Clinical outcomes of peripheral nerve interfaces for rehabilitation in paralysis and amputation: a literature review

Peripheral nerve interfaces (PNIs) are electrical systems designed to integrate with peripheral nerves in patients, such as following central nervous system (CNS) injuries to augment or replace CNS control and restore function. We review the literature for clinical trials and studies containing clinical outcome measures to explore the utility of human applications of PNIs. We discuss the various types of electrodes currently used for PNI systems and their functionalities and limitations. We discuss important design characteristics of PNI systems, including biocompatibility, resolution and specificity, efficacy, and longevity, to highlight their importance in the current and future development of PNIs. The clinical outcomes of PNI systems are also discussed. Finally, we review relevant PNI clinical trials that were conducted, up to the present date, to restore the sensory and motor function of upper or lower limbs in amputees, spinal cord injury patients, or intact individuals and describe their significant findings. This review highlights the current progress in the field of PNIs and serves as a foundation for future development and application of PNI systems.


Introduction
The development of the neurotechnology field, including establishing direct bio-interfaces and neural prosthetics with the nervous system, has impacted the treatment of many neurological and psychiatric diseases [1,2].Electrical neural interfaces are systems designed to detect or modulate the communication of signals between the central nervous system (CNS) and/or the peripheral nervous system (PNS), including through direct stimulation of peripheral nerves to augment or restore body movement in patients suffering from neurodegenerative diseases, stroke, spinal cord injury (SCI), or peripheral nerve injury.The literature is replete with research studies related to CNS neural prosthetics and comprehensive reviews focusing on brain-computer interfaces [3][4][5][6], muscle reinnervation [7], and the clinical outcomes related to each of these topics.There are several prior pre-clinical reports of peripheral nerve interfaces (PNIs) as well [2,[8][9][10][11][12][13][14][15][16][17][18][19][20].However, there is limited availability of in-depth analysis concerning the specific clinical outcomes associated with PNIs, such as the functional efficacy of restoring sensory perception, motor control, or both in patients with peripheral nerve injuries or limb loss as well as the long-term effects of these interfaces on patient quality of life, including factors such as pain management, daily activities, and social integration.
Targeting peripheral nerves may provide high resolution for recording and stimulation interfacing and carries the possibility of directly modulating functional movement in an extremity.This field requires interdisciplinary contributions to achieve the next generation of advanced neural interfaces with high selectivity, minimal invasiveness, noise mitigation, longevity, wireless capabilities, reliable power supply, and functional outcome effectiveness [21].
PNS-based control of prostheses involves interfacing of varying nerve sizes through the recording of neural signals and modulation of others.Recording of neural activity within a nerve or a section of a nerve can be challenging due to signal interference from adjacent muscle activity [22], neural injury during a procedure [23], complex nerve topology and structure, movement of the interface along the nerve, and a possible associated immune response to the implanted material itself [24].Implanting electrodes in the peripheral nerves of amputees or SCI patients enables two-way communication with robotic prosthetic devices.This process involves transferring brain-initiated motor commands, integrating them through crucial regions, and recording them with PNIs.The use of PNIs is considered ideal for embodied prosthetic control, as they allow for the natural mimicry of sensory pathways, facilitating the conveyance of sensory signals from the robotic device to the user.PNI designs, methods, electrode characteristics, their clinical applications, and long-term outcomes have all been explored to address these challenges.However, specific information regarding the safety and biocompatibility of those PNIs is widely variable [25,26], if not sparse, and further studies will be essential to demonstrate the biocompatibility and efficacy of these systems.Moreover, evaluating clinical outcomes may help better understand the challenges that those devices may face in clinical settings and allow scientists to explore avenues for improvement and development of user-friendly and effective PNIs.
Previous reviews have studied concepts pertaining to types of PNI designs [17], PNI applications [8], and strategies and best practices for optimal PNI function [27].However, there are no comprehensive studies to date that summarize the clinical studies performed with PNIs.Here, we aim to explore the utility of PNI for improved functional outcomes and to provide an updated comprehensive review of the clinical studies and trials that were conducted on PNIs, the findings reported by each one of those trials, and the significance of those findings on the treatment of neurological diseases such as stroke, SCI, and peripheral nerve injury.We commence by providing a brief description of the PNS and continue by describing the various types of electrodes and their design characteristics.We then provide a summary of the clinical trials and outline their findings.

Methods of literature search
We intend to provide a narrative review in which we report on various PNI designs and critically discuss the clinical trials conducted on those devices to help practitioners and engineers gain more understanding of the characteristics and clinical outcomes of PNIs.This review comprises studies that were encountered via search in digital libraries such as PubMed, IEEE, Xplore, and ScienceDirect.The search keywords were 'neural engineering' AND 'peripheral nerve interface' OR 'peripheral nerve electrodes' OR 'neural prosthetics' OR 'myoelectric prosthesis' OR 'neuroprosthesis' OR 'peripheral nerve interface outcomes' OR 'clinical trial on peripheral nerve interface' .The article abstracts were scanned for relevance to the proposed topic.References from the obtained articles' reference lists were used to enhance the literature search.Studies that utilized the peripheral nerve interfacing method to restore the sensory and motor function of upper or lower limbs in amputees, SCI patients, or intact individuals, published through 2023, and that meet the National Institutes of Health (NIH) revised clinical trial definition were included in our review.The NIH defines a clinical trial as 'a research study in which one or more human subjects are prospectively assigned to one or more interventions (which may include placebo or other control) to evaluate the effects of those interventions on health-related biomedical or behavioral outcomes [28].'

Anatomy of PNS
The PNS serves as a vital neural network linking the CNS to target organs, encompassing both sensory (afferent) and motor (efferent) divisions.Sensory nerves carry impulses from peripheral organs to the CNS and can be myelinated or unmyelinated.These nerves culminate in sensory receptors distributed across the skin, muscles, or deep tissues [13].On the other hand, motor nerves facilitate the transmission of messages from the brain to peripheral organs.Originating from motoneurons in the spinal cord, these messages terminate at neuromuscular junctions within skeletal muscles [29].The motor division further branches into the somatic nervous system, governing voluntary muscle control, and the autonomic nervous system, which controls involuntary functions.The autonomic nervous system, in turn, comprises sympathetic and parasympathetic divisions, relaying signals to cardiac muscle, smooth muscle, and glands [30].
Peripheral nerves composed of fascicles of nerve fibers or axons, intricately branch out to reach cutaneous, muscular, or visceral targets.These nerves are supported by three distinct sheath layers: the epineurium, perineurium, and endoneurium (figure 1) [31,32].The epineurium, the outermost layer, provides a cushion and supplies blood to the nerves.The perineurium, forming the cover of the nerve fascicle, offers mechanical support and acts as a 'bloodnerve barrier' to maintain the microenvironment and fluid pressure.The endoneurium, the innermost layer, contains nerve fibers, Schwann cells, immune cells, and capillaries, as well as fibroblast and collagen fibers [32].

Electrode types
Various electrode types have been developed to interface with the PNS, as shown in figure 2. We present a brief overview and discussion of extraneural, epineural, helical, book, cuff, intrafascicular, interfascicular, and regenerative electrodes.

Extraneural electrodes
Extraneural electrodes are the least invasive PNI electrodes and are implanted outside the nerve, which can place them at a greater distance from the nerve trunk and help conserve nerve integrity compared to intraneural or regenerative electrodes [8].However, there is still a risk of constrictive nerve injury from extraneural electrodes if the electrode is attached or applied too closely around the nerve.Constriction could restrict nutritional and metabolic supply through the blood vessels and nerve fibers [33].Extraneural electrodes may also cause damage via cabling, producing irregular nerve pressure points with lateral pulling or torquing [34].Cuff electrodes and flat interface neural electrodes (FINEs) electrodes are the most common extraneural nerve electrode interfaces [35].They wrap around the nerve and are sutured to the epineurium.They provide stability and accurate detection of nerve signals but lead migration can be a common failure mode [36][37][38].Some clinical implementations include phrenic nerve stimulation, chronic peripheral nerve pain treatment, and improving foot drop following peroneal nerve injury [39][40][41].We explore different types of extraneural PNI models and their clinical uses.

