Intractable pains are a major public health problem following nerve lesions. The underlying mechanisms are imperfectly understood. Recently, however, large sensory fibers not normally concerned with pain are shown to sprout into pain processing regions following nerve lesions. These large fibers use glutamate as a transmitter, which is neurotoxic in excess. This leads to the basic formulation underlying this proposal, namely that this novel input leads to an outpouring of glutamate from large rapidly firing fibers into a pain processing region, that this outpouring is neurotoxic , that the cells selected for destruction are predominantly inhibitory interneurons, leaving a pain processing region devoid of inhibitory control, and that prevention of this destruction will ameliorate the pain related behaviors that follow nerve lesions. These specific hypotheses test these ideas. 1. That nerve lesions lead to some neuron loss in the main spinal processing region. 2. That stimulating large sensory fibers greatly increase this loss. 3. That this loss is activity dependent. 4. That the loss can be lessened or eliminated by glutamate receptor blockers. 5. That the loss is selective for inhibitory interneurons. 6. That the loss of cells caused by nerve stimulation has behavioral consequences. 7. The stimulation induced cell loss is largely apoptotic in nature. If these hypotheses are borne out, the key factor from a clinical point of view is that the nerve fiber activity that follows a nerve lesion would result in death of inhibitory interneurons in the spinal cord. Accordingly, preventing this activity might ameliorate the loss of inhibitory control and thus have therapeutic utility.