HomeStrokeVol. 46, No. 5Neurovascular Coupling During Cortical Spreading Depolarization and –Depression Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissionsDownload Articles + Supplements ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toSupplementary MaterialsFree AccessResearch ArticlePDF/EPUBNeurovascular Coupling During Cortical Spreading Depolarization and –Depression Leif Østergaard, MD, PhD, DMSc, Jens Peter Dreier, MD, PhD, Nouchine Hadjikhani, MD, PhD, Sune Nørhøj Jespersen, PhD, Ulrich Dirnagl, MD, DMSc and Turgay Dalkara, MD, PhD Leif ØstergaardLeif Østergaard From the Center of Functionally Integrative Neuroscience and MINDLab, Department of Clinical Medicine, Aarhus University, Denmark (L.Ø., S.N.J.); Department of Neuroradiology, Aarhus University Hospital, Aarhus, Denmark (L.Ø.); Center for Stroke Research and Departments of Experimental Neurology and Neurology, Charité Universitätsmedizin, Berlin, Germany (J.P.D., U.D.); Pathophysiology and Cognition Laboratory, Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Department of Radiology, Harvard Medical School (N.H.); Department of Physics and Astronomy, Aarhus University, Aarhus, Denmark (S.N.J.); and Institute of Neurological Sciences and Psychiatry and Department of Neurology, Faculty of Medicine, Hacettepe University, Ankara, Turkey (T.D.). Search for more papers by this author , Jens Peter DreierJens Peter Dreier From the Center of Functionally Integrative Neuroscience and MINDLab, Department of Clinical Medicine, Aarhus University, Denmark (L.Ø., S.N.J.); Department of Neuroradiology, Aarhus University Hospital, Aarhus, Denmark (L.Ø.); Center for Stroke Research and Departments of Experimental Neurology and Neurology, Charité Universitätsmedizin, Berlin, Germany (J.P.D., U.D.); Pathophysiology and Cognition Laboratory, Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Department of Radiology, Harvard Medical School (N.H.); Department of Physics and Astronomy, Aarhus University, Aarhus, Denmark (S.N.J.); and Institute of Neurological Sciences and Psychiatry and Department of Neurology, Faculty of Medicine, Hacettepe University, Ankara, Turkey (T.D.). Search for more papers by this author , Nouchine HadjikhaniNouchine Hadjikhani From the Center of Functionally Integrative Neuroscience and MINDLab, Department of Clinical Medicine, Aarhus University, Denmark (L.Ø., S.N.J.); Department of Neuroradiology, Aarhus University Hospital, Aarhus, Denmark (L.Ø.); Center for Stroke Research and Departments of Experimental Neurology and Neurology, Charité Universitätsmedizin, Berlin, Germany (J.P.D., U.D.); Pathophysiology and Cognition Laboratory, Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Department of Radiology, Harvard Medical School (N.H.); Department of Physics and Astronomy, Aarhus University, Aarhus, Denmark (S.N.J.); and Institute of Neurological Sciences and Psychiatry and Department of Neurology, Faculty of Medicine, Hacettepe University, Ankara, Turkey (T.D.). Search for more papers by this author , Sune Nørhøj JespersenSune Nørhøj Jespersen From the Center of Functionally Integrative Neuroscience and MINDLab, Department of Clinical Medicine, Aarhus University, Denmark (L.Ø., S.N.J.); Department of Neuroradiology, Aarhus University Hospital, Aarhus, Denmark (L.Ø.); Center for Stroke Research and Departments of Experimental Neurology and Neurology, Charité Universitätsmedizin, Berlin, Germany (J.P.D., U.D.); Pathophysiology and Cognition Laboratory, Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Department of Radiology, Harvard Medical School (N.H.); Department of Physics and Astronomy, Aarhus University, Aarhus, Denmark (S.N.J.); and Institute of Neurological Sciences and Psychiatry and Department of Neurology, Faculty of Medicine, Hacettepe University, Ankara, Turkey (T.D.). Search for more papers by this author , Ulrich DirnaglUlrich Dirnagl From the Center of Functionally Integrative Neuroscience and MINDLab, Department of Clinical Medicine, Aarhus University, Denmark (L.Ø., S.N.J.); Department of Neuroradiology, Aarhus University Hospital, Aarhus, Denmark (L.Ø.); Center for Stroke Research and Departments of Experimental Neurology and Neurology, Charité Universitätsmedizin, Berlin, Germany (J.P.D., U.D.); Pathophysiology and Cognition Laboratory, Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Department of Radiology, Harvard Medical School (N.H.); Department of Physics and Astronomy, Aarhus University, Aarhus, Denmark (S.N.J.); and Institute of Neurological Sciences and Psychiatry and Department of Neurology, Faculty of Medicine, Hacettepe University, Ankara, Turkey (T.D.). Search for more papers by this author and Turgay DalkaraTurgay Dalkara From the Center of Functionally Integrative Neuroscience and MINDLab, Department of Clinical Medicine, Aarhus University, Denmark (L.Ø., S.N.J.); Department of Neuroradiology, Aarhus University Hospital, Aarhus, Denmark (L.