.~~		*; {{infobox\|Total processing (long term) <nowiki>=</nowiki> fit tensors + transform tensors}}
			* '''Immediate Advantage:''' when you discover the orientation problem for a single dataset, you don't have to re-fit the tensors, which means that you can quickly move on to the next step in the [[Diffusion Processing Pipeline\|pipeline]] without waiting.
			*; {{infobox\|Total processing (short term) <nowiki>=</nowiki> transform tensors}}

	== Automatic Diagnosis ==		== Automatic Diagnosis ==

Jadrian Miles at 14:45, 17 March 2009

2009-03-17T14:45:10Z

← Older revision		Revision as of 14:45, 17 March 2009
Line 5:		Line 5:
	* Flips, where the images have been mirrored in ''e.g.'' X. These are easy to detect using a tensor ellipse visualization or by looking at a tractogram, a fair bit harder to spot by looking at a field of the off-diagonal elements of the tensors, and impossible to spot in an RGB visualization of the tensor field due to the mirror symmetry of the colormap. Since the tensors have three-way mirror symmetry (that is, orientation but not direction matters), only non-axis-aligned tensors display any irregularity when flipped in one dimension.		* Flips, where the images have been mirrored in ''e.g.'' X. These are easy to detect using a tensor ellipse visualization or by looking at a tractogram, a fair bit harder to spot by looking at a field of the off-diagonal elements of the tensors, and impossible to spot in an RGB visualization of the tensor field due to the mirror symmetry of the colormap. Since the tensors have three-way mirror symmetry (that is, orientation but not direction matters), only non-axis-aligned tensors display any irregularity when flipped in one dimension.

	~~Either type can be corrected by transforming each individual tensor in the tensor field~~, ~~without changing its position. <code>$G/bin/mritensormult</code> can perform general transformations of tensors.~~		Since generating streamtubes takes a long time, it is ideal to detect orientation problems just by looking at the tensors, before generating the tubes. However, the effects of incorrect orientation are far more apparent in the tube visualization than in the tensor map.

	~~It's cheapest~~ to ~~correct~~ orientation problems by ~~transforming~~ the ~~DWIs~~ before ~~you even fit tensors~~. ~~In this case, you actually transform the entire image; that is~~, ~~you change~~ the ~~positions~~ of ~~the various voxels. Various special-purpose tools for correcting~~ orientation ~~problems at~~ the ~~DWI stage have been hacked together~~ in the ~~past, but (as far as Jadrian knows) nothing general-purpose exists yet~~.

	Since generating streamtubes takes a long time, it is ideal to detect orientation problems just by looking at the tensors, before generating the tubes. However, the effects of incorrect orientation are far more apparent in the tube visualization than in the tensor map.		== HOWTO Detect Tensor Orientation Problems ==

	== Before Generating Tubes ==		=== ... Before Generating Tubes ===
	# Generate RGB-coded tensor maps with <tt>mriatlas</tt>:		# Generate RGB-coded tensor maps with <tt>mriatlas</tt>:
	## <code>cd</code> into the directory of the processed brain you'd like to diagnose.		## <code>cd</code> into the directory of the processed brain you'd like to diagnose.
Line 24:		Line 22:
	##* Again, an incorrect color indicates a permutation.		##* Again, an incorrect color indicates a permutation.

	Lots of other groups seem to use planes of diffusion ellipses to diagnose orientation problems, but we don't currently have a way to generate those from tensors in [[MRIimage]] format.		Lots of other groups seem to use planes of diffusion ellipses to diagnose orientation problems, but we don't currently have a way to generate those from tensors in [[MRIimage]] format. If we did, we could detect dimension flips before fitting tubes.

	== After Generating Tubes ==		=== ... After Generating Tubes ===
	# Generate tubes with <code>$G/bin/tubegen</code> (check a [[diffusion processing pipeline]] Makefile to see how).		# Generate tubes with <code>$G/bin/tubegen</code> (check a [[diffusion processing pipeline]] Makefile to see how).
	# View them with [[Brainapp]].		# View them with [[Brainapp]].
Line 35:		Line 33:
	##* If they're broken, you have a flip in either the anterior-posterior direction or the left-right direction.		##* If they're broken, you have a flip in either the anterior-posterior direction or the left-right direction.
	<!--		<!--
	== Examples ==
			=== Examples ===
	Email "Brown 3T Cohen Study brain tubes (at last!)" contains screenshots from $G/data/diffusion/brown3t/cohen_hiv_study.2007.02.07/patient024/tubes.2008.07.27, a healthy brain with no orientation problems. I believe $G/data/diffusion/brown3t/cohen_hiv_study_registered.2007.02.07/patient116/tubes.2008.11.06 has an X flip. Should generate screenshots of RGB maps and streamtubes for examples.		Email "Brown 3T Cohen Study brain tubes (at last!)" contains screenshots from $G/data/diffusion/brown3t/cohen_hiv_study.2007.02.07/patient024/tubes.2008.07.27, a healthy brain with no orientation problems. I believe $G/data/diffusion/brown3t/cohen_hiv_study_registered.2007.02.07/patient116/tubes.2008.11.06 has an X flip. Should generate screenshots of RGB maps and streamtubes for examples.
	-->		-->

			== Fixing Problems ==
			There are two ways to correct tensor orientation problems: you can either transform each tensor in place, or you can transform all the DWIs and then re-fit the tensors. '''Transforming the DWIs''' is cheaper than '''transforming the tensors''', which is cheaper than '''fitting the tensors'''. Therefore there are pros and cons to each approach.

