Composite cuff-regenerative peripheral nerve interface (CC-RPNI)

Abstract

The system and method include a composite cuff tissue graft having a muscle graft portion and a dermal graft portion, surgically attached to surround a healthy, original nerve in a circular fashion. A sensing electrode is attached to the muscle graft portion, and a stimulating electrode is attached to the dermal graft portion. The implant device includes a sensing processing circuit configured to receive electrical signals from the sensing electrode, generate processed motor signal data corresponding to the electrical signals received from the sensing electrode, and transmit the processed motor signal data to an orthotic controller configured to control the orthotic device based on the processed motor signal data. The implant device also includes a stimulating processing circuit configured to receive sensory feedback data from the orthotic controller, generate processed sensory signal data based on the sensory feedback data, and transmit electrical signals corresponding to the processed sensory signal data to the stimulating electrode.

Description

[Technical Field]

[0001] This invention was developed with government support based on W81XWH-21-1-0429, awarded by the U.S. Army Medical Research and Development Command, and NS104584, awarded by the National Institutes of Health. The government has certain rights in this invention.

[0002] This application claims the benefit of U.S. Provisional Application No. 63 / 467,499, filed on 18 May 2023. The entire disclosure of the above application is incorporated herein by reference.

[0003] This disclosure relates to a regenerative peripheral nerve interface (RPNI) device. [Background technology]

[0004] This section provides background information relevant to this disclosure that is not necessarily prior art.

[0005] For example, when controlling a prosthetic limb, it is necessary to receive and record signals from nerves (e.g., human nerves) for later processing and use. Free tissue grafts can be attached to a portion of a nerve, such as a nerve bundle, and electrical signals from the nerve can be amplified by the free tissue graft. Nevertheless, systems and methods are needed to effectively and efficiently process the amplified signals from the nerve and to effectively and efficiently control the prosthetic device based on the processed signals. Furthermore, systems and methods are needed to effectively and efficiently transmit sensory feedback signals from the prosthetic device to the nerve through the free tissue graft. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] U.S. Patent Application Publication No. 2013 / 0304174 [Patent Document 2] U.S. Patent No. 10779963 [Overview of the Initiative] [Means for solving the problem]

[0007] This section provides an overview of the disclosure, but does not constitute a full or comprehensive disclosure of its features.

[0008] The system includes an implantable device having a detection processing circuit configured to receive electrical signals from a composite cuff tissue graft surgically attached to surround a subject's healthy, original nerve in a circular fashion, and a stimulation processing circuit configured to transmit electrical signals to the composite cuff tissue graft, wherein the composite tissue graft comprises a muscle graft portion and a dermal graft portion, and the composite tissue graft is surgically attached to the subject such that the composite tissue graft is completely surrounded by the subject's non-grafted tissue and in direct contact with the subject's non-grafted tissue, and additionally, the healthy, original nerve communicates electrical signals to other tissues of the subject both upstream and downstream from the location of the composite tissue graft, and the muscle graft portion and dermal graft portion are respectively excised from the subject, vascularized, nerve-removed, and then surgically attached to the subject, and the nerve is regenerated from the muscle graft portion and dermal graft portion of the composite tissue graft after the composite tissue graft has been surgically attached to the nerve. At least one sensing electrode is attached to the muscle graft portion of the composite tissue graft and communicates electrically with the muscle graft portion. At least one stimulating electrode is attached to the dermal graft portion of the composite tissue graft and communicates electrically with the dermal graft portion. The sensing processing circuit is configured to receive an electrical signal from at least one sensing electrode, process the electrical signal received from at least one sensing electrode, generate processed motor signal data corresponding to the electrical signal received from at least one sensing electrode, and transmit the processed motor signal data to the orthotic controller, which is configured to control the orthotic device based on the processed motor signal data transmitted from the sensing processing circuit of the implant device. The stimulating processing circuit is configured to receive sensory feedback data from the orthotic controller, generate processed sensory signal data based on the sensory feedback data, and transmit an electrical signal corresponding to the processed sensory signal data to at least one stimulating electrode.

[0009] The method includes the step of receiving electrical signals from a composite cuff tissue graft surgically attached to a subject so as to circumferentially surround the subject's healthy, original nerves using a detection processing circuit of an implantable device. The method further includes the step of transmitting electrical signals to the composite cuff tissue graft using a stimulation processing circuit of an implantable device, wherein the composite tissue graft comprises a muscle graft portion and a dermal graft portion, and the composite tissue graft is surgically attached to the subject so as to be completely surrounded by the subject's non-grafted tissue and in direct contact with the subject's non-grafted tissue, and additionally so as to communicate electrical signals to other tissues of the subject both upstream and downstream from the location of the composite tissue graft, the muscle graft portion and the dermal graft portion are autologous tissue grafts that are excised from the subject, vascularized, nerve-removed, and then surgically attached to the subject, and the nerves are regenerated from the muscle graft portion and dermal graft portion of the composite tissue graft after the composite tissue graft has been surgically attached to the nerves. At least one sensing electrode is attached to the muscle graft portion of the composite tissue graft and communicates electrically with the muscle graft portion. At least one stimulating electrode is attached to the dermal graft portion of the composite tissue graft and communicates electrically with the dermal graft portion. The sensing processing circuit is configured to receive an electrical signal from at least one sensing electrode, process the electrical signal received from at least one sensing electrode, generate processed motor signal data corresponding to the electrical signal received from at least one sensing electrode, and transmit the processed motor signal data to the orthotic controller, which is configured to control the orthotic device based on the processed motor signal data transmitted from the sensing processing circuit of the implant device. The stimulating processing circuit is configured to receive sensory feedback data from the orthotic controller, generate processed sensory signal data based on the sensory feedback data, and transmit an electrical signal corresponding to the processed sensory signal data to at least one stimulating electrode.

[0010] Further areas of applicability will become apparent from the descriptions provided herein. The descriptions and specific examples in the summary of the invention are intended for illustrative purposes only and are not intended to limit the scope of this disclosure.

[0011] The drawings described herein are for illustrative purposes only of selected embodiments and not of all possible implementations, and are not intended to limit the scope of this disclosure. [Brief explanation of the drawing]

[0012] [Figure 1] This is a diagram of a regenerating peripheral nerve interface according to a particular aspect of the present disclosure. [Figure 2] This is a photographic diagram illustrating three free muscle grafts extracted from a subject, and a surgical procedure for surgically attaching the free muscle grafts to the nerve bundles of the subject, according to a particular aspect of this disclosure. [Figure 3] This is a diagram of another regenerative peripheral nerve interface and orthotic device according to a particular aspect of the present disclosure. [Figure 4] This is a diagram of another regenerating peripheral nerve interface according to a particular aspect of the present disclosure. [Figure 5] This is a block diagram of an implantable device according to a particular aspect of the present disclosure. [Figure 6] This is a photographic diagram of a hardware device chip according to a particular aspect of this disclosure. [Figure 7] This is a block diagram of a processing circuit including memory and communication circuits according to a particular aspect of the present disclosure. [Figure 8] This flowchart illustrates an example control algorithm for recording amplified neural signal data according to a particular aspect of this disclosure. [Figure 9] This is a flowchart illustrating an example control algorithm for controlling a prosthetic limb according to a particular aspect of this disclosure. [Figure 10]A flowchart depicting an example control algorithm for decoding signals for prosthetic control according to certain aspects of the present disclosure. [Figure 11] A graph showing nerve signal data corresponding to the onset of a flexion movement. [Figure 12] A flowchart depicting an example control algorithm for monitoring nerves for pathological pain signals according to certain aspects of the present disclosure. [Figure 13] A flowchart depicting an example control algorithm for monitoring pathological bladder contraction signals according to certain aspects of the present disclosure. [Figure 14] A flowchart depicting an example control algorithm for stimulating nerves based on detected pressure signals from a prosthetic device according to certain aspects of the present disclosure. [Figure 15] A graph showing nerve signal data corresponding to the onset of finger flexion movement. [Figure 16] A group of graphs showing nerve signal data from an implanted regenerated peripheral nerve interface. [Figure 17] A group of graphs showing nerve signal data from an implanted regenerated peripheral nerve interface and from a controlled muscle. [Figure 18] A graph showing nerve signal data from an implanted regenerated peripheral nerve interface. [Figure 19] A graph of predicted and actual finger flexion percentages over time. [Figure 20] A diagram of a system for controlling a prosthetic based on amplified nerve signals according to the present disclosure. [Figure 21] A further diagram of a system for controlling a prosthetic based on amplified nerve signals according to the present disclosure. [Figure 22] A diagram of the electrical leads of a system for controlling a prosthetic based on amplified nerve signals according to the present disclosure. [Figure 23]Further diagrams of electrical lead wires for a system for controlling an orthotic device based on amplified nerve signals, as disclosed herein. [Figure 24A] This is a diagram of an implantable device including a header for receiving an 8-contact connector, as disclosed herein. [Figure 24B] This figure shows another embodiment of electrical lead wires for a system for controlling an orthotic device based on amplified nerve signals, according to the present disclosure. [Figure 25] This is a diagram of a protective circuit for an implantable device according to the present disclosure. [Figure 26] This is a functional block diagram for processing amplified nerve signals as disclosed herein. [Figure 27] This is a diagram of the assistive device controller and prosthetic arm as disclosed herein. [Figure 28] This is a diagram of the external programmer charger and electromagnetic induction charging pad according to this disclosure. [Figure 29] This figure shows an external programmer device that communicates with the implant device and prosthetic device controller according to this disclosure. [Figure 30] This figure shows another embodiment of the system for controlling an orthotic device based on amplified nerve signals according to the present disclosure. [Figure 31] This diagram illustrates an algorithm for generating assistive device movement commands and performing stimuli within alternating periods. [Figure 32] This diagram illustrates another algorithm for generating orthotic movement commands and performing stimulation during alternating periods, while simultaneously generating estimations of orthotic movement commands during the period in which stimulation is being performed. [Figure 33] This diagram illustrates an algorithm for generating assistive device movement commands and executing stimuli while performing artifact estimation and subtraction to remove artifacts from detection signals generated by stimuli. [Figure 34] This is a diagram of another regenerating peripheral nerve interface according to a particular aspect of the present disclosure. [Figure 35] This is a figure of a histological image showing different motor units forming a neuromuscular junction within a single RPNI, as disclosed herein. [Figure 36] This disclosure shows different regions of the RPNI on the median nerve corresponding to different movements of individual fingers and thumbs as observed under ultrasound. [Figure 37] This disclosure shows different regions of the RPNI on the ulnar nerve corresponding to different movements of individual fingers and thumbs as observed under ultrasound. [Figure 38A] This graph shows the signals for flexion and adduction of the little finger, as disclosed in this disclosure. [Figure 38B] This is a block diagram of the signal processing of the RPNI signal for generating motor commands according to the present disclosure. [Figure 39] This is a block diagram showing a processing circuit for signal processing of RPNI signals for generating motor commands, as disclosed herein. [Figure 40] This is a diagram of multiple sensing electrodes equipped with RPNI according to the present disclosure. [Figure 41] This diagram shows a processing circuit that receives inputs from multiple input channels of an RPNI according to the present disclosure. [Figure 42] This diagram shows multiple networked modules of an implantable device that receives input from multiple input channels of an RPNI, as disclosed herein. [Figure 43A] This is a photograph of an RPNI equipped with multiple sensing electrodes, as disclosed herein. [Figure 43B] This is a photograph of an RPNI equipped with multiple sensing electrodes, as disclosed herein. [Figure 44] This graph shows the signals detected from multiple sensing electrodes of the RPNI as disclosed in this disclosure. [Figure 45] This is a photograph of an RPNI having multiple sensing electrodes as disclosed herein. [Figure 46] This graph shows the signals detected from multiple sensing electrodes of the RPNI as disclosed in this disclosure. [Figure 47]This is a diagram of a composite regenerative peripheral nerve interface (C-RPNI) according to a particular aspect of this disclosure. [Figure 48] This is a block diagram of an implantable device equipped with a processing circuit and multiple C-RPNIs according to the present disclosure. [Figure 49] This disclosure shows a graph of signals collected by alternately stimulating and recording the same nerve. [Figure 50] This graph shows the patient's stable sensory detection threshold for RPNI over time, as disclosed in this document. [Figure 51] This disclosure shows a map of the hand depicting the somatotopic accuracy of the referenced sensation. [Figure 52] This is a diagram of another composite regenerative peripheral nerve interface (C-RPNI) according to a particular aspect of this disclosure. [Figure 53] This is a block diagram of the processing circuit and multiple C-RPNIs as disclosed herein. [Figure 54] This is another block diagram of the processing circuit and multiple C-RPNIs as disclosed herein. [Figure 55] This is a block diagram of a detection and stimulation implant with a large number of channels, as disclosed herein. [Figure 56] This diagram shows an implantable device and multiple networked modules of multiple C-RPNIs as disclosed herein. [Figure 57] This diagram shows an implantable device and multiple networked modules of multiple C-RPNIs as disclosed herein. [Figure 58] This is a diagram of a photograph of the C-RPNI structure. [Figure 59] The present disclosure is a photographic diagram of an electrodiagnostic setup for generating and recording afferent activity by stimulating a dermal graft and recording it using a nerve cuff. [Figure 60]This graph shows data collected by alternately stimulating and recording the same nerve, as disclosed in this disclosure. [Figure 61] This is a diagram of a composite cuff regenerative peripheral nerve interface (CC-RPNI) according to a particular aspect of the present disclosure. [Figure 62] This is a photographic diagram of the CC-RPNI structure as disclosed herein. [Figure 63] This diagram shows a photograph of the setup for generating and recording signals using CC-RPNI according to this disclosure. [Figure 64] This graph illustrates the afferent and centrifugal signal transfer using the CC-RPNI structure as described in this disclosure. [Figure 65] This diagram shows the setup for an electrophysiological experiment on a C-RPNI structure according to this disclosure. [Figure 66] This graph shows afferent nerve action potentials and electrical stimulation as disclosed herein. [Figure 67] This diagram shows a structure according to the present disclosure, having a single tissue surrounding a nerve that has been physically separated from different tissues attached to its nerve terminal. [Figure 68] This disclosure shows a photographic diagram of a structure having a single tissue surrounding a nerve that has been physically separated from different tissues attached to the nerve terminal. [Figure 69] This graph shows the afferent nerve activity of a structure having a single tissue surrounding a nerve that has been physically separated from different tissues attached to the nerve terminal, as disclosed in this disclosure. [Figure 70] This graph illustrates the afferent and efferent signal transmission of a structure having a single tissue surrounding a nerve that has been physically separated from different tissues attached to the nerve terminal. [Modes for carrying out the invention]

