Lecture 6 - MEG Flashcards

1
Q

Who used the first SQUID?

A

Cohen (1972)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What does SQUID stand for?

A

Superconducting Quantum Interference Device

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

MEG signal is ____ to EEG

A

perpendicular

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What contributes to the MEG signal?

- PSPs, APs, or secondary volume currents?

A

post synaptic potentials

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Which is faster, APs or PSPs?

A

APs

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Where do radial fields originate? Do they contribute to the MEG signal?

A

From the top of gyri

Not really

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Where do tangential fields originate? Do they contribute?

A

Folds in sulci, therefore magnetic field is perpendicular to sensors

Detectable by MEG

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What unit of measurement is detected in MEG

A

B

Magnetic Flux Density / Induction (Tesla)

Tesla= [Wb/m2])

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

When trying to detect small magnetic fields, what two conditions are requires?

A

Sensitive detectors (SQUIDs)

Low noise measurement environment (MSR)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What are SQUIDS and how do they work?

A

Superconducting Quantum Interference Device (SQUIDs)

Very sensitive detector of magnetic flux

Cooled using liquid helium

Most modern MEG systems use DC-SQUIDS

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Three configurations for signal (pickup) coils?

A

Magnetometer
Planer gradiometer
Axial Gradiometer

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Magnetometers can see ____ in the brain

Gradiometers are regarded as ____

A

deep

shortsighted

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Pros and cons of magnometers

A

can see deeper into the brain

However, they pickup more noise

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

How are SQUIDs and pickup coils connected?

A

Using a flux transformer

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What are larger, SQUIDS or pickup coils?

A

pickup coils

SQUIDS are only 1mm3 ish

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Examples of physiological artifacts?

A

Cardiac
Breathing
Eye movements
Muscle movements

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Examples of non-physiological artifacts?

A

Magnetic material in the room
Cars
Electrics

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Methods of artifact removal?

A

ICA
Epoch/Data rejection
EOG/ECG can provide information

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

MEG ___-___ data can be selected and averaged in a similar way to EEG data

A

sensor-space

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

Can choose which data points in sensor-space to include based on ____ information

A

trigger

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

What is the maximum frequency we can sample?

A

Sample rate/2
MEG systems often sample at 678.17Hz, so /2 = 339Hz
This is the Nyquist frequency

in reality, signals are limited to a maximum of 200Hz

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

Frequency resolution is determined by what?

A
The amount (s) of data we have 
Calculated as 1/(data length in seconds)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

Time-domain representation of a waveform can be turned into a frequency-domain representation by what tranformation?

A

Fourier transformation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

Time and frequency-domain representation can be transformed into what combined view?