Cuff electrodes
Cuff electrodes are widely used PNI that consist of tubes wrapped around the nerves.The size of the cuff electrode relative to the nerve is an important design consideration to ensure proper contact without damaging constrictive pressure [42,43].Cuff electrodes and FINE electrodes are the most common epineural nerve electrode interfaces [35].They wrap around the nerve and are sutured to the epineurium.They provide stability and accurate detection of nerve signals but lead migration can be a common failure mode [36][37][38].Some clinical implementations include phrenic nerve stimulation, chronic peripheral nerve pain treatment, and improving foot drop following peroneal nerve injury [39][40][41].Cuff electrodes include the following subtypes: i. FINE are electrodes that flatten the nerve into an oval shape, thus bringing deeper fascicles and axons closer to the electrodes and increasing the surface area [20,44,45].They enable selective nerve stimulation and have shown promising results in clinical studies, including applications in neural prosthetics for gait function [46,47].ii.Split cylinder cuff electrodes are tubular structures with a longitudinal cut that can be sutured or left without suturing around the nerve [17, However, C-FINE electrodes differ by incorporating soft compression, reshaping the nerve from a cylindrical to an elongated oval structure [45,47].These semi-rigid FINE electrodes bring axons in closer proximity to the electrical contacts, although concerns arise regarding chronic damage to the axons.The advantage of FINE electrodes over cylindrical cuff electrodes is attributed to their ability to allow nerve swelling [44].
iv. Spiral cuff electrodes are elastic electrodes that self-size and adapt to changes in diameter.They can snugly fit the nerve and accommodate swelling and fibrosis, perhaps offering protection against tissue encapsulation [50,51].

Book electrodes
Book electrodes consist of silicone rubber blocks with slots containing platinum foils as electrodes [52,53].They are commonly used in SCI patients for augmented bladder control [54].Complications may include motor root damage, pain, and loss of reflexes [15,55].

Helical electrodes
Helical electrodes have an adaptive fit to the shape of the nerve and are adjustable for changes in diameter.They are commonly used to stimulate the vagus nerve for epilepsy or the hypoglossal nerve for obstructive sleep apnea [56,57].Complications may include infections, vascular injury, hoarseness or voice changes, coughing, pain, and electrode dislocation [58].

Other extraneural electrodes in pre-clinical evaluation
The flexible neural clip electrode is used to insert the nerve between the clip and the springs.It has shown utility in pelvic and vagus nerves as well as some of the branches of the sciatic nerve [59].The flexible split ring electrode, similar to the neural clip, applies a slight pressure to maintain close contact of electrodes with the nerve.Split-ring electrodes were shown to be successful in stimulating the sciatic nerve to activate distal muscles [60].The neural ribbon electrode consists of a flexible stripe-like configuration with eight electrical contact sites helically implanted around the nerve, fixed with sutures [61].The neural clip, split ring, and neural ribbon electrodes are polyimidebased.The nano-clip electrode uses a 3D-printed clip to press two carbon nanotube electrodes against the nerves [62] and is directly integrated with multichannel thin-film electrodes suitable for chronic nerve interfacing [63].Zip-tie designs using parylene-based materials allow the electrode to be tied around the nerves [64,65].

Intraneural electrodes
Peripheral nerve intraneural electrodes record or stimulate inside the epineurium sheath of the nerve [66].These electrodes have higher selectivity, specificity, and signal-to-noise ratio (SNR) than extraneural electrodes because their active contacts interface more closely with individual fascicles [67].

Intrafascicular electrodes
Intrafascicular electrodes provide deep penetration into the nerve fascicles, offering robust selectivity for nerve stimulation and a high SNR in recordings [8].
They are typically made of flexible materials and positioned in close contact with targeted axons, thus reducing stimulation thresholds compared to extraneural electrodes [68].Different models of intrafascicular electrodes used in clinical trials are described below.

LIFE, think-film LIFE (tf-LIFE)
The LIFE electrode consists of polytetrafluoroethylenecoated platinum or platinum-iridium wires inserted longitudinally into a nerve fascicle, with a small uninsulated site in contact with axons.These electrodes have demonstrated high sensitivity in recording sensory signals and selective stimulation of nerve fibers within fascicles [69][70][71].The thin film LIFE (tf-LIFE) is a newer version with polyimide-based electrodes implanted transversely into fascicles, thus enabling stimulation of multiple fascicles, and has resulted in positive outcomes in clinical trials involving prosthetic arms [67,72].

Fascicle specific targeting of LIFEs (FAST-LIFE)
The FAST-LIFE electrode introduces a novel neural interfacing strategy that aims to improve the selectivity and specificity of PNIs.The array is composed of traces and thin film electrodes that are placed on a silicon substrate that has been reinforced.Six cuff electrodes are incorporated into the larger body of the array, while eight platinum intrafascicular electrodes are positioned on the thin extension of the array, separated by 0.5 mm [73].

Transverse intrafascicular multichannel electrode TIME
The TIME electrode has a similar design to tf-LIFE but is implanted transversely through the nerve, allowing access to a greater number of axons and higher selectivity [74,75].TIME implantation, combined with non-invasive somatotopic nerve stimulation, has been successful in restoring sensation in patients with transradial limb amputations [76].Clinical trials have also investigated the use of TIME for controlling neural prostheses, resulting in the restoration of touch sensation and enhanced prosthesis control accuracy [77].

Utah slanted electrode array (USEA)
The USEA electrode consists of a plane with an array of electrodes of different heights, enabling simultaneous recording or stimulation of multiple fascicles at varying distances [78].It has been used to selectively stimulate hand muscles and improve muscle activation precision [79].In trials with upper extremity amputees, USEA implantation allowed manipulation of a virtual prosthetic hand [80].

Other intrafascicular electrodes in pre-clinical evaluation
The distributed intrafascicular multielectrode (DIME) combines multiple LIFEs for access to several fascicles within the nerve [81].
The flexible penetrating microelectrode array (FPMA) is designed for increased flexibility compared to silicon, with silicon and polydimethylsiloxane (PDMS) used as substrates for the array structure supporting the silicon microelectrode needles [82].
The coiled wire intrafascicular electrodes (CWIE) are made up of nylon-coated stainless-steel wire.The flexible helical coils on the wire allow it to expand and contract inside the surrounding nerve tissue [12].The CWIE has been successfully implanted into motor nerve fascicles in trials involving rabbits, and it has demonstrated potential for application in neuromuscular electrical stimulation [83].
The adaptable intrafascicular radial electrode (AIR) was specifically designed for pudendal and sacral nerves to target sexual, bladder, or bowel dysfunctions.AIR electrodes demonstrated high selectivity, low invasiveness, faster implants, repeatable stimulation, and the ability to adapt to different nerve sizes and shapes compared to other electrodes [84].

Interfascicular electrodes
Interfascicular PNI electrodes attempt to combine the stimulation selectivity of intrafascicular electrodes with the ease of installation of extraneural electrodes.These electrodes are designed to be placed within a nerve, between the fascicles, by penetrating the epineurium without compromising the integrity of the perineurium [15,85].Slowly penetrating interfascicular nerve electrodes (SPINE) were designed to be implanted within the nerve but exterior to the fascicles [86].These electrodes consist of a siliconebased rubber tube with a blunt element attached for piercing the epineurium during implantation.The electrical contacts are brought closer to the axons within the fascicles without penetrating the fascicle itself.A pre-clinical study on felines demonstrated selective stimulation of nerve fibers after sciatic nerve implantation of a SPINE array, and selectivity varied with changes in the position of the PNI along the nerve's long axis [86].

