This study examined the effect of resistance training on exercise-induced contrast shift in magnetic resonance (MR) images. It was hypothesized that a given load could be lifted after training with less muscle showing contrast shift, thereby suggesting less muscle was used to perform the exercise. Nine males trained the left quadriceps femoris (QF) muscle 2 days/wk for 9 wk using 3-6 sets of 12 knee extensions each day. The right QF served as a "control." Exercise-induced contrast shifts in MR images evoked by each of three bouts of exercise (5 sets of 10 knee extensions with a load equal to 50, 75, and 100% of the maximum pretraining load that could be lifted for 5 sets of 10 repetitions) were quantified pre- and posttraining. MR image contrast shift was quantified by determining QF cross-sectional area (CSA) showing increased spin-spin relaxation time. One repetition maximum increased 14% in the left trained QF and 7% in the right untrained QF. Left QF CSA increased 5%, with no change in right QF CSA. Left QF CSA showing contrast shift was less after each bout of the exercise test posttraining. This was also true, to a lesser extent, for the right QF at the higher two loads. The results suggest that short-term resistance training reduces MR image contrast shift evoked by a given effort, thereby reflecting the use of less muscle to lift the load. Because this response was evident in both trained and contralateral untrained muscle, neural factors are suggested to be responsible.(ABSTRACT TRUNCATED AT 250 WORDS)