Advancing Prosthetic Hand Capabilities Through Biomimicry and Neural Interfaces

Abstract

Background and Objectives

Prosthetic hand development is undergoing a transformative phase, blending biomimicry and neural interface technologies to redefine functionality and sensory feedback. This article explores the symbiotic relationship between biomimetic design principles and neural interface technology (NIT) in advancing prosthetic hand capabilities.

Methods

Drawing inspiration from biological systems, researchers aim to replicate the intricate movements and capabilities of the human hand through innovative prosthetic designs. Central to this endeavor is NIT, facilitating seamless communication between artificial devices and the human nervous system. Recent advances in fabrication methods have propelled brain–computer interfaces, enabling precise control of prosthetic hands by decoding neural activity.

Results

Anatomical complexities of the human hand underscore the importance of understanding biomechanics, neuroanatomy, and control mechanisms for crafting effective prosthetic solutions. Furthermore, achieving the goal of a fully functional cyborg hand necessitates a multidisciplinary approach and biomimetic design to replicate the body’s inherent capabilities. By incorporating the expertise of clinicians, tissue engineers, bioengineers, electronic and data scientists, the next generation of the implantable devices is not only anatomically and biomechanically accurate but also offer intuitive control, sensory feedback, and proprioception, thereby pushing the boundaries of current prosthetic technology.

Conclusion

By integrating machine learning algorithms, biomechatronic principles, and advanced surgical techniques, prosthetic hands can achieve real-time control while restoring tactile sensation and proprioception. This manuscript contributes novel approaches to prosthetic hand development, with potential implications for enhancing the functionality, durability, and safety of the prosthetic limb.

Introduction

Researchers draw inspiration from biological systems to devise innovative strategies for developing prosthetic hands, employing biomimetic design principles to simulate the human hand’s structure, function, and control mechanisms. These approaches have revolutionized the field, results in transformative advancements.^{1 -3}

The core of this revolution is the rise of neural interface technology (NIT). An intersection where artificial devices seamlessly connect with the human nervous system. The primary goal of NIT in the context of artificial limb is to facilitate bidirectional communication, allowing for the exchange of information between the central and peripheral nervous systems (PNSs) and neuroprosthetics.^4,5

These technologies are advancing in a way that requires a multidisciplinary approach, bringing together specialists from various fields. Clinicians, who have direct patient contact, play a central role in this transformation and can leverage empathy to inspire innovative thinking. With a solid understanding of relevant scientific principles, clinicians can transform challenges into opportunities, driving groundbreaking advancements in prosthetic technology. Figure 1 illustrates a composite representation of how a multidisciplinary approach fosters the development of next-generation cyber prosthetic hands. The figure integrates key frameworks, including biological elements, neural interfaces, material sciences, closed-loop control systems, and biomechanics, highlighting their interconnections and contributions to advancing functionality, stability, and user experience in prosthetic design.

Figure 1.

The composite figure demonstrates the integration of various neural interfaces, including cuff electrodes for somatosensory feedback (A1, Raspopovic et al ⁶ ) and regenerative, flexible electrodes for nerve and muscle integration (A2, Zellmer et al ⁷ ); It also highlights biological components (eg, B1 and B2), ⁸ and advanced neural dust technology (C1, C2: US Patent 11033746B2)“ ⁹ ”, all working together to create a biomimetic cyborg hand for individuals with limb loss.

Recent advances in fabrication methods have led to a significant rise in brain computer interfaces. ¹⁰ These interfaces enable the recording of neural activity, offering valuable data for controlling prosthetic hands. Central nervous system (CNS) interfaces directly access cortical or spinal cord neurons, while PNS interfaces, capture PNS motor and sensory signals. ¹¹

From the ancient “Cairo Toe” to Götz von Berlich’s iron prosthesis, and modern developments trace back to the 19th century, several advancements in artificial limb development have been achieved.^12,13 Therefore, the future prospects for enhancing artificial hands seem exceedingly optimistic. ¹ A significant opportunity lies in integrating bidirectional peripheral nerve interfaces and regenerative peripheral nerve interfaces (RPNI) within an implant device, as depicted hypothetically in Figure 2.