Regenerative electrodes
Regenerative electrodes represent a notable advancement in PNIs, designed to foster nerve fiber regeneration.These electrodes feature channels or pores where transected nerve fibers can grow and integrate within and across the electrodes, thus allowing for stimulation, recordings, and regeneration within the microchannels [59].Regenerative electrodes offer high spatial resolution and stability but require nerve transection, limiting their use to applications where distal innervation is unnecessary or already with a functional deficit.
The traditional sieve electrode is the earliest model of regenerative electrodes and consists of polytetrafluoroethylene-coated wires with channels for nerve fiber growth [60].Early versions were made of silver wires drilled into an epoxy wafer, while more recent versions utilize silicon and polyimidebased electrodes [87][88][89][90][91]. Studies have shown successful axon conduction and reinnervation of target organs with polyimide sieve electrodes in transected nerves [92].Nerve tape, a recently emerged technology, provides a way to reliably align transected peripheral nerves for regeneration using microhooks to bind firmly to connective tissue [93].It is also the first FDA-approved medical device to surgically repair transected nerves.
The regenerative multielectrode interface (REMI) embeds an array of microwires in a collagen scaffold inside a polyurethane tube [94].REMI allows for stable single-unit recordings for extended periods and has demonstrated successful recordings for almost a full year using biodegradable agarose scaffolds [95].The tissue-engineered electronic nerve interface (TEENI) consists of thread-like arms in a biodegradable hydrogel scaffold, showing promising results with an excellent SNR in recording trials [96][97][98][99].
The microchannel roll electrode is a unique design that facilitates axon recording by guiding their regrowth through microchannels embedded with metal electrodes [12].It aims to maximize the interaction between regenerated axons and the metal electrodes.The electrode microchannel arrays are created using different methods: gold microelectrodes on a polyimide substrate with photosensitive polyimide microchannels [100], or gold electrodes on a PDMS elastomer with SU-8 photoresist microchannels [12].These two-dimensional arrays are then rolled into a three-dimensional structure, similar to a Swiss roll, fitting the transected peripheral nerve at both ends [12].This design enables effective interface and recording capabilities.
A recent innovation, the magnetically aligned regenerative TEENI (MARTEENI), combines various technologies in a single device [101][102][103].MARTEENI employs a microchanneled hydrogel as a structural support and matrix for tissue regeneration.It allows for the customization of electrode density and size to accommodate nerves of different diameters.Implantation of MARTEENI devices in rat sciatic nerves demonstrated successful electrophysiology recordings and analysis of the channel-isolated activity [103].
Finally, regenerative PNIs (RPNIs) graft nerves to small groups of muscles to boost nerve signals through muscle contractions.Ultrasound assessments of RPNIs in upper limb amputees showed prominent contractions during phantom finger flexion and generated electromyography (EMG) signals with a high SNR, enabling real-time control of hand prostheses for extended periods [104].

PNIs for clinical translation
PNIs generally aim to address two main application spaces in functional recovery: to restore sensorimotor function through prosthetics and to improve learning outcomes through neuroplasticity-based feedback.These functional recovery applications have the shared challenges of optimizing task-specific form factors, specificity of neural recruitment, time resolution, efficacy, and longevity of the device [105].By translating PNI systems to human amputees, researchers could decode motor signals [59,60] and provide sensory feedback [47] via neural stimulation.This feedback allows the user to adjust the prosthesis grasping force and perceive different object characteristics, thus improving the functionality and realism of hand prostheses and bringing us closer to creating lifelike replacements for missing hands [77].Research in PNI device development has yielded large advances in signal extraction mechanisms [106].It has also led to the development of biocompatible, flexible, and implantable PNIs [101], and considerations for body augmentation without restricting physical abilities [107].
The design characteristics and evaluation metrics achieved in the PNI field reflect the goals of major funding agencies that sponsored their development.The 2010-2011 DARPA reliable neuralinterface technology (Re-NET) program saw first-inhuman demonstrations of the PNI technology [108,109] and muscles [110].Critically, the translation to human clinical trials opened the possibility of collecting verbal feedback from research participants about induced somatosensory perception [19,111,112] and psychosocial benefits [112,113].
The 2014 DARPA hand proprioception and touch interface (HAPTIX) program aimed to develop fully implantable bidirectional neuro-prosthetics for amputees, to develop motor decoders to translate neural activity from residual nerves or muscles into motor commands for the prosthesis, and to produce proprioceptive feedback [105,114].In 2017, the targeted neuroplasticity training program aimed to improve neuro-prosthetic control and cognitive skills learning by pairing peripheral nerve stimulation with task training and embodied feedback [115].In 2019, under the Horizon 2020 program in Europe, the NeuTouch project aimed to improve artificial tactile perception in robotic limbs and prostheses by understanding how to best extract, integrate, and exploit tactile information at the system level [116].

Biocompatibility and safety
It is essential to consider the biocompatibility and safety of PNIs in clinical trials and for subsequent translation into clinical practice.To encourage longterm tissue integration, the materials utilized in PNIs should have biocompatible qualities [117] such as surface enhancements and coatings to improve electrode-tissue contact, enhancing stability and lowering tissue inflammation [117].PNIs have frequently been developed using biocompatible materials such as titanium, polyimide, and medical-grade silicone [118].Positive tissue response and compatibility with brain structures have been demonstrated for these materials [119].Susceptibility to electromagnetic field interference and tissue reaction, such as with foreign body reactions, inflammation, and fibrosis, are important considerations for PNIs [120].Peripheral nerve injury, which can lead to Wallerian degeneration and fibrosis causing signal loss, is also a design concern [10].Strategies to reduce tissue reaction involve coating devices with anti-inflammatory agents, surface modification, and variation in implantation sites [121].Sterilization methods, such as dry warmth, steam, ethylene oxide, radiation, hydrogen peroxide, and ozone sterilization, are also employed to prevent microbiological colonization during implantation [122].The use of hydrogels in PNIs has been introduced to enhance biocompatibility and mitigate high interfacial impedance.As an example, incorporating polypyrrole and hydrogel into neurostimulation electrodes reduces charge transfer resistance and shows promise in improving performance [123].Wireless bioelectronic neural interfaces have also been developed to prevent local and neural damage caused by conventional neural interfaces [124].
Surgeons with expertise in the technique should be involved to guarantee adequate interface implantation, fixation, and management [15].Adhering to stringent sterility measures during surgery lowers the risk of infection.
Clinical studies ought to establish robust surveillance procedures.Participants are watched in realtime to assist in identifying any negative occurrences or safety issues as soon as they arise.By establishing explicit methods for reporting adverse events, investigators can make sure that any possible hazards or issues are reported and handled properly.
Electrodes must be built to offer safe, regulated electrical stimulation without going above tissue damage limits.Unintentional tissue heating is prevented, and the danger of thermal damage is reduced by monitoring and regulating current density or charge injection levels.Regular evaluations of the mechanical integrity, signal quality, and electrode impedance can assist in spotting possible problems and trigger quick action [15].Redundancy and fail-safe features in the system design help reduce possible damage in the event of failures [15].The safety of PNI electrodes in clinical trials must be guaranteed by close cooperation with medical specialists, adherence to ethical and regulatory norms, and continual attention to safety monitoring.

