Brain metabolism and blood flow are tightly coupled with neuronal activity. Changes in
neuronal activity result in the modulation of glucose consumption by neurons. Both
glucose and lactate levels return to their baseline instantly as neuronal activity
ceases, a phenomenon known as neurometabolic coupling. Given the limited energetic
reserves in the central nervous system, neuronal activity heavily relies on the finely
regulated supply of glucose from the bloodstream. However, the dynamic increase in
cerebral blood flow (CBF) during neuronal activation far exceeds the increase in
oxidative metabolism. This relative hyperemic response ensures an increased oxygen
gradient between blood vessels and tissue, providing ample oxygen supply. The close
temporal and regional link between changes in neuronal activity and CBF increase is
referred to as neurovascular coupling (NVC) and involves a complex cascade of events.
Neurotransmitters, such as glutamate, released at synapses bind to receptors on neurons
and astrocytes, leading to the release of various chemical mediators, like nitric oxide
and prostaglandins, which directly act on arterial smooth muscle tone. More complex and
incompletely understood signaling pathways, including Na+ and Ca2+-mediated astrocyte
signaling mechanisms, are also presumed to contribute to NVC.
The tight relationship between neuronal activity and both regional blood flow and
metabolism has provided the basis for non-invasive functional brain imaging methods,
including positron emission tomography (PET) and functional magnetic resonance imaging
(fMRI). PET using [18Fluor ]-fluorodeoxyglucose (FDG) is a technique based on the
accumulation of metabolized FDG (i.e., FDG-6-phosphate) in the astrocyte-neuron complex,
reflecting the level of glucose consumption. Since the seminal works of Sokoloff et al.,
glucose utilization is considered a valid, accurate, and quantitative indicator of the
level of local neuronal activity within the brain. In contrast, fMRI, which relies on the
blood oxygen level-dependent (BOLD) signal, provides indirect information about neuronal
activity by investigating perfusion-related changes coupled with neuronal activity. In
areas of increased CBF due to modulations in neuronal activity, oxygen delivery exceeds
the rate of oxygen utilization, inducing a local increase in the oxy-/deoxy-hemoglobin
ratio. This leads to a detectable increase in the magnetic-susceptibility weighted MRI
signal.
One of the earliest and still recognized clinical applications of fMRI has been
preoperative functional mapping of the primary sensorimotor cortex in patients with brain
tumors. This technique has significantly impacted surgical planning, often enabling more
aggressive approaches than those considered without functional localization. fMRI has
also been increasingly used in the presurgical evaluation of patients with vascular or
epileptogenic lesions. However, despite the growing use of BOLD fMRI in patients with
brain lesions, this technique has major limitations that must be considered when
interpreting fMRI results in such populations. The main limitation is the impairment of
BOLD signal changes due to lesion-related loss of normal vascular coupling with neuronal
activity, a phenomenon referred to as neurovascular uncoupling (NVU). This can result in
false-negative or false-positive results in critical eloquent cortex. If neuronal
activity is preserved in diseased but viable cortex, NVU is presumed to occur due to
astrocytic, neurotransmitter, or vascular dysfunction.
NVU has been mainly reported in patients with high-grade glial tumors and meningiomas. In
such patients, the volume of task-based fMRI signal increases has been shown to be
reduced adjacent to the tumor compared to homologous fMRI signal changes in the
contralesional hemisphere, despite the absence of neurological deficit. In line with
experimental data in healthy subjects showing that BOLD signal may decrease as cerebral
blood volume (CBV) increases, impaired cerebrovascular reactivity (CVR) in brain tumor
patients may be explained by changes in local perfusion. In hypervascularized tumors such
as high-grade gliomas and meningiomas, local hyperperfusion has been suggested to explain
the decreased BOLD signal on task-based fMRI. However, recent studies have demonstrated
that NVU may also occur in low-grade gliomas. Given the absence of hyperperfusion in this
tumor type, different mechanisms need to be considered. In low-grade gliomas, the
observed NVU is currently thought to be, at least in part, due to disruption of
astrocyte-vascular coupling (gliovascular uncoupling). Patients with arteriovenous
malformations may exhibit impaired peri-nidal cerebrovascular reserve due to high-flow
shunting, making perfusion-dependent mapping signals unreliable. Epilepsy patients may
also exhibit regional impairment of CVR due to dramatic increases in brain metabolism and
CBF during the ictal period, disruption of the brain-blood barrier, and an acute loss of
cerebral pressure autoregulation.
According to previous research, CVR can be studied through the "hypercapnia challenge"
during fMRI recordings, including breath-hold fMRI (BH fMRI) and carbogen inhalation
fMRI. Hypercapnia is a potent vasodilator that increases the BOLD baseline signal by
detecting an increase in tissue oxygenation resulting from increases in CBF while
oxidative metabolism demands are considered to remain constant. However, the influence of
hypercapnia on neural activity and neurometabolic/neurovascular couplings is not well
understood and remains debated. In practice, areas of reduced or absent
hypercapnia-induced increase in fMRI signal on CVR maps compared to homologous
contralateral activation are assumed to indicate NVU. Recent studies suggest potential
advantages in using resting-state (rs) fMRI as a preoperative technique. rs-fMRI is a
functional neuroimaging technique that allows the measurement of spontaneous brain
activity in patients at rest. Spontaneous BOLD signal fluctuations are highly correlated
in distinct and long-ranged brain regions, indicating functional connectivity within
specific and highly organized neuroanatomical networks. Functional connectivity studies
have also demonstrated a high degree of spatial correlation between rs-fMRI functional
brain connectivity and those studied during a hypercapnia challenge. Interestingly,
recent research suggests that rs-BOLD signal may be impaired in patients in whom
task-based increases in fMRI signals are reduced or absent due to NVU. Therefore,
alterations in functional brain connectivity studied with rs-fMRI might provide insights
into the presence of NVU as studied with CVR during hypercapnia. Such findings would be
of interest in clinical practice as they could avoid the need for CVR-mapping with a
hypercapnia challenge.