Figure 2.

A schematically designed implanted device for prosthetic hand control. The device can be made of silicon-based organic polymers, housing tissue-engineered skeletal muscles, which are vascularized through methods artificial membranes. The muscle structures are innervated by regenerative nerves, and bidirectional neural interfaces. Neural dust devices wirelessly connect to antagonist muscles, facilitating proprioception through agonist–antagonist interactions. The system integrates real-time sensory feedback, motor control, and protective mechanisms to enhance the precision, safety, and functionality of the prosthetic hand.

In an embodiment, merging RPNI and perineural interfaces could provide a novel setting for controlling prosthetic hands. This promotes nerve growth and sprouting through the grafted muscles, while restoring proprioception through wireless or Bluetooth or direct connection between agonist and antagonist regenerative myoneural interfaces (ARMI). The Massachusetts Institute of Technology has introduced the agonist–antagonist myoneural interface (AMI) in case of lower limb loss. The AMI is a method to connect nerves within their residuum to external bionic prosthesis. It comprises 2 muscles that are surgically connected. An agonist and an antagonist. When the agonist muscle contracts upon electrical activation, it stretches the antagonist one. This muscle dynamic interaction causes sensors within the muscle tendons to send the information through the nerve to the CNS relating information on the muscles and tendons’ length, speed, and force. Similar techniques can be used to restore proprioception through an Agonist–ARMI.^14,15

This paper offers a comprehensive framework that begins with an in-depth exploration of anatomical and biomechanical concepts, followed by discussions on the integration of NIT and biomimicry in prosthetic hand design. This approach could lead to a transformative innovation that dynamically adjusts and speed, decodes intended motor commands and even enables ballistic movements and reflexes. This system not only mimics the natural hand’s function but also opens doors to the next generation of the artificial prosthetic hands.^16,17

Exploring Biomimicry in Prosthetic Hand Design

The development of a biomimetic prosthetic hand requires a deep understanding of both gross and microscopic anatomy, along with the biomechanical functions of the musculoskeletal system. This knowledge is crucial for designing a prosthesis that replicates the functionality of a natural hand and integrates advanced control systems using the remaining anatomical structures.

Connective Tissues

The body’s ability to retain structural integrity depends on connective tissue. It may be generally divided into 2 categories: specialized connective tissues and connective tissue proper. Developing biomimetic prostheses that mimic natural tissue qualities and incorporate cutting-edge control systems requires a deep understanding of these tissues. Connective tissue proper interacts with nerve and muscle tissues, offering structural support and regenerative capabilities. This category includes loose connective tissue, such as areolar tissue, which provides flexibility and dynamic support, making it beneficial for prosthetic interfaces. Dense connective tissue—comprising regular tissues (eg, tendons and ligaments), irregular tissues (eg, dermis), and elastic tissues—delivers tensile strength, multidirectional resilience, and stretch-recoil ability. In biomimetic designs, incorporating hydrogels that emulate these connective tissue properties can significantly enhance the functionality and durability of implanted devices. ¹⁸

By mimicking the properties of connective tissues with conductive or non-conductive biomaterials, such as hydrogels, biomimetic designs can optimize and miniaturize interfaces and circuits, significantly improving the efficiency, functionality, and durability of prosthetic systems.^{18 -21}

Specialized connective tissues—such as bone, cartilage, blood, and adipose tissue—play crucial structural and physiological roles, including support, protection, nutrient transport, and storage. These factors are vital considerations in the development of advanced prosthetics. Using biomaterials and bioengineering processes, these specialized tissues can be engineered, incorporating grafts or preserving functional structures at the site of limb loss.^22,23

Nerves, Muscles, and Tendons: Functional (Control) Units

A thorough understanding of the microscopic anatomy of nerves, muscles, tendons, and their associated connective tissues is essential for developing advanced prosthetic interfaces that effectively replicate the functions of a natural limb.

