Enhancing plasticity in central networks improves motor and sensory recovery after nerve damage

Abstract

Nerve damage can cause chronic, debilitating problems including loss of motor control and paresthesia, and generates maladaptive neuroplasticity as central networks attempt to compensate for the loss of peripheral connectivity. However, it remains unclear if this is a critical feature responsible for the expression of symptoms. Here, we use brief bursts of closed-loop vagus nerve stimulation (CL-VNS) delivered during rehabilitation to reverse the aberrant central plasticity resulting from forelimb nerve transection. CL-VNS therapy drives extensive synaptic reorganization in central networks paralleled by improved sensorimotor recovery without any observable changes in the nerve or muscle. Depleting cortical acetylcholine blocks the plasticity-enhancing effects of CL-VNS and consequently eliminates recovery, indicating a critical role for brain circuits in recovery. These findings demonstrate that manipulations to enhance central plasticity can improve sensorimotor recovery and define CL-VNS as a readily translatable therapy to restore function after nerve damage.

Conflict of interest statement

M.P.K. has a financial interest in MicroTransponder, Inc., which is developing VNS for stroke and tinnitus. R.L.R. is a co-owner of Vulintus, Inc., which makes rodent behavioral testing systems, and owner of Teliatry, which is developing a VNS device. All other authors declare no conflict of interest.

Figures

Fig. 1. Median and ulnar nerve injury and delivery of vagus nerve stimulation during rehabilitation.

a A rat performing the volitional forelimb isometric pull task. b Schematic of the nerve injury. The median and ulnar nerves in the trained forelimb, innervating the digit flexor muscles required for grasping, were individually transected. The nerve stumps were then sutured into a guide conduit leaving a 6 mm gap between stumps. Reinnervation takes place, but the procedure results in chronic deficits in nerve architecture distal to the injury site. The radial nerve innervating the wrist and digit extensors of the forelimb was spared. Scale bar indicates 10 µm. c Illustration of the VNS device. A cuff electrode was placed on the left cervical vagus nerve in the neck with subcutaneous leads connected to a head-mount. d Timeline of rehabilitative training after nerve injury. Rehabilitation began 6 weeks after nerve injury, and rats received either closed-loop VNS paired with forelimb movement during rehabilitation (CL-VNS), equivalent training without VNS (Rehab), or a matched amount of VNS delivered after equivalent daily training sessions (Delayed VNS).

Fig. 2. Closed-loop VNS restores cortical motor maps after nerve damage.

a Intracortical microstimulation reveals that, despite extensive rehabilitative training, nerve injury results in a substantial reduction in motor cortical area evoking movements of the denervated digit flexors and an increase in motor cortical area that evokes extension movements (Rehab, n = 11). CL-VNS (n = 8) reverses these lesion-induced cortical map changes, restoring the digit flexion representations and reducing the aberrant expansion of extensor representations. Delayed VNS (n = 7) failed to restore motor map representations, demonstrating that VNS relies on timed engagement with rehabilitation. b Bubble plots detailing the cortical locations of digit flexion and wrist extension movements across animals in each experimental group. The size of each bubble represents the proportion of subjects that the stimulation evoked a digit flexion or wrist extension movement at each cortical site. Note that CL-VNS significantly increases the denervated digit flexion representation and reduces the lesion-induced expansion of extension compared to Rehab and Delayed VNS. Same group sizes as in (a). c Due to misdirected reinnervation, nerve damage results in a substantial increase in the percentage of sites that generate multi-joint movements involving simultaneous digit and elbow flexion. CL-VNS reduced the percentage of sites that simultaneously elicit both movements compared to Rehab or Delayed VNS, suggesting the restoration of independent muscle control. Same group sizes as in a. d Bubble plots illustrating the stimulation site locations that generate the simultaneous digit and elbow flexion movements. Same group sizes as in a. Circles depict individual subjects in a, c. Error bars indicate S.E.M. All comparisons represent Bonferroni-corrected t-tests to CL-VNS, ***p < 0.0005, **p < 0.005, *p < 0.025. Source data are provided as a Source Data file.

Fig. 3. Closed-loop VNS delivered during rehabilitative training improves motor and sensory function after nerve damage.

a, b Closed-loop VNS paired with rehabilitation (CL-VNS, n = 10) significantly enhances recovery of volitional forelimb function compared to equivalent rehabilitation without CL-VNS (Rehab, n = 13). Consistent with the failure to restore cortical motor maps, Delayed VNS (n = 11) fails to improve recovery compared to CL-VNS. c Plots depicting the proportion of subjects in each group sorted by percentage of pull force recovery. All CL-VNS subjects recovered at least 80% of pre-injury function by the end of therapy. Subjects receiving either Rehab or Delayed VNS exhibited substantially reduced levels of recovery, with many subjects demonstrating < 50% recovery. Same group sizes as in a, b. d CL-VNS (n = 7) significantly improved tactile sensation in the denervated forepaw compared to both Rehab (n = 7) and Delayed VNS (n = 7) groups. In a, b asterisks indicate significant differences using t-tests across groups at each time point. The color of the asterisk denotes the group compared to CL-VNS (blue indicates CL-VNS v. Rehab, and black indicates CL-VNS v. Delayed VNS). Circles depict individual subjects in d. Error bars indicate S.E.M. All comparisons represent Bonferroni-corrected t-tests to CL-VNS, ***p < 0.0005, **p < 0.005, *p < 0.025. Source data are provided as a Source Data file.

