Lecture 3: MEG Flashcards
MEG stands for
magnetoencephalography
Magneto in MEG means
magnetic
encephalo in MEG stands for
relating to the brain
graphy in MEG stands for
measurement
MEG uses sensors to measure small changes
in magnetic fields
The brain produces magnetic fields via (2)
- The brain produces electrical current
- Whenever we have electrical current flow (in pyramidal cells down dendrites), a magnetic field is also generated at right angles (perepndicular ; around the neuron)
Magnetic fields are also known as
‘B’ field
MEG is a non-inasive method as
changes in magnetic fields are detectable outside the head if the field is big enough
Like EEG, MEG signal gets weaker as it moves from the source
and gets further away
What are the 2 things that will affect the size of the magnetic field? - (2)
- Size of electrical current (more elec current , bigger magnetic field)
- Radius/Distance (further away from magnetic field, less able to detect it)
MEG also measures the activity of populations of neurons firing together (as still need many aligned cells to make a
dipole)
MEG is not measuring
action potentials
MEg also measures the activity of populations of neurons firing together as still need
many aligned cells to make a dipole
MEG still needs many
pyramidal cells lined up near the cortical surface to create a large enough dipole, to make a measurable magnetic field
The same activity looks different when measured with MEG vs EEG as looking at pyramdial and chuck of them being active and looking at brain activity but… - (2)
magnetic field is casued by electrical current
But has different orientaiton (90 degrees apart/perpendicular)
MEG measures the magnetic flux density which is measured in units called
Teslas (T)
MRI typically produces a 3T
magnetic field
Since the magnetic fields in the brain are very small, MEG usually measured in
femto Teslas (fT) - 10 to the power minus 15
Diagram shows different types of Teslas
There are 2 types of MEG systems that measure the small magnetic fields with two different kinds of sensors,
which is - (2)
- Traditional cryogenic MEG helmet with SQUIDS
- Wearable optically pumped magnetometers
Traditional MEGs uses SQUIDS which stands for
Superconducting QUantum Interference Devices
Superconductors are materials which display very
unusual properties when cooled down below their critical temperature
Superconductors lose all electrical resistance (thus allowing electricity to flow through them) when they get
very cold
A current in a superconducting loop will keep
flowing forever
When you apply a magnetic field to a superconductor, it will
affect the current flowing in the superconductor
Diagram of squid magnetomer - (3)
- Kept it cold and electricity flows straight through it
- Same amount of electricity at the end as the start as no friction and nothing lost
- Expect, the black bits which are layers of an insulator (which doesn’t let electricity through) which are called Josephson junctions
The layers of a superconductor separated by insulator allows to convert
magnetic field fluctuations to voltage (Josephson junction)
Josephson junction allows to measure changes in magnetic field by
converting it into voltage - turn back to electrical signal as works great with computers
What does this diagram show of Josephson junction? - (3)
- Current moving through superconductor but when it hits the insulator
- Inside of all current going through, it will reduce a bit
- How much current reduces at insulator portion depends on how strong the magnetic field is acting on it
Invention of (cryogenic) MEG - (3)
- Josephson (1962) – worked out how to use superconductors to convert magnetic field fluctuations to electric signals so we can measure their strength
- Cohen (1968) – recorded the magnetic signal using an induction coil
- Cohen (1972) – did first single SQUID recording behind someone’s ear in magnetically shielded room (photo)
Cohen’s first single SQUID recording on occipital lobe where there was
difference in alpha osciliations in MEG and EEG when eyes closed and open
SQUIDs are ‘cryogenic’ – need to be very cold as they (2)
Requires liquid helium to cool to ~4 Kelvin (~ minus 270 degrees Celsius)
Can’t be right next to the head but signal drops off as get further away
Inside the MEG helmet they need pick up coils which relays the signals
closer to a coil neares the SQUIDs since they are so cold
The SQUIDs pick up the signal in
pick up coils
Different types of pick up coils which record and examples - (2)
signals at different depths and angles of the magnetic field
(e.g., Magnetomers, planar gradiometers, axial gradiometers)
Disadvantages of the traditional MEG - (3)
- Needs very expensive liquid helium
- Need to be fixed in place and eated meaning cognitive tasks you can do is limited
- Single size helmet - coverage dependent on head size - bad for children
Traditional MEG: Cryogenic helmet wirh SQUIDs require many
sensors e.g., 248
The traditional MEG the cryogenic MEG helmet wirh SQUIDs is.. bad for.. - (2)
fixed in place, seated
small heads - children and doing certain tasks
Wearable Optically Pumped Magnetometers , like traditional MEG helmet with SQUIDs , detects magnetic field and turn into signal but in a
different way
Wearable Optically Pumped Magnetomers uses
Optically Pumped Magnetomers
Diagram of optically pumped mangetomers and how does it work? - (5)
- Theres a laser projected at the top
- ‘Pump’ light from a laser (1) (light pushed through tube) through a channel in the sensor and measure it with a photodiode (3)
- In this channel, there is an area containing rubidium gas (black gas)
- This affects how much light comes through it the channel
- Photodiode produces an (electircal) signal which is a measure of magentic field (i.e., how strong is it)
Whats a photodiode?