A comprehensive understanding of nerve fiber types and their functions is essential for developing effective interfaces in sophisticated prosthetic devices. There are 3 main types of nerve fibers: A, B, and C, each with a distinct role. ²⁴

Type Aα fibers (specifically Ia and Ib) are critical for proprioception, which senses muscle length and tendon tension through muscle spindles and Golgi tendon organs. These primary motor fibers are responsible for transmitting motor signals from the CNS to the muscles. While there are various other types of sensory fibers, Aα fibers play a key role in controlling motor instructions. Notably, proprioceptive feedback involving these fibers is crucial for fine-tuning motor control systems and enhancing feedback mechanisms in movement.^25,26

This highlights the importance of integrating motor control commands and signals from these agonist–antagonist interfaces to manage various types of movement, including discrete, cyclic, and ballistic motions.

Layers of connective tissue play a vital role in maintaining the structural integrity of nerve fibers. The endoneurium surrounds each individual nerve fiber, facilitating the safe transmission of electrical impulses, while the perineurium encloses specific fascicles, and the epineurium provides overall protection. To enable effective 2-way communication between the prosthetic device and the user’s nervous system, it is essential to replicate these connective tissue layers in prosthetic designs using biomimetic materials.^{27 -29}

Additionally, the regenerative capacity of nerves presents transformative potential for prosthetic integration. Neural sprouting, in which axons develop new branches to repair damaged networks, can be stimulated to reestablish motor and sensory connections, enhancing the control and functionality of prosthetic devices.^30,31

Conduits can be utilized to guide nerve regeneration and to position intrafascicular interfaces, allowing for the integration of wires within specially designed hydrogels. This approach directs signals from the interface toward circuits and microcontrollers while preserving the integrity of the nerve fascicles. Another effective strategy is autologous nerve grafting, which involves using the patient’s own nerve tissue to bridge gaps and facilitate axonal regeneration.^32,33

This method is essential for reattaching damaged nerves to the muscle scaffolds and for creating a biological framework for nerve regeneration.^{34 -36} Figure 3 illustrates how distal portion of cut nerves could be regenerated toward the target points.

Figure 3.

The regeneration process of the distal portion of severed nerves, showing how natural nerve sprouting can be directed toward target points. Depending on the level of amputation and anatomical features, nerve grafts or conduits can be employed to guide the regenerating nerves and promote functional recovery.

Furthermore, advanced microscopic imaging techniques, such as magnetic resonance microscopy and high-frequency ultrasonography, provide invaluable insights into fascicular topography. These methods enable physicians to precisely map nerve fascicles, allowing for targeted nerve regeneration toward specific motor units and the careful placement of intraneural interfaces. Additionally, particular fascicles can be selected from the distal end of the severed nerve. By encasing non-conductive fascicles with non-conductive materials like hydrogels, the nerve can be guided more effectively toward the desired muscle scaffold, enhancing both sensory and motor recovery.^{37 -40} Theoretically, single-fascicle nerve grafting, as highlighted by Tzou et al, ³⁹ allows for more targeted nerve repair, can be used in these cases as well. By utilizing non-conductive materials to guide the axonal growth toward desired functional outcomes, clinicians can optimize the integration of prosthetic devices with the patient’s nervous system.

Even with these advanced methods available, the optimal approach for each patient must be personalized. Clinical decisions should consider factors such as the patient’s individual functional goals and the extent of amputation. For instance, the needs of an elderly patient may differ significantly from those of a painter or musician. Therefore, the best outcomes can only be achieved by tailoring the technique to each patient’s specific requirements and desired level of functionality.

While the nervous system regulates movement, muscles act as the primary functional units responsible for that movement. In prosthetic design, replicating this structure and function involves creating actuators and control systems that emulate natural muscle movements and force generation. Muscle fibers can be classified into 3 primary types: Type I fibers (slow-twitch, endurance), Type IIa fibers (fast-twitch, versatile), and Type IIb fibers (fast-twitch, power). The functional units of muscle fibers, known as sarcomere-based myofibrils, are responsible for contraction. ⁴¹

The transmission of impulses that trigger muscle contraction largely depends on the neuromuscular junction, the connection between motor neurons and muscle fibers. This contact is maintained by the connective tissue layers surrounding both muscles and nerves. To achieve effective control in prosthetic design, especially for tasks requiring fine motor skills, it is essential to replicate the neuromuscular junction to enable accurate and coordinated movements. Figure 4 illustrates a schematic of 3D muscular scaffolds designed to be innervated by regenerating nerves. These structures may incorporate artificial membranes intended to simulate the functionality of arterial walls.