Fig. 4. Closed-loop VNS increases putative central synaptic connectivity of injured networks without altering peripheral reinnervation.

a Representative images depicting muscle fibers from the denervated digit flexors. Muscle fibers were visualized using non-specific background fluorescence. Scale bar is 100 µm. b No differences were observed in fiber area or other metrics of muscle morphology (Rehab, n = 5; CL-VNS, n = 8; Delayed VNS, n = 7; uninjured, n = 6; Kruskal–Wallis test, see Supplementary Fig. 14). Histology on nerve and muscle sections was collected from subjects at week 13. c Representative images of nerves harvested from subjects at the conclusion of rehabilitation illustrating nerve morphology proximal and distal to the injury site in both a Rehab and CL-VNS subject. The distal sections show some reinnervation, but substantially impaired nerve architecture in subjects from all groups. Scale bar is 10 µm. d Nerve injury substantially disrupts nerve architecture distal to the injury site in the median nerve (Rehab, n = 3; CL-VNS, n = 3; Delayed VNS, n = 3, Kruskal–Wallis test). However, no differences in myelin area or other measures of nerve health were observed across groups (see Supplementary Figs. 10–13), indicating that VNS does not influence peripheral reinnervation. e Pseudorabies virus (PRV) tracing of putative central synaptic connectivity reveals that CL-VNS delivered with rehabilitation (n = 5) significantly increased the number of neurons in layer 5 (LV) motor cortex synaptically coupled to forelimb digit flexors compared to Rehab (n = 5), consistent with the expansion of digit flexion movement representations in the cortical maps. Same group sizes in e–g. f No differences were observed in the connectivity of the spared extensor networks. g CL-VNS significantly reduced the percentage of neurons displaying connectivity to both digit flexion and extensor muscles. Circles depict individual subjects in b, d–g. Error bars indicate S.E.M. All comparisons in e–g represent unpaired t-tests, ***p < 0.001, **p < 0.01, *p < 0.05. Source data are provided as a Source Data file.

Fig. 5. Depleting acetylcholine in the brain prevents the reversal of pathological plasticity and blocks motor and sensory recovery after nerve injury.

a Depletion of acetylcholine in rats that receive equivalent VNS paired with rehabilitation (ACh-:CL-VNS, n = 4) blocks the VNS-dependent restoration area of motor cortex evoking movements of the denervated digit flexors (CL-VNS, n = 4; Rehab, n = 4). Additionally, CL-VNS reverses the aberrant expansion of evoked extension movements, while acetylcholine-depleted subjects display an enduring expansion of extension representations. Compare to Fig. 2. b Bubble plots displaying the cortical locations of digit flexion and extension movements across all animals in each experimental group. The size of each bubble represents the proportion of subjects for which ICMS evoked digit flexion or extension movements at each cortical site. Note that depletion of acetylcholine prevents CL-VNS-dependent restoration of digit flexion representations and also prevents the reduction of extensor representations. Same group sizes as in a. c Consistent with blocking the reversal of pathological plasticity, depletion of acetylcholine prevents the VNS-dependent enhancement of recovery of motor function after nerve injury (Rehab, n = 9; CL-VNS, n = 9; ACh-:CL-VNS, n = 5). d ACh-:CL-VNS subjects exhibit substantially reduced levels of motor recovery compared to CL-VNS. Same group sizes as in c. e Depletion of acetylcholine also prevents VNS-dependent recovery of tactile thresholds (Rehab: n = 5; CL-VNS: n = 5; ACh-:CL-VNS: n = 5). Circles depict individual subjects in panels (a, e). In c asterisks indicate significant differences using t-tests across groups at each time point, and the color of the asterisk denotes the group compared to CL-VNS. All comparisons represent Bonferroni-corrected t-tests to CL-VNS, ***p < 0.0005, **p < 0.005, *p < 0.025. Source data are provided as a Source Data file.

Fig. 6. Proposed mechanism underlying VNS-mediated enhanced plasticity and recovery after nerve damage.

a Prior to injury, the majority of motor cortex evokes movements of the digit flexors through the median and ulnar nerves, and a small area evokes movements of the extensor muscles through the radial nerve. Subjects are able to grasp and pull a handle to generate around 150 g of force. b Damage to the median and ulnar nerves generates maladaptive changes in central networks. Cortical drive and synaptic connectivity within the injured digit flexion networks is reduced. In conjunction, networks innervated through the spared radial nerve demonstrate a pathological cortical expansion. Despite reinnervation and rehabilitative training, subjects continue to exhibit maladaptive central changes and long-term deficits in force generation. c Closed-loop VNS paired with forelimb movement during rehabilitation generates timed activation of neuromodulatory networks, including the cholinergic nucleus basalis (NB). This precisely timed neuromodulation enhances synaptic connectivity and cortical drive to increase output to muscles via the reinnervated median and ulnar nerves. Correspondingly, CL-VNS reverses the pathological expansion of extensor networks controlled via the spared radial nerve. Motor function is recovered in the absence of large scale peripheral changes in the nerves or muscle, indicating that central changes can compensate and restore function. d Lesion of the nucleus basalis (NB) prevents acetylcholine release and consequently CL-VNS-dependent central plasticity. No reorganization of central networks was observed in subjects with NB lesions that received CL-VNS. NB lesions prevented the enhanced recovery seen with CL-VNS, providing a causal link between enhanced central reorganization and improved recovery after nerve damage. e CL-VNS is based on precise timing between the activation of neuromodulatory networks and neural activity during rehabilitation. A matched amount of VNS delayed by 2 h from equivalent rehabilitation degrades the temporal association and prevents CL-VNS-dependent plasticity in central networks. Consequently, subjects that receive Delayed VNS fail to demonstrate enhanced recovery of function. This illustrates that timing-independent effects of VNS cannot account for enhanced recovery, and reinforces the notion that plasticity in central networks directed by CL-VNS supports the recovery of motor and sensory function after nerve damage.