[0013] The corresponding reference number indicates the corresponding portion across several figures in the drawing.

[0014] It should be noted that the figures described herein are intended to illustrate the general characteristics of methods, devices, and materials within the disclosure for the purpose of illustrating specific embodiments. These figures may not accurately reflect the characteristics of any given embodiment and are not necessarily intended to fully define or limit any specific embodiment within the scope of this disclosure.

[0015] Specific examples are described more fully hereby with reference to the accompanying drawings. A non-limiting discussion of terms and phrases intended to aid in understanding this disclosure is provided at the end of the descriptions of embodiments for carrying out this invention.

[0016] In various embodiments, the Disclosure provides methods for amplifying and receiving signals from a portion of a nerve, such as an individual nerve bundle, at levels greater than those produced by any prior methods or techniques. Specifically, as will be described in more detail below, the Disclosure provides methods for amplifying and receiving signals from a portion of a nerve, such as an individual nerve bundle, at levels of approximately 150 μVpp or higher, and in some examples, approximately 250 or 500 μVpp or higher, and up to, for example, approximately 1,000 μVpp or higher. As described above, signals detected by previous neural interface systems were typically less than 100 μVpp when recorded from within the nerve and less than 10 μVpp when recorded from a cuff around the nerve. In certain embodiments, the Disclosure provides implantable neural interface devices, also referred to without distinction as regenerative peripheral nerve interface (RPNI) devices, that facilitate the amplification of signals from individual nerve bundles to about 150 μVpp or higher, and in some examples to about 250 or 500 μVpp or higher, and to a maximum, for example, about 1,000 μVpp or higher.

[0017] Referring to Figure 1, a neural interface system 4 within a subject or patient is shown. The subject may be an animal with a complex nervous system, such as a mammal, such as a human, primate, or companion animal. A portion of a nerve 6, such as a nerve terminal of the subject, may be injured or severed, for example, such as a nerve terminal that is completely or partially damaged due to injury, disease, or surgery. In certain embodiments, the method may include surgically dividing, cutting, slicing, and / or transversely incising the portion of the nerve 6 into one or more individual branches or fiber bundles 8. Note that in certain variant embodiments, the method may include isolating the portion of the nerve 6 of interest to produce one or more individual branches or fiber bundles 8. Each of the one or more individual branches or fiber bundles 8 is placed within a free tissue graft 10. In certain embodiments, the free tissue graft 10 may be an autologous graft of muscle or dermal tissue previously excised from the subject. In certain preferred embodiments, the free tissue graft 10 is muscle tissue. The free tissue graft 10 is excised or removed to have a standard predetermined volume or size depending on the size of the branch or fiber bundle 8. When the free tissue graft 10 is excised, the tissue graft is vascularized, and the original blood vessel is no longer functional. The predetermined volume of the free tissue graft 10 may be selected to be small enough to allow for adequate vascular regeneration by parallel blood flow, while providing an area or volume large enough for the terminal branch or fiber bundle 8 of the nerve 6 to grow, as will be described in more detail below.

[0018] For example, over a period of several months, the nerve bundle 8 can regenerate the free tissue graft 10 and sprout nerve fibers 12 in search of new nerve targets. Once the free tissue graft 10 is regenerated, action potentials from neurons traveling down the nerve then generate muscle-level signal amplitudes rather than nerve-level amplitudes. Thus, the free tissue graft 10 (e.g., a free muscle graft) acts as an amplifier of signals generated by terminal branches or fiber bundles 8 of nerve 6, and signals from a single nerve bundle 8 have voltage amplitudes of approximately 150 μVpp or more, and in some examples, approximately 250 or 500 μVpp or more, and at most, for example, approximately 1,000 μVpp or more.

[0019] The nerve interface system 4 can be used with any damaged, severed, or injured portion of a nerve within a subject (e.g., a nerve terminal), but is particularly well suited for use with peripheral nerves. The nerve interface system 4 may therefore be used for damaged or injured peripheral nerves, such as those resulting from amputation surgery. Nevertheless, the methods described herein may also be used with a variety of different nerves. Thus, in certain embodiments, the methods of this disclosure are particularly beneficial in the case of peripheral nerves, but the discussion of peripheral nerves and peripheral nerve interface devices is illustrative and non-limiting.

[0020] As shown in Figure 1, the free tissue graft 10 may be configured with a conductor, such as an electrode 14, that communicates with the free tissue graft 10. The electrode 14 is then in electrical communication with a wire 18a that communicates with an implant device 20, which includes a processing circuit 22 and an amplifier 24, as will be described in more detail below. In this case, signals from the nerve bundle 8 are received by the electrode 14 and communicated via the wire 18a to the processing circuit 22 of the implant device 20, for example, through the amplifier 24. Alternatively, the electrode 14 may be omitted, and the conductor communicating with the free tissue graft 10 may be a wire 18b placed directly in or on the free tissue graft 10. In this case, signals from the nerve bundle 8 are received by the wire 18b itself through direct or indirect electrical communication with the free tissue graft 10 and communicated via the wire 18b to the processing circuit 22 of the implant device 20, for example, through the amplifier 24. As a further alternative, the conductor communicating with the free tissue graft 10 can be a wire grid 17 having a number of electrode sites and a number of conduction channels placed in or on the free tissue graft 10. The wire grid 17 is then communicating with a number of wires 18c that communicate with the implant device 20. In such a case, signals from the nerve bundle 8 are received by the wire grid 17 and communicated via the wires 18c to the processing circuit 22 of the implant device 20, for example, through an amplifier 24. The amplifier 24 is a high-impedance amplifier. Furthermore, although a single amplifier 24 is shown in Figure 1, additional amplifiers, including additional / separate amplifier circuits for one or more individual nerve bundles 8 or groups of fiber bundles 8, may be further included in the implant device 20. Additionally, multiple implant devices 20, or implant devices with additional processing circuits 22, may be further used.As referred to herein, the neural interface system 4 may also be an implantable neural interface device or RPNI device, and generally includes an implantable device 20 comprising a free tissue graft 10, associated wires 18a, 18b, 18c, an electrode 14 or a wire grid 17 having multiple electrodes if any, and a processing circuit 22.

[0021] For example, the implantable device 20 may be an implantable medical device within a subject that is similar to an automated cardiac defibrillator but has the processing capability to receive, process, record, and / or communicate nerve signals received from a free tissue graft 10, as described in this disclosure. Since signals from individual nerve bundles 8 are amplified by the free tissue graft 10 (e.g., a free muscle graft) to a level of, for example, about 150 μVpp or higher, the electronics contained within the implantable device 20 are smaller, less expensive, do not require processing power, and / or consume battery power than the electronics that would need to adequately and meaningfully receive, record, and process nerve signals detected by previous systems, which are typically less than 100 μVpp when received from within a nerve and less than 10 μVpp when received from a cuff around a nerve, as discussed above. Furthermore, since the signals from individual nerve bundles 8 are amplified by the free tissue graft 10 to a level of, for example, about 150 μVpp or higher, the signals are less affected by noise and interference, have a higher signal-to-noise ratio, more accurately represent the actual nerve signals produced by the individual nerve bundles 8, and match the actual nerve signals. For example, the signals can have a signal-to-noise ratio of 4 or higher. In particular, electrical signals of such a level may be produced in certain embodiments by an implantable nerve interface system 4 which essentially consists of the free tissue graft 10 and one or more conductors (e.g., wires 18a, 18b, 18c, electrodes 14, and / or wire grids 17), along with one or more portions of the regenerated and regenerated nerves 6 in the free tissue graft.

[0022] In certain embodiments, a method for amplifying nerve signals within a subject includes placing a portion of nerve 6 (e.g., nerve bundle 8) within a free tissue graft 10 and fixing the portion of nerve 6 (nerve bundle 8) within the free tissue graft 10. For example, the free tissue graft 10 can be attached to the nerve bundle 8 by suture, glue, tension, or other suitable attachment method or mechanism. At least a conductor (e.g., an electrode 14, wires 18a, 18b, 18c, and / or a wire grid 17) may then be introduced into the free tissue graft 10. It should be noted that at least one conductor may be introduced into the free tissue graft before fixing the portion or branch of nerve to the free tissue graft. At least one conductor provides electrical communication with nerve 6. The conductor may have a maximum thickness of about 5 mm or less. One or more portions of nerve 6 (nerve bundle 8) are therefore regenerated within the free tissue graft, which regenerates the tissue. Such nerve regeneration may include the growth of sprouting nerve fibers 12. Thus, nerve 6 has the ability to produce amplified electrical signals of approximately 150 microvolts or more, even without any external electrical input, as previously discussed above. In particular, the ability to amplify and generate electrical signals from nerves reflects the generation of voluntary, spontaneous electrical signals from subjects at previously unseen voltage levels. Such voluntary, spontaneous electrical signals (e.g., spontaneously generated from motor nerves) can be distinguished from stimulated nerve signals (e.g., stimulation by combined compound action potentials (CMAPs) resulting from external nerve activation) generated by introducing an external electrical input into the nerve for activation.