Figure 4.

Schematically represents 3D muscular scaffolds designed to be innervated by regenerating nerves. These structures need to be supported by artificial membranes intended to mimic the properties of arteriole walls, aiding in the integration and functionality of the scaffolds in regenerative medicine applications.

Furthermore, techniques such as muscle grafts and bioengineered muscle scaffolds can significantly enhance the functionality of prosthetic systems. Depending on the level of amputation and the available tissues, clinicians can utilize remaining muscles or tendons to improve function, including methods like target muscle reinnervation. Additionally, bioengineered scaffolds that mimic tendon structure can be surgically attached to the remaining muscle, integrating with other interfaces, such as mechanical or osseointegrated systems, to enhance connectivity and overall function.^8,42

It is important to note that a biomimetic approach aims to achieve the best outcomes for each patient by utilizing their existing tissues. In this context, clinicians play a vital role by selecting appropriate strategies and making informed decisions based on each patient’s specific needs and functional goals. Their expertise is essential for addressing the challenges of tissue integration and prosthesis design, ensuring optimal results.

Tendons connect muscles to bones, transmitting the force needed for joint movement. They consist of hierarchical connective tissue layers: the epitendon provides overall structural integrity, the endotendon separates fascicles to facilitate smooth movement, and collagen bundles contribute to strength and flexibility. This highly organized structure enables tendons to withstand repetitive mechanical loads effectively. It is essential to comprehend this structure in order to build prosthetic devices that mimic the mechanical characteristics of tendons, improving the device’s suppleness, robustness, and bodily integration. A more functional and lifelike performance may be obtained for prostheses by employing biomimetic materials that mimic these characteristics. ⁴³ Within this context, hydrogels—which are widely used in prosthetic development as biomimetic materials—are engineered to mimic the thick, regular connective tissue present in tendons. These water-rich polymer networks align structurally to emulate the mechanical characteristics of natural tendons, allowing for the requisite flexibility to be maintained while successfully absorbing mechanical stress. Because they act similarly to real tendons, they are especially well-suited for conveying forces from artificial muscles to a mechanical interface, which might be utilized to change the direction of the motions.^{44 -46} Furthermore, their innate biocompatibility facilitates better integration with biological tissues, enhancing prosthetic device efficacy and comfort while reducing the likelihood of mechanical failure. ⁴⁷ Sun et al ⁴⁸ developed a notable tendon-mimetic hydrogel using the anisotropic structure of aramid nanofiber composites, contributing to ongoing advancements in this field. These hydrogels are promising for applications in tissue engineering and prosthetic technologies, as they effectively mimic the structural interactions between aligned collagen fibers and proteoglycans.

Bones, Ligaments, and Joints: Structural and Supportive Units

Bones, ligaments, and joints are important structural components of the upper limb that allow for coordinated movements, stability, and guidance of movement. As hard, mineralized connective tissues, bones sustain the body’s structure, safeguard internal organs, and enable movement by acting as sites of attachment for tendons and muscles. Ligaments composed of thick regular connective tissue stabilize joints where bones articulate by joining the bones together and keeping the joints stable.^43,49

Ligaments provide sensory feedback on joint tension and position, which allows them to play an active part in proprioception even though they are passive. For the purpose of preventing injuries, fine motor control, and coordinated muscle action, this input is crucial. Ligaments also aid in dynamic stability by controlling the activation of muscles during movement and guaranteeing that joint movements remain within a safe range, preventing damage to tendons and muscles. To create bionic hands that move naturally, steadily, and safely, it is essential to replicate these structural and proprioceptive processes.^50,51

