Neurogenesis exists in adult brains of mammalian, which can be stimulated by hypoxic/ischemic brain injury. Stroke-induced newborn striatal neurons can develop long axons to project into the substantia nigra and integrate with preexisted neural networks. Rehabilitative training improves striatal neurogenesis in brains after stroke. However, whether it can improve axon-regeneration of new striatonigral projection neurons remains unknown, which is critical anatomical basis for recovering motor function. Therefore, we investigated the effects of exercise on axon-regeneration of newborn projection neurons in rat brains following ischemic stroke. Adult male Sprague-Dawley rats (250-280g, n=25) were divided into stroke (MCAO) and stroke+exercise (MCAO+EX) groups. Rats were anesthetized with 10% chloral hydrate (360 mg/kg, i.p.) and then subjected to a transient middle cerebral artery occlusion (MCAO) to induce ischemic stroke, followed by 30 minutes of exercise training daily for 5 to 28 days after MCAO (MCAO+EX). Motor function was tested using the rotarod test. We used fluorogold (FG) nigral injection to trace striatonigral and corticonigral projection neurons, and GFP-targeting retroviral vectors combined with FG double labeling (FG+-GFP+) to detect newborn projection neurons. The results further showed that such exercise could improve recovery of motor function of rats after MCAO. Under this experimental condition, we observed that FG+-GFP+ cells were detected in ipsilateral cortex and striatum of rats to MCAO, suggesting existence of newborn striatonigral and corticonigral projection neurons in ischemic injured brain, which was consisted with our previous reports (1,2). Interestingly, exercise poststroke could significantly increase the number of FG+-GFP+ neurons in ipsilateral striatum and cortex. The number of GFP+-FG+ cells in MCAO+EX rats increased 2.52-fold and 1.78-fold of MCAO rats in the cortex and striatum, respectively. This data suggested that exercise enhanced the capacity for axon-regenration of newborn neurons. Moreover, the tyrosine hydroxylase immunostaining results showed that exercise significantly increased dopaminergic neurons in the substantia nigra, a region remote from the ischemic territory, compared with the controls. In summary, present study provides the first evidence that passive exercise poststroke can effectively improve axon-regeneration of newborn projection neurons and accelerate the reestablishment of new neural circuitry within the basal ganglia after ischemic injury, which should provide very important anatomical foundation for the recovery of motor behavioral function. Our results illustrate mechanisms, at cellular level, for passive rehabilitative treatment poststroke in the clinic.