Resolution and specificity in stimulation and recording
The spatial and temporal resolution of PNIs is crucial for measuring and regulating neural activity.Ongoing efforts focus on improving the design of interfaces, and stimulation protocols, and addressing application-specific resolution requirements [125][126][127].High temporal resolution, generally within the millisecond range, is necessary for real-time neuro-prosthetic interfaces, while fascicle-level spatial resolution is ideal for devices interacting with multiple objects and sensory stimulation patterns.Inflammatory tissue reaction or fibrosis may decrease spatial resolution in vivo [128].Intrafascicular electrodes provide selective stimulation of focused groups of nerve fibers, and using multiple extraneural electrodes simultaneously can enhance selectivity compared to a single electrode [27].FINE and multicontact cuff electrodes have shown increased spatial resolution [129,130].To illustrate the variability in electrode resolution, different electrode types (TIME, LIFE, multipolar cuff) were implanted on a rat sciatic nerve to activate target muscles [124,131].Selectivity of stimulation was evaluated at the interfascicular and intrafascicular levels, with TIME demonstrating selective stimulation at both levels, LIFE being selective only at the intrafascicular level, and cuff electrodes being selective only at the interfascicular level.Further advancement of PNIs will require improved electrode resolution and specificity to allow for increased functionality.
PNIs can be used both to stimulate and to record from the contact site at the nerve.Extraneural electrodes record from the epineural space, which captures the activity from a population of fascicles.Intrafascicular electrodes record from inside individual fascicles.In extraneural electrodes, amplitudes are small compared to intraneural amplitudes, they have a lower SNR, and increased artifacts due to surrounding EMG activation and electrode movement [132].Extraneural and regenerative electrodes have been demonstrated to have long-term recording stability in animal studies [22,133] and human studies [47,134,135].
FINE electrodes flatten the nerve and the contacts are closer to the individual fascicles, improving selectivity in recording compared to cuff electrodes [136].Signal processing algorithms inspired by electroencephalography (EEG) source localization provide the capability of distinguishing activity from individual fascicles in animal studies.These algorithms include linear regression [137], beamforming approaches to reconstruct signals [138], standardized low-resolution brain electromagnetic tomography [139,140], Bayesian source filter for signal extraction [141], independent component analysis [136,142], and hybrid Bayesian signal extraction [136].
Intraneural intrafascicular USEA electrodes have been found to lead to degradation in signal amplitude and increased electrode impedance over time in humans [80,143].The insertion of these electrodes into nerve fascicles leads to an inflammatory response that leads to signal degradation [144].Intrafascicular USEA electrodes have been used to report the effectiveness of electrical stimulation to elicit sensory feedback for 14 months, but not for recording over long periods.TIME electrodes in humans have also been found to present with foreign body response over time, in a study where TIME electrode stimulation was performed for 3 months [145].Time-series recordings of electrode potentials were not explicitly reported.
Extraneural electrodes have been used for stimulation rather than recording in human prosthetic users.Longevity and signal stability limitations in intrafascicular electrodes for human use have found more use as sensory feedback and motor output interfaces than long-term recording devices.
Evaluation of PNI efficacy includes long-term stability, stimulation, and recording performance.Metrics such as the recruitment curve, compound action potentials (CAPs), and correlative signals have been used to measure stimulation and recording capabilities [140,[156][157][158][159][160][161][162][163][164].Recruitment curves, which plot the output signal versus the stimulation charge or other parameters, are frequently used to assess stimulation performance.The effectiveness of stimulation may be evaluated using a variety of metrics, including joint torque, limb motion, muscle force, EMG amplitude, and neural CAP area.Metrics like the selectivity index (SI) or the device SI may be used to measure the selectivity of stimulation, which denotes the capacity to stimulate specific muscles or targets.
PNI recording performance has less detailed characterization [165].Electrically induced CAPs, naturally evoked signals, or spontaneous activity can all be used to record signals [109,164,166,167].
Proprioceptive muscle afferents and volitional motor control signals are examples of meaningful messages that are desired [165].Analogous to those employed for stimulation characterization, correlative measures can help understand recorded signals.It is possible to localize signals inside the nerve trunk by activating distinct distal branches and limiting afferent CAPs to particular fascicles [140,165,168].SNR and confirmation of the neural nature of recorded signals are important considerations [62,164,169,170].The SNR, which is often the ratio of the peak-topeak amplitudes of a spike and the background noise, is frequently cited but not universally defined [169].Verification of neural origin can be performed by applying local anesthetics to the nerve or using techniques to reject interference signals [62,164,170].
Long-term stability is crucial, and changes in parameters and histology should be examined to identify potential issues [108,[171][172][173]. Changes in the input-output characteristics and signal capture might be brought on by tissue reorganization, host reaction, or aging of the device [125].It is crucial to keep track of benchmarks such as electrode impedance, stimulation threshold, and stimulated or recorded objectives while conducting longer experiments [125].Examining explanted devices and doing histological investigations after the study can offer insights into alterations and potential problems that were not previously seen.

Innovation in wireless power delivery
Wireless stimulation systems offer advantages in lead migration and surgical routing, utilizing technologies such as ultrasound and magnetoelectric power [174][175][176].Battery-free bioelectronics offers promising alternatives to bulky standalone PNIs.Using an alternating magnetic field for wireless power transfer, recent advances allow for precision-timed peripheral nerve stimulation [177] a five-fold increase in power density compared to radiofrequency electromagnetic waves for mm-sized wireless bioelectronic implants [178], and in vivo proof of concept for the stimulation of peripheral nerves [175].As an alternative to PNIs that receive external wireless power, battery-free PNIs could also power themselves by harvesting energy from the body itself through kinetic motion [179].Triboelectric nanogenerators (TENGs) have been lately introduced as a promising technology that translates human body mechanical energy into usable electric power, based on triboelectrification and electrostatic induction.To develop battery-free neutral electrodes, scientists connected a sling interface with TENGs which was implanted on a rat sciatic nerve to selectively activate the tibialis anterior (TA) muscle [179].They also demonstrated stimulation of the common peroneal (CP) nerve using the TENGs combined with a pair of platinum/iridium wires to control a TA muscle.Their results showed direct stimulation of the sciatic nerve and a branch nerve in live animals using neural interfaces combined with TENGs.

Longevity
Efforts are underway to improve the longevity of PNI interaction, specifically considering stimulation effects in PNIs [27,126,180].It is crucial to note that different aspects contribute to longevity, with some intricacies more pertinent to stimulation compared to recording situations.Adverse events such as micromotion, wire breakage, and insulation breakdown can shorten the lifespan of implanted interfaces, necessitating frequent monitoring and appropriate action.This includes medical interventions such as device repositioning, repair, replacement, or removal, as well as monitoring the patient's condition, administering necessary medical treatments, and adjusting the implantation technique or materials to improve longevity and functionality [181].
Notably, studies comparing various PNIs have shown varying performance and impedance rates over time [166,182,183].However, due to FDA regulations, longevity studies in human PNI implants are limited to 30 d or fewer due to concerns about patient safety, ethical considerations, and the need for iterative development in a rapidly advancing field [182].

Clinical outcomes 3.10.1. Embodiment and sense of self
One of the most important aspects of the development of neuroprosthetics is embodiment, or the smooth integration of a prosthesis into a person's body image.With a focus on motorized and sensorized prostheses, ongoing endeavors seek to improve prosthetic functioning and lower healthcare costs.Notably, haptic input is prioritized above peripheral nerve stimulation, which increases the adaptability of everyday tasks [184].A precedent-breaking experiment assessed the effects of a sensory restoration prosthesis with direct nerve interfaces on prosthetic usage, functional performance, and psychosocial feedback in a home context by autonomously improving the everyday lives of two patients [113].To evaluate the prosthesis's integration with body image, a rubber hand illusion embodiment survey was carried out using FINEs for electrical stimulation and sensors within a myoelectric hand.Prosthetic embodiment increased significantly, according to the participants, when they used the prosthesis with enabled feeling at home.Improved self-efficacy, embodiment, body image, prosthesis efficiency, and social contact were further validated by patient experience measure questionnaires [185][186][187].
Applications of an electronic cuff to nerves in the upper arm showed promising results for amputees struggling with loss of sensation.The hand somatosensory cortex was directly triggered by this cuff, which was attached to a replacement neuroprosthetic limb, allowing the prosthetic hand to experience unique feelings [188].The amputees accomplished complex motor actions using a range of stimulation patterns over 24 months, converting the usual 'tingling sensation' into a variety of natural sensations.Better control over the prosthesis's grasping force and handling of delicate objects was made possible by these sensations, which were described as vibration, light-moving touch, constant pressure, and natural tapping [188].Peripheral nerve cuff electrodes were used to demonstrate the long-term effectiveness of artificial touch sensation in augmenting prosthesis users' sensory experiences.

Sensory feedback
PNIs are essential for providing sensory feedback to people who have lost limbs or suffered nerve injury.PNIs aid in the restoration of tactile sensations, proprioception, and temperature perception by enhancing the transmission of electrical stimulation to sensory pathways.Through direct connection with peripheral nerves, this sensory feedback allows amputees and SCI patients to duplicate touch sensations, which is extremely important to them.
The influence PNIs have on touch, proprioception, and temperature perception demonstrate the importance of PNIs in sensory feedback.USEA electrodes, which simulate touch sensations on a prosthetic hand or fingers by electrically stimulating residual nerves, have been beneficial for amputees of upper limbs [77].Participants in clinical studies, both amputees and intact, were able to distinguish between textures with prosthetic fingers thanks to high-density electrode arrays delivering intraneural microstimulation [189].PNIs have been successful in restoring proprioceptive and kinesthetic feedback, which is necessary for motor control and coordination [188,190].Furthermore, it has been shown that PNIs can produce temperature sensations selected from a list, including 'cold' or 'airbrush' percepts [80].Designers can incorporate temperature perception by providing electrical stimulation to peripheral nerves and integrating temperature-sensitive materials [191][192][193] into the interface design [194].This capacity is especially important for awareness of thermal safety and environmental adaptation.Thus, PNIs are vital instruments for people who have lost limbs, providing not only functional advantages but also a significant influence on their sensory perception and general quality of life.