			=== ... by Transforming the DWIs ===
			<code>$G/bin/mritransp</code> can perform dimension permutations and flips on DWIs.
			* '''Long-Term Advantage:''' the next time you get a dataset from the same protocol, you can fix the orientation problem cheaply at the DWI stage and then the tensors that you fit to the DWIs will be correct. Total processing: '''transform DWIs''' + '''fit tensors'''.
			* '''Immediate Disadvantage:''' when you discover the orientation problem for a single dataset, you have already fit the tensors, so fixing it by modifying the DWIs means you will have to re-fit the tensors, which takes time. Total processing: '''transform DWIs''' + '''fit tensors'''.

			=== ... by Transforming the Tensors ===
			<code>$G/bin/mritensormult</code> can perform general transformations on all the tensors in a DTI, leaving the tensors in the same positions but changing their values in place.
			* '''Immediate Advantage:''' when you discover the orientation problem for a single dataset, you don't have to re-fit the tensors, which means that you can quickly move on to the next step in the [[Diffusion Processing Pipeline\|pipeline]] without waiting. Total processing: '''transform tensors'''.
			* '''Long-Term Disadvantage:''' the next time you get a dataset from the same protocol, the same orientation problem will probably be present, and so you will have to correct it, and transforming the tensors is expensive. Total processing: '''fit tensors''' + '''transform tensors'''.

	== Automatic Diagnosis ==		== Automatic Diagnosis ==

Jadrian Miles at 21:40, 16 March 2009

2009-03-16T21:40:17Z

← Older revision		Revision as of 21:40, 16 March 2009
Line 3:		Line 3:
	There are two types of mis-orientations:		There are two types of mis-orientations:
	* Permutations, where ''e.g.'' X has been exchanged with Y. These are easy to detect using an RGB-mapped visualization of the tensor field, or even by examining the DWIs if you know the b-vectors.		* Permutations, where ''e.g.'' X has been exchanged with Y. These are easy to detect using an RGB-mapped visualization of the tensor field, or even by examining the DWIs if you know the b-vectors.
	* Flips, where the images have been mirrored in ''e.g.'' X. These are easy to detect using a tensor ellipse visualization or by looking at a tractogram, a fair bit harder to spot by looking at a field of the off-diagonal elements of the tensors, and impossible to spot in an RGB visualization of the tensor field due to the mirror symmetry of the colormap.		* Flips, where the images have been mirrored in ''e.g.'' X. These are easy to detect using a tensor ellipse visualization or by looking at a tractogram, a fair bit harder to spot by looking at a field of the off-diagonal elements of the tensors, and impossible to spot in an RGB visualization of the tensor field due to the mirror symmetry of the colormap. Since the tensors have three-way mirror symmetry (that is, orientation but not direction matters), only non-axis-aligned tensors display any irregularity when flipped in one dimension.

	Either type can be corrected by transforming each individual tensor in the tensor field, without changing its position. <code>$G/bin/mritensormult</code> can perform general transformations of tensors.		Either type can be corrected by transforming each individual tensor in the tensor field, without changing its position. <code>$G/bin/mritensormult</code> can perform general transformations of tensors.

Jadrian Miles: /* After Generating Tubes */

2009-03-16T21:37:57Z

After Generating Tubes

← Older revision		Revision as of 21:37, 16 March 2009
Line 30:		Line 30:
	# View them with [[Brainapp]].		# View them with [[Brainapp]].
	# Check for orientation problems:		# Check for orientation problems:
	## Look at a front view of the tubes through the callosum and make sure they complete a U the goes to the superior cortex.		## Look at a front view of the tubes through the callosum and make sure they complete a U the goes all the way to the superior cortex. Since the CC is almost always axis-aligned, it rarely visibly breaks.
	##* If it's broken, you have a flip in either the superior-inferior direction or the left-right direction.		##* If they're broken, you have a flip in either the superior-inferior direction or the left-right direction.
	## Look at a top view of tubes and make sure that ~~splenium~~ and ~~forceps~~ complete a U that ~~goes~~ anterior and posterior from the callosum.		## Look at a top view of tubes and make sure that forceps major and minor complete a pair of U shapes that go anterior and posterior from the callosum.
	##* If it's broken, you have a flip in either the anterior-posterior direction or the left-right direction.		##* If they're broken, you have a flip in either the anterior-posterior direction or the left-right direction.
	<!--		<!--
	== Examples ==		== Examples ==