Intermolecular double-quantum coherences (iDQCs) are well known to be sensitive to magnetic-field perturbations inside tissues. However, the exact relation between iDQC contrast in magnetic resonance imaging (MRI) and the underlying physiology is less well understood. To investigate parameters that influence iDQC signal changes observed during neuronal activation, carbogen-inhalation experiments were performed to produce a pure hemodynamic response without affecting oxidative metabolism. Eight human volunteers were studied at 2.9 T using gradientrecalled echo (GRE) and spin-echo (SE) variants of a single-shot sequence selecting iDQCs. Results were compared with conventional recordings of the blood oxygen level-dependent (BOLD) effect. Maps of voxels responding to the carbogen challenge showed similar distributions for iDQC and conventional MRI after adjustment for different sensitivities. In magnetic resonance imaging (MRI) experiments sensitized to intermolecular multiple-quantum coherences (iMQCs), contrast originates in long-range dipolar couplings on a spatial scale from tens of micrometers to millimeters (1,2). Such interactions between nuclear spins in different molecules separated by the so-called "correlation distance" can be observed by suppressing the signal contributions from conventionally detected single-quantum coherences (SQCs) in the imaging sequence. Previous work has demonstrated a strong susceptibility of intermolecular double-quantum coherences (iDQCs) to local magnetic field perturbations. This suggests a potential for oxygenation studies (for example, in tumor evaluation and detection [3]), and has already been exploited for functional magnetic resonance imaging (fMRI) experiments (4 -6).Currently, the origin of activation-induced signal changes in iDQC imaging is not completely understood. While conventional fMRI employing T 2 *-weighted or T 2 -weighted sequences reflects the average strength of the internal field gradient within an imaging voxel, the iDQC signal is a function of the resonance frequency difference of the two interacting spins. It was thus speculated if previously observed differences in the activation maps recorded with iDQC-based and conventional fMRI might reflect true differences in the physiological parameters influencing the signal (4). In this case, iDQC contrast would be fundamentally different from the conventional blood oxygen level-dependent (BOLD) effect. However, differences in the activation maps may alternatively just reflect differences in the sensitivities and signal stabilities of the two techniques.The BOLD-weighted fMRI signal depends on multiple physiological parameters, including changes in the cerebral blood flow (CBF), the cerebral blood volume (CBV), and the cerebral metabolic rate of oxygen (CMRO 2 ) (7). For further investigation of parameters influencing iDQC contrast, it would be useful to manipulate such physiological parameters more selectively as achieved in simple taskactivation studies. Previously, iDQC imaging was shown to be sensitive to...