Quality of life
The quality of life of people who have suffered nerve injury or limb loss is greatly impacted by peripheral nerve contacts.These interfaces help to enhance different facets of daily life and overall well-being by restoring motor function, providing sensory feedback, and boosting prosthesis control.PNIs enable individuals to regain independence and functionality in their daily activities.By recovering motor control and providing sensory feedback, these interfaces allow individuals to perform tasks such as grasping objects, manipulating tools, and engaging in self-care activities, improving overall functional capabilities [77].PNIs can help improve psychological and emotional health by restoring motor function and sensory input [8].Gaining back control over and engaging with one's surroundings, engaging in sensory experiences, and improving one's body image can boost one's self-esteem, lessen psychological discomfort, and enhance one's general quality of life [195].The incorporation of prosthetic devices into people's bodies and their embodiment are improved via PNIs as discussed earlier [113].These interfaces provide a more natural and seamless contact between the person and their prosthetic limb by offering intuitive control and sensory input, increasing acceptance, and integrating the prosthesis into one's sense of self [184,187].PNIs can enhance prosthesis control and restore motor function, which can have a good effect on social involvement [196].Individuals are more inclined to socialize and partake in hobbies and pursuits of work if they have regained functional capacities and increased independence, which improves social integration and overall quality of life [197].

Clinical trials
PNIs have seen rapid expansion in the field due to an increase in in vivo assessments on a variety of animal models, including rats, pigs, and monkeys [79,157,198].Although there was a significant shift over the last two decades toward human clinical trials, with an emphasis on sensorimotor function in limbs and amputees' control over robotic arms [172,199] (see figure 3), we aim to provide a nuanced analysis of the most pertinent human clinical trials involving diverse PNI designs.We will prioritize important clinical outcomes while focusing on methodological insights (table 1).Accompanying this, a distribution of clinical trials for each target peripheral nerve is illustrated in figure 4.

Extraneural cuff electrodes
Extraneural cuff electrodes have shown promise in regaining functional motions in paralyzed hands due to their effective use in SCI circumstances [200].This discovery presents a technological leap as well as an early look at the revolutionary potential of neural interfacing.Their dependability is further validated by the 28 day assessment period, whereby at least one patient had a 100% EMG detection rate and compliance rate over 90% [201].These results highlight the potential scalability and rapid effect of epidural electrodes in the context of SCI.The investigation of neural stimulation using epineural electrodes in patients with high tetraplegia shows encouraging trends in surgical success, electrode stability, and the induction of selective movements, offering chances to enhance the functioning of people with restricted hand mobility [200].But problems still exist as some studies point out [200,201].
These include the complexity of reaching the ideal synergy adjustments and the need for further surgical operations to address problems such as wrist extension during finger flexion.Regulatory need for a brief trial period is a common barrier that restricts the thorough evaluation of long-term impacts [201].While both studies acknowledge the need for further research to consolidate the proof of concept and refine electrode designs, they also recognize opportunities for expanding inclusion criteria, potentially benefiting patients with diverse tetraplegia profiles [178,201].The constraints imposed by regulatory frameworks, such as the European medical device regulation [201], necessitate careful consideration of study durations and pathways for clinical relevance, creating opportunities for innovation in adaptive technologies.Ultimately, the combined insights from these studies contribute to shaping the trajectory of neural stimulation research for upper limb functionality in tetraplegic patients, pointing towards avenues for improved surgical techniques, electrode design, and comprehensive long-term solutions.
Various notable trends appear from the analyzed literature on neural control systems and extraneural cuff electrodes, showing the dynamic nature of research and development.Previous research [130] has suggested that the focus of investigations into nerve cuff electrodes is on obtaining selectivity, improving stimulation settings, and investigating therapeutic applications.Peripheral nerve cuff electrodes can provide somatosensory neuromodulation and prolonged sensory input, opening up possibilities for improved prosthesis design and modulation pattern investigation [188].Although difficulties with at-home usage and the effects of muscle deconditioning are clear, the long-term stability and dependability of non-penetrating spiral nerve cuff electrodes, as reviewed in one of the investigations [172], imply potential therapeutic implications.Meanwhile, research on thumb force functional electrical stimulation (FES) [202] highlights the need for better electrode insertion methods and FES interfaces that are easy to use.A potential approach to robotic hand control is the extraction of EMG data from supra-lesional muscles [203].Nevertheless, issues like co-contraction and muscle weakening still exist.Studies examining the long-term durability of nerve cuff electrodes highlight the need for predicted intraoperative testing and implant concerns [204].Despite surgical and functional constraints, implantable neuroprosthetic devices provide prospects for improved functional results and extended applications [205].Implanted drop-foot stimulators show promise for long-term therapeutic investigation because of their beneficial effects on walking parameters [206].While there are still issues with surgical complexity and the need for comparative -The nerve signal included data that may be utilized to identify slips when they occur and to intensify stimulation of the thumb flexor/adductor muscles to prevent the slide.
-The system offered a grip that could grab an object if it began to slip for any reason, such as waning muscular force or changes in load forces tangential to the object's surface.
-The technique made it feasible to automatically lower the stimulation level to the lowest setting without losing one's grasp or needing to know in advance an object's weight or surface roughness or the muscular power required to lift it.-The stimulation parameter values necessary to create these sensations did not alter much over time.
-The rate and pattern of progress in the amputees' ability to enhance volitional control over motor neuron activity was comparable to that found with practice in healthy people doing motor activities.b.This method enables amputees to assess and control grip force and joint position in an artificial arm without the use of visual input, creating a foundation for a more seamless integration of the prosthetic limb with the amputee's overall appearance.

Dhillon
(Continued.)At least 6 months -Fifteen participants were implanted and 13 completed the trial.
-Walking speed and the distance covered in 4 min improved over time when stimulated, and the stimulation's positive effects on orthotics improved.
-No significant device-related adverse effects were identified, and the device did not impair the speed of nerve conduction.
-Technical issues were addressed by the long-term follow-up assessment at which improvement in walking was obtained.-Peak isokinetic knee extension moments generated by individual contacts on the nerve-cuffs were equivalent to the response from a single epimysial electrode.

Polasek
-Additional knee extension moment was observed by stimulating multiple contacts within the left nerve-cuff.
-The IST-16 system outperformed the IRS-8 system in terms of maximum standing time and %BW distribution on the legs at 16-, 40-, and 72-weeks post-implantation.
-Even without the additional hip extension provided by certain components, the IST-16 system surpassed the maximum standing time achieved with the IRS-8 system over the five years of use.
-Similar to muscle-based electrodes utilized in implanted functional electrical stimulation systems, the activation variability across time was seen.
-Single-contact stimulation allowed for the selective activation of a single muscle from each electrode, and field steering methods improved the selectivity.
-Selectivity evaluated during the implant operation was consistent with selectivity at three years.-About 85% of trials were correctly classified in real-time by artificial intelligence.
-To enable repeatable and localized hand/finger sensations, a variety of current stimulation types were employed.
-Transcranial magnetic stimulation (TMS) was used to investigate cortical organization, with partial disappearance of symptoms due to phantom-limb syndrome (PLS).--Demonstrating minimal inadvertent activation of non-target areas, a minimum of three out of the four designated targets within the tibial nerve, encompassing two of the three muscles comprising the triceps surae group, were consistently and autonomously engaged in all participants.