[0023] In other specific embodiments, the method may include cutting a portion of a nerve in a subject, such as cutting the nerve terminal, to create one or more branches or fiber bundles. In specific embodiments, cutting may include cutting the nerve terminal into multiple parts, such as branches / fiber bundles. Thus, the placement of the nerve within the free tissue graft and the introduction of the conductor into the aggregate of free tissue grafts may be repeated for each portion of the nerve. The method may further include excising the free tissue graft from the subject's tissue prior to the placement of the severed terminals. In specific embodiments, the tissue is muscle tissue. In alternative embodiments, the tissue may be dermal tissue. In specific embodiments, as will be discussed in more detail below, the maximum dimensions of the free tissue graft are about 10 cm or less. In other embodiments, the maximum dimensions of the free tissue graft are about 5 cm or less.

[0024] In certain embodiments, methods according to certain embodiments of the present disclosure may include further stimulating one or more parts of a nerve (e.g., a branch / fiber bundle) with a stimulus signal transmitted through one or more conductors that are electrically communicating with a free tissue graft. This provides the ability to transmit sensory feedback to the subject's brain via the neural interface system 4.

[0025] Referring to Figure 2, three free tissue grafts 10 of free muscle tissue are shown after being extracted from a subject but before being surgically attached to the subject's nerve bundle. Figure 2 further includes photograph 28 showing the surgical procedure for surgically attaching the free tissue grafts 10 to the subject's nerve bundle. Further explanations for surgically attaching free tissue grafts 10, such as muscle grafts, also called autologous grafts, which are free-grafted pieces of autologous muscle tissue from a subject, to a nerve bundle are provided in U.S. Patent Application Publication No. 2013 / 0304174, published on November 14, 2013, by the same applicant. The entire disclosure of U.S. Patent Application Publication No. 2013 / 0304174 is incorporated herein by reference.

[0026] For example, a free tissue graft 10, such as a muscle graft, can be surgically removed from a non-essential donor muscle within the subject. Therefore, after removal, the free tissue graft 10 undergoes a complete denervation process, and thus, any previously existing nerve development within the free tissue graft 10 is terminated. As discussed above, this removal process also causes vascular resection of the original cells of the free tissue graft 10. Once the free tissue graft 10 is surgically attached to a nerve bundle 8, the free tissue graft 10 undergoes a nerve regeneration process. Thus, the attached nerve bundle 8 causes nerve regeneration in the free tissue graft 10, germinating nerve fibers 12, which then grow within the free tissue graft 10 in search of new nerve targets. Because the denervation process has been previously performed, signals from the newly attached nerve bundle 8 and newly germinated nerve fibers 12 do not need to compete with the remaining nerve signals from the nerve bundle and nerve fibers that previously neuratinated the free tissue graft 10.

[0027] Furthermore, rather than simply dying and being reabsorbed by the subject's body, if surgically reattached to the subject, the free tissue graft 10 can acquire nutrients through the process of aspiration. Thus, even without the blood supply from the original blood vessels, if the free tissue graft 10 is within an optimal volume / size range, the free tissue graft 10 can absorb nutrients and blood through the surrounding tissues and fluids to support the process of nerve regeneration. Finally, as the free tissue graft 10 is reintegrated with the subject's body, a new blood supply network may be established. This process of nerve removal of the free tissue graft 10 prior to nerve regeneration of the free tissue graft 10 by the attached nerve bundle through newly sprouted nerve fibers 12, linked to the processes of aspiration and vascular regeneration, results in areas of muscle or other tissue where highly specific electrical signals from individual nerve bundles 8 at approximately 150 μVpp or higher can be received, for example, by an implant device 20.

[0028] As described above, to facilitate the processes of nerve regeneration and aspiration, the free tissue graft 10 is preferably within an optimal volume / size range. For example, the volume / size of the free tissue graft 10 may be selected to be small enough to allow rapid vascular regeneration by parallel blood flow, while providing an area or volume large enough for nerve growth without forming a non-histological neuroma. The maximum dimensions of the free tissue graft 10 may be about 10 cm or less in certain preferred embodiments. For example, in certain modified forms, the free tissue graft 10 can have a maximum dimension in any direction of about 10 cm or less. For example, in certain modified forms, the length of the free tissue graft 10 may be about 10 cm or less, or more preferably about 5 cm or less. Furthermore, the width of the free tissue graft 10 may be about 10 cm or less, or more preferably about 5 cm or less. The thickness of the free tissue graft 10 may optionally be about 2 to 3 cm or less. Furthermore, the optimal dimensions of the free tissue graft 10 may include a length of about 5 cm or less and a diameter of about 2 cm or more and about 3 cm or less. For example, the optimal and preferred dimensions of the free tissue graft 10 could include a length of approximately 3.5 cm and a diameter of approximately 2 cm. It should be noted that the free tissue graft 10 can have a variety of entirely different dimensions and / or geometric shapes, and those described herein are illustrative. In addition, a discussion of the dimensions of the autologous muscle tissue piece to be free-grafted, from the subject, is included, for example, in paragraphs

[0082] to

[0088] of U.S. Patent Application Publication No. 2013 / 0304174, published on November 14, 2013, which is incorporated herein by reference in whole.

[0029] Referring to Figure 3, the example embodiment 100 shows multiple free tissue grafts 10 (e.g., free muscle grafts) and electrodes 14 connected to an implant device 20. In addition to receiving signals from individual nerve bundles through the free tissue grafts 10 and electrodes 14, the implant device also controls a terminal device, in this case a prosthetic hand 110. Specifically, as shown in Figure 3, the radial nerve 102, median nerve 104, and ulnar nerve 106 are each separated into multiple individual nerve bundles attached to the corresponding free tissue grafts 10 and communicate with the processing device 22 of the implant device 20 via electrical communication with the electrodes 14.

[0030] As will be discussed in more detail below, the processing circuit 22 of the implant device 20 monitors signals from various fiber bundles and, based on the analysis of the received signals, controls, for example, the flexion and extension of the prosthetic hand 110. For example, as will be discussed in more detail below, training data can be acquired through a calibration process, and thus the subject is asked to perform specific movements while nerve signals are monitored and recorded by the processing circuit 22 of the implant device 20 and communicated to an external computing device such as a desktop computer or laptop. The training dataset is then analyzed and used to estimate the parameters used by the processing circuit 22 to move the prosthetic hand 110, and the parameters are then downloaded from the external computing device to the processing circuit 22 of the implant device 20. For example, as shown in Figure 3, the implant device 20 is communicating with actuators 112 that flex and extend the individual fingers of the prosthetic hand 110, as will be described in more detail below.

[0031] Referring to Figure 4, in addition to detecting or reading signals generated by nerve bundles 8 amplified through a free tissue graft 10, the RPNI device of this disclosure can also be used to stimulate individual nerve bundles 8 or individual nerve fibers. For example, as shown in Figure 4, the free tissue graft 10 comprises three electrodes 30a, 30b, and 30c that communicate with the processing circuit 22 of the implant device 20 through amplifiers 32, 32b, and 32c and wires 34a, 34b, and 34c. Thus, as will be discussed in more detail below, the processing circuit 22 can stimulate individual nerve bundles 8 using, for example, negative voltage stimulation signals or positive voltage suppression stimulation signals. Generally, negative voltage signals cause nerves to fire, while positive voltages suppress nerve firing.

[0032] Existing clinical applications, such as vagus nerve stimulation, typically use a cuff around the entire nerve. Therefore, a large portion of the nerve is usually stimulated. Using an RPNI device as shown in Figure 4, however, the processing circuit 22 can still address stimulation of specific fiber bundles by directing signals to electrodes 30a, 30b, and 30c through wires 34a, 34b, and 34c. More specifically, using a free tissue graft 10, it is possible to allow a single fiber bundle 8, which may have a diameter of approximately 1 mm, to expand into a 1 cm × 3.5 cm structure for stimulation. Stimulating the entire structure, i.e., the entire free tissue graft 10, can specifically address a single corresponding fiber bundle 8.

[0033] Furthermore, current manipulation can be used to enable stimulation of even more specific areas, such as subsections of the fiber bundle 8 or individual fibers, through the use of multiple electrical contacts, such as multiple electrodes. For example, when only a single negative contact is used for stimulation, the negative voltage can diffuse and dissipate. By using current manipulation, on the one hand, the negative voltage can be surrounded by a positive voltage so that it is concentrated on a single location. Referring further to Figure 4, a negative voltage may be applied to electrode 30b, while on the other hand, a positive voltage may be applied to electrodes 30a and 30c to more greatly concentrate the location for the application of the negative voltage from electrode 30b.

[0034] As will be discussed in more detail below, for example, nerve stimulation can be used so that a sensory orthosis stimulates nerves in response to pressure detected by a pressure sensor on the prosthesis. Additionally, nerve stimulation can be used to suppress pathological pain signals. Additionally, nerve stimulation can be used to suppress pathological contractions of the bladder, for example. Additionally, nerve stimulation can be used for sphincter control, erectile dysfunction, and / or to control nerves associated with internal organs such as the liver, adrenal glands, stomach, pancreas, kidneys, and similar organs. For example, such nerve stimulation may be used on the renal artery to interrupt and treat abnormal nerve signals of the kidneys that could otherwise cause hypertension.

[0035] Referring to Figure 5, further details about the implant device 20 are shown. As discussed above, the implant device 20 is capable of receiving amplified nerve signal inputs 52 from the free tissue graft 10 via wires 18a, 18b, and 18c (shown in Figure 1). As further discussed above, the implant device 20 is capable of communicating nerve stimulation signal outputs 54 to stimulate nerves at electrodes 30a, 30b, and 30c via wires 34a, 34b, and 34c (shown in Figure 4). As further discussed above, the implant device 20 is capable of communicating orthotic control signal outputs 56 to control the movement of the orthotic device. For example, the implant device 20 is capable of communicating orthotic control signal outputs 56 via a data bus such as a controller area network (CAN) bus to control the flexion and extension of the prosthetic hand 110 (shown in Figure 3). For example, the implant device 20 can communicate an orthotic control signal output 56 to control individual actuators 112 (shown in Figure 3) of the prosthetic arm. Furthermore, as discussed above, the implant device 20 can receive an orthotic sensor input signal 58 generated by one or more pressure sensors, corresponding to, for example, the pressure detected by the pressure sensors of the prosthesis. As discussed above, the implant device 20 can generate a nerve stimulation signal output 54 based on the pressure sensor signal input received by the implant device from the pressure sensors of the prosthesis.

[0036] As shown in Figure 5, the implant device includes a processing circuit 22, as well as a communication circuit 50 and memory 62 with which the processing circuit 22 communicates. The communication circuit 50 enables the implant device, and specifically the processing circuit 22 of the implant device 20, to communicate wirelessly with a computing device outside the implant device 20, such as a desktop or laptop computer, for example, a computing device located outside the subject. Thus, the processing circuit 22 can communicate data such as amplified nerve signal input data received via the amplified nerve signal input 52. Such communication can be used during the calibration process to receive training and calibration data from the implant device 20 on an external computing device for review and analysis, and to communicate estimated operating parameters and configuration data used by the implant device 20 to move a prosthesis, for example, or to generate nerve stimulation signal outputs. The communication circuit 50 may include an antenna, as well as a receiver and transmitter or transceiver, for communication via wireless radio frequency (RF) 60. For example, the communication circuit 50 can communicate via a wireless protocol such as the CEN ISO / IEEE11073 communication protocol for communication between a medical device and an external information system. Alternatively, the communication circuit 50 can communicate via other wireless protocols such as WiFi(R) or Bluetooth(R).

[0037] Memory 62 can be used by the processing circuit 22 to store amplified neural signal input data received via amplified neural signal input 52 before communication to an external computing device via the communication circuit 50. Memory 62 can also be used to store estimated operating parameters and configuration data received from an external computing device and used by the implant device during operation. Memory 62 can also be used by the processing circuit 22 to store event or operation history data, or any other data associated with various inputs and outputs received or generated by the processing circuit 22.

