Neurovascular Coupling

Units 1 and 2 established how changes in the concentration of oxy- and deoxy-haemoglobin can be recovered from light attenuation, using the Modified Beer-Lambert Law. Neither unit, however, addressed why those concentrations change in the first place. This lesson introduces neurovascular coupling, the physiological chain linking neural activity to the vascular response that follows it, and the reason light attenuation in tissue carries any information about brain activity at all.

The lesson builds this chain and uses it to resolve a point that trips up almost everyone learning fNIRS for the first time: activation increases oxygen consumption, yet the fNIRS signal most associated with oxygen loss, HbR, decreases rather than increases. By the end of this lesson, that result will follow directly from the physiology rather than needing to be taken on faith.

1. The Neurovascular Unit

The brain has essentially no local energy reserve. It accounts for roughly 2% of body mass but consumes around 20% of the body’s resting oxygen supply, and it cannot store meaningful amounts of glucose or oxygen for later use. Every increase in local neural activity has to be met, in near real time, by an increase in local oxygen and glucose delivery. The system responsible for this matching is the neurovascular unit: neurons, astrocytes, the endothelial cells lining blood vessels, and the smooth muscle and pericytes that control vessel diameter, all working as a coordinated local circuit.

In outline, the sequence runs as follows. Synaptic activity releases neurotransmitters and requires ion pumps to restore resting gradients, both of which consume ATP. This local rise in metabolic activity triggers signalling, substantially mediated by astrocytes, whose processes contact both synapses and the walls of nearby blood vessels. This signalling relaxes the smooth muscle around local arterioles, which dilate; blood flow into the capillary bed downstream increases within one to two seconds of the triggering activity, and continues to rise for several seconds more. This is the process most people mean when they say “neurovascular coupling,” and it is the reason a hemodynamic signal can be used as a proxy for neural activity, despite not measuring neurons directly.

2. Delivery, Consumption, and Extraction of Oxygen

Reasoning precisely about what happens when blood flow increases requires three quantities. The first two were introduced informally in Unit 1: cerebral blood flow (CBF), the volume of blood delivered to a tissue per unit time, and cerebral metabolic rate of oxygen (CMRO₂), the rate at which that tissue consumes oxygen. The third, which links them, is the oxygen extraction fraction (OEF): the proportion of the oxygen delivered by the blood that the tissue actually removes and consumes.

These two quantities are related through Fick’s principle, briefly introduced in Unit 1. Stated precisely, Fick’s principle says that the rate of oxygen consumption equals blood flow multiplied by the difference in oxygen content between the blood arriving (arterial) and the blood leaving (venous) the tissue:

OEF is the quantity that connects delivery and consumption to oxygenation, which is what fNIRS actually measures, and to oxygen saturation (StO₂) in particular, which represents the fraction of haemoglobin carrying oxygen. Venous oxygen saturation is, to a close approximation, the complement of OEF, since a higher extraction fraction leaves less oxygen in the blood as it exits the tissue. OEF is therefore the variable that sets how oxygenated the blood in a vascular bed is at any moment, and consequently how much of the local haemoglobin pool is HbO₂ versus HbR.

Before we conclude, it is worth noting that Fick’s principle and the extraction fraction derived from it are steady-state relationships, i.e., they describe a tissue mass balance once inflow, outflow, and consumption have reached equilibrium, not an instantaneous relationship valid at any moment during, e.g., a rapidly changing transient. During the first several seconds after stimulus onset, CBF, CBV, and OEF are all changing together, and blood requires a finite transit time to move through the vascular bed. The comparison used throughout this lesson, resting state versus sustained activation, is a comparison of two such steady states, which is exactly how the Fox and Raichle measurements below were collected: each PET value is a block average, not an instantaneous snapshot. Modelling the dynamic transition between these two states requires a more elaborate framework. Two well-known examples are the balloon model (Buxton, Wong & Frank, 1998), which models blood volume and deoxyhaemoglobin content directly as functions of time, and the Windkessel model (Mandeville et al., 1999), which treats the venous compartment as a passive, compliant vessel that inflates under arteriolar pressure and drains more slowly than it fills, accounting for the delayed return to baseline often seen in blood volume after a stimulus ends. Both are widely used in the fNIRS literature specifically, since fNIRS’s high temporal resolution is well suited to testing exactly this kind of dynamic prediction. We will not need this level of detail here, but it is worth knowing these models exist before we discuss the time course of the hemodynamic response itself.

3. Changes in Blood Flow And Metabolism During Activation

Given the relationship above, a natural assumption is that CBF and CMRO₂ increase by roughly the same proportion during activation, leaving OEF, and therefore oxygenation, largely unchanged. However, this assumption is incorrect, and the size of the mismatch between CBF and CMRO₂ during activation is key for interpreting the fNIRS signal.

The physiological reason the brain over-delivers oxygen during activation is not fully settled. One account, formalised by Buxton and Frank (1997), holds that oxygen diffusion from capillary blood into mitochondria is itself a limiting step: diffusion depends on the concentration gradient between capillary blood and tissue, so a large increase in flow is needed to raise that gradient enough to support even a modest increase in consumption. A second line of evidence, from Fox, Raichle, Mintun and Dence (1988), found that during activation, glucose uptake and blood flow both rose far more (51% and 50%, respectively) than oxygen consumption (5%), consistent with a transient contribution from non-oxidative glycolysis: glucose metabolised without a matched increase in oxygen use. This remains an active area of study. What matters for interpreting fNIRS is that the CBF/CMRO₂ mismatch itself, whatever its ultimate explanation, is well established and highly reproducible.