Schiefer
-As the acceptable levels of unintended activation slightly increased, the recruitment of the intended muscles correspondingly expanded.
-The precise and exclusive activation of muscles innervated by the peroneal nerve posed a relatively more intricate challenge.
-Of notable significance, estimated joint moments imply the feasibility of achieving adequate plantarflexion for propulsion during the stance phase of gait, as well as sufficient dorsiflexion to counteract foot drop during the swing phase.
(Continued.) -This has facilitated a range of movements encompassing the hand, wrist, forearm, elbow, and shoulder.
-The integration of a mobile arm support was requisite to manage the arm's weight during functional tasks.
-Notably, one participant achieved a degree of independence in various activities of daily living, notwithstanding certain limitations attributed to spasticity.
-The second participant attained partial completion of two specific activities of daily living.10 years -Semi natural sensory information can be transferred to an amputee during the real-time decoding of several grasping tasks to control a dexterous hand prosthesis.
-The participant consistently employed and distinguished between three different force levels.
-By utilizing many aspects of the elicited sensations, a high level of perceptual complexity was attained, enabling the patient to differentiate between the stiffness and shape of three separate items.
(Continued.) 21 years -Despite being scarce at these distant implant locations, isolated action potentials from nerve motor fibers were triggered during fictitious motions of the phantom hand.
-Different digits of the virtual hand were individually controlled with one or two degrees of freedom using online and offline Kalman-filter decodes of thresholded neuronal or EMG pulses.
-Up to 106 separate percepts, spanning a large portion of the phantom hand, were induced by microstimulation through individual electrodes of the two USEAs.
-The individual made distinctions between two somatosensory submodalities at one location and five perceived stimulus locations.
-The virtual hand was controlled by USEA recordings that were entirely or primarily made up of EMG.USEA-evoked percepts were utilized to stop the motions, simulating contact with either a close or distant virtual target.-For the length of this ongoing, chronic investigation, consistent threshold, impedance, and percept regions have shown that the brain interface is robust.
-By exhibiting distinctive percept regions and thresholds for each contact, the investigators were able to produce a selective nerve response using multi-contact cuff electrodes.
-Two upper-limb amputees' selective sensory responses remained constant for 1 and 2 years, making it the longest multi-contact sensory feedback system ever.-More than 80 sensory perceptions were elicited by electrical activation of one USEA.
-Varying the stimulation parameters modulated percept quality.
-For the entire length of the investigation, devices remained intrafascicularly implanted with no appreciable alterations to the SNRs or percept thresholds.
Subject 2: 1.5 years (Continued.) 2 years -The participants responded to electrical stimulation of their somatosensory neurons by performing each of the three psychophysical tasks of intensity matching, magnitude scaling, and intensity discrimination.
-The perceived tactile intensity of stimulation was systematically and cooperatively influenced by stimulation pulse width and pulse frequency, and the artificial tactile sensations could be accurately matched to skin indentations on the intact limb.
-To anticipate the strength of artificial tactile perceptions under all testing situations, the investigators found a variable they called the activation charge rate (ACR), obtained from stimulation parameters.
The ACR is a representation of the overall population spike count in the active neural population based on nerve fiber recruitment principles.
-The results show that sensory amplitude may be consistently changed by changing a single stimulus quantity and that population spike count determines the magnitude of tactile percepts.-All four fibular nerve cuffs in the active individuals produced dorsiflexion moments greater than what was necessary to prevent foot drop.
-For up to 6.3 years after implantation, multi-contact spiral cuffs were extremely selective for various populations of motor units.
(Continued.) -Independent control was achieved for a 5-DOF real-time decode that included flexion/extension of the thumb, index, middle, and ring fingers, and the wrist.
-Proportional control was achieved for a 4-DOF real-time decode -One participant completed a 1-DOF closed-loop virtual-hand movement challenge using input from a USEA-evoked hand sensation.
-The USEA implants did not appear to cause any long-term functional consequences.-The missing toes, foot, ankle, and residual limb were all stimulated by separate connections within the nerve cuffs, evoking repeated feelings of different modalities.
-It is possible for more proximal places on the nerve to trigger the same sensory reactions, according to the substantial overlap in reported locations between distal and proximal cuffs.
(Continued.)  -With an average maximal moment of 0.83 Nm kg −1 , three of the five knee extension contacts activated distinct populations of quadriceps motor units.
-Recruitment patterns stabilized 1-3 weeks after surgery and stayed the same after that.
-Six months after surgery, 14 of 16 contacts continued to recruit the same muscles as they had upon implantation.
- -The application of these strategies facilitated discrimination during object manipulation by the robotic hand, encompassing variations in compliance, shape, and spatial arrangement on the prosthesis.Similar commendable performances were evident when the subjects were tasked with exerting varying force levels in a randomized sequence on a dynamometer through the utilization of either LAM or LFM.
-Exclusive employment of the LAM approach enabled subjects to maintain continuous modulation of grip pressure on the dynamometer.
-The continuous application of LFM generated a rapid adaptation phenomenon when subjected to prolonged stimulation, ultimately leading to the cessation of perceived sensations.In contrast, the LAM strategy, when employed for encoding, allowed for an enduring perception of restored sensations.
-Both encoding methodologies elicited perceptions distinctly dissimilar from the tactile sensations associated with a healthy limb, underlining their artificial nature.  2 years -Despite significant clinical variations among them, every individual reported stimulation-induced sensation from the phantom hand during the whole study period.
-They were able to successfully incorporate sensory feedback into their motor control methods while carrying out experimental assessments that mimicked real-world tasks (both with and without the use of eyesight).
-They reported a reduction in the pain from their phantom limb and an increase in mood overall.-When the motor fascicles were stimulated, the subjects reported deeper proprioceptive experiences as well as tactile and cutaneous sensations.
-Inside sensory fascicles, stimulation thresholds and strength-duration time constants were frequently higher.-Despite the employment of patterned modulation techniques, the quality of these sensations predominantly retained an artificial aspect.

Ortiz
-Sensations emanating from the absent limb, albeit inherently intuitive and consistent in their spatial attribution, retained an artificial quality.
-This quality, indispensable for functional reinstatement, appeared to remain resistant to transformation from paresthesia to more authentic sensory qualia, thus suggesting that the attainment of such natural perceptions entails considerations beyond the scope of patterned stimulation.-The study approach provided enhanced mobility, fall prevention, and agility as seen by the feedback's use in active tasks.

Petrini
-Improved lower limb prosthesis (LLP) embodiment as measured by electroencephalographic recordings, phantom leg displacement perception, and questionnaires, as well as ease of the cognitive effort during a dual-task paradigm.
-Throughout all assessments, sensory nerve action potentials (SNAPs) remained constant.
-In three of the four externalized contacts per cuff, there were only minor alterations to the order of muscle activation, and CMAP initial charge thresholds remained below 50 nC.
-Peak tetanic moments were steady, and over the course of a year, there were bilateral increases in thigh and calf circumference of 5% and 14%, respectively.
(Continued.) -These outcomes affirm the potential utility of neural signals captured through multi-channel intraneural electrodes for accurately decoding diverse movement intentions.
-The ability to consistently employ the same temporal channels for decoding purposes over a span of one week within the initial month underscores promising stability.5 years -Over the course of a year of rehabilitation with standing and quadriceps workouts, the amplitude and area of femoral CMU action potentials increased by 31% and 34%, respectively, while elicited knee extension moments significantly increased by 79%.
-Charge thresholds averaged 19.7 nC 6.2, and they were modest and steady.
-Modifications in the action potentials of the saphenous nerve and needle electromyography showed mild nerve irritation postoperatively.
(Continued.) -This precise approach resulted in the induction of robust and functional grasping motions, which exhibited remarkable stability throughout the entirety of the 28 day implantation period.
-Findings elucidated the participants' ability to initiate the activation of movements within their paralyzed limbs, leveraging an intuitive interface governed by voluntary actions.
-Notably, these individuals demonstrated proficiency in executing practical functional actions, including the adept handling of objects such as a can and the ability to consume fluids through a straw.-Implanted electrodes are incorporated into the neuromuscular constructs, allowing for bidirectional communication with the prosthesis and direct skeletal attachment through a permanent osseointegrated interface.
-A patient with a trans-humeral amputation achieved individual flexion and extension of all five fingers of a prosthetic hand, indicating improved control over the prosthesis.
-The enhanced control resulted in improved functionality of the prosthetic device during daily life tasks.'-In contrast to conventional methods, biomimetic neurostimulations increased movement while requiring less mental effort.research, the integration of osseointegrated prostheses with neural control systems is a notable achievement [207].Selective motions can be induced using multi-contact cuff electrodes for tetraplegia, offering the potential for cost-effectiveness and further testing [208].Together, these investigations open new avenues for study in the field of neuroprosthetics, highlighting the need for continued investigation and improvement in the technological as well as clinical domains.