Skin and Soft Tissue Management in Prosthetic Design

The role of skin in prosthetic design extends far beyond physical coverage, acting as a critical element in the user’s psychological well-being, rehabilitation, and overall prosthetic success. This is especially true for finger amputees, where bone anchors restore finger function. In such cases, skin protection against infection is vital, greatly enhancing the long-term durability of bone-anchored prostheses. Proper soft tissue coverage ensures secure attachment and prevents complications. Biomaterials like RTV silicones are essential in mimicking the natural properties of skin, providing both aesthetic realism and mechanical flexibility, addressing the challenge of creating prosthetics that look natural while functioning effectively and durably.^52,53

Neural Interface Technology

The current state-of-the-art in prosthetic hand technology has made remarkable strides in the development of actuators and robotic systems, allowing modern prostheses to replicate a wide range of hand movements. However, despite these advancements, the control systems that detect and translate signals from the body to the prosthetic limb remain a significant challenge. While the mechanical and robotic components of prosthetic hands are evolving rapidly, the ability to accurately harness and interpret neural signals is lagging behind.

To overcome this challenge, it is essential to understand the intricacies of movement patterns and the functioning of both feedforward and feedback control systems. Feedforward control enables rapid, pre-planned movements without immediate sensory feedback, while feedback control relies on real-time sensory input to refine and correct those movements. Research indicates that effective motor control involves a dynamic interplay between these 2 processes, with specific brain regions—such as the motor cortex and cerebellum—reacting to changes in task difficulty and movement precision. As tasks become more challenging, feedback mechanisms become increasingly critical for error correction, while feedforward control dominates in simpler tasks. ⁵⁴

Sufficient signals from muscles, nerves, and the feedback loop between agonist and antagonist muscle groups can be detected by neural interfaces to replicate natural control systems in prosthetic devices. For example, in ballistic motions like tossing a ball, specific muscles contract first against the direction of motion, followed by the activation of agonist muscles to complete the movement toward the target. This sequence highlights the importance of antagonist feedback before initiating abrupt agonist activity. By recording and analyzing these signals using pattern recognition, feedforward commands and feedback-driven signals can be distinguished, enhancing the system’s ability to anticipate and respond to real-time movement requirements.

Neural interface technologies, such as electromyography (EMG), electroneurography (ENG), and antagonist–agonist myoneural interfaces, play a crucial role in simulating the complex feedforward and feedback loops of muscle control. These technologies have the potential to significantly enhance the precision and responsiveness of prosthetic devices.

US patent (US8676334B2) outlines a peripheral nerve interfaces (PNIs) system integrating a nerve conduit with a prosthetic limb, facilitating direct communication between the amputated nerve and the prosthesis. Another patent application (US 2013/253606) proposes a similar system, contingent on nerve regeneration post-injury.

The utilization of PNI offers both bidirectional control and sensory feedback in limb prosthetics.^17,55,56

Bradley Greger and Mark Mahan (US Patent No.15/044,469) described a neural prosthesis device designed for interfacing with peripheral nerves, aiming to overcome limitations in current neural interfaces. The device utilizes mechanically compliant materials, such as silicone elastomers, to ensure compatibility and minimize nerve damage during implantation. It features arrays of micro-wires with conformal sheaths, allowing for precise positioning of microelectrodes on the nerve. The device can be configured for both electrophysiological recording and micro-stimulation, providing spatially distributed microelectrodes for comprehensive nerve signal sampling. The implantation technique involves opening the nerve’s outer layer (epineurium) to visualize fascicles, minimizing damage. The device can be integrated into prosthetic limbs for neural control and sensory feedback.

Studies have shown that longitudinal intrafascicular electrodes electrodes can record peripheral nerve activities in amputees, allowing graded control of a robotic arm but with limited degrees of freedom. To implement PNI systems, neural activity signals from amputees are recorded, processed, classified using wavelet de-noising, and supported by a spike classification stage. A support vector machine classifier relates motor intent to neural firing, enabling the control of a powered prosthesis. ⁵⁷ The embodiment disclosed in this article aims to suggest that a direct signal connection between RPNI and a neural interfaces can be used for mimicking a broad range of movement control patterns.

