Background: To determine the feasibility and validity of using blood oxygen level-dependent (BOLD) magnetic resonance imaging (MRI) to evaluate the effects of back extension exercise on core lumbar paraspinal muscle strength. Methods: In this prospective study, R2* and T2 mapping of paraspinal muscles of 100 healthy volunteers were performed before and after back extension exercises in different recovery sessions (session I, II, III or IV). Volunteers use the Roman chair to complete the back extension exercises. The cross-sectional area (CSA), R2* and T2 values were measured and analyzed in 3 muscles (iliocostalis, longissimus, and multifidus muscles) of the lower back before and after exercise. Results: The CSA and T2 values of iliocostalis, longissimus, and multifidus muscles at L3 and L4 levels were higher in recovery sessions I and II than in the resting-state (P<0.05); however, compared to that in the resting-state, the R2* value was significantly reduced in session I but increased in sessions II?IV (P<0.05). Furthermore, the CSA and T2 values in recovery session I were higher than those in the resting-state, whereas the R2* value was lower (P<0.05). After exercise, the recovery tendency of R2* and T2 value was consistent in both males and females, but a significant sex difference in R2* value was observed between recovery sessions III and IV (P<0.05). Conclusions: R2* mapping and T2 mapping are effective and feasible for assessment of the effects of back extension exercises on lumbar paraspinal muscle strength.