Extraneural FINE electrodes
Across the studies on FINE electrodes, consistent patterns emerge in the quest for enhanced neural stimulation technologies.Tendon transfer surgery, multi-contact cuff electrodes, and 3D-shaping current configurations showcase current advancements, laying the foundation for restoring functional movements in individuals with complete tetraplegia [113].
Integration of EEG and tf-LIFE signals demonstrates promise in refining movement classification and promoting neural plasticity, despite challenges in charge limitations and temporary sensory elicitation [209].FINE electrodes exhibit potential in muscle selectivity and joint moment estimation, offering avenues for functional restoration, yet subject-specific challenges and limited exploration of stimulus space pose obstacles [210].Selective muscle recruitment through tibial and peroneal nerve stimulation presents opportunities for functional improvements, contingent on addressing challenges in achieving complete selectivity and minimizing undesired moments [211].Bilaterally implanted tibial and peroneal nerve cuffs above the knee showcase stable neuromuscular function enhancement, emphasizing the feasibility and effectiveness of cuff electrodes [212].In the realm of neuroprosthetics, C-FINEs exhibit promising electrophysiological changes, stimulation charge thresholds, and selective muscle activation, opening avenues for clinical applications and rehabilitation potential, albeit with considerations for limited sample size and the need for long-term impact assessment [213].These studies collectively offer a nuanced understanding of the evolving landscape in FINE electrode research, underscoring the potential for significant advancements in neural stimulation technologies for rehabilitation.

Intrafascicular LIFE electrodes
The LIFE electrodes as described in the literature show recurring patterns that point to advancements in improving signal quality and decoding power.
Among the research, one that stands out highlights the application of wavelet denoising techniques to improve the SNRs of recorded LIFE motor signals [72].Furthermore, it is proposed that the investigation of spike sorting algorithms for neural signal decoding can provide better performance than conventional techniques that depend on power-related data.Building on this trend, a different study integrates afferent stimulation to elicit distinct tactile sensations, demonstrating the regulation of intensity with stimulus frequency, in addition to seeing a gradual improvement in the SNR of tf-LIFE4s post-surgery [67].Improved motor control is also facilitated by the combination of training, sensory input, and sophisticated signal-processing techniques [67].Nevertheless, these developments encounter obstacles including the learning curve associated with user adaptation, difficulties with signal temporal variability [72], and the brief sensation elicitation length [67].Apart from impediments, other study gaps include the short experiment durations [72] and the requirement for histological confirmation of tissue responses following electrode removal [67].The latter creates a crucial void in our knowledge of the long-term impacts on tissue and its ramifications for the functioning of electrodes.Multichannel recording [72] presents research opportunities going forward, providing a technique to improve classification performance by obtaining more data.
There are opportunities to expand the capabilities of neuroprosthetic systems, such as the ability to decode not just basic control signals but also more complicated information about grip kinds [72] and investigate closed-loop control based purely on brain control and feedback [214].Furthermore, the FAST-LIFE array's selectivity and specificity [73] present prospects for a range of neural interface applications, including possible uses in lower limb nerves and treatments such as vagal nerve stimulation.The invasiveness of the FAST-LIFE array surgical process is recognized as a possible disadvantage, highlighting the need for more research into long-term consequences and optimization.Notwithstanding these challenges, these investigations' overall implications are encouraging as they suggest improvements in signal quality, the decoding of complex data, and possible practical uses for prosthetic control.Bioelectric brain interfaces may serve as the foundation for future human-machine symbiosis, according to investigations on bidirectional, closed-loop neural interfaces [73] and the integration of intrafascicular interfaces [72].However, the literature highlights the need for more development and thorough research in subsequent studies to address issues with user adaptation, electrode placement variability, and the long-term stability of neuroprosthetic technologies in practical settings.The Scorpius system is a paradigm for attaining high-accuracy control in prosthetic hands with impressive degrees of freedom, integrating neural recording, FAST-LIFE microelectrode arrays, and AI models [215].In addition to being a significant technological advancement, this combination of cutting-edge technology has the potential to close the gap that now exists between the human mind and machines by using the peripheral nerve route.Future studies should focus on improving AI models and enhancing the Scorpius system to make it more suitable for wider uses in neuroprosthetics.

Intrafascicular TIME electrodes
When it comes to sensory input for prosthetic limbs, in particular, the integration of TIME electrodes in neuroprosthetics poses both continuous challenges and state-of-the-art advances.Clinical trials using TIME electrodes are the source of this data.With an emphasis on precise force modulation and bidirectional control, the goal of achieving lifelike control over prosthetic hands has made great progress [77].The limitations of generalizability and the durability of user performance over extended periods, however, emphasize the need for robust longitudinal studies and larger participant cohorts [77].Furthermore, the potential use of neuromorphic approaches in prosthetic design is made possible by the encouraging outcomes of needle microstimulation and TIME-based stimulation [189].While this indicates promise, a study acknowledges the limits of a limited sample size and suggests deeper investigation into more complex circumstances and stimuli to fully understand its use [189].Further studies on the stability of TIME electrodes have demonstrated positive benefits on prosthesis control, mood improvement, and phantom limb pain relief [198].However, issues with infection, strange sensations, and connection reliability persist, requiring efforts to enhance the naturalness of sensory feedback and enhancements to connector design [198].The study of proprioceptive acuity restoration and efficient information exploitation shows how functional performance may be enhanced by sensory replacement [172].While admitting the limits imposed by delays in prosthetic limb controllers and the cognitive effort involved with sensory replacement, a prior study [172] provides directions for further research.This is particularly true for jobs that need for feedback for all five fingers and finer position sensitivity.Furthermore, a study highlights the need to incorporate real-time sensory feedback into prosthetic limbs [199].However, the challenges posed by a limited sample size and a short testing time raise doubts about the neuroprosthetic system's long-term sustainability and practical applicability.The ongoing efforts to bridge these gaps are highlighted by the requirement for more clinical research and the development of novel encoding methods.All of these findings contribute to the fascinating advancement of intraneural electrodes in neuroprosthetics and offer a glimpse into the ground-breaking potential for prosthetic limb functioning.However, to optimize these technologies for broader acceptability, prolonged usage, and increased user experience, continued research and development are necessary owing to persisting hurdles and uncertainties.

Intrafascicular USEA electrodes
USEAs are crucial components that have contributed significantly to the advancements that are discussed in a few of the clinical research.Effective decoding of movements and stimulation-evoked sensory impressions opens up possibilities for potential integration into the subjective physical representation and proportional control of virtual digits [80].Nevertheless, those studies recognize the challenges posed by crosstalk, mechanical failures, and impedance fluctuation and emphasize the necessity of greater mechanical durability and crosstalk reduction for clinical use.A previous work using USEAs to demonstrate multi-DOF control and advanced neural decoding advances the discussion [216].The research opens the idea of a USEA-enabled wireless system with closed-loop control, offering opportunities for useful prosthetic applications.Yet stability problems and implant-related challenges, such as closed-loop evaluation and infection control, show how important it is to get rid of these roadblocks to allow for practical use.Nonetheless, several studies focus on the long-term effectiveness of USEAs, indicating the emerging trend of assessing their robustness and validating their enhanced functioning [143].While technological innovations like extending the helical wire bundle show promise, electrode stability issues and performance variations over time remain major obstacles [143].These findings demonstrate the need for more research to completely understand longevity-related factors and to continuously advance USEA technology.The integration of USEAs into a neuromyoelectric interface for transradial amputees [217] demonstrates the trend toward the development of effective interfaces for controlling and sensing prosthetic limbs.In dexterous sensorized prostheses, experiments demonstrate enhanced fine motor control and the potential for biomimetic sensory feedback [217].However, stimulation artifacts pose a problem, and further research is required to close gaps in the literature caused by insufficient quantitative metrics and a lack of studies on long-term effects.All things considered, these studies demonstrate the positive advancements in USEAs.They highlight the potential for more proportional control, realistic feedback, and motor intent evaluation.However, problems like as instability, artifacts of stimulation, and unpredictability in impedance necessitate concerted efforts to advance design, refine technology, and conduct comprehensive, long-term studies.