[0038] Referring to Figure 6, a hardware device chip 64 is shown which may include, or be used to implement, the processing circuit 22, communication circuit 50, and memory 62 of the implant device (shown in Figure 5). The hardware device chip 64 may also include, or be used to implement, amplifiers 24 (shown in Figure 1) and amplifiers 32a, 32b, and 32c (shown in Figure 4). The hardware device chip 64 may preferably include a high input impedance bioamplifier configured to process high output impedance biological signals received from the free tissue graft 10. Furthermore, the hardware device chip 64 may preferably include functionality for rejecting large common-mode signals, such as specifically recording from a pair of electrodes, relating the received signal to a local reference near the RPNI device, or using a strong high-pass filter in the first stage. The hardware device chip 64 may preferably include an amplifier for amplifying the signal received from the nerve bundle 8 by, for example, 1,000 times. Preferably, the hardware device chip 64 has very low noise, although this feature may not be very important considering the relatively high amplitude of the amplified nerve signal input received from nerve bundle 8. Additionally, the hardware device chip 64 may preferably include a bandpass filter for filtering the received amplified nerve signal input between 10,000 and 2,000 Hz before further processing. Additionally, the hardware device chip 64 may preferably take the absolute value in the analog domain of the received amplified nerve signal input.

[0039] Referring to Figure 7, further details are shown of the processing circuit 22, which is shown to be communicating with the communication circuit 50 and memory 62 of the implant device 20. The processing circuit 22 includes a nerve input signal adjustment circuit 70 for receiving and adjusting the amplified nerve signal input 52 received from the nerve bundle through the free tissue graft 10. The function and operation of the nerve input signal adjustment circuit 70 will be discussed in more detail below. The processing circuit 22 also includes a nerve input signal decoding circuit 72 for processing and decoding the signal data after it has been adjusted by the nerve input signal adjustment circuit 70. The function and operation of the nerve input signal decoding circuit 72 will be discussed in more detail below. The processing circuit 22 also includes a nerve stimulation signal output circuit 74 for generating a nerve stimulation signal output 54. The function and operation of the nerve stimulation signal output circuit 74 will be discussed in more detail below. The processing circuit 22 also includes an orthotic control circuit 76 for generating an orthotic control signal output 56 for controlling the prosthesis. The function and operation of the orthotic control circuit 76 will be discussed in more detail below. The orthotic control signal output 56 for controlling the prosthesis may be communicated to the prosthesis via a data bus such as a CAN bus. The processing circuit 22 also includes an orthotic sensor receiver circuit 78 for receiving an orthotic sensor signal input 58 from the pressure sensor of the prosthesis. The function and operation of the orthotic sensor receiver circuit 78 will be discussed in more detail below.

[0040] Referring to Figure 8, a control algorithm 800 for receiving and recording amplified nerve signal data from a free tissue graft 10 is shown. The control algorithm 800 may be implemented by the processing circuit 22 of the implant device 20. More specifically, the control algorithm 800 may be implemented at least in part by the nerve input signal adjustment circuit 70 (shown in Figure 7) of the processing circuit 22. The control algorithm 800 begins at 802.

[0041] In 804, the nerve input signal adjustment circuit 70 of the processing circuit 22 receives an amplified nerve signal from a conductor that communicates with the free muscle graft. As described in detail above with reference to Figure 1, the conductor may be an electrode 14, a wire grid 17, or a wire 18b that communicates with the free tissue graft 10. As further described above, the electrical signal can have a voltage amplitude of about 150 μVpp or more, and in some examples about 250 or 500 μVpp or more, and a maximum of, for example, about 1,000 μVpp or more. Furthermore, in embodiments in which an amplifier 24 (shown in Figure 1) is used, the signal may be amplified before being received by the processing circuit 22. In such cases, for example, the signal may be pre-adjusted by being amplified 1,000 times by the amplifier 24 before being received by the nerve input signal adjustment circuit 70 of the processing circuit 22.

[0042] In 806, the neural input signal conditioning circuit 70 of the processing circuit 22 adjusts and extracts features from the received signal from the free tissue graft 10. For example, in an embodiment that does not include the amplifier 24 (shown in Figure 1), the neural input signal conditioning circuit 70 can amplify the received signal, for example, by 1,000 times. The neural input signal conditioning circuit 70 can then filter the received signal using a predetermined analog frequency range. For example, the predetermined analog frequency range may be between 10 Hz and 1,000 Hz, or between 10 Hz and 2,000 Hz. After filtering, the neural input signal conditioning circuit 70 can digitize the filtered signal using a predetermined sampling rate. For example, the neural input signal conditioning circuit 70 can digitize the filtered signal using a sampling rate of 30,000 samples per second. The digitized signal may then be digitally filtered using a predetermined digital frequency range. For example, the predetermined digital frequency range may be from 100 to 500 Hz. After digital filtering, the signal may then be downsampled to 1,000 samples per second. Due to the relatively large voltage amplitude of the initial received signal, the adjustments and feature extraction performed in step 806 result in a robust and clearly defined signal that is less affected by noise or interference compared to a system that does not utilize free tissue grafts 10 for amplification of the signal from the nerve bundle.

[0043] In the 808, the processing circuit 22 records the resulting signal data in the memory 62 and / or communicates the resulting signal data to an external computing device using the communication circuit 50. For example, the resulting signal data can be stored in the memory 62 of the implant device 20 and then communicated to an external computing device via the communication circuit 50 through a batch communication process. Alternatively, the memory 62 can act as a buffer to receive and store the resulting signal data for further processing by the processing circuit 22 or for communication to an external computing device via the communication circuit 50. Alternatively, the resulting signal data can be streamed in real time to an external computing device via the communication circuit 50.

[0044] After recording or communicating the resulting signal data in 808, the processing circuit 22 loops back to 804 and continues to receive the amplified neural signal. Although the control algorithm 800 is shown as a series of steps for illustrative purposes, it is understood that each step can be continuously generated in parallel by the processing circuit 22 at the same time that the amplified neural signal is continuously received in real time.

[0045] Referring to Figure 9, a control algorithm 900 for controlling a prosthetic limb, such as an artificial hand 110 (shown in Figure 3), based on signals received from a free tissue graft 10 is shown. The control algorithm 900 may be implemented by a processing circuit 22 of the implant device 20. More specifically, the control algorithm 900 may be implemented at least in part by a nerve input signal adjustment circuit 70 (shown in Figure 7), a nerve input signal decoding circuit 72, and an orthotic control circuit 76 of the processing circuit 22. The control algorithm 900 begins at 902.

[0046] In step 904, the nerve input signal adjustment circuit 70 of the processing circuit 22 receives amplified nerve signals from a conductor that is electrically communicating with the free muscle graft. The function of step 904 has been described above with respect to step 804 in Figure 8 and will not be repeated here.

[0047] In step 906, the neural input signal adjustment circuit 70 of the processing circuit 22 adjusts and extracts features from the received signal from the free tissue graft 10. The function of step 906 has been described above with respect to step 806 in Figure 8 and will not be repeated here.

[0048] In 908, the neural input signal decoding circuit 72 decodes the resulting signal data to determine whether the resulting signal data corresponds to, for example, flexion or extension of the prosthesis. While the control algorithm 900 in Figure 9 is described in terms of decoding the resulting signal data to determine whether it is a flexion or extension action, it is understood that other prosthesis control movements can be similarly decoded from the resulting signal data as needed. The neural input signal decoding circuit 72 can use a 1 / 2 classifier such as a Naive Bayes classifier or regression analysis to determine whether the resulting signal data over a predetermined period, such as 25 milliseconds (ms), indicates flexion or extension. Further details on decoding the resulting signal data are described below with respect to the control algorithm 1000 for decoding signals for prosthesis control, shown in Figure 10.

[0049] In step 910, the processing circuit 22 determines whether the resulting signal data for a portion of a predetermined period corresponds to flexion or extension of the prosthesis. If the resulting signal data in step 910 corresponds to extension, the processing circuit proceeds to 912, where the prosthetic control circuit 76 of the processing circuit 22 moves the prosthesis in the extension direction. If the resulting signal data in step 910 corresponds to extension, the processing circuit proceeds to 914, where the prosthetic control circuit 76 of the processing circuit 22 moves the prosthesis in the flexion direction. For example, in the case of a prosthetic hand 110 as shown in Figure 3, the processing circuit 22 can move the actuator 112 of the prosthetic hand 110 in the flexion or extension direction as needed. After moving the prosthesis in step 912 or 914, the processing circuit 22 loops back to 904.

[0050] Referring to Figure 10, a control algorithm 1000 for decoding signals for the control of a prosthetic limb is shown. The control algorithm 1000 may be implemented by the processing circuit 22 of the implant device 20. More specifically, the control algorithm 1000 may be implemented at least partially by the neural input signal decoding circuit 72. The function of the control algorithm 1000 shown in Figure 10 is encompassed in step 908 of Figure 9. The control algorithm 1000 begins at 1002.

[0051] In step 1004, the neural input signal decoding circuit 72 determines whether the current sample group for a portion of a predetermined period is complete. For example, the portion of the predetermined period may be 25 ms, and the sample interval may be 1 ms. In such a case, the neural input signal decoding circuit 72 can wait in step 1004 until a complete sample group of 25 samples at 1 ms intervals is completed. If the sample group is not yet complete, the neural input signal decoding circuit 72 loops back to 1004. If the sample group is complete, the neural input signal decoding circuit 72 proceeds to 1006.

[0052] In 1006, the neural input signal decoding circuit 72 classifies each sample within the sample group using a 1 / 2 classifier. For example, when the neural input signal decoding circuit 72 is decoding the resulting signal data to determine whether it is bent or stretched, the neural input signal decoding circuit 72 can classify each sample within the sample group as either a bent sample or a stretched sample. For example, the neural input signal decoding circuit 72 can classify each sample within the sample group as either a bent sample or a stretched sample using a 1 / 2 Naive Bayes classifier or regression analysis.

[0053] The 1 / 2Naive Bayes classifier can utilize training data previously collected from subjects during calibration procedures and routines. For example, a subject may be instructed to perform a flexion or extension action, and the resulting neural signal data can be recorded by the processing circuit 22 and communicated to an external computing device for analysis. Based on the collected training data, a Gaussian distribution can be estimated or calculated for each of the flexion and extension movements based on the received neural signal data. For example, the Gaussian distributions for flexion and extension movements will therefore have different medians and variances. Parameters and data for the 1 / 2Naive Bayes classifier can be estimated by an external computing device based on the collected training data, then communicated to the processing circuit 22 for use by the neural input signal decoding circuit 72 when decoding the neural signal data, and stored in memory 62.

[0054] During step 1006, the neural input signal decoding circuit 72 compares each sample in the sample group with a previously determined Gaussian distribution having different medians and variances for flexion and stretching movements, and can calculate the probability that a particular sample was derived from each of the two distributions. Each sample is then classified based on which of the two movements has a higher probability for that particular sample. For example, if a particular sample has a higher probability of corresponding to a flexion movement, the sample is classified as a flexion sample. If a particular sample has a higher probability of corresponding to a stretching movement, the sample is classified as a stretching sample. Once all samples in the sample group have been classified, the neural input signal decoding circuit 72 proceeds to step 1008.

[0055] At 1008, the neural input signal decoding circuit 72 determines whether a particular sample group has more flexion samples or more stretching samples, and classifies the entire sample group based on this determination. For example, if a sample group has more flexion samples, the sample group is classified as a flexion sample group, and if a sample group has more stretching samples, the sample group is classified as a stretching sample group. In this way, the neural input signal decoding circuit 72 predicts, for example, whether a group of samples within a particular sample group indicates a flexion movement or a stretching movement. It is understood that other movements may be similarly included in the classification and prediction process. After classifying the sample groups, the neural input signal decoding circuit 72 loops back to 1004.

[0056] Thus, referring to both Figures 9 and 10, the processing circuit 22 can determine or predict whether nerve signal data during a specific predetermined period, such as 25 ms, corresponds to flexion or extension, or whether it instructs flexion or extension. Furthermore, the prosthetic control circuit 76 can send control commands to the prosthesis every 25 ms based on the classification of the current or most recent sample group. In this way, the processing circuit 22 can continuously monitor and decode nerve signal data and send corresponding commands to operate the prosthesis. Although the above example is illustrated using a portion of a predetermined period, shorter or longer portions of a predetermined period can be used.

