Introduction: Phantom limb pain (PLP) is a common and poorly understood pathology of difficult medical control that progressively takes place after amputation occurs. Objective: This article discusses the multifactorial bases of PLP. These bases involve local changes at the stump level, spinal modifications of excitability, deafferentation, and central sensitization, leading to the development of maladaptive plasticity, and consequentially, defective processing of sensory information by associative neural networks. These changes can be traced by neurophysiology and imaging topographical studies, indicating a degree of cortical reorganization that perpetuates pain and discomfort. Intervention: Noninvasive brain stimulation can be an alternative way to manage PLP. This article discusses two techniques-transcranial direct current stimulation (tDCS) and repetitive transcranial magnetic stimulation (rTMS)-that have shown promising results for controlling PLP. The modulation that both techniques rely on is based on synaptic mechanisms linked to long-term potentiation and long-term depression phenomena. By applying tDCS or rTMS, clinicians can target processes associated with central sensitization and maladaptive plasticity, while promoting adequate sensory information processing by integrative cognitive behavioral techniques in a comprehensive rehabilitation program. Conclusions: Understanding PLP from a dynamic neurocomputational perspective will help to develop better treatments. Furthermore, Bayesian analysis of sensory information can help guide and monitor therapeutic interventions directed toward PLP resolution.
Keywords: Bayesian model; maladaptive plasticity; noninvasive brain stimulation; phantom limb pain.