A hybrid two-dimensional relaxometry (2DR) sequence was used to simultaneously measure both the spin-spin (R 2 ) and spin-lattice relaxation rates (R 1 ) of skeletal muscle in vivo. The 2DR sequence involved a 180°inversion pulse followed by a variable delay time (30 values from 40 to 7000 ms); a projection presaturation (PP) scheme to localize a 16-ml cylindrical voxel; and a CPMG sequence (950 even echoes, effective echo spacing ؍ 1.2 ms, equilibrium time ؍ 12 s). The 2DR data were collected at 3.0 Tesla from the flexor digitorum profundus of eight healthy males, 26 ؎ 2 years old. Analysis was performed with a 2D version of the non-negative least-squares algorithm and a one-way ANOVA. All subjects exhibited at least three spin-groups (R 2 < 200 s Spin-lattice (R 1 ) and spin-spin relaxation rates (R 2 ) do not reflect instantaneous molecular configurations; they are average values that arise from a system of molecular environments (spin-groups) over a finite time period. A proper interpretation of relaxometry data requires a detailed understanding of the nature and timescale of processes that contribute to the averaging phenomenon. The exchange of entire water molecules between spin-groups is a predominant process (1,2). The exchange model of Zimmerman and Britten (3) predicts that fast exchange between m spin-groups would produce an averaged relaxation rate R i,avg :where P j and R i,j is the population fraction and relaxation rate of the jth spin-group respectively (⌺P j ϭ 1). Equation[1] yields monoexponential relaxation that would mask information about the individual spin-groups. On the other hand, relaxation rates would not be averaged at the slow-exchange limit, and multiexponential relaxation would be observed. The characterization of fast and slow exchange depends on the relaxation timescale defined by a particular relaxation parameter. For two spin-groups a and b, the exchange rate between them (k ab ) must be greater than, equal to, or less than the quantity R i,b -R i,a to meet the conditions of fast, intermediate, and slow exchange, respectively (4). The variable R i could represent either a spinlattice or a spin-spin relaxation rate, so it is possible for a system to be in fast exchange on a spin lattice relaxation (T 1 ) timescale (T i ϭ R i -1 ), but in slow or intermediate exchange on a spin-spin relaxation (T 2 ) timescale.This would be consistent with data from NMR studies of excised muscle, using (non-imaging) relaxometry techniques with optimal experimental parameters (5). For several decades, T 2 of excised muscle tissue has been described as multiexponential. About 80 -90% of the signal decay is characterized by one spin-group (T 2 ϭ 20 -40 ms), while 10 -15% of the signal arises from another (T 2 Ն 100 ms) (6 -9). These results are contrary to what would be expected for a fast-exchange model of muscle tissue (10,11). On a T 1 timescale, however, there is no apparent distinction between these spin-groups, indicating fast exchange on that timescale (8,(12)(13)(14)(15). Due to technica...