We examined the effects of voluntary (16 weeks of wheel running) and forced (16 weeks of treadmill running) exercise on memory-related behavior, hippocampal volume, thioflavine-stained plaque number, and soluble Abeta levels in brain tissue in the Tg2576 mouse model of Alzheimer's disease (AD). Voluntary running animals spent more time investigating a novel object in a recognition memory paradigm than all other groups. Also, voluntary running animals showed fewer thioflavine S stained plaques than all other groups, whereas forced running animals showed an intermediate number of plaques between voluntary running and sedentary animals. Both voluntary and forced running animals had larger hippocampal volumes than sedentary animals. However, levels of soluble Abeta-40 or Abeta-42 did not significantly differ among groups. The results indicate that voluntary exercise may be superior to forced exercise for reducing certain aspects of AD-like deficits - i.e., plaque deposition and memory impairment, in a mouse model of AD.