Pitfalls in NIR-spectroscopy and functional MRI for measuring brain oxidative metabolism with special remarks on A-V shunts

Minoru Tomita, Yutaka Tomita, Takashi Osada, Miyuki Unekawa, Haruki Toriumi, Norihiro Suzuki

Research output: Contribution to journalArticlepeer-review


The interplay between function and oxidative metabolism of the brain has been extensively investigated by spectroscopy (NIRS) and fMRI of metabolic changes responding to somatosensory stimuli. However, concerning optical spectroscopy, we have pointed out that the Lambert-Beer law does not hold during brain activation, since the extinction coefficient of the blood alters due to flow-dependent changes in light scattering by RBC (flow effect: Tomita et al. (2006) NeuroImage 33,1; ibid., 13). In addition, the extinction coefficient per se varies with RBC oxygenation: the concave shape of RBC tends to be spherical with a diameter of ca. 7.2 micrometer, whereas the same RBC become flat and flaccid with a diameter of ca. 7.7 micrometer when deoxygenated. Light scattering at such membranes is significantly affected. The inconstant extinction coefficient jeopardizes quantification of oxy-/deoxyhemoglobin in the tissue. In addition, the Bohr effect in arterial blood produced by respiratory changes in human subjects, e.g., breath-holding or hyperventilation with powerful somatosensory stimuli, shifts the oxygen dissociation curve of hemoglobin, changing cerebral blood oxygenation without any changes in brain metabolism. This paper examines the effect of arterio-venous shunting blood, which could augment dissociation between blood oxygenation and brain oxygenation in both NIRS and fMRI, since the channels transport oxygenated blood from the arteries to the veins without gas exchange. We used 50 rats with a closed cranial window and measured velocities of FITC-labeled RBC in the 50 micrometer deep cortical microvasculature with high speed laser scanning confocal fluorescence microscopy. We found labeled RBC velocity varied widely in capillaries selected with the aid of FITC-dextran staining. The RBC velocity distribution revealed a group of outlier higher velocities (p<0.01) from the normal group which must belong to a different category of vessels: the RBC velocities were extraordinarily high, nevertheless the channels had the same diameter as capillaries (Unekawa, Brain 07). Dynamic observation in vivo revealed the channels run straight and are richly invested with astrocyte endfeet, which were stained in vivo with sulforhodamine 101. The amount of shunted blood without gas exchange through these channels was estimated from the velocity, number of capillaries, and theoretical extraction as 22% of the total perfusion, which agrees with our previous observation of 20% based on the compartmental analyses of inert gas clearance from human brain. In the human study we found that the shunted blood held in reserve was recruited with CO2 inhalation. With CO2, cerebral tissues became equally perfused by blood. When the brain is activated, CO2 in the tissue increases, and the direct shunting blood decreases. NIRS and fMRI falsely estimate the oxidative metabolism to be increased. In conclusion, the shunting blood constitutes a serious pitfall for quantifying brain oxidative metabolism. However, this paper does not intend to deny the possible clinical usefulness of NIRS and fMRI as tools to investigate brain function, since their signals are surely biological, but rather, to emphasize that the interpretation must be cautious.

Original languageEnglish
Pages (from-to)BO03-04
JournalJournal of Cerebral Blood Flow and Metabolism
Issue numberSUPPL. 1
Publication statusPublished - 2007 Nov 13
Externally publishedYes

ASJC Scopus subject areas

  • Neurology
  • Clinical Neurology
  • Cardiology and Cardiovascular Medicine


Dive into the research topics of 'Pitfalls in NIR-spectroscopy and functional MRI for measuring brain oxidative metabolism with special remarks on A-V shunts'. Together they form a unique fingerprint.

Cite this