Regenerative electrodes
The development of RPNIs, which exhibits their potential for precise PNI, functional selection during volitional movements, and long-term stability, is the overarching trend [104].However, a crucial obstacle is found in the discomfort caused by the needle during implantation and the difficulties in isolating RPNI motions, especially in distal forearm regions, which may prevent general adoption [104].A notable deficiency is that there is not a direct comparison with current surgical techniques, including targeted muscle reinnervation (TMR), and there are safety concerns about the number of intramuscular electrodes used [104].The aforementioned trends and obstacles present opportunities for improving prosthetic control, mitigating neuroma pain, and advancing a bidirectional closed-loop prosthetic paradigm [104].A study that highlights pattern-recognition algorithms like the Hidden Markov Model with Naive Bayes (HMM-NB) classifier [218], also focuses on high accuracy and speed.Although demonstrating a promising use of real-time control in functional prosthetics, obstacles include difficulty in adapting to various circumstances, constraints in continuous speed modulation, and issues in optimizing feedback mechanisms [219].Prospective avenues for further investigation encompass investigating osseointegration, refining the system's capacity for generalization, accomplishing continuous speed modulation, and including upgraded feedback systems to provide more organic prosthesis control [219].This emphasizes how crucial it is to improve user experience and handle environmental differences in neuroprosthetic technology.Another investigation [220] highlights decreased calibration demands and increased grip selection performance utilizing RPNIs.Realworld obstacles, on the other hand, include things like hardware reaction times and efficient object interaction during physical activities [220].Limitations are demonstrated in its low generalizability because it only involved one person.Another is the lack of validated tests to measure prosthetic grip transitions.Possibilities include improving multi-grasp control with dynamic wrist motion and lowering cognitive burden in comparison to traditional commercial control systems [220].This emphasizes how important it is to carry out more studies in real-world contexts to confirm and improve the suitability of RPNIs and intramuscular EMG in a variety of circumstances.In summary, the combined research highlights the complex environment around RPNIs and highlights their potential benefits for neuroprosthetics.To improve the usefulness and application of RPNIs in the field, more research, comparison analysis, and real-world validations are necessary, as shown by the hurdles and gaps that have been uncovered.

Integrating cuff, epimysial, and intramuscular electrodes
Recent groundbreaking research investigated a novel surgical technique that aims to transform the control of myoelectric prosthetic limbs in patients who have had upper limb amputations [221].The study tracks one transhumeral amputation patient through a multi-stage surgical repair that includes electrode placement into free muscle grafts and reinnervated muscles as well as nerve transfer and transection [221].The main advantage is the stability the neuromusculoskeletal interface showed over two years.Undesired cross-talk between myoelectric sites is one difficulty, indicating challenges with obtaining the best possible signal separation and specificity [221].
The findings present exciting new directions for sophisticated and skillful prosthesis control that might enhance amputees' quality of life.Before widespread clinical use, the study highlights the necessity for more research, comparative investigations, and the disentanglement of components leading to functional improvement.In essence, this research marks a major advancement in the smooth integration of engineering and surgical technologies to improve prosthetic results.
In conclusion, the wide range of neural interfaces that have been demonstrated in clinical trials highlights both the advancement of technology and its potential to significantly improve the lives of those who have limb abnormalities.To ensure that neural interfaces grow into more than simply tools, but smooth extensions of the human experience, further research and development efforts should focus on issues of long-term stability, biocompatibility, and integration into daily life as these technologies advance.

Conclusions and future directions
In summary, a great deal of research has been done on different PNIs in preclinical and clinical contexts, which highlights the increasing potential for improving neuroprosthetic devices.The extraordinary success in recovering sensorimotor function and enhancing the quality of life for people with amputations and SCIs has been demonstrated by the clinical studies discussed here.These investigations span from one month up to more than six years of follow-up postimplantation [172].
Critical insights into the viability and effectiveness of these technologies are provided by the human clinical trials that have been carried out, particularly those that deal with epineural, extraneural cuff, extraneural FINE, intrafascicular LIFE, intrafascicular TIME, intrafascicular USEA, regenerative electrodes, the Scorpius electrode system, and integrated electrode approaches.Even while encouraging results have been documented, it is critical to recognize that there are still certain restrictions and knowledge gaps that need to be filled.
A noteworthy finding is that PNI longevity varies depending on the application.Prolonged research has indicated potential benefits for upper limb applications, namely in terms of regaining functional motions and offering sensory input.Still, there are difficulties in maintaining sustained performance after two years, particularly with neuroprosthetic devices.Future studies should focus on determining the variables affecting PNIs' long-term stability and creating plans to increase their robustness for ongoing therapeutic benefit.Moreover, several studies have shown the advantages of incorporating tactile sense into functional rehabilitation.The potential for improving natural function in patients with SCIs and neuroprosthetics is highlighted by the increased neuroprosthetic control and functional outcomes attained in both upper and lower extremity applications.Future directions should focus on refining techniques to incorporate tactile sensation effectively, thereby contributing to increased user satisfaction and overall device acceptance.
The development of FINE electrodes has shown encouraging outcomes, especially when it comes to giving amputees of the upper extremities sensory feedback.Biocompatibility, long-term stability, selectivity, and spatial resolution should all be further improved.It makes sense to extend the use of FINE electrodes to lower limb prostheses to fully use their potential advantages for a wider range of neuroprosthetic therapies.Patients with upper extremity amputations have successfully produced sensory perception with intrafascicular USEA electrodes.But longevity-related issues still exist, therefore more research on ways to boost biocompatibility for better long-term stability is necessary.Regenerative electrodes have great potential for improving upper extremity motor control, providing dependable and consistent control over prosthetic limbs.To enhance functional results, future research should focus on enhancing the efficacy of regeneration electrodes by delving further into the processes behind the observed plasticity.With its combination of brain recording, microelectrode arrays, and AI models, the Scorpius electrode system offers a viable path toward high-accuracy control of prosthetic hands.Further investigation and advancement in this area may provide significant advances towards the construction of resilient and dexterous control mechanisms for contemporary neuroprostheses.The integration of intramuscular, epimysial, and cuff electrodes is a new method that may help improve motor neuron instructions.Research on the division of severed nerves into fascicles to simultaneously innervate different muscle targets offers promising opportunities for enhancing control over prosthetic limbs.Subsequent efforts need to persist in refining these electro-neuromuscular structures, with a focus on their utilization in augmenting motor neuron instructions.
In conclusion, while substantial progress has been made in the field of PNIs, ongoing research endeavors should address challenges related to longevity, stability, and broader applicability.To advance the discipline, future research must not only improve on current technology but also investigate cutting-edge strategies and interdisciplinary collaborations.By systematically addressing these challenges and building upon the foundation laid by current clinical trials, the field of PNIs holds great promise for revolutionizing neuroprosthetics and improving the lives of individuals with limb loss and spinal cord injuries.

Figure 2 .
Figure 2. Classification of the different spatial designs of neural electrodes.

Figure 3 .
Figure 3. Evolution of PNI trial over the years.
first instance of amputees receiving direct cerebral control over and feedback from an artificial arm.
2010) -Three actions were controlled in real time by motor activity.
up to 11 years.-After2-4.5 years, all four subjects wearing spiral cuffs on their femoral nerves were able to generate enough moment to keep their knees locked while standing.
post-injury -Three connections in each C-FINE produced hip flexion while three contacts in each were selected for knee extension.
years -FAST-LIFE arrays allow for targeted, long-term electrical stimulation of specific peripheral nerve fascicles.
unveiled that the perceived location of sensations was characterized by distal reference and somatotopic congruence.
leg neuroprosthesis that stimulated nerves to give three transfemoral amputees proprioceptive sensation and tactile feedback in real time.
the stimulation contact points of the TIMEs were effective and stable electrically during the whole implant time.-Afterexplanation, optical investigation showed that at the metallization-PI interface, 62.5% of the stimulation contacts to other interfascicular PNIs, USEAs offer the best functionally recording and functionally stimulating electrodes.
spatio-temporal neural dynamic was more reminiscent to the naturally elicited one because of biomimetic stimulation, which caused a neural activity to transit consistently along the neuroaxis.

Figure 4 .
Figure 4. Distribution of clinical trials by nerve type.

Table 1 .
Summary of clinical trials investigating PNIs.

Table 1 .
Notably, in each instance, the muscle that demonstrated selective activation was the first to branch distally from the location of the cuff. -

Table 1 .
Only muscles triggered by neighboring connections were affected by changes in recruitment in the remaining 2 contacts.