Echo‐planar brain images (1.5 T, 1‐shot GRE, TR/TE=3000/45) were acquired during 30–60 s cycles of repetitive (1.3–1.7 Hz) finger flexion exercise against light (0.24) vs heavy (1.40 kg) weights. In protocols during which exercise was alternated with periods of rest, active voxels were identified in the contralateral motor and somatosensory areas by cross‐correlation against an on–off waveform. However, there was no significant difference between the responses to light vs heavy weights. In a continuous repetitive exercise protocol, in which only the weights were cycled, no force‐correlated voxels were identified. Inasmuch as force‐correlated neurons are known to be present in the primate cortex, the results illustrate a potential limitation of functional MRI studies based on the BOLD effect.