Ø.); Center for Stroke Research and Departments of Experimental Neurology and Neurology, Charité Universitätsmedizin, Berlin, Germany (J.P.D., U.D.); Pathophysiology and Cognition Laboratory, Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Department of Radiology, Harvard Medical School (N.H.); Department of Physics and Astronomy, Aarhus University, Aarhus, Denmark (S.N.J.); and Institute of Neurological Sciences and Psychiatry and Department of Neurology, Faculty of Medicine, Hacettepe University, Ankara, Turkey (T.D.). Search for more papers by this author Originally published16 Apr 2015https://doi.org/10.1161/STROKEAHA.114.008077Stroke. 2015;46:1392–1401Other version(s) of this articleYou are viewing the most recent version of this article. Previous versions: January 1, 2015: Previous Version 1 Cortical spreading depolarizations (CSDs) are self-propagating waves of transient loss of neuronal transmembrane ion gradients, followed by prolonged suppression of neuronal activity (spreading depression). CSDs emerge spontaneously in animal models of traumatic brain injury,1 subarachnoid hemorrhage,2 and in focal ischemia3 where they are associated with infarct growth.4 In humans, CSDs have been demonstrated in traumatic brain injury,5 subarachnoid hemorrhage,6,7 and malignant hemispheric stroke,8 and they are believed to be the brain mechanism underlying migraine aura.9CSDs are associated with dramatic changes in cerebral blood flow (CBF). During the depolarization phase of CSDs induced in healthy and well-perfused brain tissue in animal models, an early hyperemic response is observed, typically followed by prolonged oligemia after the neuronal repolarization. Despite the initial CBF increase, tissue hypoxia may develop in more distant territories of capillary supply.10,11 Similar CBF changes have been observed in patients with migraine aura.12–14In the injured brain, CSDs can be accompanied by severe initial CBF reduction instead of a CBF increase during the depolarization phase, termed spreading ischemia.15 When this inverse hemodynamic response is observed, the energy-dependent recovery from CSD is delayed in a characteristic fashion, indicating a severe mismatch between oxygen supply and demand2 and a high risk of tissue damage.16 In rat and cat models of focal ischemia, CSD-related CBF transients range from monophasic, positive CBF responses in peri-ischemic tissue, over biphasic transients in mildly ischemic tissue, to negative CBF transients in more severe ischemia.17–19During CSD in animal models, capillary flow patterns become severely disturbed.10,20,21 The passage of a CSD causes erythrocytes in some capillaries to reduce their speed, whereas other capillaries reveal 4-fold increases in flow or higher.21 Flow cessations during CSD have been observed to affect those capillaries, where pre-CSD erythrocyte velocity and flux were already low, more than those with high pre-CSD velocity and flux,20 again suggesting severe maldistribution of blood flow among capillaries during CSD.We recently showed that capillary transit time heterogeneity (CTH) reduces oxygen extraction efficacy, and thereby limits the oxygen utilization that can be supported for a given CBF and tissue oxygen tension (PtO2).22 Therefore, flow-metabolism coupling mechanisms would be expected to adjust CBF and CBF responses to compensate for the reduced oxygen extraction efficacy downstream, see Figure 1A. Below, we describe this capillary dysfunction and its implications for net oxygen extraction and ATP production. We then analyze how CBF would be expected to change to maintain flow-metabolism coupling if CTH increases in relation to a CSD and then remains high during the subsequent cortical depression. Although changes in capillary patency/resistance are likely to affect capillary flow patterns in the injured brain (see Figure 1B), data regarding CTH during CSD are sparse. The purpose of this report is therefore to present predictions that can be tested in future experiments and clinical studies to support or rule out CTH-related effects in the coupling of CBF to tissue metabolism during CSD in the normal and injured brain.Download figureDownload PowerPointFigure 1. A, The importance of capillary flow patterns (yellow arrows) for the efficacy of oxygen extraction. Intravascular colors indicate blood saturation (red, fully oxygenated; darker blue, more deoxygenated) and surrounding darker blue colors indicate lower tissue oxygen tension—adapted from Østergaard et al.23 Copyright ©2013, The Authors (see: http://creativecommons.org/licenses/by-nc-sa/3.0/). In the resting, normal brain (top), erythrocyte velocities vary greatly among capillaries, with little oxygen being extracted from fast-flowing blood.22 As cerebral blood flow (CBF) increases (right), capillary flow patterns homogenize in parallel.22 This phenomenon reduces the functional shunting that otherwise