[0057] Referring to Figure 11, an example graph of neural signal data over a 5-second period corresponding to the initiation of flexion movement is shown. Similarly, referring to Figure 15, an example graph of neural signal data over a 400-millisecond period corresponding to the initiation of finger flexion movement is shown.

[0058] As described above with respect to Figures 9 and 10, training data collected from subjects during the calibration routine can be used to generate mappings of, for example, received nerve signals to corresponding prosthetic movements or actions. In the case of individuals who have undergone radial resection surgery, individual nerve bundles can be successfully mapped to individual hand muscles simulated by the prosthesis. In the case of individuals who have undergone humeral resection surgery, training data can be used to determine which nerve bundles, or sets of nerve signal inputs, best correlate to which individual hand muscles simulated by the prosthesis.

[0059] Additionally, neural signal data such as average signal power, the number of zero-crossing events, or the count of detected spikes can be monitored, recorded, and analyzed for each signal from each nerve bundle and used to calculate desired velocities, for example, all five fingers of a prosthetic hand, to be sent to the prosthetic hand with a single command at each time step, such as every 25ms. There is no one-to-one correspondence between specific muscles and the velocities of individual fingers. For example, to bend only the little finger, the subject may need to extend their index finger simultaneously. Therefore, finger velocities can be traced back to muscle activity across all neural signal channels to determine a consistent overall map. Various algorithms are available to estimate instantaneous velocity from various signals, including, for example, linear filters, Kalman filters, and particle filters.

[0060] Additionally, individual states, such as grasping and pointing, can be predicted using a linear discriminant, a Naive Bayes classifier, or a support vector machine. In each case, the subject is asked to perform various movements or actions, and a training dataset is acquired through a calibration process by monitoring and recording the resulting neural signal data using the implant device 20 and the processing circuit 22. The training dataset can then be used to estimate operating parameters used by the processing circuit 22 and, for example, the prosthetic control circuit 76, for controlling the prosthetic hand. The estimated operating parameters can then be downloaded to the processing circuit 22 via the communication circuit 50 and stored in memory 62 for use by the processing circuit 22, for example, to perform real-time estimation of finger velocity to move the prosthetic hand.

[0061] Referring to Figure 12, a control algorithm 1200 for monitoring nerves for pathological pain signals is shown. The control algorithm 1200 may be implemented by the processing circuit 22 of the implant device. More specifically, the control algorithm 1200 may be implemented at least in part by the nerve input signal adjustment circuit 70, the nerve input signal decoding circuit 72, and the nerve stimulation signal output circuit 74. The control algorithm 1200 begins at 1202.

[0062] In step 1204, the nerve input signal adjustment circuit 70 of the processing circuit 22 receives amplified nerve signals from a conductor that is electrically communicating with the free muscle graft. The function of step 1204 has been described above with respect to step 804 in Figure 8 and will not be repeated here.

[0063] In step 1206, the neural input signal adjustment circuit 70 of the processing circuit 22 adjusts and extracts features from the received signal from the free tissue graft 10. The function of step 1206 has been described above in relation to step 806 in Figure 8 and will not be repeated here.

[0064] In step 1208, the neural input signal decoding circuit 72 decodes the resulting signal data to determine whether the resulting signal data indicates, for example, a pathological pain signal. The decoding performed in 1208 is similar to the decoding described above with respect to step 908 in Figure 9 and steps 1004 to 1008 in Figure 10, which described decoding the resulting signal data to determine whether the resulting signal data indicates a flexion action or a stretching action. As with the decoding described above with respect to Figures 9 and 10, the decoding performed in 1208 can similarly use a 1 / 2 classifier such as a Naive Bayes classifier, based on the training dataset collected during the calibration procedure in the subject, to determine whether the resulting signal data corresponds to a state in which a pathological pain signal is being generated. After decoding the resulting signal in 1208, the processing circuit 22 proceeds to 1210.

[0065] At 1210, the processing circuit 22 determines whether a pathological pain signal has been detected based on the decoding of the resulting signal data. If a pathological pain signal is detected, the processing circuit 22 proceeds to 1212 and stimulates the appropriate nerve bundle with an inhibitory stimulus. Specifically, the nerve stimulation signal output circuit 74 of the processing circuit 22 stimulates the appropriate nerve bundle with a positive voltage to suppress nerve activity, thereby suppressing or reducing pathological pain signal activity in the nerve bundle. In this way, pain signals in the subject can be reduced without permanent loss of sensation in the specific nerve or nerve bundle in question. After stimulating the nerve with an inhibitory stimulus at 1212, or after determining at 1210 that no pathological pain signal was detected, the processing circuit 22 loops back to 1204.

[0066] Referring to Figure 13, a control algorithm 1300 for monitoring pathological bladder contraction signals is shown. The control algorithm 1300 may be implemented by the processing circuit 22 of the implant device. More specifically, the control algorithm 1300 may be implemented at least in part by the nerve input signal adjustment circuit 70, the nerve input signal decoding circuit 72, and the nerve stimulation signal output circuit 74. The control algorithm 1300 begins at 1302.

[0067] In step 1304, the nerve input signal adjustment circuit 70 of the processing circuit 22 receives amplified nerve signals from a conductor that is electrically communicating with the free muscle graft. The function of step 1304 has been described above with respect to step 804 in Figure 8 and will not be repeated here.

[0068] In step 1306, the neural input signal adjustment circuit 70 of the processing circuit 22 adjusts and extracts features from the received signal from the free tissue graft 10. The function of step 1306 has been described above in relation to step 806 in Figure 8 and will not be repeated here.

[0069] In step 1308, the neural input signal decoding circuit 72 decodes the resulting signal data to determine whether the resulting signal data indicates, for example, a pathological bladder contraction signal. The decoding performed in 1308 is similar to the decoding described above with respect to step 908 in Figure 9 and steps 1004 to 1008 in Figure 10, which described decoding the resulting signal data to determine whether the resulting signal data indicates a flexion action or an extension action. As with the decoding described above with respect to Figures 9 and 10, the decoding performed in 1308 can similarly use a 1 / 2 classifier, such as a Naive Bayes classifier, based on the training dataset collected during the calibration procedure in the subject, to determine whether the resulting signal data corresponds to a state in which a pathological bladder contraction signal is being generated. After decoding the resulting signal in 1308, the processing circuit 22 proceeds to 1310.

[0070] At 1310, the processing circuit 22 determines whether a pathological bladder contraction signal has been detected based on the decoding of the resulting signal data. If a pathological bladder contraction signal is detected, the processing circuit 22 proceeds to 1312 and stimulates the appropriate nerve bundle with an inhibitory stimulus. Specifically, the nerve stimulation signal output circuit 74 of the processing circuit 22 can stimulate the appropriate nerve bundle with a positive voltage to suppress nerve activity, thereby suppressing or reducing the pathological bladder contraction signal activity in the nerve bundle. After stimulating the nerve using an inhibitory stimulus at 1312, or after determining at 1310 that no pathological pain signal was detected, the processing circuit 22 loops back to 1304.

[0071] Although described in the context of monitoring and suppressing pathological bladder contraction signals, the control algorithm 1300 described with respect to Figure 13 can be similarly adapted to other applications. For example, the control algorithm 1300 can be appropriately adapted for sphincter control or erectile dysfunction. Similarly, the control algorithm 1300 can be adapted to control nerves associated with internal organs such as the liver, adrenal glands, stomach, pancreas, and kidneys. In each case, training data is collected and analyzed from the subject to generate appropriate monitoring parameters, which are then downloaded to the implant device 20. The implant device 20 then monitors nerve signal activity to determine whether nerve stimulation is appropriate.

[0072] Referring to Figure 14, a control algorithm 1400 for stimulating nerves based on pressure signals detected from an assistive device is shown. For example, an assistive device such as the prosthetic hand 110 shown in Figure 3 may be equipped with a pressure sensor, for example, placed on the fingertip of the assistive device. The pressure sensor can detect pressure and communicate the pressure signal back to the processing circuit 22 of the implant device 20. The control algorithm 1400 may be implemented by the processing circuit 22 of the implant device. More specifically, the control algorithm 1400 may be implemented at least in part by the assistive sensor receiver circuit 78 and the nerve stimulation signal output circuit 74. The control algorithm 1400 begins at 1402.

[0073] In 1404, the orthotic sensor receiver circuit 78 receives a pressure signal from the orthotic pressure sensor corresponding to the pressure detected at the pressure sensor location. In 1406, the nerve stimulation signal output circuit 74 stimulates individual nerve bundles based on the received detected pressure signal. Calibration procedures in the subject can be used to generate training data to determine which individual nerve bundles should be most appropriately mapped to which pressure sensors. Furthermore, the firing rate or level of stimulation can be matched to the level of pressure detected by the pressure sensor. In this way, the implantable device can communicate tactile feedback signals from the orthotic to the appropriate nerve bundles.

[0074] The following specific examples are provided to illustrate how the components, devices, and methods of this technology may be made and used, and are not intended to indicate whether or not a given embodiment of this technology has been made or not, or whether or not it has been tested, unless otherwise expressly stated.

[0075] "Example 1" RPNI research in non-human primates In the following example, RPNIs were surgically implanted in the forearms of two non-human primates, monkeys R and L. Specifically, three RPNIs were implanted in monkey R and four in monkey L. The muscle grafts were attached to small branches of the median and radial nerves, providing independent finger flexion / extension and thumb flexion signals. The surgery followed a standard surgical procedure checklist, and the animals were monitored daily in cages for 10 days postoperatively, and then observed in primate chairs during daily experiments thereafter.

[0076] No major problems were identified, and the animals regained normal use of their limbs within one week post-surgery. In the second surgery in both animals, evident vascular regeneration was observed in the muscle grafts. In response to electrical stimulation, RPNI generated large-amplitude compound muscle action potentials (CMAPs) that directed nerve regeneration of the muscle grafts by the implanted nerve bundles.

[0077] In the third surgery on monkey L, four bipolar "IM-MES" intramuscular electrodes manufactured by Ardiem Medical were implanted in two matured RPNIs and in a healthy, unmodified muscle (ECRB, wrist stretcher) for comparison. One electrode was placed in a muscle graft of a newly created RPNI structure, which was then matured for three months to produce a high-amplitude signal. The presence of the electrode during the maturation phase did not negatively affect RPNI regeneration, nerve regeneration, or maturation. The electrode leads were routed subcutaneously through tunnels from the monkey's forearm to its back and emerged through the skin on the back for connection to recording equipment. Daily recordings from these implanted electrodes were taken during task behavior. The percutaneous sites where the leads emerged were lightly washed weekly with betadine solution, and no infection was noted. The sites appeared clean with minimal inflammation and did not cause any apparent discomfort to the animals.

[0078] Referring to Figure 16, Graph 1600 shows the voluntary RPNI signal in μV recorded by a semi-chronic Ardiem IM-MES electrode in monkey L, along with the calculated percentage of flexion corresponding to the RPNI signal. Graph 1602 shows the voluntary RPNI signal in μV recorded by a transdermal thin wire electrode in monkey L, along with the calculated percentage of flexion corresponding to the RPNI signal. Graph 1604 shows the voluntary RPNI signal in μV recorded by a transdermal thin wire electrode in monkey R, along with the calculated percentage of flexion corresponding to the RPNI signal.

[0079] Referring to Graph 1600, the signals recorded by the IM-MES electrodes vary between the animal and the RPNI graft with amplitudes ranging from 50 to 500 μVpp. Graphs 1600, 1602, and 1604 show representative signals that appear similar to sparse electromyographic (EMG) signals, typically displaying multiple distinct single motor units. The far right portion of Graph 1600 shows an enlarged portion of the voltage signal indicating individual muscle contractions. All observed inferential single units reliably correspond to flexion events, with a length of approximately 4 ms and a variable firing frequency. The high signal-to-noise ratio (SNR) of the RPNI signal enabled automated detection of voluntary RPNI activation with 95+ accuracy using a linear discriminant classifier. The RPNI signal is used to control the prosthetic arm in real time while the monkey L performs behavioral tasks.

