BOLD Flashcards
the spin of which protons is usually measured in MRI?
a oxygen
b haemoglobin
c hydrogen
c hydrogen
BOLD signal reflects spin of protons in water molecules
is deoxygenated or oxygenated blood more magnetic?
deoxygenated
what happens to the T2* signal if blood is deoxygenated?
less oxygen (more deoxygenated blood) -> more magnetic -> spin stops faster -> shorter and therefor weaker T2* signal -> region appears darker
more oxygen -> slower relaxation -> brighter image
which relaxation coefficient shows the higher difference between oxygenated and deoxygenated blood?
T2*
What causes the BOLD signal (or contrast)?
susceptibility changes between oxygenated and deoxygenated blood
(changes in proton spin / relaxation)
how does a BOLD signal reflect neural activity?
a) more activity -> higher neural metabolism -> more oxygen is metabolized -> more deoxygenated blood -> more magnetic -> faster relaxation of hydrogen spin -> weaker signal -> darker image
b) more activity -> higher neural metabolism -> more oxygen is metabolized -> initial increase of deoxygenated blood -> neurovascular coupling -> inflow of fresh and oxygenated blood -> less magnetic -> slower relaxation of hydrogen spin -> region apprears brighter
b)
otherwise all my oxygenated blood is used up now and I st
what’s neurovascular coupling again?
copy paste:
term used to describe alterations in local perfusion that occur in response to changes in neuronal activity
refers to the relationship between local neural activity and subsequent changes in cerebral blood flow
name the steps of the canonical haemodynamic response function (HRF)
(temporally extended response to nerve firing)
- HRF baseline
- increased neural acticity after stimulus presentation:
- initial dip in oxygenated blood ( 0 - 2 s)
- rise / overshoot (4-6 s to peak) -> lasts longer if the stimulus stays present
- decay (~ 10 s)
- post stimulus undershoot (10 - 20 s)
- back to baseline
whole thingy takes 20-30 s
(in one slide he says decay takes 10 s but nothing about undershoot, in the other he says undershoot takes 10 - 20)…
1mm³ cortex contains how much neurons? synapses? densdrites (in m) ? axons (in m)? capillaries (in m)?
numbers taken from visual cortex I think
10000 - 100000 neurons 1000000000 - 10000000000 synapses 300 m dendrites 4 km axons 40 cm capillaries
what s the typical voxel size in
a) structural MRI
b) functional MRI
a) 1x1x1 mm
b) 3x3x3 mm (possible in 3 Tesla scanner)
spatial resolution in fMRI?
reliable / stable up to submilimeter scale e.g. in ocular dominance columns (= very precise :)
temporal resolution?
not too good
onset of signal from fMRI does not directly reflect neural firing, as described before. - it takes a while for the BOLD contrast to set in (and the contrast rather reflects that some activity happened/oxygen metabolism raised, than actual neuron firing).
BUT.. comparing e.g. stimuli presented first in one hemifield and then after onset latency presented in the other hemifield, signals reliebly reflected the onset latencies in a linear relationship between stimulus onset dealy and measured onset latencies, tho temporally delayed to actual stimulus presentation
how variable is HRF?
- intersubject variability
- no correlation between peak latency and amplitude
linearity of bold response?
- scaling -> signal gets stronger over trials
- superposition -> overlapping curve from multiple trials
- temporal linearity -> high overlap between predicted and measured BOLD response
- non-linearity at < 6s
what is a boxcar regressor good for in fMRI?
boxcar regressor for modelling the HRF depending on RT or processing time under the assumption on linearity
