EPI Imaging of Global Increase of Brain MR Signal with Breath-hold Preceded by Breathing 0, zy Kenneth K. Kwong, Isabel Wanke, Kathleen M. Donahue, Timothy L. Davis, Bruce R. Rosen zyxwvu Brain MR signal has been observed to decrease during ces- sation of breathing due to the increase of deoxyhemoglobin in the blood. However, for both animal and human studies, we have demonstrated that if the subjects breathed 100% oxygen in advance of apnea for a short time, T,*-weighted MR brain signal increased when breathing was stopped for a period of zyxwvu 30-60 s. This demonstrates the possibility of measuring re- sponses to hemodynamic change throughout the entire brain with a single respiratoty perturbation in a rapid, reliable, and robust manner. Key words: EPI; brain MRI; breath-hold; 0,. INTRODUCTION It has been reported that T,*-weighted signal decreases in the brain during breath-hold for humans and animals (1-4). This has been attributed to an increase in blood deoxyhemoglobin (5). However, a confounding factor in these experiments is that CO,, which also increases dur- ing breath-hold, acts to increase blood flow and therefore T,*-weighted signal. This suggests that in previously reported breath-hold experiments where the T,*- weighted signal intensity decreased, the increased de- oxyhemoglobin effect was greater than the hypercapnia- induced flow effect. Here we present experiments in which we controlled for the effect of increased deoxyhe- moglobin, enabling the flow effect to dominate. Specifi- cally, we have observed in animals and in people that MR brain signal increases during breath-hold provided the subjects first breathed 100% oxygen instead of air. The objectives of this paper are to: 1) Apply the spe- cific protocol of breath-hold with prebreathing oxygen to both animals and human subjects. 2) Sample the arterial pC0, in animals for comparison with the MR signal response. 3) Measure the blood oxygen saturation level of humans during breath-hold with oximeters for compari- son with the MR signal response. 4) Measure the effect of breathing 0, after breathing air in both animals and human subjects in order to study the effect of breathing 0, alone. 5) Compare the animal's MR signal response during breath-hold to that when breathing CO,. zyxwvut ~ ~~ MRM 33:448-452 (1995) From the MGH-NMR Center, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, and Harvard-MIT Division of Health Sciences & Technology, Charlestown. Massachusetts (T.L.D.). Address correspondence to: Kenneth Kwong, Ph.D.. Massachusetts Gen- eral Hospital-NMR Center (2301), Building 149, 13th Street, Charlestown, MA 02129. Received July 27, 1994; revised October 31, 1994; accepted November 22, 1994 0740-31 94/95 $3.00 Copyright zyxwvutsrqpon 0 1995 by Williams & Wilkins All rights of reproduction in any form reserved. METHODS All experiments were performed on a 1.5 Tesla GE Signa clinical system retrofitted with echo planar imaging (EPI) from advanced NMR. The quadrature head coil used in these studies had a signal to noise of 90 to 1 for the experimental parameters reported here. We studied three rabbits and five healthy human sub- jects who breathed pure oxygen before the breath-hold exercise began. A single slice T,*-sensitive echo planar (EPI) asymmetric spin echo (ASE) sequence (6) was used for all animal experiments (TR = 2s, TE = 77 ms, 180° offset = -25 ms). (ASE can reduce unwanted signals by 40% from large veins because fast flowing spins in draining veins experience the 90" but not the 180' of the spin echo component of ASE (7).) In plane resolution was 3 x 1.5 mm and slice thickness was 7 mm. At the higher resolution required to study the small rabbit's brain, the complete coverage of k-space required a minimum TE of 77 ms on our EPI system. For the animal studies, New Zealand white rabbits weighing 2-3 kg were anesthetized, tracheostomized, and ventilated with 100% 0,. Catheters were inserted into the femoral artery for blood sampling. MR images were acquired continuously while the rabbits were breathing 100% 0,. Breathing was then interrupted by turning the respirator off for a period ranging from 60s to 70s. Then the animals breathed 100% 0, for the remain- der of the imaging period. Blood samples were taken at baseline when the animal was breathing 100% 0,, dur- ing breath-hold, and at the resumption of 0, breathing. The MRI response to breath-hold was compared with the response when transiently breathing CO,. Specifi- cally, MR images were acquired prior to, during, and after the delivery of a bolus of 10% CO, (20% 0,) for 20 and 30 s. To identify the effect of 0, alone, the MR response to breathing 0, and then normal room air was also mea- sured in three rabbits. No interruption or cessation of breathing was performed in these experiments. Parallel studies were performed in five human sub- jects. All subjects were studied with a T,*-weighted echo planar (EPI) asymmetric spin echo sequence (TR = 2 s, TE = 70 ms, 180' offset = -25ms) with 10 to 12 slices (3-mm in-plane resolution, 7-mm thickness, and 2-mm gap) covering the whole head. The use of a somewhat lower resolution image in the human studies, relative to the animal studies, allowed us to reduce data acquisition time and thus reduce TE to 70 ms. To reduce motion artifacts, subjects were secured with pillows and pad- dings in the head coil. Human subjects breathed 100% 0, through a mask (Baxter Healthcare) for 5-10 min before MR images were 448