[0080] "Example 2" Human RPNI research In the following example, three RPNIs were surgically implanted in a human for neuroma control. The patient underwent distal radial transection just below the wrist. Muscle grafts approximately 1 x 3 cm were taken from the surrounding tissue and sutured separately to the distal ends of the median, ulnar, and radial nerves. At this level, the radial nerve (and therefore the RPNI graft) contains only the sensory fibers that originally stimulated the dorsal skin of the hand. The median and ulnar nerves, as well as the RPNIs, contain a mixture of sensory and motor fibers that originally stimulated the intrinsic muscles of the hand. Electromyographic (EMG) activity was recorded from the median and ulnar RPNIs using thin percutaneous wire electrodes while the patient performed several hand movements. As expected, the RPNIs produced EMG in response to movements that engaged the muscles originally stimulated by the transected nerves.

[0081] Referring to Figure 17, graphs 1700 and 1702 show signals recorded from the median RPNI, ulnar RPNI, and the muscles of a healthy wrist (flexor carpi ulnaris - FCU) during two different hand movements. Specifically, graph 1700 shows the signal recorded during a thumb-little finger opposition action, corresponding to touching the tip of the thumb to the tip of the little finger. Graph 1702 shows the signal recorded during a thumb opposition action, corresponding to touching the tip of the thumb to the base of the little finger without bending the little finger. These graphs illustrate that physiologically correct signals were obtained from the RPNI. In other words, the nerve is activated during correct movement.

[0082] As expected, the median RPNI signal (shown in the top row of graphs 1700 and 1702) shows similar activity amplitudes during both thumb-little finger opposition and thumb-only opposition. This is because the median nerve originally stimulated the thumb muscles but not the little finger muscles. The ulnar RPNI signal (shown in the bottom row of graphs 1700 and 1702) is activated more during thumb-little finger opposition than during thumb-only opposition because the ulnar nerve originally stimulated the little finger muscles more than the thumb-only muscles. Finally, healthy FCU muscle activity is clearly present but is not correlated with either movement as it is entirely dedicated to wrist flexion.

[0083] In summary, since it would be impossible to achieve the same pattern of activity through the healthy, untouched muscles surrounding the RPNI, graphs 1700 and 1702 show that the RPNI is being stimulated by the expected nerve.

[0084] Referring to Figure 18, Graph 1800 shows the signal recorded from the ulnar RPNI during a key-pinch motion, corresponding to making a fist with the thumb pressed against the outer surface of the index finger, the shape made when inserting a key into a car's ignition switch. Graph 1800 illustrates the high signal amplitude achievable through the RPNI technique. For example, the signal-to-noise ratio (SNR) of the data in Graph 1800 is 8.65.

[0085] "Example 3" Research on continuous positional control in non-human primates In the following example, the nerve signals of a monkey were detected and monitored while the monkey flexed and extended its fingers. The nerve signals were processed using the techniques described above for this instruction, including the use of a Kalman filter, and the percentage of flexion was predicted based on the nerve signals. Furthermore, the actual percentage of flexion of the monkey's fingers was monitored and compared to the predicted percentage of flexion.

[0086] Referring to Figure 19, Graph 1900 shows the predicted percentage of finger flexion 1902 graphed over time, along with the actually observed percentage of finger flexion. In Graph 1900, the predicted percentage of flexion 1902 correlates with the actual percentage of flexion 1904 by a correlation coefficient of 0.87. Graph 1900 illustrates the accuracy of the techniques described above in relation to this instruction, including the use of the Kalman filter, when predicting the actual percentage of flexion based on monitored nerve signals. Thus, the techniques of this instruction can be used for continuous positional control of orthotic devices. In other words, this instruction can be used to process nerve signals and control orthotic devices through various flexion positions, as opposed to individual orthotic device positions or states such as flexion or extension.

[0087] Referring to Figures 20 to 30, additional embodiments of the present disclosure are shown, including systems and methods for controlling an orthotic device based on amplified nerve signals. Referring to Figure 20, a system 200 for controlling an orthotic device such as a prosthetic hand 110 is shown. Although system 200 is described using the example of a prosthetic hand 110, other suitable orthotic devices can be used. Similar to the systems and methods described above, system 200 includes a neural interface system 4 with an implant device 20 that receives nerve signals from a nerve 6, such as a peripheral nerve, amplified by a free tissue graft 10, as described above. The signals are communicated from the free tissue graft 10 to the implant device 20 via electrical leads 18, such as wires 18a, 18b, and 18c, as described above with reference to Figure 1. The implant device 20, wires 18a, 18b, and 18c, the free tissue graft 10, the prosthetic hand 110, and the neural interface system 4 for amplifying nerve signals from the nerve 6 are described above, for example with reference to Figures 1 to 7.

[0088] System 200 is similar to the systems described above with reference to Figures 1-7, except that the implant device 20 of System 200 wirelessly communicates data relating to amplified nerve signals, such as processed EMG data, to the prosthetic controller 220, as will be described in more detail below. The prosthetic controller 220 controls the actuator 112 (shown in Figure 3) of the prosthetic hand 110 based on signals received from the implant device 20, in the same manner in which the implant device 20 controls the prosthetic hand 110 as described above with reference to Figures 1, 3, 5-7, and 9.

[0089] Thus, this disclosure provides a system and method for controlling a prosthetic hand 110 using amplified nerve signals from a free tissue graft 10. The system and method of this disclosure provides intuitive functional control of a multi-jointed prosthetic hand for upper limb amputation patients who have undergone a regenerative peripheral nerve interface (RPNI) surgical procedure. As discussed above, the RPNI surgical procedure places individual small muscle grafts or free tissue grafts 10 at the ends of nerve bundles from the remaining limb. The nerve bundles regenerate their respective muscle grafts, forming a healthy, stable, and long-lasting neuromuscular connection. Intramuscular bipolar electrodes, further described below, can be used to record directly from the RPNI. While this disclosure describes an example of utilizing RPNI, the system and method of this disclosure can also be used in patients who have undergone targeted muscle reinnervation (TMR) surgical procedure. The TMR procedure involves transferring remaining nerves from an amputated limb to regenerate new target muscles from the subject that would otherwise have lost their function. The regenerated target muscle then acts as a biological amplifier of motor signals from the nerve that has undergone transection. Electrodes can therefore be attached to or implanted in the target muscle and connected to the implant device 20 of this disclosure, which can receive nerve signals generated by the regenerated nerve in the target muscle and process the signals according to this disclosure to control an orthotic device. Furthermore, the implant device 20 can communicate signals to the regenerated nerve in the target muscle to provide sensory feedback stimuli, as described below. Thus, the system and method of this disclosure, including the implant device 20, can be used in patients who have undergone RPNI and patients who have undergone TMR procedures.

[0090] As described above, the device is a fully implantable recording system capable of wirelessly transmitting electromyography (EMG) signals to an external device, such as an orthotic controller 220, for upper limb orthotic control. For example, the system may include (1) implantable electrodes and electrical leads 18, (2) an implantable sensing unit, such as an implantable device 20, (3) a wireless transmitter, such as one included in the communication circuit 50 of the implantable device 20, (4) a wireless receiver, such as one included in the communication circuit 222 of the orthotic controller 220, (5) an external smart link controller, such as the orthotic controller 220, and (6) a charging unit, such as an external programmer charger 280 and an electromagnetic induction charging pad 282, which will be described in more detail below with reference to Figures 20 to 30. In a further example, the implantable device 20 may be implemented by a modified PMA-approved spinal cord stimulator (Nuvectra Algovita, PMA P130028). However, instead of stimulation, the unit can be configured to record EMG signals from both the remaining muscle and RPNI using electrode leads. The sensing unit or implantable device 20 can wirelessly transmit signals to an external smart controller, such as an orthotic controller 220, which then decodes or interprets the EMG signals and sends commands to a terminal device, such as a myoelectric prosthesis, such as an artificial hand 110. Additionally or alternatively, the implantable device can be configured as a stimulating device with a receiver for receiving signals from the transmitter of the orthotic controller 220 and for outputting signals to implantable electrodes / electrical leads 18, which are then amplified by a neural interface system 4 to stimulate nerve 6.Thus, in addition to detecting nerve signals and communicating signals from the implant device to the orthotic controller, the system and method of this disclosure can also be used to wirelessly communicate signals from the orthotic controller 220, which will subsequently be used for stimulating nerve 6, to the implant device 20.

[0091] Referring further to Figure 21, the implant device 20 includes the amplifier 24, processing circuit 22, and communication circuit 50 described above. The implant device 20 includes a battery 230 connected to the amplifier 24, processing circuit 22, and communication circuit 50. Although a single amplifier 24 is shown in Figure 21, additional amplifiers may be used, as will be discussed below. The implant device 20 receives amplified nerve signals 240 from the free tissue graft 10 via electrical leads 18. The implant device 20 processes the amplified nerve signals, as will be described in more detail below, and wirelessly communicates data relating to the amplified nerve signals, such as processed EMG data, to the orthotic controller 220 via the communication circuit 50. The communication circuit 50 includes, for example, a wireless transmitter configured for wireless communication. The orthotic controller 220 includes, for example, a communication circuit 222, which includes a wireless receiver configured for wireless communication. Wireless communication between the communication circuit 50 of the implant device 20 and the communication circuit 222 of the prosthetic device controller 220 can be performed using the Medical Implant Communication Service (MICS) protocol and / or the Medical Device Wireless Communication Service (MedRadio) protocol. Additionally or alternatively, any other suitable wireless communication protocol may be used, including, for example, Bluetooth, Bluetooth Low Energy (BLE), WiFi, Near Field Communication (NFC), or other suitable wireless communication protocols.

[0092] The processing circuit 224 of the prosthetic controller 220 is configured to process signals received from the implant device 20 by the communication circuit 222 and to communicate prosthetic control signal outputs 56 for controlling the actuator 112 (shown in Figure 3) of the prosthetic hand 110 via a data communication bus such as the CAN bus 226. For example, the processing circuit 224 of the prosthetic controller 220 can implement the control algorithm 900 described above with reference to Figure 9 for controlling the prosthesis based on signals received from the free tissue graft 10. Additional or alternative, as discussed above, the prosthetic controller 220 is further configured to process stimulus feedback signal inputs, such as prosthetic sensor input signals 58 generated by one or more pressure sensors 225, corresponding to pressure detected by the pressure sensors 225 of the prosthetic hand 110. For example, as the prosthetic hand 110 grasps an object, the pressure sensors 225 detect the pressure generated from grasping the object. The assistive device sensor input signal 58 is received via a data communication bus such as a CAN bus 226 and can be processed by the processing circuit 224 of the assistive device controller 220. The processing circuit 224 can then communicate the stimulus / sensory feedback signal to the implant device 20 via wireless communication transmitted from the communication circuit 222 to the communication circuit 50.

[0093] Referring to Figure 22, four electrical leads 18 are shown, each containing two wires connected to a positive electrode 252 and a negative electrode 254, respectively. Figure 23 shows photographs of the electrical leads 18, the positive surface of the positive electrode 252, and the negative surface of the negative electrode 254. In the example of Figure 22, eight wires of the electrical leads 18 are connected to an 8-contact connector 250. In particular, the eight contact wires of the 8-contact connector 250 are connected to the exposed wires of the electrical leads 18, which have platinum crimps, for example, that are then sealed in a silicone tube, as shown in Figure 22. In the example of Figure 22, platinum crimps are used together with a silicone tube to connect the contact wires of the 8-contact connector 250 to the exposed wires of the electrical leads 18, but any other suitable method can be used to connect the contact wires of the 8-contact connector 250 to the exposed wires of the electrical leads 18. Alternatively, the exposed wires of the electrical leads 18 can be connected directly to the contacts of the 8-contact connector 250. Furthermore, while the example in Figure 22 utilizes four 2-wire electrical leads 18 and an 8-contact connector 250, any other suitable number of wires, electrical leads, and contact connectors can be used. For example, the contact connector may include contacts to accommodate 2, 4, 6, 10, 12, or any other suitable number of electrical wires, and the electrical leads may include 2, 4, 6, 10, 12, or any other suitable number of electrical wires.