RPNIs represent an innovative approach involving the reinnervation of peripheral nerve fascicles within freely transplanted skeletal muscle grafts. This pioneering technique not only generates high-amplitude EMG signals but also holds promise in preventing neuromas. Moreover, RPNIs function as biological amplifiers, providing a reliable implantation site with demonstrated robust nerve regeneration, neurotization, and revascularization.

The US patent US9299248B2 presents a muscle interface device designed for interacting with a controllable connected device. It includes a sensor worn on the user’s forearm, capable of recognizing various gestures made by the user. The device employs multiple sensors, such as capacitive EMG sensors and an inertial measurement unit sensor, to detect these gestures, and it may also incorporate other sensors like mechanomyography sensors. The information from these sensors is processed to create control signals, enabling users to interact with content on the connected devices. ⁵⁸

Recent advances in muscle interface technology have improved the recording of EMG data, benefiting amputees by utilizing residual muscles and regenerative peripheral nerve techniques. Implantable myoelectric sensors for EMG recording offer minimally invasive methods for obtaining strong signals to control prosthetic hands effectively. ⁵⁹

In the context of nerve signal detection and stimulation, various interfaces exist, such as depth type, planar type, sieve type, and cuff type. Figure 5 illustrates various neural interfaces and compares them in terms of their invasiveness and sensitivity.

Figure 5.

The diagram provides a visual representation of different neural interface technologies. (A) Standard cuff electrode, (B) flat interface nerve electrode, (C) slowly penetrating interfascicular nerve electrode, (D) longitudinal intrafascicular electrode, (E) transverse intrafascicular multichannel electrode, (F) Utah slanted electrode array, and (G) regenerative multielectrode interface.

The planar type, also known as microelectrode array, focuses on measuring nerve signals through nerve cell cultivation. The planar type is mainly for studying nerve system signals And not suitable for connecting a nerve system and an artificial device.

The depth type directly inserts an electrode into a nerve tissue but, prolonged use of insertion-type electrodes may lead to necrosis or cell accumulations, potentially affecting the stable measurement of nerve signals.

The sieve type, as described in U.S. Patent No. 6908470, is termed a nerve-regenerating electrode. This type utilizes the regenerative ability of nerves by placing a sieve-shaped electrode between cut nerves to facilitate the regeneration of efferent axons, enabling nerve signal measurement. However, its application is limited to situations where nerves are alive at both ends.

The cuff type focuses on directly measuring nerve signals by surrounding nerves. Since the cuff type measures nerve signals from outside the nerves, which are surrounded by an insulator, accurate signal measurement and differentiation of afferent and efferent signals prove challenging. ¹⁷ Yet, by integrating this minimally invasive technique with another method like RPNI, accuracy, sensitivity, and specificity of pattern recognition can be significantly enhanced while minimizing additional damage to the nerve.

Despite challenges, some patents describe systems for receiving electrical signals from a tissue graft connected to a nerve. These signals are strong, with a voltage amplitude of at least 150 µV, but issues like scarring and interference persist. Current systems struggle with low-level signals from nerves, requiring extensive hardware and processing power for signal detection and analysis. ¹⁷

Peripheral nerve electrodes have exhibited biocompatibility and stability during in vivo testing over extended periods. This remarkable progress could lead to the formation of new neuromuscular junctions and a tangible improvement in the functionality of prosthetic limbs for individuals with limb loss. Another innovative idea in the future generations of the prosthetic hands is a novel framework that comprises multiple signal transceivers including ENG signals originating from the nerves, EMG signals from RPNI, and signals from microfluidic device designed to monitor Ion fluctuations within the neuromuscular junction connecting between the regrown nerves and tissue engineered muscles at the cut nerve’s distal end (Figure 3). Furthermore, RPNIs offer the exciting potential to restore proprioceptive sensation, which could substantially enhance the usability of prosthetic devices and foster a stronger sense of embodiment in users.^60,61 The previously mentioned concepts affirm that the amalgamation of bidirectional PNI and RPNI yields substantial benefits. This fusion offers rapid movement, tactile sensory feedback, controllable force generation, and proprioception.

