Lecture 5 Neurophysiology of fMRI signals

Question 1

Briefly, in basic physics how is the BOLD signal produced?

Answer 1

subject in MRI scanner, radio signal pulses emitted which flips water molecules in blood until they decay back to original energy. Decay is T2, and it is affected by blood and whether it is oxygenated or not.
decreased blood oxygen (more deoxy-haemoglobin) increases this decay/relation rate in time and gives a lower signal, vice versa for increased blood oxygen (more oxyhaemoglobin)

Question 2

how does oxyhaemoglobin change BOLD

Answer 2

larger/more positive BOLD

Question 3

does oxyhaemoglobin or deoxyhaemoglobin dominate BOLD?

Answer 3

deoxy (even though it attenuates/makes it smaller)

Question 4

when neurons are activated, what is happening to change the BOLD signal

Answer 4

there’s a large increase in blood flow to this region, increasing volume, or HbO2 blood which washes away Hbr, causing increased BOLD signal

Question 5

what are the 3 elements to the BOLD tri-phasic signal

Answer 5

initial dip
main positive BOLD signal
post-stimulus undershoot

Question 6

describe the discourse on main positive BOLD signal

Answer 6

using Positron Emission Tomography (PET) which measures quantitative metabolic changes produced by stimulation
Fox and Raichle 1986- used PET to look at responses to sensory stimulation, but found very little increase in O2 in region. They conducted another study finding there was an increase in glucose consumption, but not a change in oxygenation. , suggesting that glucose produces energy through anaerobic respiration, however, MRI field looked into this and found different conclusion

early BOLD fMRI experiment in the 90s used hypercapnia in rats to produce a positive BOLD signal (vasodilation) and in humans indicating BOLD response to visual stimulation- hence suggesting use of O2.

Question 7

describe research into oxygen consumption (which came out after Raichle & Fox’s ideas that brain didn’t use much oxygen when active)

Answer 7

research using arterial spin labelling (ASL) in order to label blood in the neck and measure spin in the brain to measure blood flow
Richard Hoge 90s- combined ASL and fMRI BOLD and found both visual stimuli and hypercapnia in rats caused the same change in blood flow, however, they found the BOLD response was greater for the high hypercapnia, showing a greater washout of deoxyhaemoglobin, indicating the stimulation was using more oxygen - this did not support conclusions of Raichle and Fox PET studies

2000s research by Berwick et al in sheffield found the same patterns in whisker stimulation and barrell cortex activity with hypercapnia

Devnor et al 2011- indicated that this large increase in oxyhaemoglobin is required to maintain baseline tissue oxygenation at locations distal to blood vessels

Question 8

at its most basic level, what are the 2 types of neural activity?

Answer 8

synaptic activity
spiking activity

Question 9

describe synaptic activity

Answer 9

shown as local field potentials (LFPs)- reflecting the synchronised input into area and inter-cortical processing by sinks and sources of current

Question 10

describe spiking activity

Answer 10

shown as multiunit activity (MUA)- recorded activity of populations of neurons output

Question 11

describe research into whether main positive BOLD signal is a reflection of spiking or synaptic activity

Answer 11

Nikos Logothetis research in anaesthetised primates using visual checkerboard stimulation found BOLD response has a large time delay compared to neuronal response. MUA is short/transient, whereas LFP waves continue over longer time LFP is thought to be more closely correlated with BOLD signal, in particular gamma LFP waves (although anaesthetics can affect NV coupling, however later Logothetis found he same in awake primates).

a further study into cats found similar relationship between haemodynamics and gamma oscillations

a more recent study found relationship between oscillatory EEG activity and laminar specific BOLD signal, where gamma oscillations were associated with superficial OLD and alpha with both deep and BOLD- suggesting the relationship between LFP and BOLD vary at cortical layers

Question 12

overall, what is positive BOLD caused by

Answer 12

a large wash in of oxyhaemoglobin

Question 13

what briefly is the initial deoxy-dip

Answer 13

a small drop in BOLD signal before it shoots up

Question 14

what did Malonek and Grinwald find in 1996 (initial dip)?