4. Consequences to Haemoglobin Concentrations

If a brain region is consuming more oxygen during activation, CMRO₂ has gone up. HbR reflects deoxygenated haemoglobin, the product of oxygen being unloaded from blood into tissue. The standard fNIRS finding, however, is that HbR falls during activation rather than rising.

The explanation for this is the mismatch discussed in the previous section. CMRO₂ rises, but only modestly. CBF rises far more. Because OEF is the ratio between how much oxygen is consumed and how much is delivered, and delivery increases much faster than consumption, OEF falls during activation, even though absolute oxygen consumption is higher than at rest. A smaller fraction of a larger oxygen supply is being extracted. The blood leaving the activated tissue is therefore more oxygen-rich than it was at rest, and the local haemoglobin pool tips further toward HbO₂ and away from HbR.

Using Fox and Raichle’s own figures as a worked illustration: if OEF at rest is around 40%, and CMRO₂ rises 5% while CBF rises 29%, the new OEF is approximately 0.40 × (1.05 / 1.29) ≈ 0.33. Oxygen extraction drops by roughly a fifth in relative terms, purely because delivery outran consumption. Combined with the accompanying increase in local blood volume, as more of the local vascular bed fills with fresh, oxygenated blood, this produces the canonical fNIRS response: HbO₂ rises, HbR falls, and both changes are consistent with a genuine increase in oxygen consumption.

5. From Physiology to a Measurable Signal

The chain described in this lesson, from synaptic activity, through astrocyte-mediated vascular signalling, to a CBF increase that outpaces the CMRO₂ increase, to a falling OEF and a rising local blood volume, together determine the balance of HbO₂ and HbR in the tissue beneath an fNIRS probe. These are exactly the concentration changes that Units 1 and 2 showed how to recover from light attenuation through the Modified Beer-Lambert Law.

This is why fNIRS functions as a neuroimaging tool at all. A local increase in neural activity triggers a well-characterised, reproducible physiological chain that changes local haemoglobin concentration in a specific and predictable direction. Measuring that concentration change, which is what fNIRS does, therefore provides a usable, if indirect, window onto local neural activity.

📚 Further Reading & Key References

1. Fox, P.T. & Raichle, M.E. (1986). “Focal physiological uncoupling of cerebral blood flow and oxidative metabolism during somatosensory stimulation in human subjects.” Proceedings of the National Academy of Sciences, 83(4), 1140–1144. [doi] — The founding demonstration of the CBF/CMRO₂ mismatch that this lesson is built around.
2. Attwell, D. & Iadecola, C. (2002). “The neural basis of functional brain imaging signals.” Trends in Neurosciences, 25(12), 621–625. — A compact, accessible entry point into the signalling mechanisms behind neurovascular coupling.
3. Attwell, D., Buchan, A.M., Charpak, S., Lauritzen, M., MacVicar, B.A. & Newman, E.A. (2010). “Glial and neuronal control of brain blood flow.” Nature, 468(7321), 232–243. [doi] — The major mechanistic review of astrocyte- and neuron-mediated vascular signalling.
4. Iadecola, C. (2017). “The neurovascular unit coming of age: a journey through neurovascular coupling in health and disease.” Neuron, 96(1), 17–42. — The most comprehensive modern review; also covers what changes in disease and ageing, relevant background for Unit 6.
5. Buxton, R.B. (2010). “Interpreting oxygenation-based neuroimaging signals: the importance and the challenge of understanding brain oxygen metabolism.” Frontiers in Neuroenergetics, 2, 8. — A thoughtful discussion of why CBF outpaces CMRO₂, and what remains genuinely unresolved about this question.
6. Leenders, K.L., Perani, D., Lammertsma, A.A., et al. (1990). “Cerebral blood flow, blood volume and oxygen utilization: normal values and effect of age.” Brain, 113(1), 27–47. — Source for the normative resting OEF range cited in Section 2.
7. Buxton, R.B. & Frank, L.R. (1997). “A model for the coupling between cerebral blood flow and oxygen metabolism during neural stimulation.” Journal of Cerebral Blood Flow & Metabolism, 17(1), 64–72. — The diffusion-limitation account referenced in Section 3.
8. Buxton, R.B., Wong, E.C. & Frank, L.R. (1998). “Dynamics of blood flow and oxygenation changes during brain activation: the balloon model.” Magnetic Resonance in Medicine, 39(6), 855–864. — The standard dynamic (non-steady-state) generalisation referenced in Section 2, relevant again in Lesson 3.3.
9. Fox, P.T., Raichle, M.E., Mintun, M.A. & Dence, C. (1988). “Nonoxidative glucose consumption during focal physiologic neural activity.” Science, 241(4864), 462–464. — The non-oxidative glycolysis account referenced in Section 3.
10. Mandeville, J.B., Marota, J.J.A., Ayata, C., Zaharchuk, G., Moskowitz, M.A., Rosen, B.R. & Weisskoff, R.M. (1999). “Evidence of a cerebrovascular post-arteriole Windkessel with delayed compliance.” Journal of Cerebral Blood Flow & Metabolism, 19(6), 679–689. — The Windkessel model referenced in Section 2, extensively developed in later fNIRS dynamic-modelling work.