Neurophysiology of fMRI signals

Question 1

What is the BOLD signal and what is the physics of the signal?

Answer 1

BOLD signal measures inhomogeneities in the magnetic field due to changes in the level of O2 in the blood

After an initial radiofrequency pulse, protons return to their original orientation and emit energy in the form of radio waves
T2 = time constant of how quickly the protons emit energy when recovering to equilibrium
Oxygenated blood is diamagnetic = no signal loss
Deoxygenated blood is paramagnetic = signal loss

Increase in neural activity = increase in blood flow = increase oxyhaemoglobin = increase in T2 = increase in MR signal

Question 2

What causes a bigger BOLD signal?

Answer 2

The more oxyhaemoglobin there is

Question 3

What drives the BOLD signal?

Answer 3

Changes in deoxyhaemoglobin (Hbr)

Large increase in blood flow and volume washes away the Hbr
A decrease in the level of Hbr means less signal attenuation resulting in an increased MRI signal

Question 4

Describe the tri-phasic BOLD signal

Answer 4

1. The initial dip
2. Positive BOLD signal
3. Post-stimulus undershoot

Question 5

How was the BOLD signal first predicted?

Answer 5

Using Positron Emission Tomography (PET) imaging to measure metabolic changes produced by sensory stimulation
PET is quantitative - using radiation labelled isotopes you can measure blood flow, glucose uptake and oxygen metabolism in the brain

Question 6

What did early experiments using PET tell us about the BOLD signal?

Answer 6

Fox and Raichle (1986)
Proposed that during functional tasks, blood supply to the brain was not to supply oxygen but was instead glucose - large increase in blood flow but little oxygen consumption

Predicted that blood saturation had to increase after sensory stimulation - predicted the positive BOLD signal in an active area

Led to the lactate shuffle hypothesis - glutamate taken up by astrocytes and released lactate which was oxidised by neurons for energy - negating the need for oxygen - explains why oxygen consumption was so low

Question 7

How did fMRI researchers contest the idea of little oxygen consumption?

Answer 7

Richard Hoge
Used an fMRI method called arterial spin labelling to measure blood flow
Experiments which combined ASL and BOLD imaging with titrated visual stimulation and hypercapnia
Hypercapnia and stimulation produced exactly the same change in blood flow, the BOLD response was far stronger for hypercapnia
This strongly suggests that the washout of deoxyhaemoglobin in the hypercapnia condition was far greater indicating that the stimulation was using up more oxygen

Question 8

What was launched in response to the findings of Hoge?

Answer 8

Launched a new field called calibrated fMRI
Attempt to quantify cerebral metabolism with fMRI BOLD and vascular reactivity methods

Important for disease

Question 9

What are some other suggestions for the large increase in blood flow?

Answer 9

1. Blood flow response acts to only keep tissue oxygen constant to ensure there is no hypoxia
2. The large increase in blood flow is to ensure regions far away from vessels do not become hypoxic known as lethal corners

Question 10

Explain the research by Devor et al. (2011)

Answer 10

Used 2-photon oxygen phosphorescence imaging
Shows the increase in oxygenation was smaller for regions that had a lower baseline oxygen saturation
Arguing that this is the reason the blood flow needs to go up as large as it does - to make sure distal regions get enough oxygen

Question 11

What types of neural activity drive the BOLD signal?

Answer 11

Synaptic activity = local field potentials (LFPs)
- LFPs reflect the synchronised input into the area and intra-cortical processing
- LFPs are generated by current sinks and current sources

Spiking activity = multi-unit activity (MUA)
- Multi-unit activity represents the recorded activity of populations of neurons
- This is the output of the stimulated region

Question 12

What frequency wavelength dominates the BOLD signal?

Answer 12

Evidence that within the FP range, GAMMA frequencies were dominant
Other research found evidence that GAMMA power correlated with superficial BOLD responses

Question 13

Is the BOLD signal a reflection of synaptic activity or spiking activity?

Answer 13

Synaptic activity

Some research suggested that BOLD activation may actually reflect the neural activity related to the input and the local processing in any given area, rather than the spiking activity commonly thought as the output of the area

Question 14

What did Malonek and Grinvald show?

Answer 14

Used slit-optical imaging spectroscopy in the cat visual cortex

Watering the entire garden for the sake of one thirsty flower
Strongly suggested oxygen was consumed before neurovascular coupling could start
If this produced a response in BOLD fMRI it would be better localised than the main positive BOLD response - would lead to better mapping of active regions
It also challenged other work, suggesting oxygen consumption may drive the coupling

Question 15

What are the arguments for and against the deoxy dip?

Answer 15

Berwick et al. (2008) saw a Hbr dip to single whisker stimulations

Primate research showed that the dip was due to a fast increase in blood volume - not Hbr

It has been seen in fMRI but not often

Question 16

What is the post-stimulus undershoot?

Answer 16

Represents an increase in Hbr after the stimulation has ended

Question 17

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

Answer 17

1. Slower return of CBV (total blood volume) to baseline compared to CBF - main feature of the balloon model put forward
2. At the end of stimulation CBF drops below control levels potentially due to interneuron involvement
3. Even though CBF returns to baseline quickly after the stimulus the neurons are still consuming oxygen more than baseline resulting in increased amounts of Hbr

Question 18

What does evidence suggest is the correct theory?

Answer 18

Paper looking at calculated changes to CMRO2 (cerebral metabolic rate of oxygen) - provided a strong answer

In humans, measured BOLD, CBF, CBV - used visual stimulation and breath holding stimuli (similar to hypercapnia challenge)
Breath hold had no post-stimulus undershoot
Visual stimulation did have an undershoot

CMRO2 changes to stimulation remained elevated long after the stimulus had finished
CMRO2 changes to breath hold returned to baseline quickly
They estimated that 20% of the post stimulus undershoot was due to CBF/CBV differences but 80% was due to elevated oxygen consumption after the stimulation had stopped

Question 19

What is the practical importance of the post stimulus undershoot?

Answer 19

You need to leave enough time between scans to allow for the post stimulus undershoot to recover or you could really effect your average trial response
Leave twice as long as your stimulation period

Question 20

Why is understanding the negative BOLD signal important?

Answer 20

Assumed that negative BOLD is equal and opposite to positive BOLD in that it represents a decrease in neuronal firing
Negative BOLD could be an important biomarker of disease progression - especially if the disease leads to a breakdown of neurovascular coupling

Question 21

What causes the negative BOLD response?

Answer 21

Caused by an increase in deoxyhaemoglobin (Hbr)
The tacit assumption is that it is caused by neuronal inhibition
However, increased Hbr could be generated in a number of circumstances:
1. Decrease in neural activity
2. No change in neural activity ‘vascular steal’
3. Breakdown in neurovascular coupling
4. Increase in neural activity