Brain connectivity
Task 2

Do we need to study connectivity to understand brain function

Why are there so many connectivity studies lately

How does neural tracing work

How can we study white matter

Why can't we use it in humans

How can we see what is connected to what

How can we study human wiring with MRI

What is resting state functional connectivity

What is DTI

If 2 structures correlate, does that mean they are connected?

invasive and destructive

Based on signals from 1H(proton) nuclei

limitations

spatial resolution

contrast

10 mumeter

data size

too big or small?

Water

if these molecules are the same then regions indestinguishable

Contrasts

dependent on

Proton density (PD)

T1 and T2

Diffusion coefficient (D)

decay times after excitation

motion of water molecules

formula for singal

S=PD(1-e2TR=T1 )e2TE/T2e-bD

b

diffusion weigning factor

can be changed

important for DTI

D

diffusion term

motion of water molecules

3 facts

water molecules move

DTI uses this motion to infer neuroanatomy

DTI is mostly about static anatomy

diffusion

isotropic

circular ink stain

anisotropic

when higher fiberdencity is oriented in one direction

estimate axonal organization

B0 field

strong magnetic field along bore

can be linearly altered by gradient pulse

Gradients

X Y Z

changing strength and timing of gradient pulses

click to edit

detects water along gradient axis

Combine X, Y, Z axes

fiber orientations mostly oblique to axes

FA

Fractional anisotrophy

0(isotropic) to 1(anisotropic)

limitations

prone to artifacts caused by physiological motion

limited spatial resolution

Advantages

2 types of new contrasts

diffusion anisotropy

fiber orientation

can display the complexity of white matter

cannot differentiate directionality of axons

diffusion can be blocked by obstacles

structures are assumed homologues within pixel

Application

check myelinization

T2 contrast

reveal locations of white matter bundles

click to edit

studying the level of coactivation between functional time-series of anatomically separated brain region during rest

Patterns usually reflect functional correlation

How to process this data

model dependent methods

seed method

model free methods

correlate resing state of 1 region against the rest of the brain

functional connectivity map

must be defined a priori

simple analyses and straight foreward result

difficult to examine whole brain scale

Examines whole brain

Principal component analysis

independent component analysis

searches mixture of underlying sources that can explain pattern

add clustering strategies

all methods have high overlap

Resting state networks

anatomically separated but functionally linked regions of functional connectivity during rest

8 subnetworks

motor

visual

lateralized superior parietal /frontal

default mode

bilateral temporal/insular and ACC

Analyses

small world principle

click to edit

graph theory

click to edit

Disorders

Alzheimers

decreased default mode

decreased resting activity in PCC and hippocampus

schizophrenia

aberrant default mode connectivity

decreased connectivity in medial frontal cortex and precuneus

altered integrity of the cingulum

MS

click to edit

ALS

inject a tracer to see where it goes

retrograde and antrograde tracing

cons

brain has to be taken out

Pros

good technique

Signal loss between pulses = diffusion

many different gradients

for each direction

ellipsoid

around the points were it diffused to

the pointier the more diffusion

limitations

influenced by things like breathing