occurs when erythrocytes pass through capillaries at short transit times. Although the mean transit time (MTT) of blood through the capillary bed is related to CBF through the central volume theorem (MTT=CBV/CBF, where CBV is the capillary blood volume), capillary transit time heterogeneity (CTH) indicates the distribution of capillary transit times relative to this mean (eg, in terms of their standard deviation). Both MTT and CTH are measured in seconds, and the degree of functional shunting generally increases as CTH approaches MTT.24Bottom, Capillary dysfunction, characterized by elevated CTH relative to MTT, and failure of capillary flow patterns to homogenize during hyperemia. The conditions may be the result of pericyte dysfunction, changes in blood viscosity or capillary wall morphology, or external capillary compression. These changes hinder the redistribution of blood across the capillary bed, with less affected capillary paths tending to act as functional shunts for oxygenated blood. Biophysically, the only means of attenuating the accompanying oxygen loss is to reduce transit times across all capillaries, that is, to attenuate CBF and CBF responses.25 The accompanying reduction in net oxygen supply reduces tissue oxygen tension as cells continue to use oxygen, increasing blood-tissue concentration gradients and net oxygen extraction.25 With flow responses attenuated, oxygen extraction can be increased from the 30% of normal brain up to near unity, and normal brain function maintained although capillary dysfunction becomes more severe.25 It is important to note that current state-of-the-art algorithms to generate maps of transit time-related metrics based on perfusion-weighted magnetic resonance imaging or computed tomography cannot distinguish changes in tracer retention caused by prolonged MTT (reduced CBF) from changes caused by capillary flow disturbances (capillary dysfunction).25 The effects of CTH must therefore be separately modeled to ascertain whether clinical signs of ischemia are indeed caused by limited blood supply or by capillary dysfunction.26B, The acute changes in capillary morphology that accompany conditions in which CSD are common. In traumatic brain injury (TBI; i), massive swelling of the perivascular astrocytic end feet (AE) and flattening or compression of the capillary lumen (L, indicated by red arrows) is observed.27 A cortical capillary (top) and a white matter capillary (bottom) from 2 TBI victims undergoing surgery.27 Reproduced from Bullock et al27 with the permission of BMJ Publishing Group Ltd. Copyright ©1991, British Medical Journal. In experimental subarachnoid hemorrhage (SAH; ii), swelling of astrocytic end-feet (AE) and endothelial protrusions (arrows) are observed here as early as 1 hour after hemorrhage.28 Reprinted from Sehba and Friedrich28 with the permission of Springer Science and Business Media. Copyright ©2013, Springer-Verlag Wien. Ischemia (iii) can cause pericytes to constrict and compress the capillary lumen. Top, A horseradish peroxidase (HRP)–filled capillary segment interrupted by a constriction, merged with an image that highlights pericyte α-smooth muscle actin in green (bar indicates 10 μm). Bottom, Multiple disruptions of a HRP-filled capillary, highlighting that pericyte constrictions may prevent reperfusion (no-reflow) or disturb capillary flow patterns after recanalization (bar indicates 20 μm).29 Ischemia data were obtained in mice. Reprinted from Yemisci et al29 with the permission of Macmillan Publishers Ltd. Copyright ©2009, Nature America, Inc.Predicted Changes in CBF During Episodes of Increased CTH in Normal BrainIn the following, we assume that flow-metabolism coupling adjusts CBF to meet the metabolic needs of tissue whenever possible. We therefore examine which compensatory CBF changes are required to support various types of tissue metabolism if some erythrocytes pass through the capillary bed with flow velocities that are too high to permit complete extraction of their oxygen load. Direct observations of cortical capillaries in normal brain show that capillary flow velocities are heterogeneous during rest, but homogenize during functional activation and the accompanying hyperemia.22 Such reductions of CTH would be expected to secure efficient oxygen extraction during hyperemia22 as illustrated in Figure 1A (top right panel). A recent study suggests that both CBF and CTH are indeed controlled by capillary pericytes during functional activation.30Figure 2A illustrates how CBF must be adjusted to compensate for an increase in CTH to support a constant rate of oxygen metabolism. The figure also indicates the parallel change in oxygen extraction fraction (OEF). We first consider a mild increase in CTH, defined here as a CTH increase for which the resulting OEF reduction can be compensated by an increase in CBF.Download figureDownload PowerPointFigure 2. A and B, The changes in