Fiber Tracking & DTI

Question 1

What is the Connectome?

Answer 1

Term invented in 2005 independently by Sporns and Hayman. It is the total map of all the connections present in the human brain. (Or actually mostly white matter connections but not so much intra cortical finer scale connections?) Yes how different gray eras are connected d by white matter

Question 2

Which property of water in the brain is Diffusion Tensor Imaging leveraging?

Answer 2

The fact that free roamng water will diffuse in an isotropic (i.e. towards all directions equally) way, while when constrained by myelin fibers it would do in an anisotropic (i.e. directional) way

Question 3

What is a tensor?

Answer 3

It is a 3 dimensional shape that can range in DTI from almost round to cigar shaped, it measures the orientation of anisotropic water diffusion. It is measured in every single voxel.

Question 4

What are principal eigen vectors for a tensor?

Answer 4

They are the three main vectors that indicate the orientation of a specific vector

Question 5

How does a tracking through a streamline works?

Answer 5

You select seed voxel(s) and from there make the streamline go in the direction of the tensor to adjacent voxels.

Question 6

What are stop criteria for streamline fiber tracking?

Answer 6

Going outside of ROI, going outside of brain, making a sharp turn (e.g. 90 degrees, biologically implausible)

Question 7

What is Fractional Anisotropy?

Answer 7

It is a scalar value (from 0 to 1) measuring the directionality of a tensor, it is derived from its eigen vectors

Question 8

What is a Continuous Tracking Algorithm for?

Answer 8

It serves the purpose of analyzing fiber paths at a finer level than voxel wise, it raises the resolution when fiber tracks have to be modelled

Question 9

How can FA help us infer the strength of some connections?

Answer 9

The higher the FA of voxels in a streamline the higher the connection strength will probably be

Question 10

What is roughly the work flow for analyzing connections inside the brain with DTI?

Answer 10

1. High res. struct. MRI is taken
2. Multiple diffusion tensor images for same subject are taken and weighted.
3. The two images are superimposed and we can then see how many connections are there between two brain areas

Question 11

Based on the human connectome project data, what are the amounts of (possible) connections that everyone has and the ones that no one has?

Answer 11

7% are shared by everyone, 20% not present in anyone. There is then a big variability in possible connections inside peoples’ brains.

Question 12

Connection weight measures: Number of Streamlines

Answer 12

How many streamlines are connecting two different areas

Question 13

Connection weight measures: Fractional Anisotropy/ Generalized FA

Answer 13

The mean weighted FA of a single tract the brain is proportional to its strength. Generalized accounts for non gaussian distribution.

Question 14

Connection weight measures: Streamline Volume Density/Streamline Surface Density

Answer 14

The number of connections is scaled comparing the size of the brain areas

Question 15

Is Diffusion MRI a functional MRI measurement?

Answer 15

No it is structural, it is measuring anatomy not the brain functioning. It can be performed by any MRI scanner

Question 16

How long Diffusion MRI takes?

Answer 16

Around 15 minutes, it is preferrable to perform several times for maximum accuracy but not so common

Question 17

How many diffusion directions are needed?

Answer 17

At least six but more is better (e.g. 98)

Question 18

Image acquisition techniques.

Answer 18

Radio pulse that flips spins 90 degrees and then another pulse in opposite direction, 180 degrees. The time between signal and readout is 90 ms. (not clear is they do both signals and then measure of if they do first one signal then the other and then they measure)

Question 19

Graph theory. What does the formula G=(V,E) means?

Answer 19

A Graph G is described by the relationship between its vertices (nodes) and edges (connections)

Question 20

What is the degree distribution and how is it in the brain?

Answer 20

It measures how are the degree of vectors in the networks distributed. The brain has an exponential degree distribution, few areas are hyperconnected but most are not.

Question 21

How is local clustering in the brain and path length?

Answer 21

There is a high degree of local clustering and a short path length across the whole brain. There are more brain areas not so connected to each other than very connecte d to each other.

Question 22

Which models are used to tets whetherT the brain has the stricture of a small word network?

Answer 22

Two null hypotheses: Th ebrain is a totally random network (poor local clustering but very short path) or to a “lettuce”(?) path, made by connecting each cnode to the ones to its sides, high in local clustering but with long paths. DS Bassett research showed how neural networks in different animals (humans comprised) are instances of small world networks

Question 23

What is the “rich club” and what is its role in the connectome?

Answer 23

A series of brain areas highly interconnected. (frontal lobes, putamens, precunei, thalami and hippocampi). It links the brain together, every long path has to pass through it. Some researches also think it correlates with cognition and mental disorders.