Neurorehabilitation as Network Perturbation: Shaping Neuroplasticity with Robotics, Virtual Reality, and Neuromodulation.

Neurological injury induces widespread neuroplastic changes that extend well beyond focal structural damage, altering synaptic function, circuit dynamics, and large-scale network organization. While these processes provide the biological substrate for recovery, they can also drive the stabilization of maladaptive network states that constrain long-term functional improvement. Traditional neurorehabilitation has largely emphasized compensation and task practice, often without explicitly targeting the neural dynamics that underlie persistent disability. In this Opinion, we propose that contemporary rehabilitation technologies, including robotics, virtual reality, and neuromodulation, should be conceptualized as mechanistically grounded interventions that actively perturb neural networks and interact with the pathobiology of post-injury reorganization. Drawing on advances in systems and network neuroscience, we examine key molecular, synaptic, and network-level mechanisms that govern adaptive and maladaptive plasticity, and discuss how these technologies modulate error processing, sensory context, and excitability landscapes to reshape recovery trajectories. We argue that when interventions are appropriately structured, timed, and combined within adaptive and closed-loop frameworks, technology-assisted rehabilitation can move beyond compensation and toward principled modulation of neuroplasticity, aligning therapeutic innovation with the biological rules that govern recovery. This perspective highlights the need for network-informed biomarkers and longitudinal approaches to translate technological advances into durable functional gains.