BOLD-Based fMRI or “The Stuff We Do With The 4T” Part I Chris Thomas April 27, 2001

Outline Lecture 1 – The BOLD Response –What is it? –How does it come about? –What are its characteristics and where do they come from? Lecture 2 – The fMRI Experiment & Pitfalls –Temporal & spatial resolution. –Noise. –Artifacts. –MRI specifics.

Vascular Network Arterioles –Y=95% at rest. –Y=100% during activation. –25  m diameter. –<15% blood volume of cortical tissue. Venules –Y=60% at rest. –Y=90% during activation. –25-50  m diameter. –40% blood volume of cortical tissue. Red blood cell –6  m wide and 1-2  m thick. –Delivers O 2 in form of oxyhemoglobin. Capillaries –Y=80% at rest. –Y=90% during activation. –8  m diameter. –40% blood volume of cortical tissue. –Primary site of O 2 exchange with tissue. Transit Time = 2-3 s

BOLD-Based fMRI BOLD = Blood Oxygenation Level-Dependent –Ogawa et al Based on physiological responses related to brain activation (physiology). Relies on detection on the macroscopic level of changes in the microscopic magnetic fields surrounding red blood cells (physics). Red blood cell (RBC) goes from oxygenated to deoxygenated state during functional activation: –Change in its magnetic properties. –Oxyhemoglobin (HbO 2 ) to deoxygemoglobin (Hb). We’ll get to this

From the Biophysics POV “Magnetic susceptibility” (  ) refers to magnetic response of a material when placed in a magnetic field. RBC has change in  when HbO 2 becomes Hb: –Becomes paramagnetic. –Produce an additive internal field that increases local magnetic field. –They become small magnets distorting the field around them. –Susceptibility difference between venous vasculature and surroundings (susceptibility induced field shifts).

What we measure is interaction of: –  of RBCs. –And protons diffusing through these field distortions. 2 pools of water protons that can sense magnetic changes in blood: –Intravascular Protons moving through fields of RBCs. –Extravascular Protons moving thrrough fields that extends into tissue set up by large number of RBCs in vessels. Shortens T 2 * and T 2 of tissue and blood. MRI sequences sensitive to this show a change in image intensity as signal disappears (eg.-EPI).

BUT what causes all this to take place in the first place??

From A Physiology POV Neural Activity Local Consumption of ATP CMRO 2 Local Energy Metabolism CBV CMRGlc CBF BOLD signal results from a complicated mixture of these parameters

CBF, CBV, and CMRO 2 have different effects on HbO 2 concentration: Interaction of these 3 produce BOLD response –They change [Hb] which affects magnetic environment. (delivery of more HbO 2 -> less Hb on venous side if excess O 2 not used) CMRO 2 CBV CBF Local Hb Content Local Hb Content Local Hb Content (extraction of O 2 -> HbO 2 becomes Hb) (more Hb in a given imaging voxel)

But  CBF >>  CMRO 2 (30% >> 5%) –Appears to be more than is necessary to support the small increase in O 2 metabolism. –This “imbalance” is primary physiological phenomenon that gives BOLD signal change. –Results in higher [HbO 2 ] on venous side as compared to [Hb]. –Results in increase in image intensity Less Hb translates to less paramagnetism. 2 models for imbalance of CBF and CMRO 2 –Both assume CBF and CMRO 2 are coupled at some level.

Turner-Grinvald Model CMRO 2 occurs on fine spatial scale. CBF controlled on a coarser scale. CBF & CMRO 2 linearly coupled. Similar magnitude over small spatial region. Focal CMRO 2 change is diluted in measurements of larger tissue volumes –leads to small average increase in CMRO 2. CBF change is more spatially widespread and so is not diluted. Results in apparent imbalance.

Oxygen Limitation Model Brain metabolizes almost all of O 2 that leaves capillary and goes extravascular –Only fraction leaves capillary though. O 2 delivery is limited at rest. When CBF increases:  CBF and CMRO 2 must be non-linearly coupled to counteract this –Large  CBF needed to support small  CMRO 2. Smaller Fraction of O 2 Available for Metabolism O 2 Extraction Capillary Transit Time Blood Velocity (NOT capillary recuitment)

Caveat Assumption: –Spatial extent of alterations in CMRO 2, CMRGl, and CBF is consistent with location of increased neuronal activity. MAYBE (area of metabolic response) > (area of electrical activity). Has been shown in cat auditory cortex –CMRGl increased during synaptic activity without the necessity of electric discharge. –(area of synaptic activity) > (area of electrical activity). –Metabolic- and hemodynamic-based techniques that indirectly measure synaptic activity would conclude larger “active area”. Also, excitatory and inhibitory activity are both metabolically active processes.

Small Vessels Vs. Big Vessels Microvasculature (Arterioles, capillaries, and venules) –We are concerned with signals in and around capillary bed and venules. –Lie more on grey matter adjacent to fissures and in deeper brain structures (sub-cortical). –Shows weaker fMRI signal because of small change in O 2 saturation. –Area seems to be more spatially specific to increased neuronal activity.

Small Vessels Vs. Big Vessels Macrovasculature (large arteries and veins). –We are concerned with signal from veins. –Response seen shifted in time with respect to response seen in small vessels. –Signals are stronger and often dominate fMRI signal Show larger O 2 saturation changes. More likely that veins dominate relative blood volume in any voxel that they happen to pass through because of their size. –Potential for spatial mismatching of site of activation Results in contrast further “downstream”. –Drains active and inactive regions of cortex Hence dillution of Hb occurs. Reduces contribution.

(1)Hyperoxygenation phase (2)Fast response (3)Post-response undershoot (a)Microvascular signal (b)Macrovascular signal Generally %  S micro < %  S macro, BUT not always. Phase micro < Phase macro from stimulus onset

Hyperoxygenation Phase Tightly coupled to neural metabolic activity (theoretically). Due to the imbalance of  CBF >>  CMRO 2 –Results in increase in [HbO 2 ] on venous side of capillary bed. Seen in both small and large vessels. Majority of fMRI studies are based on mapping of this response.

Fast Response Amplitude is independent of stimulus duration –Except for stimuli <2s. Correlates with optical measurements. Could reflect: 1) initial +  CMRO 2 and O 2 extraction from vasculature (  ). 2) decrease in blood flow. 3) or rapid increase in capillary and venous blood volume. Seen in small vessels and capillary bed –Spatially closer to site of increased electrical activity and cellular metabolism. –Mapping of fast/early reponse has potential to overcome some spatial specificity problems in fMRI.

Post-Response Undershoot Amplitude is dependent on duration of stimuli. Very slow to recover (eg.-1 minute!). Likely due to either: 1) unbalanced metabolic energetics (CMRO 2 still elevated). 2) or slow return of increased CBV fraction to basal state after end of stimulus (  ). Larger than fast response. Spatially, correlates more with hyperoxygenation phase. Seen in both large and small vessels (more so small). Does not reflect directly on energy metabolism.

Summary  in local neural activity  CMRO2,  CBF, and  CBV  of RBCs  in local magnetic fields Scanner used to detect these changes Write Nature paper Go to Grad Club …

Next Week … Putting it all in place to do an experiment –MR specifics. Important things to keep in mind when designing an experiment: –Noise. –Artifacts. –How will data be analyzed?