Spinal cord injury (SCI) disrupts the communication between the brain and the spinal circuits responsible for movement, thereby causing severe motor deficits. Current strategies to restore function to paralyzed limbs have separately investigated electrical stimulation of the spinal cord or of the peripheral neuromuscular system. Various neuromodulation strategies, for instance electrical epidural stimulation (EES) of the spinal cord, reactivate spinal circuits below the lesion and enable the generation of locomotor activity. EES targets muscle synergies rather than specific muscles or joints, and can therefore be limited by low selectivity. Accessing distal muscles individually is key to restore refined movement. Peripheral nerve stimulation (PNS) offers this possibility by selectively recruiting fibers innervating distinct muscles. Here, I developed a hybrid electrical stimulation paradigm, concomitantly targeting the spinal cord and the peripheral nerves for a global activation of coordinated multi-joint leg movements and a selective activation of distal muscles respectively. This approach combines two highly complementary stimulation paradigms into one refined neuroprosthetic system that could improve functional restoration after paralysis. The first part of this work addressed the validation of intraneural electrodes for selective and stable PNS. Albeit highly promising, incomplete characterization of long-term usability and biocompatibility has so far restricted their widespread use. To bridge this gap, I conducted a longitudinal assessment and comprehensively characterized their functional properties in light of their bio-integration in rats. Results showed that i) stimulation thresholds increased moderately during one month after implantation and then stabilized, ii) these changes correlated with progressive implant encapsulation, and iii) selectivity in muscle recruitment was retained in spite of the encapsulation, permitting precise control over ankle kinematics in anesthetized experiments. Overall, these results demonstrated the potential for long-term usability of intraneural implants. In the second part of this work, I developed and characterized a hybrid PNS-EES paradigm that concomitantly stimulated the spinal cord and the sciatic nerves in rat models of severe SCI, and validated it in a pilot study with a human SCI. I showed that i) muscle recruitment obtained by EES and PNS was highly complementary, ii) PNS enabled controllable adjustments in leg movements during locomotion, and iii) the hybrid PNS-EES paradigm permitted refined movements that increased functionality during locomotion in rats and a human pilot subject. This thesis provides evidence about the long-term functionality of intraneural implants and demonstrates their potential for stable interfacing with peripheral nerves. The hybrid PNS-EES paradigm reveals how the complementarity of both strategies effectively improved functional outcomes for paralyzed lower limbs. These findings open promising perspectives for the development of hybrid neuroprosthetic systems to restore functional and refined movements to paralyzed limbs.