Negative Dip In BOLD fMRI Is Caused By Blood Flow— Oxygen Consumption Uncoupling In Humans Joachim Ro ¨ther, Rene Knab, Farsin Hamzei, Jens Fiehler, Ju ¨ rgen R. Reichenbach,* Christian Bu ¨ chel, and Cornelius Weiller Department of Neurology, University Hospital Hamburg Eppendorf, Hamburg; and *Institute of Diagnostic and Interventional Radiology, Friedrich-Schiller-University, Jena, Germany Received January 16, 2001 The sensitivity of MRI for local changes in the de- oxyhemoglobin concentration is the basis of the blood oxygen level dependent (BOLD) effect. Time-resolved fMRI studies during visual activation show an early signal intensity (SI) decrease indicating a short last- ing uncoupling of oxygen consumption and cerebral blood flow (CBF) before a SI increase due to the over- compensating hemodynamic response occurs. Normal neuronal activity may be preserved despite absent vascular responsiveness. Here we show that a nega- tive BOLD effect occurs during motor activation in an asymptomatic patient with severely disturbed cere- bral autoregulation due to extracranial artery disease. This is thought to be due to oxygen consumption in the absence of a hemodynamic response. This rare case of a persisting uncoupling of oxygen metabolism and CBF serves as a model that supports changes of the cerebral blood oxygen saturation as the major contrib- utor of the BOLD effect. © 2002 Elsevier Science Key Words: functional magnetic resonance imaging; blood oxygen level-dependent (BOLD) effect; cerebral blood oxygen saturation; cerebral autoregulation; ce- rebrovascular reserve capacity. INTRODUCTION The mechanism of fMRI is based on the intrinsic contrast of deoxyhemoglobin that acts as a paramag- netic particle (Pauling and Coryell, 1936). Changes in its concentration within blood induce signal intensity (SI) changes in T2*-weighted MR sequences (Ogawa et al., 1990; Thulborn et al., 1982; Turner et al., 1991). The sensitivity of MRI for this intrinsic contrast me- dium is the basis of the BOLD effect (Ogawa et al., 1992), where SI changes mirror cerebral blood oxygen saturation changes during neuronal activation, thus reflecting neuronal activity (Kwong et al., 1992). A decrease of deoxyhemoglobin due to the washout by an increased influx of fresh, oxygenated blood or, alterna- tively, a blood volume effect, i.e., a relative increase of the intravascular water fraction in a given voxel, or both are thought to be the basic mechanism of BOLD signal changes (Frahm et al., 1994; Kwong et al., 1992; Ogawa et al., 1990). This model is based on the physi- ological principal that neuronal activity is tightly cou- pled to a local increase in blood flow and energy me- tabolism (Roy and Sherrington, 1890; Sokoloff et al., 1977). The time course of the BOLD effect is characterized by a triphasic response with an initial negative dip, a subsequent positive SI change and a poststimulus un- dershoot (Buxton et al., 1998; Menon et al., 1995; Ya- coub and Hu, 1999). Under physiological conditions this initial negative dip is thought to represent the deoxyhemoglobin increase due to a short lasting un- coupling between blood flow and oxidative metabolism (Fox and Raichle, 1986; Frahm et al., 1996; Magistretti and Pellerin, 1999). This interpretation is supported by optical imaging studies (Frostig et al., 1990) and stud- ies applying magnetic resonance spectroscopy showing a 4 s lasting deoxyhemoglobin increase (Ernst and Hennig, 1994; Prichard et al., 1991). Here we predicted that this negative dip should per- sist during neuronal activation under conditions of se- verely impaired cerebrovascular reserve capacity (CVRC). The rational for this assumption is delivered by positron emission tomography studies in humans that have shown abnormal regional CBF responses to physiological stimulation of the sensorimotor cortex despite normal brain function (Powers et al., 1988). We therefore studied the BOLD effect in response to a motor paradigm (finger tapping) in a patient with oc- clusion of both internal carotid arteries and one verte- bral artery and severely disturbed CVRC. MATERIAL AND METHODS The patient under study was a 74-year-old male who presented with a transient ischemic attack 6 months before the first study. Neurological symptoms with a weakness of the right arm resolved within one day NeuroImage 15, 98 –102 (2002) doi:10.1006/nimg.2001.0965, available online at http://www.idealibrary.com on 98 1053-8119/02 $35.00 © 2002 Elsevier Science All rights reserved.