Furthermore, agonist–AMIs are a novel surgical approach that connects agonist and antagonist muscles to maintain proprioceptive input. The CNS may now receive natural force, velocity, and position input thanks to this repair. According to research, prosthesis users can feel more accurate motor control, proprioception, and functional connection across sensory areas when using AMI approaches, which preserve muscle-tendon dynamics. These results imply that in future neuroprosthetic systems, AMI could improve both movement coordination and sensory feedback. ¹⁵

Finally, neural dust, a groundbreaking interface introduced by Bertrand et al ⁶² in 2014, consists of millimeter-sized wireless recording devices made up of numerous tiny dust motes. Initially designed for measuring brain activity, these crystal-based neural dust interfaces can be strategically implanted throughout the body, including in muscles and nerves, to provide detailed, real-time feedback. One of the key advantages of neural dust is its wireless, battery-free design, making it ideal for long-term implantation.^63,64

These neural dust interfaces transmit data wirelessly to an external device, where it can be processed by a microcontroller or interface, offering precise feedback on muscle stretch, joint angles, and overall movement. When multiple neural dust devices are utilized, their triangulated arrangement in 3-dimensional space could theoretically enable users to perceive not only the position of the limb but also its orientation and movement in space. Once implanted, neural dust devices can both receive and transmit signals, modulating the target structures as needed. This capability closely aligns with the concepts discussed in this point of view, and can direct future research in this field.^9,65,66

Innovative System Design for Neural Prosthesis in Prosthetic Hand Control

To closely resemble a normal hand’s functioning, a unique control system for prosthetic hands must include various closed-loop systems. For real-time control and protection, the optimal architecture would have 2 sensory feedback systems and 1 motor control system, or at least 3 linked loops. Together, these loops produce sustain precise motor motions that allow the prosthesis to dynamically adjust to the user’s requirements and surroundings.

The motor control loop in this system generates commands based on input signals from brain interfaces, residual muscles, or nerves to initiate movement. Sensory feedback from proprioceptive inputs and force sensors enables real-time comparison and correction, enhancing the responsiveness of the prosthetic. A secondary sensory loop monitors external factors, such as excessive pressure or overstretching, providing protective input to ensure the safety of both the prosthesis and the user.

The effectiveness of a continuous closed-loop system relies on its ability to analyze information and resolve signals. High-resolution data from sensors and interfaces facilitate the accurate identification of movement patterns and sensory inputs. The system’s computing power determines how quickly and accurately these signals are decoded and translated into motor output. Collectively, these components ensure that the prosthetic hand closely mimics normal hand function, continuously adjusting and responding to sensory inputs with fluidity and precision.

Figure 2 visualizes a device that could be built based on silicon based organic polymer. Inside the device, tissue-engineered skeletal muscle can be vascularized using techniques similar to those used for dialysis fistulas. The muscles inside the device should be coated by some hydrogels and barriers to mimic different muscular channels with different origins and insertions. Pluronic F127 which is a common hydrogel in bioprinting can be used to create channels within the devices. Those channels are innervated by regenerative nerves. The distal cut point of the nerve can be divided into 2 or 3 divisions based on the level of the injury and place each of them into one of the muscular channels. This arrangement stimulates sprouting and nerve regeneration. Moreover, growth factors can be added to enhance and direct the stimulation. ⁸

Outside the device there are 2 or more neural dusts that are wirelessly connected to the antagonist side. When tissue engineered control units are stimulated in 1 side, the adjacent neural dust will be stimulated and it can generate signals. These signals stimulate the antagonist neural dust, allowing proprioception to be restored through the antagonist pathway. The radio frequency transmit-receive system can be implemented using multiplexing techniques, such that those signals make zero interference with each other. Alternatively, those signals can be transferred through microwires. The signals from nerves, whether extra-, intra-, or regenerative neural interfaces, EMG signals, and Agonist–ARMIs need to be collected and sent to a microcontroller for further processing.