its a sensor that measures light
When there is a magnetic field around rubidium it changes its
transparency
If we apply a strong magnetic field, rubidium will become … if we had weaker magnetic field rubidium will stay - (2)
darker, less light
light, more light goes straight through channel and bigger signal
OPM systems each have sensor that are
very small (~2.5 CM long)
OPM systems don’t need any cooling so sesnors can be next to
scalp (and no helium costs)
OPM system sensors are wearable meaning
move around more (still use with coils that null magnetic fields or a shielded room)
OPM systems sensors are mounted in
age-appropraite cap or custom 3D printed head cast
OPM system is now avaliable with whole
head coverage (128 sensors)
MEG has better spatial resolution than EEG as - (3)
- Magnetic signals aren’t affected differently by the different materials (skull,skin) they must pass through
- May be easier to approx where the MEG signal from sensor space is orginated from in the brain
- Means it is easier to estimate the source of brain activity in MEG than EEG
MEG like EEG struggles to get signals from
deep sources
OPMS have better spatial resoltuion than SQUID as (5)
- SQUIDs are further from the head then OPMs
- The further you are away from head, lower the magnetic signal
- The signal has reduced and dispered more in SQUIDs than OPMS
- As OPMS more closer, they are more sensitive (get more signal) and have better spatial resolution
- Have fewever OPMs but give similar peformance
Advantages of wearable optically pumped magnetomers - (4)
- Cheap to run (no colling)
- Similar detection ability, perhaps better spatail resolution than cryogenic MEG helmet with SQUIDs (Traditional MEG)
- Different size caps -good for children
- Can move around
Wearable Pumped Magnetomers traditionally have fewever sensors but can now get
whole head coverage e.g., 128
Magnetic fields are produced by the
brain
Magnetic fields are produced by the brain and detected by a sensor and converted to an analog electrical signal in MEG
For SQUIDS it involves
For OPMS it involves - (2)
- For SQUIDs this involves a Superconductor
- For OPMs it involves a laser (O for optical)
After the Magnetic fields are detected by a sensor and converted to an analog electrical signal in MEGs they are then
sample and digitize this electrical signal, usually at a rate of 1000Hz or higher (one sample every 1ms)
MEG outputs are similar to EEG and looking at …
Looking at when there are magnetic changes in magnetic signal over time (ms)
MEG is getting better at where are these
magneit changes - source space to brain space
In source space, brain produces electromagnetic
signals
We can detect the electromagnetic signals in MEG outside the head with one signal per
sensor = sensor space
We often want to know which areas of the brain were active and not
which sensor
In MEG to know where the areas of brain (the ;source;) are active, not sensors, we can estimate the activity
in source space
To know where they are from in the brain (the ‘source’) we can estimate the activity in source space is called
source localisation
source localisation is more reliable and popular with
MEG than EEG as good spatial resolution and magnetic signals does not care for other materials
fMRI asks question of
where
EEG asks the question of
when
MEG asks the question of - (2)
where and when?
i.e., when do different regions play a role?
Advantages of EEG/MEG - (9)
- Non-invasive (unlike intracortical electrodes)
- More direct measure of brain activity than fMRI (activity produces electric potentials and magnetic fields, whereas activity is coupled to changes in blood oxygenation in ways that depend on the vascular system, although still correlation not causation).
- Excellent temporal resolution (msecs; unlike fMRI and PET; seconds+)
- Can measure dynamic (changing) connectivity
- Give frequency information so we can look at oscillations and particular frequency bands
- EEG and OPM use is cheap (<£2 per subject) and EEG is cheap (and OPMs relatively cheap) to setup
- EEG and OPMs are okay for kids, patients, etc.
- Are silent techniques (unlike MRI which is quite noisy), so good for studies with auditory stimuli
- Clinical applications (can use MEG but EEG cheaper and more widely available)
Electric potential/magnetic field strength plotted over time gives us
osciliation
Osciliations have cycles as shown below:
A location in a cycle is called the phase shown below
Number of cycles per second is
frequency (Hz)
We can identify brain activity at different frequencies with
EEG and MEG
We can see different brain locations produce activity at different frequences with an
MEG
Disadvantages of MEG/EEG - (6)
Have to estimate where in the brain the signals are coming from, instead of getting ‘images’ directly from the scanner as in fMRI. MEG better at this than EEG.
EEG has poor spatial resolution (cms)
MEG has okay spatial resolution (~1cm) so we can localise activity to specific brain regions but worse than fMRI/PET
Miss deep signals (subcortical, hippocampal, etc)
EEG has problems with sulci cancelling each other out
Even in MEG get signal from gyri > sulci and depends on orientation of neurons
MEG can sometimes detect signals that are lost to EEG as - (3)
Human brains are very convoluted (they have many folds)
Opposite potentials across a sulcus can cancel each other out
But the magnetic field should be OK because it is orthogonal (at right angles)
Can see different brain locations produce activity at different frequencies with MEG
and
Identify brain activity at different frequencies with MEG or EEG
But not with
fMRI
