We have conducted studies to determine the potential of exercise to benefit the injured spinal cord using neurotrophins. Adult rats were randomly assigned to one of three groups: (1) intact control (Con); (2) sedentary, hemisected at a mid-thoracic level (Sed-Hx), or (3) exercised, hemisected (Ex-Hx). One week after surgery, the Ex-Hx rats were exposed to voluntary running wheels for 3, 7, or 28 days. BDNF mRNA levels on the lesioned side of the spinal cord lumbar region of Sed-Hx rats were approximately 80% of Con values at all time points and BDNF protein levels were approximately 40% of Con at 28 days. Exercise compensated for the reductions in BDNF after hemisection, such that BDNF mRNA levels in the Ex-Hx rats were similar to Con after 3 days and higher than Con after 7 (17%) and 28 (27%) days of exercise. After 28 days of exercise, BDNF protein levels were 33% higher in Ex-Hx than Con rats and were highly correlated (r=0.86) to running distance. The levels of the downstream effectors for the action of BDNF on synaptic plasticity synapsin I and CREB were lower in Sed-Hx than Con rats at all time points. Synapsin I mRNA and protein levels were higher in Ex-Hx rats than Sed-Hx rats and similar to Con rats at 28 days. CREB mRNA values were higher in Ex-Hx than Sed-Hx rats at all time points. Hemisection had no significant effects on the levels of NT-3 mRNA or protein; however, voluntary exercise resulted in an increase in NT-3 mRNA levels after 28 days (145%). These results are consistent with the concept that synaptic pathways under the regulatory role of BDNF induced by exercise can play a role in facilitating recovery of locomotion following spinal cord injury.