Neurophysiology Of fMRI Signals Flashcards
Positron Emission Tomography
Used radiation labelled isotope to measure blood flow, glucose uptake and oxygen metabolism in the human brain.
Limitations of PET
Very invasive
Spatial and temporal resolution is pretty poor compared to fMRI
Pioneering experiments for blood flow and oxygen consumption
Fox and Raichle (1986) - large increase in blood flow (rushing to active region when stimulated) but oxygen remained the same?
Fox et al (1988) - looking at glucose consumption as well and suggested blood flow was to provide more glucose not oxygen - fit with positive BOLD signal cos it explains the decrease in deoxyH.
Both experiments predicted blood saturation had to increase after sensory stimulation, predicting positive BOLD signal in an active area.
Lactate Shuttle hypothesis
This stated that glutamate (main excitatory neurotransmitter) was taken up by astrocytes which released lactate which is oxidised by neurons for energy, negating the need for oxygen.
Explains why oxygen consumption was so low following stimulation.
Arterial spin labelling
Can measure blood flow in the brain at the same time as measuring BOLD signal
Combined ASL and BOLD imaging with titrated visual stimulation and hypercapnia
Alternated visual stimulation at different levels and the hypercapnia (90% oxygen, 10% co2) at different levels to get different blood flow responses - can look at ASL and BOLD together to see blood flow and BOLD
Hypercapnia and visual stimulation produced same change in blood flow but BOLD response was far stronger for hypercapnia - it would be the same if brain wasn’t using oxygen like PET argued.
Strongly suggested the washout of deoxyH (producing BOLD) was far greater indicating the stimulation was using up more oxygen
Why do we have oxygen in NVC?
One idea is when blood flow increases, it’s harder for oxygen to come off the blood vessels so need to increase oxygen in the ridge the blood flow needs to increase more by dilating vessels.
Another idea is the blood flow response acts to keep tissue oxygen constant and ensure there is no hypoxia, especially the regions far from the blood vessels.
Devor et al (2011)
2-photon oxygen phosphorescence imaging
Shows the increase in oxygenation was smaller for regions with lower baseline oxygen saturation. So we get a big overshoot to make sure the distal regions get enough oxygen.
What type of neural activity drives BOLD signal and the dilation of blood vessels?
Synaptic activity is most correlated with results of positive BOLD signal.
However, lots of studies have been done in anaesthetised animals and this is an issue as it can impair NVC.
2 types of neural activity
Synaptic activity - local field potentials (LFPs) - reflect synchronised input into the area and its cortical processing.
Spiking activity - multiple-unit activity (MUAs) - represents recorded activity of populations of neurons as the output of a stimulated region.
Positive BOLD signal
Caused by large washout of deoxyH
Blood flow needs to go up by large amount to maintain the tissue oxygen gradient to provide it to the active neurons and avoid hypoxia.
Most related for synaptic activity with gamma power being the best marker of positive BOLD response.
What did Malonek and Grinvald show?
Strongly suggested oxygen was consumed before NVC could start.
Challenged Fox and Raichle suggesting oxygen consumption may drive NVC.
Deoxy dip
2D-OIS showed it is produced by a fast increase in blood volume, not a decrease in deoxyH.
fMRI hardly sees this deoxy dip so can’t be used to map human brain.
Post stimulus undershoot
Seen more frequently than deoxy dip even at lower MRI field strengths.
Represents an increase in deoxyH after stimulation due to:
- slower return of blood volume to baseline compared to blood flow
- blood flow drops below baseline after stimulation as arteries constrict below their baseline
- neurons are still consuming oxygen more than baseline, increasing deoxyH after stimulation
Hua et al (2011) - why is there a post stimulation undershoot?
Measured at 3T BOLD, blood flow, and blood volume using visual stimulation and breath holding stimuli.
Only visual stimulation had an undershoot but little difference in return of blood flow and volume returning to baseline.
Carbon dioxide remained elevated after visual stimulation has stopped.
This estimated 20% of post stimulus undershoot was due to blood flow and volume difference but the rest was due to elevated oxygen consumption after stimulation had stopped.
Simultaneous EEG and fMRI suggested a neuronal marker
Post stimulus undershoot was predicted by magnitude of alpha-power EEG in post stimulation condition:
Stronger alpha power (low frequency EEG) = bigger post stimulus undershoot = strongest alpha power after stimulation drives this undershoot (linked to oxygen consumption)
Summary of post stimulus undershoot
Most likely caused by prolonged uptake of oxygen post stimulation.
Indicates enough time between scans needs to be left to allow this to recover or could effect average trial response.
Could be important biomarker in disease - maybe enhanced or disappear (same for all tri-phasic BOLD signals)
Negative BOLD signal
Assumed this is equal and opposite to positive BOLD signal in that it represents a decrease in neuronal firing.
Could be an important biomarker in disease progression especially if it leads to breakdown of NVC.
Caused by an increase in deoxyH after- maybe caused by neuronal inhibition?
General stimulation paradigm for experiments
Anaesthetised rats - well controlled and able to do experiments for longer
Provide electrical whisker stimulation - long stimulation to drive a blood response to see a BOLD signal
Where are the BOLD signals appearing?
Negative BOLD appears deeper in the cortex than positive BOLD which is in shallow layers where barrel regions are
Can use simultaneous data to refine the model of BOLD data
Take blood volume and BOLD signals to predict what BOLD would look like if it could be measured
= better quality data in fewer animals
Negative BOLD signal is due to inhibition of neuronal activity in deep cortical layers
Frequency power results in surround region showed blood flow response is following a neural inhibition process including a bounceback
BOLD response can change due to anaesthesia
Deep anaesthetised rats couldn’t see a negative BOLD
Aroused state - anaesthetised but alert = see a distinct negative BOLD
Baseline changed dramatically which can affect haemodynamic signals
Sirotin and Das (2009)
Optical imaging study on 2 awake monkeys
They fixated ready for stimulus which never appeared but animal was responding as if it was there
They stated this went against theories of NVS that animals have strong BOLD signals with no neural activity - has preemptive blood flow response.