cerebral blood flow (CBF) that must accompany mild (A) and moderate (B) capillary dysfunction to compensate for the changes in oxygen extraction fraction (OEF) caused by the parallel increase in capillary transit time heterogeneity (CTH). Mild capillary dysfunction is operationally defined by the property that the accompanying reduction in OEF can be compensated by an increase in CBF alone, whereas the CTH increase of moderate capillary dysfunction, by definition, exceeds the threshold for which flow suppression becomes necessary to secure sufficient net oxygen extraction.25 The horizontal dotted lines indicates this flow suppression threshold (FST). We assumed that capillary flow disturbances, such as those described at the onset of cortical spreading depolarizations (CSDs) in Refs. 10,20,21, develop over a brief period of time. During an increase to moderate CTH (B), capillary dysfunction is therefore initially mild according to the definition above, and transient hyperemia is therefore expected as CTH pass the threshold of flow suppression (highlighted by red circles). C and D, The modulation of CBF responses required to meet metabolic challenges before and after the onset of mild and moderate capillary dysfunction, respectively. The equally sized gray boxes before and after the increase in CTH are meant to illustrate how CBF responses to a repeated stimulus are modulated to maintain flow-metabolism coupling. Note that CBF responses are amplified by mild capillary dysfunction, but attenuated by moderate capillary dysfunction, relative to CBF responses before the onset of capillary dysfunction. The larger gray box is meant to represent the energy expenditure of repolarization, assuming that this occurs immediately after a depolarization-induced increase in CTH.A compensatory CBF increase shortens capillary transit times, and thereby increases CTH-related functional shunting of oxygenated blood. Hyperemia can therefore only compensate for mild increases in CTH. For higher CTH, attenuation of CBF becomes mandatory to meet the metabolic demands of the tissue.22,25 Note that an initial, transient hyperemia is expected as CTH increases beyond the threshold at which CBF starts to be attenuated.Predicted Changes in CBF to Support Increased Metabolic Demands When CTH Is ElevatedAbove, we considered how neurovascular coupling supports resting brain metabolism as CTH increases. We now consider how CBF must be adjusted to meet the additional metabolic demands of functional activation or repolarization after CSD. Figure 2C and 2D illustrates this for CTH increases to levels below and above the threshold of flow suppression, respectively (compare Figure 2A and 2B).Figure 2C shows how functional hyperemia is exaggerated during a mild increase in CTH: by definition, metabolic needs can still be met by hyperemia alone, and larger CBF responses hence compensate for the lower OEF. In Figure 2D, CBF responses are illustrated both in the normal state and after a moderate increase in CTH. Note how both baseline CBF and CBF responses to certain metabolic challenges are reduced after such a CTH increase. Note that flow suppression is expected as a result of flow-metabolism coupling and therefore should depend on the rate of oxygen utilization in the tissue. Accordingly, the suppression of activation-related CBF is expected to occur at a lower CTH level than the CTH level at which resting CBF must be suppressed. We review evidence of this phenomenon in migraine below.Observed Changes in CBF, Metabolism, and Blood Oxygenation During CSD in Normal Mouse and Rat Brain: Migraine AuraFigure 3A shows typical CBF responses during the passage of a CSD in mouse brain.31 Depolarization is accompanied by a small increase, followed by a drop, in CBF. If we assume that the long-lasting capillary flow disturbances caused by a CSD10,20,21 give rise to elevated CTH during the course of the initial depolarization, then the observed course of CBF changes (a slight hyperperfusion followed by hypoperfusion) is consistent with the prediction in Figure 2D. The subsequent repolarization is energetically demanding, and the transient increase in CBF in the mouse model of CSD indeed coincides with the restoration of membrane potentials, after which CBF returns to a level consistent with high CTH, that is, continued capillary flow disturbances. The prediction that CSD-related CBF-changes are coupled—through CTH—to the metabolic needs of the tissue may help explain the distinct appearance of CBF response to subsequent CSDs,11 elicited while CTH remains high; see Figure 3B. Note that the shapes of the first and second CBF response (here in mice) are nearly identical except for the CBF drop during the first CSD, which we ascribe to neurovascular coupling during a gradual increase in CTH.Download figureDownload PowerPointFigure 3. A, Recorded changes in electric potentials (top) and cerebral blood flow (CBF; bottom) in mice.31 Note the resemblance of the CBF trace to the prediction Figure 2D, including signs of early hyperemia during the depolarization phase. Reprinted from Ayata et al31 with the permission of Macmillan Publishers Ltd. Copyright ©2004, The International Society for Cerebral Blood Flow and Metabolism. B, The CBF responses to 2 subsequent cortical spreading depolarizations (CSDs) elicited in mice. Reprinted from Yuzawa et al11 with the permission of Macmillan Publishers Ltd. Copyright ©2011, ISCBFM. Note that the CBF drop observed during the first CSD is consistent with a compensatory vasoconstriction to compensate for a moderate increase in capillary transit time heterogeneity (CTH).10,20,21 Reported changes in capillary flow patterns are long-lasting, and the attenuation of CBF would therefore be expected to outlast the subsequent repolarization phase and be present at the time of the next CSD being induced. Indeed, CBF responses during 2 subsequent repolarizations seem identical, attributable to the metabolic demands of restoring membrane potentials. C and D, CBF changes during CSD in mouse and rat, respectively.31 The dark gray boxes highlight the amplitude of CSD-related CBF changes in the 2 species, whereas the light gray boxes highlight the amplitude and duration of CBF changes during the subsequent repolarization phase. Adapted from Ayata et al31 with the permission of Macmillan Publishers Ltd. Copyright ©2004, The International Society for Cerebral Blood Flow and Metabolism. As discussed in the text, the differences in CBF over time may be as a result of the extent of capillary flow disturbances (CTH) induced by the CSD in the 2 species or to differences in their oxygen utilization during anesthesia. Note that the extent of CBF suppression after CSD (60% in mouse, 25% in rat) is reflected in the subsequent repolarization-related hyperemia. The CBF response in mouse is attenuated and prolonged, consistent with a large and possibly critical CTH increase (see Figure 6), whereas the flow response in rats is much less suppressed.The extent of CBF suppression required to maintain flow metabolism during and after CSD is predicted to depend on both CTH and oxygen utilization. The prolonged 60% CBF suppression after repolarization in mouse (Figure 3C) is therefore predicted to be the result of a pronounced increase in CTH after CSD, high oxygen utilization in the anesthetized mouse brain, or both. The 25% reduction in CBF before and after repolarization in rat (Figure 3D),31 in turn, is predicted to reflect a more modest CSD-related CTH increase, a lower oxygen utilization in the anesthetized rat brain, or both. During repolarization, the extent of CBF suppression is expected to depend mainly on oxygen extraction efficacy, and thereby the degree of capillary flow disturbance. In normal brain, an inverse relationship is therefore expected between the amplitude of the CBF response during repolarization and the CBF reduction before and after repolarization, as indeed observed in Figure 3C and 3D.31The duration of post-CSD oligemia varies. This CBF reduction may reflect reduced cortical activity (see below), but may also reflect adjustments of CBF and CBF responses to account for prolonged capillary flow disturbances. This prediction is consistent with findings that responses to increased CO2 levels, basal forebrain stimulation, and direct arteriolar application of vasoactive substances are reduced or disappear12 during post-CSD oligemia and that evoked PtO2 changes are inversely correlated with the degree of baseline hypoxia.32 As capillary flow patterns normalize, metabolic needs can be met at a higher CBF for a given PtO2. Such indirect evidence of improved capillary flow patterns was observed by Fordsman et al who reported that blockage of 20-hydroxyeicosatetraenoic acid synthesis ameliorated post-CSD hypoperfusion, although having modest effects on the degree of hypoxia.32 These results are consistent with a role of 20-hydroxyeicosatetraenoic acid in the capillary flow disturbances and neurovascular coupling after CSD, possibly through its action as a powerful pericyte constrictor.30Migraine AuraIn patients with migraine aura, CBF responses to vasodilatory stimuli, such as CO2, are attenuated during the spreading oligemia,12 consistent with neurovascular coupling during a state in which CTH is elevated (Figure 2D). Hypoperfusion typically persists as patients develop headache, until patchy areas of hyperperfusion appear in previously hypoperfused areas after 2 to 6 hours.33 The late hyperperfusion is consistent with the CBF change predicted to occur as a large increase in CTH resolves.Blood oxygen level–dependent (BOLD) functional magnetic resonance imaging localizes brain activity via the accompanying reductions in blood deoxyhemoglobin concentration [dHgb]. During functional activation, this effect gives rise to increased signal