[0094] To record EMG signals from the remaining muscle and RPNI, the system 200 utilizes an electrode and lead wire assembly to connect muscle tissue, such as a free tissue graft 10, to the implant device 20. The assembly can include a modified version of an existing bipolar electrode and a modified version of an existing cable lead wire to connect to the implant device 20. Each implant device 20 can be connected to, for example, 12 bipolar electrodes, and each lead wire contains 8 contacts with 3 leads per implant device 20. The bipolar electrodes can be modified from PermaLoc(TM) electrodes, which were approved by the FDA as part of the long-term implant system NeuRX Diaphragm Pacing System. Compared to the unipolar PermaLoc(TM) electrode, the bipolar electrode has an additional deinsulated lead wire surface, as shown in Figure 23, and the plastic-plated anchor at the distal tip of the electrode has been removed. The basic design, all material parts, and sterilization protocol remain the same. The electrodes can be percutaneously implanted, for example, in patients who have undergone upper limb amputation, to record EMG signals from remaining muscle and RPNI.

[0095] Several modifications and additions were made to ensure that the bipolar electrodes were fully implanted and connected to the implant device 20. Industry-standard Bal Seal connector leads manufactured by Cirtec Biomedical, Inc. were used for connection to the implant device 20. A single Bal Seal connector lead has eight contacts with the lead body and eight conductors in parallel. Other numbers of contacts, such as twelve conductors, are possible. The lead diameter is 0.053 mm, and the estimated impedance is 4 ohms per 30.48 cm (1 foot). The conductor material used is MP35N containing 28% silver, and the outer insulator is 55D Pellethane with ETFE conductor insulator. A platinum-iridium connector ring is used for the electrical connection to the Bal Seal. The bare cable is exposed 10 to 15 cm from the proximal end of the Bal Seal connector lead, and the bare cable is exposed at the proximal end of the PermaLoc bipolar lead. The distal and proximal ends of the two lead wires are permanently joined to form a single lead wire with four bipolar electrodes. The interlocking joint will connect each individual bare wire using platinum crimps and will be covered with a silicone tube for protection. The protected joint will be positioned as close as possible to the implant device 20 to avoid crossing the thicker portion of the lead wire with the anatomical joint in the patient's body. The total length of the assembled lead wire from proximal to distal is approximately 100 to 105 cm. Thus, the distal and proximal ends of the two previously approved electrode lead wires are permanently interlocked without altering the electrical or material properties of the lead wires.

[0096] Referring to Figure 24A, the implant device 20 includes a header that receives one or more contact connectors connected to electrical leads. For example, as shown in Figure 24A, the implant device includes electrical ports 256, each configured to receive an 8-contact connector 250. Each port 256 may include a Bal Seal connector, such as a Bal Seal connector manufactured by Cirtec Biomedical, Inc., which is configured to seal the implant device 20 when the 8-contact connector 250 is inserted into the electrical port. The implant device also includes a header having parallel conductors 254 with contacts, positioned alongside and spaced apart, corresponding to the contacts of the 8-contact connector 250. In this example, the Baal Seal conductors 254 are manufactured by Bal Seal Inc. and are components of the header manufactured by Cirtec. In the example in Figure 24A, the implant device is configured with a header having three parallel conductors 254 to receive three 8-contact connectors 250. Thus, the implant device 20 shown in Figure 24A can accommodate signals from up to 24 wires, i.e., signals from three sets of eight electrical wires based on signals from 24 electrodes, including 12 positive electrodes 252 and 12 negative electrodes 254. The 12 positive electrodes 252 and 12 negative electrodes 254 allow the implant device 20 to receive 12 bipolar channels of EMG activity based on amplified nerve signals 240 from the free tissue graft 10. While the example in Figure 24A utilizes 24 wires and electrodes, the implant device 20 can be configured with headers and parallel conductors 254 to accommodate any number of wires and electrodes. For example, the implant device 20 can include a header comprising two sets of parallel conductors 254, each containing 12 contacts for receiving a 12-contact connector.Alternatively, the implant device 20 may include four sets of parallel conductors 254 to receive four 8-contact connectors 250, so that signals from four sets of electrical wires can be received based on signals from 32 electrodes, including 16 positive electrodes 252 and 16 negative electrodes 254. Thus, the implant device 20 may include a header having any appropriate number of parallel conductors 254, each having any number of contacts to receive contact connectors with any number of wires.

[0097] The implantable device 20 can be implemented using a modified implantable recording device based on the Algovita platform developed by Cirtec Biomedical, Inc. The Algovita platform (PMA P130028) is an FDA-approved implantable pulse generator for pain reduction via spinal cord stimulation. The electrode leads are connected to the implantable device 20, which has a header with three rows, each containing eight spring platinum-iridium Bal seals. The leads are secured after implantation with titanium set screws, which are inserted through a partition and tightened with a provided torque wrench. The Algovita is internally powered by a 4.1V lithium-ion battery with a nominal capacity of 215mAh. This has been tested to maintain its capacity after 1000 discharge cycles. It also has a low self-discharge rate, and therefore the shelf life is not an issue for this application, as the battery will exceed the shelf life of the sterile packaging. The implant device 20 communicates wirelessly with external components via a medical implant communication system (MICS) comprising a communication circuit 50, which may include, for example, a Microsemi ZL70323MNJ, a newer version of the MICS chip used in the original Algovita. MICS communication uses the 2.45 GHz band for wake-up and the 402 to 405 MHz band for data transfer. Aside from the updated MICS chip, the battery, power management, and communication circuitry remain unchanged.

[0098] The external physical design of the implantable device 20 remains the same as that of the Algovita device. Changes to the internal components to convert Algovita into a sensing device are limited. The Saturn stimulator ASIC was replaced with an Intan RHD2216 amplifier to sense biopotentials rather than stimulate nerve tissue. The Intan amplifier has low power consumption and has been previously used in surgically invasive clinical studies. No adverse events were noted in these acute trials. As an example, a linear voltage regulator (TPS7A2033, Texas Instruments) can utilize a 4.1V battery voltage as input, converting it to a 3.3V signal to power the RHD2216. The MICS module (ZL70321MNJ, Microsemi) on Algovita was replaced with a newer version, the ZL70323MNJ, from Microsemi.

[0099] Referring to Figure 24B, another embodiment of an alternative arrangement is shown, which includes six electrical leads 18' for six bipolar electrodes connected to a 12-contact connector 250'. The 12-contact connector 250' can be received by a header of the implant device 20, similar to the configuration described above with respect to Figures 22–24A. As stated above, the port 256 of the implant device 20 can include a Bal Seal connector, such as the Bal Seal connector manufactured by Cirtec Biomedical, Inc., which is configured to seal the implant device 20 when the 12-contact connector 250' is inserted into the electrical port of the implant device 20. The contacts of the 12-contact connector 250' are spaced sufficiently apart to match the spacing of the corresponding connectors on the header of the implant device 20. Additionally, in the embodiment of Figure 24, a triangular coupling device 251 is used to receive the six electrical leads 18' and to connect the six electrical leads 18' to the 12-contact connector 250'. As shown in Figure 24B, the coupling device 251 includes a parallel connection for six electrical leads 18' in a broad portion of the triangular coupling device 251, connecting twelve wires for six bipolar electrodes into a single cable, with each of the twelve wires connected to an individual contact of the 12-contact connector 250'.

[0100] Referring to Figure 25, the implant device 20 includes a protective circuit to protect the electrical components of the implant device from surge voltages during an electrostatic discharge (ESD) event. For example, the implant device may experience an ESD event while the implant device 20 is being handled during a surgical procedure to implant the implant device 20 in the patient's body. As shown in Figure 25, to provide protection against surge voltages that may intensify in the event of an ESD event, for example, a 220-ohm resistor 262 is placed before the input to the amplifier 24. In practice, the 220-ohm resistor 262 could be an AC0603JR-07220RL resistor available from Yageo. In further practice, as shown in Figure 25, the amplifier 24 could be an Intan RHD2216 amplifier for sensing biopotentials. In yet another example, an ESD diode 260 could be placed opposite the resistor 262, facing the input to the amplifier 24. The diode 260 can be, for example, an SMS24T1G ESD diode available from On Semiconductor. While an ESD protection circuit is illustrated in Figure 25 as one input to the implant device 20, the same protection circuit, including a 220-ohm resistor 262 and an ESD diode 260, can be included for each input to the implant device 20. Thus, in the case of an implant device 20 configured with three sets of 8-contact connectors 250, i.e., three sets of contactors 254 for a total of 24 inputs, as shown in Figures 21 to 24, the implant device 20 can include 24 sets of the protection circuits shown in Figure 25. In other words, the implant device 20 can include one 220-ohm resistor 262 and one ESD diode 260 for each of the two inputs of the implant device 20.Specific examples of resistor 262 and diode 260 are shown in Figure 25, but any other suitable resistor and diode can be used to protect the implant device 20 from ESD events.

[0101] Referring to Figure 26, a functional block diagram illustrating the system and method for processing amplified neural signals according to this disclosure is illustrated. As shown in Figure 26, the raw amplified neural signal 240 is received by the implant device 20. For example, as shown in Figures 21 to 24 and as discussed above, the implant device 20 receives 12 bipolar channels of EMG activity from the implanted electrodes and electrical leads 18. In 270, the amplified neural signal is filtered using a bandpass filter from 100 to 500 Hz, as shown to make accurate measurements of EMG activity. In 272, the filtered signal is then sampled by the processing circuit 22 of the implant device 20 at a Nyquist rate of approximately 1 kHz, resulting in approximately 1 kilosample per second (1 KSps) of EMG data. In 274, the processing circuit 22 of the implant device 20 then calculates the mean-absolute-value (MAV) for each activity of the bipolar channels. In 276, the processing circuit 22 uses the communication circuit 50 of the implant device 20 to wirelessly transmit the bipolar channel activity-dependent MAV to the communication circuit 222 of the orthotic controller 220 using the MICS protocol.

[0102] To wirelessly send and receive data to and from a non-implanted device, such as an orthotic controller 220 or another external communication device, the implantable device 20 can use the communication circuit 50 of an existing Microsemi MICS communication device from the Algovita platform, as discussed above. Following implantation, the implantable device 20 can be turned on / off by shaking a magnet on the device. When the implantable device 20 is operating, it can interact with two external devices: a programmer charger, as discussed below, and an orthotic controller 220, also known as a Smart Link Controller (SLC). The implantable device 20 is configured to wirelessly stream EMG to the orthotic controller 220, as discussed above. The implantable device 20 includes a bootloader for wireless firmware updates. The microcontroller of the implantable device 20 is the Texas Instruments MSP430F2618 (MSP430), which is also used in the Algovita platform. The microcontroller is programmed to record 12 bipolar channels of EMG activity from the implanted electrodes using an Intan RHD2216 amplifier.

[0103] Thus, instead of sending raw EMG data to the orthotic controller 220, the implant device 20 first filters, samples, and processes the raw EMG signals from the bipolar channels to calculate the MAV of the sampled EMG signals for each bipolar channel, and then wirelessly transmits only the processed MAV data for each bipolar channel to the orthotic controller 220. This configuration offers a technical advantage in maximizing the battery life of the implant device 20's battery 230 and the power source of the orthotic device itself, because the implant device 20 would require far more power to directly stream raw EMG signals with a larger bandwidth from the implant device 20 to the orthotic controller 220. In this way, the system and method of the present disclosure maximize the battery life of the device by first filtering, sampling, and processing the raw EMG data in the implant device 20 and then wirelessly transmitting only the processed EMG signals for each bipolar channel to the orthotic controller 220.

[0104] Referring to Figure 27, examples of the prosthetic controller 220 are shown both installed on the outside and inside the prosthetic hand 110. To illustrate the relative size and scale of the prosthetic controller 220, a 25-cent coin is also shown in Figure 27. As shown in Figure 27, the prosthetic controller 220 can be implemented using a printed circuit board assembly (PCBA).

