Spatially dissociated flow-metabolism coupling in brain activation Manouchehr S. Vafaee * and Albert Gjedde Center of Functionally Integrative Neuroscience, University of Aarhus, and PET Center, Aarhus University Hospitals, Aarhus, Denmark Received 31 March 2003; revised 8 September 2003; accepted 3 October 2003 The relationships among cerebral blood flow (CBF), oxygen con- sumption (CMRO 2 ) and glucose use (CMR glc ) constitute the basis of functional brain-imaging. Here we report spatially dissociated changes of CMRO 2 and CBF during motor activity that lead us to propose a revision of conventional CBF-CMRO 2 coupling models. In the left primary and supplementary motor cortices, CBF and CMRO 2 rose significantly during finger-thumb tapping. However, in the right putamen CBF did not rise, despite a significant increase in CMRO 2 . We explain these observations by invoking a central command mechanism that regulates CBF in the putamen in anticipation of movement. By this mechanism, CBF rose in the putamen before the measurements of CBF and CMRO 2 while CMRO 2 rose when actual motion commenced. D 2003 Elsevier Inc. All rights reserved. Keywords: Flow-metabolism; Brain activation; Functional brain-imaging It is generally accepted that most of the work performed by neurons active in the brain is fueled by the oxidative metabolism of glucose. Because neurons contain minimal energy reserves, the additional glucose and oxygen needed during activation are supplied by elevated cerebral blood flow (CBF). Thus, CBF is believed to fluctuate according to neuronal energy demands (Olesen, 1971; Sokoloff, 1991). The concept of a fixed relation- ship between CBF and neuronal work forms the basis of functional imaging using positron emission tomography (PET) and functional magnetic resonance imaging (fMRI). the results of techniques that measure changes only in CBF would have to be reevaluated should this concept prove to be unreliable or have less than general applicability. Although numerous reports confirm that steady-state elevations in oxygen and glucose consumption are accompanied by propor- tionally increased CBF (Gjedde, 2001), the relationship between CBF and oxygen consumption (measured as cerebral metabolic rate for oxygen; CMRO 2 ) appears to depend on the type and 1053-8119/$ - see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2003.10.003 Abbreviations: CBF, cerebral blood flow; CMRO 2 , cerebral metabolic rate of oxygen. * Corresponding author. Center of Functionally Integrative Neuro- science, University of Aarhus, and PET Center, Aarhus University Hospitals, 44 Norrebrogade, 8000 Aarhus C, Denmark. Fax: +45-8949- 4400. E-mail address: manou@pet.auh.dk (M.S. Vafaee). Available online on ScienceDirect (www.sciencedirect.com.) www.elsevier.com/locate/ynimg NeuroImage 21 (2004) 507 – 515 duration of stimulation and the brain region studied (Gjedde et al., 2002; Vafaee and Gjedde, 2000). During low-intensity somatosen- sory stimulation in humans, CMRO 2 rarely rises significantly in the cerebral cortex whereas CBF increases substantially (Fox and Raichle, 1986; Fox et al., 1988; Fujita et al., 1999; Ribeiro et al., 1993). By contrast, directly measured values of CMRO 2 increase significantly in human striate cortex during visual stimulation, but the change is always accompanied by a disproportionately greater elevation in CBF (Mintun et al., 2002; Vafaee and Gjedde, 2000; Vafaee et al., 1999). These observations show that changes of blood flow are not uniquely coupled to changes of oxidative metabolism. Current models of nonlinear flow-metabolism coupling (Buxton and Frank, 1997; Gjedde, 1997; Hoge et al., 1999; Vafaee and Gjedde, 2000; Zheng et al., 2002) all predict approximately fixed ratios between changes of oxidative metabolism and blood flow, generally in the range 0.25– 0.5. To explain variable coupling, one alternative is to introduce a recruitment index, which reflects the hypothetical adjustment of the capillary diffusion capacity for oxygen (Hyder et al., 2000) to the required degree of coupling. However, it is not a simple task to conjoin an increase of blood flow with a reduction of capillary diffusion capacity in a single mechanism of flow-metabolism coupling. A second alternative is to assign the changes of blood flow and oxidative metabolism to separate mechanisms of input processing and output generation in functional subdivisions of the brain. It is the latter alternative that we test in the present study. Motor activity, such as moving the fingers at different rates, is associated with frequency-dependent increases in regional CBF (rCBF) in the motor cortex (Blinkenberg et al., 1996; Sabatini et al., 1993; Sadato et al., 1996; Seitz et al., 1990) but not in the putamen (Frackowiak et al., 1997; Jenkins et al., 1994). Regional CMRO 2 (rCMRO 2 ) rises significantly in the motor cortex during hand movements (Iida et al., 1993; Kastrup et al., 2002; Raichle et al., 1976), but the relationship between rCMRO 2 and rCBF elsewhere in the brain during this activity is not known. Because the proposed coupling between CBF and CMRO 2 is crucial for the interpretation of functional imaging maps, it is important to identify accurately the relationships between these two variables in all areas of the brain that respond functionally to a motion stimulus. Following a command to initiate skeletal muscle contractions, heart rate and cardiac output increase, also when no work is performed (Nowak et al., 1999). To explain the differences in coupling between CBF and CMRO 2 in different brain regions, we