Hemodynamic scaling of fMRI-BOLD signal: validation of low-frequency spectral amplitude as a scalability factor

Functional magnetic resonance imaging blood-oxygenation-level-dependent (fMRI-BOLD) signal representing neural activity may be optimized by discriminating MR signal components related to neural activity and those related to intrinsic properties of the cortical vasculature. The objective of this study was to reduce the hemodynamic change independent of neural activity to obtain a scaled fMRI-BOLD response using two factors, namely, low-frequency spectral amplitude (LFSA) and breath-hold amplitude (BHA). Ten subjects (age range, 22-38 years) were scanned during four task conditions: (a) rest while breathing room air, (b) bilateral finger tapping while breathing room air, (c) rest during a four inspirational breath-hold, and (d) rest during moderate hypercapnia (breathing 5% CO2, 20% OZ and 75% N-2). In all subjects who breathed 5% CO2, regions with significant BOLD response during breath-hold correlated significantly with the percent signal increase during 5% CO2 inhalation. Finger-tapping-induced responses in the motor cortex were diminished to a similar extent after scaling using either LFSA or BHA. Inter- and intrasubject variation in the amplitude of the BOLD signal response reduced after hemodynamic scaling using LFSA or BHA. The results validated the hemodynamic amplitude scaling using LFSA with the earlier established BHA. LFSA free from motor-task contamination can be used to calibrate the fMRI-BOLD response in lieu of BHA or hypercapnia to minimize intra- and intersubject variation arising from vascular anatomy and vasodilative capacity. (c) 2007 Elsevier Inc. All rights reserved.