Answer 14

using slit-optical imaging in cat visual cortex using known absorption spectra of oxy and deoxy-haemoglobin. HbO2 and Hbr were visualised in barrel cortex during whisker stimulation they found this slight deoxy increase (BOLD-dip) before HbO2 washout. they posited that it was due to neurons taking O2 out of vessels before they dilate

Question 15

what did Berwick et al 2008 find regarding initial dip?

Answer 15

saw dip even when applying stimulation to individual whisker barrels- showing it’s a very specific response

Question 16

what did Sirotin and Das find in 2009 relating to initial dip?

Answer 16

in awake primates they found no evidence of the initial dip, just a fast increase in blood volume
(maybe difference between cats/rats and primates due to brain sizes?)

Question 17

What did Ralph-Freeman group find regarding initial dip?

Answer 17

looking at cat visual cortex and grating stimulation found increases in neuronal spike rate were accompanied by decreases in oxygenation of tissue- perhaps dips associated with neighbouring inactive neurons around activated ocular dominance neurons

Question 18

what are the 2 main posited hypothesis for initial BOLD dip?

Answer 18

increase in Hbr or fast localised increase in blood volume

Question 19

what are the 3 main theories of post-stimulus undershoot?

Answer 19

1) slower return of cerebral blood volume (CBV) to baseline compared to CBF- balloon theory of veins remaining swollen so can hold more volume for a time after stimulation
2) after stimulation CBF drops below control levels- potentially due to interneuron involvement
3) CBF returns to baseline after stimuli, however, neurons are still consuming more O2, like an O2 payback- resulting in a short increase of Hbr

Question 20

discuss evidence surrounding post-stimulus undershoot

Answer 20

Peter Van Zijl’s group used a scan to quantify blood volume and used visual stimulation and breath holding (hypercapnia) and found that from the hypercapnia did not display undershoot, hence evidence against the balloon vasodilation theory , they also found O2 consumption remained high after stimulation so support the after stimulus O2 payback theory

Richard Buxton’s Group used multimodal fMRI and EEG and found post-stimulus undershoot was predicted by alpha-power EEG (similar to Gamma power predicting positive BOLD signal) - clear neural- vascular link

Question 21

which theory is most likely correct for after-stimulus undershoot?

Answer 21

continued oxygen consumption/payback

Question 22

how could looking at tri-phasic patterns in people be useful

Answer 22

as a disease biomarker

Question 23

what causes a negative BOLD signal?

Answer 23

assumed to be equal and opposite of positive
caused by increase in Hbr

Question 24

what did Nikos Logothetis show about the location of negative BOLD

Answer 24

he found is corresponds with decreases in neuronal activity in monkey V1 in areas surrounding positive BOLD

Question 25

what are the 4 theories of what causes increased Hbr/negative BOLD

Answer 25

1- decreased neural activity (constricts vessels)
2- vascular steal (if several branches of artery supply regions and one area is active, the increased blood in that area causes negative BOLD in other artery branch regions)
3- breakdown in neurovascular coupling- when an increase in neural activity doesn’t increase BF, so Hbr increases due to increased oxygen consumption with lack of O2 supply
4- massive increase in neural activity e.g. epilepsy, where BVs cannot supply enough blood to areas of high activity so Hbr builds up

Question 26

describe the findings of unpublished work in Sheffield on negative BOLD

Answer 26

used fMRI on rats during electrical whisker stimulation- found strong positive BOLD signal in corresponding barrel cortex, which negative BOLD surrounding and on contralateral cortex (being on contralateral suggests this isn’t due to vascular steal)
they also found negative BOLD comes from deeper cortical layers than positive
they then combined with 2D- OIS (measuring blood volume) and found stimulation was eliciting robust neurovascular response with a reduced blood supply in surround with Hbr decrease in stimulated region but increase surround
they then used electrodes and laser-doppler flowmetry and found increase in BF in positive BOLD and decrease BF in negative BOLD
hence close links with flow and volume
they also found associations between negative BOLD and reduced MUA (reduced spiking)
they then looked at individual stimulations to better understand dynamics and found strong GAMMA power= weak negative BOLD and vice versa
so Gamma associated both with positive and negative inversely