However, in distal amputees, only a perineural interface might be adequate, in mid and proximal amputee, a more complex approach to the peripheral nerve, including the combination of both extraneural and regenerative neural interfaces, could offer several advantages more complex approach to the peripheral nerve, including the combination of both extraneural and regenerative neural interfaces, could offer several advantages. For example, signals from perineural interface can be obtained, converted into digital signals, and used as a go-no go command and if there is no signal from perineural interface, EMG signals cannot start and action. on the other, hand the delay between detecting neural interfaces and EMG signals can be recognized by pattern recognition and applying a threshold that if EMG does not reach that, the command works as a discrete or cyclic movement according to the pattern. On the other, hand if signals are strong enough to contract adequate muscles, a continuous movement would be performed and the analog signals from EMG will determine the force. Moreover, an upper limit threshold can be used as a protective feedback loop to stop the movement and prevent extra stretching of the muscle. Finally, feedback from agonist–ARMIs can be used to address ballistic movements by recognizing the patterns, by the processor.

Closed loop system for sensory function, bidirectional cuff electrodes or intraneural electrodes have demonstrated trustable connection with sensory electrodes on the surface of the prosthetic hand. In this embodiment, proprioception can also be understanded by the brain through agonist–antagonist interface and muscle spindles feedback loops. These, together can be used for achieving a real cyborg interface that can provide tactile sensation as well as fast motor signals to start the movements or even reflexes. Moreover, adding some extra interfaces for example a thermoreceptor with a loop connected with motor control can be used to increase the safety and it can stop movement if the temperature has raised and it is danger for the device.

Figure 6 represents the closed loop systems and their feedback on each other to achieve optimum function of the prosthetic hand. This approach could lead to a transformative innovation that not only mimics real hand function but also opens doors toward the next generation of the cyber prosthetic hands. So, this is completely different than everyone has done before. Furthermore, the synergistic integration of RPNIs and PNIs brings notable advantages, enhancing sensitivity in the control of prosthetic hands with minimal invasiveness compared to alternative approaches at this level of precision. Interestingly, RPNIs, with their potential in preventing neuroma formation, alleviating phantom pain, and continuing brain’s regular function contribute to the sustained maintenance of normal nerve function, providing robust signals for force control. The amalgamation of these technologies has the potential to significantly impact patients’ lives, offering enduring benefits and substantial improvements in long-term outcomes.

Figure 6.

(A) Depicts closed-loop systems designed to integrate various neural interfaces, enabling real-time control of a prosthetic limb. This system combines motor control with sensory feedback loops to enhance precision and responsiveness. Interfaces, such as intraneural and extraneural electrodes, connect to regenerative muscle units and sensory pathways, allowing dynamic adjustments based on the user’s neural inputs. Proprioception is achieved through agonist–antagonist pathways, while additional sensors monitor external factors, ensuring the prosthetic limb adapts to both user needs and environmental changes in real time. (B) Electroneurogrphy (ENG), electromyography (EMG), mechanoreceptors, and agonist–antagonist myoneural interfaces (AARMI) signals are first amplified with low-noise amplifiers (LNAs). After that, the signals as sampled at a rate fs using a N-bit analog-to digital converter. This ditital stream is then upconverted for wireless transmission. After being upconverted, the modulated signal goes through a power amplifier and is fed to the transmitter antenna. The reciever antenna picks upthis signal, and after amplification by an LNA, the signal is downconverted and filtered. A digital-to analog converted is used to recover the original analog signals wich are then sent to microcontroller for further processing. The microcontroller follows a finite state machine to detect the type and pattern of the movement based on the acquired ENG, EMG, and AARMI signals.

Conclusions

Prosthetic hand development undergoes a revolutionary shift, led by biomimicry, mimicking natural hand functions. Advanced techniques and neural interfaces offer remarkable control, dexterity, and sensory feedback. This amalgamation of biomimicry, neural interfaces, and historical insight reflects human ingenuity, promising enhanced mobility and integration for amputees.