A

time-frequency representation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
What is oscillatory activity thought to arise from?
Assembles of neurones acting together
26
What do the red and blue parts of an equivalent current dipole?
Red shows fields coming "out" of the brain, blue shows fields going back "into" the brain
27
The source of activity in equivelent current dipoles is most likely found at what point?
the 0-point between dipole peaks
28
The distance between "peaks" in a dipole pattern gives a clue as how ___ the dipole is
deep
29
Dipole strength drops off as a ___/___ with increasing depth
1/r^3
30
A focal deep source and distributed superficial source will produce ___ magnetic fields
similar
31
ECD is composed of...
A position in the head An orientation A strength
32
What is the unit of an ECD?
Am (ampere-meters i.e. current)
33
From an ECD, we can predict what the MEG sensors would show. This is called the ____ model.
forward
34
An example of phase-locked response is...
Evoked responses to stimuli
35
ERP in MEG is known as...
ERF - evoked response fields
36
Analysis of phase-locked responses involves...
averaging the time representation
37
An example of a non-phased-locked response is...
Induced responses
38
Analysis of non-phase-locked responses involves...
Averaging the frequency domain representation Changes in power are not time locked and may be jittered
39
Primary sensory responses tend to have more interesting ___-___ components, where as higher level functions and inter-area communications tend to involve ______-___ activity.
phase-locked | non phase-locked activity
40
What is an inverse problem?
When we attempt to make inferences about a system from a set of observations
41
In regards to the inverse problem in MEG, the things we know are...?
MEG sensor measurments and their positions Structural information about the brain Knowledge of how electrical currents relate to magnetic fields (via approximations of Maxwell's equations)
42
Von Helmholtz (1853) showed there is ___ unique solution to explaining a set of magnetic fields outside a conducting sphere
NO The problem is "ill-posed"
43
To "solve" the inverse problem, we need to apply ___
constraints
44
There are two main groups of constraint based inverse-problem solutions...
Fitting/Localization methods (Dipole Fitting) Scanning or Imaging Methods (Minimum Norm Estimates; Beamforming)
45
To solve the inverse problem, we can model the ___ problem
Forward problem
46
The forward model describes...
Given a set of currents in the brain (position, orientation, strength), what would we expect to see on the sensors? Most models use an ECD model for this.
47
When forward modelling, we need to model the brain. What are some brain models used?
Single Sphere Multiple spheres Some models can tesselate: Boundary Element Models (BEM) Finite Element Models (FEM)
48
After modelling the brain and ECDs in the forward model, we get a ____. This shows what the ___ ___ would 'see' for a given ____ at each ___ and ___ in the brain.
``` Leadfield MEG Sensors dipole location orientation ```
49
Simplest form of inverse solution is to fit a ___ ___ ___
single dipolar source
50
Single dipole source (inverse solution) minimises the ____ ___ between ____ pattern and ____ pattern on the sensors. This is known as ____-____ minimisation.
Squared errors predicted observed least-squares minimisation
51
Although single dipole fitting is the simplest form of the technique, what are two alternative and more complex methods?
Multiple dipole fitting | Spatio-temporal dipole fitting
52
What does the algorithm RAP-MUSIC do?
Estimates how many dipoles should be fitted
53
Do you need to decide the number of dipoles to fit a-priori ?
yes
54
Minimum norm estimate?
Similar to dipole fitting, however, finds the most likely distributed source rather than a single point.
55
Beamforming
Does not try and minimise the error with respect to the MEG data for all source location at one Takes each source location in turn and estimates activity at that point Closely linked to radar
56
In relation to beamforming, what does the spatial filter do?
Attempts to focus on a particular spatial location. The filter tries to attenuate responses from other locations.
57
In beamforming, once we have the weights for each location in the brain, what can we produce?
A virtual electrode
58
What is a virtual electrode time series? What are the units?
An estimate of current moment at each "voxel" in the brain VE units are the same as ECDs: Am
59
Methods to combat multiple comparisons?
Bonferroni Correction Non-parametric Random Field Theory
60
Advantages of MEG
Doesn't rely on indirect measure of brain activity (e.g. haemodynamics) Good temporal resolution Doesn't suffer from same spatial smearing of EEG Good for network connectivity
61
Disadvantages of MEG
Complex analysis Poorer resolution than fMRI Expensive
62
Working in source space involves solving the ____ problem
inverse
63
What neural activity is thought to be the primary origin of the responses recorded in MEG?
Postsynaptic electrical activity from Pyramidal Cells
64
Whether an MEG channel is a magnetometer or gradiometer setup is determined by:
The configuration of the pickup (sensor) coils
65
What is the full name for a SQUID
superconducting quantum interference device
66
Individual MEG sensors can see ___ but not ___ sources
tangential | radial
67
What is the Nyquist freq?
sample rate / 2
68
Minimum norm is a ____ analysis method
Soure-space
69
Evoked activity is analysed by _____ of the time-series MEG data
taking a direct average
70
What did Josephson do?
Describes Josephson Junctions
71
Who recorded alpha rhythms using an induction coil?
Cohen (1968)
72
Electrical current flow (I) in a wire produces a magnetic field (B) _____ to the current flow
perpendicular
73
Why do secondary volume currents not really contribute to MEG signal?
Tend to cancel out
74
Ferromagnetic?
Retain magnetism after being exposed
75
Paramagnetic?
Amplify magnetic fields but do not stay magnetized
76
Diamagnetic?
Weaken applied fields
77
What is magnetic permeability?
How much does a material magnetize | when exposed to an external magnetic field?
78
Superconductors are perfect ___
diamagnets
79
MEG signal acquisition overview?
1. Magnetic fields pass through the pickup coil • Either magnetometer or gradiometer configuration 2. A current is induced in the pickup coil, and via the flux transformer, it is coupled to the coil next to the SQUID 3. The measurement of this current tells us about the magnetic field
80
Reference channels are automatically used to help ___ the data and we normally ___ ___ for data analysis purposes
'de-noise' | ignore them
81
What is a butterfly plot?
Shows magnetic flux/field VS time for all the channels
82
'high' gamma?
70-150Hz - cognitive processing
83
Very high frequency oscillations?
150Hz+ Low level perceptual processing/epilepsy
84
Need ABOUT ____ neurons to detect signal (estimation ranges)
50,000
85
How much space do 50,000 neurons take up? ABOUT
50,000 neurons would need a patch about 0.63mm2 in area
86
The spatial resolution of MEG is _ ____ ____
a complex question
87
Evoked activity is analysed by taking a ____ of the ____ MEG data.
direct average of the time-series data