intensity in activated brain areas compared with resting conditions. Both increases in CBF and reductions of OEF tend to reduce [dHgb], and small CTH increases (Figure 2C) are therefore predicted to elevate BOLD signal amplitudes for a given tissue metabolism. Conversely, reductions in CBF and elevated OEF (Figure 2D) both attenuate BOLD signal amplitudes for a given tissue oxygen metabolism. Small increases in CTH would therefore be predicted to increase both baseline BOLD levels and the amplitude of task-related BOLD signal changes, relative to a condition of normal CTH. As CTH increases further, flow suppression is expected to affect task-related hyperemia (where metabolic demands are higher) first, then the baseline BOLD amplitude. Figure 4 illustrates how task-related BOLD changes observed during the onset and early phases of migraine aura14 are remarkably consistent with these predictions.Download figureDownload PowerPointFigure 4. A, A functional magnetic resonance imaging (fMRI) slice oriented perpendicular to the calcarine fissure in a subject with inducible migraine aura that affects the right hemisphere.14B and C, Blood oxygen level–dependent (BOLD) signal changes to a visual stimulus (alternating, 32 second epochs of radial flickering checkerboard vs black screen with a fixation point) recorded in single image voxels (brown arrows) within homologous areas of the right (B) and left (C) occipital lobe during the onset of migraine aura (onset time indicated by white arrow). In the right hemisphere voxel, note the initial increase in activation-induced BOLD signal amplitude (time period indicated by green line) followed by an increase in baseline BOLD signal level (indicated by the left-most yellow line). These changes are consistent with a mild increase in CTH, in that both elevated cerebral blood flow (CBF) and reduced oxygen extraction fraction (OEF; cf. Figure 2A) cause tissue deoxyhemoglobin levels to fall and, hence, BOLD fMRI signal amplitude to increase. C, Simultaneous signal changes in the unaffected hemisphere: these illustrate typical signal fluctuations over time. Next, coinciding with the onset of aura symptoms,14 stimulus-induced BOLD changes become attenuated (indicated by blue line)—and even disappear during the time period assumed to mark the passing of a cortical spreading depolarization and -repolarization.14 In B, note that BOLD signal intensities during stimulus-off periods (indicated by the right-most yellow line) return to the baseline recorded at the onset of the imaging session. Although stimulus-induced BOLD responses remain suppressed, consistent with either suppressed neuronal activation or moderately elevated CTH, these findings do not support that resting CBF is suppressed to compensate for a moderate increase in CTH as defined in Figure 2B. D, The stimulus-off BOLD signal intensities fell below preaura levels in areas V3a and V3 in the same subject, consistent with a larger (moderate) CTH increase and flow-suppression. Signal source analysis revealed that the spreading BOLD changes had in fact originated in V3a in this subject,14 and our analysis thus suggests that the aura phenomenon arises under conditions with a more severe oxygen supply–demand imbalance than observed as the wave subsequently spreads across the visual cortex. It should be noted that neuronal responses to the visual stimulus are likely to be affected by the spreading depression of activity. It is therefore difficult to ascertain to which degree stimulus-induced BOLD responses are diminished as a result of changes in neuronal or in capillary function, respectively. Adapted from Hadjikhani et al14 with permission of the Authors. Copyright ©2001, National Academy of Science, USA. Authorization for this adaptation has been obtained both from the owner of the copyright in the original work and from the owner of copyright in the translation or adaptation.ATP Production at Elevated CTH: Differential Effects of Capillary Dysfunction on Oxygen and Glucose ExtractionAfter the initial, near-complete breakdown of ion gradients during a CSD, large amounts of ATP are needed to reestablish normal ion distributions.34Figure 5 shows the course of CBF, and the tissue levels of glucose and ATP, during the passage of CSDs in rat brain.35 Autoradiographic measurements during CSD have revealed 100% increases in the uptake of glucose analogs into the affected tissue.38 These data are widely interpreted as indicating a 100% increase in neuronal energy turnover during the restoration of ion gradients after CSD. The ATP production during the repolarization phase is, however, heavily dependent on oxygen availability, keeping in mind that a 100% CBF increase does not ensure a 100% increase in tissue oxygen availability. Under aerobic conditions, one glucose molecule generates 29 to 30 ATP molecules by oxidative phosphorylation, but if oxygen availability is limited, lactate pr