Caffeine alters the temporal dynamics of the visual BOLD response Thomas T. Liu, a, * Yashar Behzadi, a,b Khaled Restom, a Kamil Uludag, a Kun Lu, a Giedrius T. Buracas, a David J. Dubowitz, a and Richard B. Buxton a a Center for Functional Magnetic Resonance Imaging and Department of Radiology, University of California San Diego, La Jolla CA, USA b Department of Bioengineering, University of California San Diego, La Jolla, CA, USA Received 6 April 2004; revised 19 July 2004; accepted 22 July 2004 The blood oxygenation level-dependent (BOLD) responses to visual stimuli, using both a 1-s long single trial stimulus and a 20-s long block stimulus, were measured in a 4-T magnetic field both before and immediately after a 200-mg caffeine dose. In addition, resting levels of cerebral blood flow (CBF) were measured using arterial spin labeling. For the single trial stimulus, the caffeine dose significantly ( p b 0.05) reduced the time to peak (TTP), the time after the peak at which the response returned to 50% of the peak amplitude (TA 50 ), and the amplitude of the poststimulus undershoot in all subjects (N = 5). Other parameters, such as the full-width half-maximum (FWHM) and the peak amplitude, also showed significant changes in the majority of subjects. For the block stimulus, the TTP, TA 50 , and the time for the response to reach 50% of the peak amplitude (T 50 ) were significantly reduced. In some subjects, oscillations were observed in the post- stimulus portion of the response with median peak periods of 9.1 and 9.5 s for the single trial and block responses, respectively. Resting CBF was reduced by an average of 24%. The reproducibility of the results was verified in one subject who was scanned on 3 different days. The dynamic changes are similar to those previously reported for baseline CBF reductions induced by hypocapnia and hyperoxia. D 2004 Elsevier Inc. All rights reserved. Keywords: BOLD responses; Hypocapnia; Hyperoxia Introduction The blood oxygenation level-dependent (BOLD) signal meas- ured in functional magnetic resonance imaging (fMRI) experi- ments is a complex function of cerebral blood flow (CBF), cerebral blood volume, and oxygen metabolism (Buxton et al., 1998). The shape of the BOLD hemodynamic response (HDR) exhibits significant variability between subjects and somewhat less variability in repeated scans of a single subject (Aguirre et al., 1998; D’Esposito et al., 2003; Handwerker et al., 2004). This variability in the BOLD temporal dynamics has a direct impact on the interpretation of fMRI experiments, especially for event-related experiments in which the shape of the HDR is often of interest (Rosen et al., 1998). An understanding of the factors that contribute to the variability of the HDR is therefore important when comparing responses between populations or experimental conditions. Recent studies have suggested that the temporal dynamics of the BOLD response are strongly dependent on the baseline CBF state. Kemna and Posse (2001) modified CBF levels by adjusting the arterial partial pressure of carbon dioxide (Paco 2 ) with either controlled hyperventilation (hypocapnia) or CO 2 breathing (hyper- capnia). Using a 1.5-T system, they found that the temporal width and time to peak (TTP) of the BOLD response to a short visual stimulus increased with the baseline level of Paco 2 . In a later study at 7 T, Cohen et al. (2002) observed a similar dependence on Paco 2 and noted that the peak amplitude of the visual BOLD signal decreased with hypercapnia and increased with hypocapnia. In addition, the poststimulus undershoot in the response resolved more quickly with hypocapnia and appeared to be abolished with hypercapnia. The carbon dioxide studies suggest that similar dynamic effects should be observed with other vasoactive agents that modify the baseline CBF state. One such agent is caffeine, which is the most widely consumed neural stimulant in the world with most of the intake coming from dietary sources such as coffee and tea (Fredholm et al., 1999). In addition to its neurostimulant effects, caffeine acts to reduce CBF, primarily by inhibition of the adenosine A 2A receptors (Cameron et al., 1990; Ngai et al., 2001). Several studies have examined the effect of caffeine on the amplitude of the BOLD response. In a study using a 1.5-T system, Mulderink et al. (2002) found that a 200-mg caffeine dose increased the amplitude of the BOLD response to a brief 2-s stimulus by 37% and 26% in the visual and motor cortices, respectively, and suggested that caffeine could be used as a contrast booster for BOLD studies. Laurienti et al. (2002, 2003), however, 1053-8119/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2004.07.061 * Corresponding author. UCSD Center for Functional MRI, Depart- ment of Radiology, 9500 Gilman Drive, MC 0677, La Jolla, CA 92093- 0677. Fax: +1 858 822 0605. E-mail address: ttliu@ucsd.edu (T.T. Liu). Available online on ScienceDirect (www.sciencedirect.com.) www.elsevier.com/locate/ynimg NeuroImage 23 (2004) 1402 – 1413