[0105] The orthotic controller 220 receives processed EMG packets and decodes them into motor commands for the orthotic device. The orthotic controller 220 includes a communication circuit 222, which may include, for example, a ZL70123 chip for wireless and secure communication with the implant device 20 using a standard MICS protocol. All components of the orthotic controller 220 are soldered onto a PCBA, and the device is enclosed in a water-resistant housing within the orthotic device. The materials used in the orthotic controller 220 are typical of modern PCBAs. The circuit board is a flat laminated composite material with internal copper circuits. The components mounted on the circuit board are application-specific integrated circuits (ASICS) used as intended by the manufacturer. ASICS are attached to the board with lead-free solder using industry-standard surface mount technology.

[0106] The prosthetic controller uses state-of-the-art algorithms, including machine learning algorithms, to decode EMG signals into motor commands, as described above. The choice of algorithm depends on the mode of motion of the prosthetic hand. A classifier may be used to decode grasping for an intuitive gripping selection, with further advancements for corresponding control. Alternatively, a regressor may be used to control individual fingers simultaneously and independently.

[0107] The orthotic controller can be connected to a laptop, such as an external programmer device 290, as described below, to record processed EMGs from the implant device 20. The implant device 20 can also be wirelessly connected to and communicate with the laptop, such as the external programmer device 290, as discussed below. The recorded EMGs can be used to measure the signal intensity from each electrode pair and calculate the decode parameters. Configuration mode can be used to load new parameters into the orthotic controller 220. Test mode is available to verify the orthotic controller functionality of the implant device 20. In test mode, the external programmer device 290 can transfer pre-recorded EMG packets to the orthotic controller, which will respond with decoded movement commands. Otherwise, the device operation is fixed, including receiving signals, decoding intended movements, and sending outputs to the orthotic.

[0108] Power for the orthotic device controller 220 is supplied by the orthotic device's battery. Upon startup, the orthotic device controller 220 automatically connects to the prosthetic hand 110 and waits to receive a signal from the implant device 20. When the orthotic device controller 220 does not receive a signal from the implant device 20, it stops sending movement predictions to the prosthetic hand 110. This prevents unintended or unpredictable movements of the orthotic device.

[0109] Referring to Figure 28, an external programmer charger 280 and an electromagnetic induction charging pad 282 for charging the implant device 20 are shown. To charge the implant device 20 after it has been surgically implanted in the patient, the electromagnetic induction charging pad 282 can be placed on the patient near the location of the implant device 20 within the patient's body. For example, if the implant device 20 is implanted in the patient's rib cage, as shown in Figure 20, the electromagnetic induction charging pad 282 can be placed on the patient's rib cage over the approximate location of the implant device 20 within the patient's rib cage. The external programmer charger 280 and / or the electromagnetic induction charging pad 282 generate an output, such as emitting a light-emitting diode (LED) or generating an audible sound, to indicate that the electromagnetic induction charging pad 282 and the implant device 20 have been placed for electromagnetic induction charging. Once placed and communicating, the external programmer charger 280 and the electromagnetic induction charging pad 282 can charge the battery 230 of the implant device 20 while the implant device 20 is implanted in the patient's body. The same patient programmer charger (PPC) from the Algovita platform can be used to wirelessly recharge the ISU's internal battery using electromagnetic induction coupling at 40kHz with 0.65W. The PPC can also be used to check the ISU's battery level.

[0110] Referring to Figure 29, an external programmer device 290 is shown that communicates with the communication circuit 50 of the implant device 20 and the communication circuit 222 of the orthotic controller 220. Communication between the external programmer device 290 and the communication circuit 50 of the implant device 20 can be performed wirelessly while the implant device 20 is implanted in the patient's body. Communication between the external programmer device 290 and the communication circuit 222 of the orthotic controller 220 can be performed wirelessly or via a wired communication connection. The external programmer device 290 can be implemented in a computing device such as a laptop computer, desktop computer, tablet device, mobile device, or any other suitable computing device configured with an application configured to communicate with the implant device 20 and the orthotic controller 220. The external programmer device 290 can receive EMG signals of the amplified nerve signal 240 from the implant device 20, including, for example, raw EMG signals or filtered, processed, and sampled EMG signals, as described above. As a further example, the external programmer device 290 can determine the configuration parameters of the orthotic controller 220 using filtered, processed, and sampled EMG signals from the implant device 20. Alternatively, the external programmer device 290 may also receive filtered, processed, and sampled EMG signals from the communication circuit 222 of the orthotic controller 220 after the filtered, processed, and sampled EMG signals have been received from the implant device 20. The external programmer device 290 can then determine the configuration parameters that will be used by the orthotic controller 220 to control the prosthetic hand 110 based on the filtered, processed, and sampled EMG signals from the implant device 20.The external programmer device 290 and / or orthotic controller 220 can also use machine learning algorithms to determine and adjust configuration parameters used by the orthotic controller 220 to control the prosthetic hand 110 based on filtered, processed, and sampled EMG signals from the implant device 20.

[0111] Referring to Figure 30, another embodiment of the present disclosure is shown. System 300 in Figure 30 is similar to System 200 described above, except that System 300 does not include an implantable device 20 implanted in a patient. Instead, System 300 in Figure 30 is configured for testing and configuration before the implantable device 20 is implanted in a patient. Similar to System 200 described above, System 300 includes a neural interface system 4 with nerve signals from nerves 6, such as peripheral nerves, amplified by a free tissue graft 10, as described above. In System 300 in Figure 30, however, electrical leads 18 extend outside the patient's body through port 302 and connect to interface 304. Interface 304 connects to an external programmer device 290 that communicates with an orthotic controller 220 of the prosthetic hand 110. System 300 can be used to test and calibrate the parameters of the prosthetic controller 220 before the implant device 20 is implanted in the patient's body, and whether or not the prosthetic hand 110 is attached to the patient. Thus, before the implant device is implanted in the patient, System 300 can be used to review and analyze nerve signals generated by nerve 6 and amplified by free tissue graft 10 in order to determine whether the signals should be sufficient for use in the prosthetic controller 220 to control the prosthetic hand 110.

[0112] As described above, the implantable device 20 of this disclosure can be used both to generate an orthotic control signal output 56 used to control an orthotic device such as a prosthetic hand 110, and to receive stimulation feedback signals such as an orthotic sensor input signal 58 generated by one or more pressure sensors 225 used to provide stimulation and sensory feedback signals to a nerve 6. For example, the implantable device 20 may include one or more Intan RHS2116 stimulator / amplifier chips configured to stimulate nerve tissue and detect biopotentials. The RHS2116 stimulator / amplifier chip provides current-controlled stimulation pulses to individual electrode contacts up to 2.55 milliamperes (mA), which have been shown to be sufficient to elicit sensory representations. The stimulation pulses can be transmitted in a two-phase manner for charge equilibrium. The RHS2116 stimulator / amplifier chip has a fast amplifier reset feature for removing stimulation artifacts before detecting biopotentials. The RHS2116 stimulator / amplifier chip can be powered by a 3.3-volt linear voltage regulator, such as the TPS7A2033 from Texas Instruments. Additionally or as an alternative, a bipolar output power supply can be used to provide the RHS2116 stimulator / amplifier chip with positive and negative voltage sources for stimulation, up to + / - 7 volts (+ / - 7V).

[0113] In one embodiment, the system and method of the present disclosure can be used for both (1) receiving signals from a free tissue graft to generate an orthotic control signal output 56 used to control an orthotic device such as a prosthetic hand 110, and (2) receiving stimulus and sensory feedback signals such as an orthotic sensor input signal 58 generated by one or more pressure sensors 225 of an orthotic device such as a prosthetic hand 110. However, one problem associated with performing simultaneous motor control and sensory perception / stimulation functions is that stimulus signals can produce artifacts or noise when signals are detected and recorded for motor control of the orthotic device. Therefore, the implant device 20 of the present disclosure is configured to mitigate and / or eliminate signal corruption and artifact problems using one or more of the approaches and algorithms described below with reference to Figures 31-33, which illustrate different approaches and algorithms for mitigating and / or eliminating signal corruption and artifacts when signals are detected and recorded for motor control of the orthotic device.

[0114] For example, referring to Figure 31, an algorithm for generating orthotic movement commands and performing stimulation within alternating periods is illustrated. In Figures 31-33, periods t-3, t-2, t-1, and t are illustrated from left to right along the horizontal axis. The periods can be, for example, 10 to 50 milliseconds, but any appropriate period length can be used. Furthermore, X represents the orthotic movement commands generated by the implant device 20 during the period indicated by the subscript. Furthermore, Y represents the EMG signals received from the electrodes implanted or attached to the RPNI, the remaining muscle, and / or the target muscle after the TMR procedure during the period indicated by the subscript. In period t-3, the stimulation function is initially turned off, and the implant device 20 receives the EMG signal Y t-3 Based on the Orthotic Movement Command X t-3This generates a stimulus event that continues throughout the entirety of periods t-2, t-1, and 5 of the example. The stimulus event could be, for example, receiving a signal from the pressure sensor 225 of the assistive device. In other words, a stimulus can be initiated and used when it is detected that there is a need to provide a stimulus to sensor perception, such as when the pressure sensor 225 detects an increase in pressure on the assistive device and generates a pressure signal.

[0115] In the approach and algorithm illustrated in Figure 31, the implant device 20 alternates between performing sensory stimulation and generating orthotic motor commands during a continuous period. For example, sensory stimulation is performed when the stimulation function is on during periods t-2 and t, and orthotic motor commands X are generated. t-2 and X t-1 However, the EMG signal Y t-3 and Y t-1 When generated during periods t-3 and t-1 based on this, the stimulation function is off. Thus, the orthotic motor command is generated during periods t-3 and t-1, and the stimulus for sensory perception is performed during periods t-2 and t. Thus, the generation of the orthotic motor command is temporarily paused during periods t-2 and t while the stimulus is being performed. Similarly, the stimulus is temporarily paused during periods t-3 and t-1 while the orthotic motor command is being generated. In the example in Figure 31, after applying the stimulus and then moving to the next period for generating the orthotic motor command, such as t-1, the position of the orthotic and / or the state of the orthotic command are in the same position or state as they were at the end of the previous period for generating the orthotic motor command, such as t-3. In other words, in the approach and algorithm of Figure 31, the implant device 20 alternates between time windows of (A) generating orthotic motor command X with the stimulus off and (B) pausing orthotic control with the stimulus on. In this way, the implant device 20 switches back and forth between receiving stimuli and sending control signals at intervals of 10 to 50 milliseconds.

[0116] In the approach and algorithm of FIG. 31, the prosthetic movement commands are not updated during the stimulation period, e.g., at t-2 and t. In other words, if at the end of period t-3, the prosthetic 110 is commanded to approach at the speed indicated by command X t-3 then at the beginning of period t-1, the prosthetic will still be approaching at the same speed indicated by the previous command X t-3 This approach and algorithm can be implemented in several control algorithms, including a regressor for estimating kinematics, such as linear regression or a neural network, or a classifier for estimating individual states, such as linear discriminant analysis, Naive Bayes, support vector machine, and / or a neural network

[0117] Referring to FIG. 32, another algorithm for generating prosthetic movement commands and performing stimulation is illustrated. The approach and algorithm illustrated in FIG. 32 are similar to the approach and algorithm illustrated with reference to FIG. 31, except that during the period when the stimulation is being performed, the approach and algorithm illustrated in FIG. 32 estimate the prosthetic movement commands. For example, during periods t-2 and t, the implant device 20 estimates the prosthetic movement commands X t-3 and X t-1 based on previously generated prosthetic movement commands such as X t-2 and X tThis generates the following. For example, instead of simply pausing the generation of orthotic motor commands during the stimulation period, as done by the approach and algorithm in Figure 31, the approach and algorithm in Figure 32 instead provides estimated orthotic motor commands in periods t-2 and t, etc., to make a smoother transition to the generation of orthotic commands in periods t-3 and t-1. This approach and algorithm can utilize recursive or regression algorithms for estimating kinematics, such as Kalman filters or particle filters, or classifiers for estimating individual states, such as Markov models. In this way, the implant device 20 recursively estimates the motor state over time using both incoming signal measurements and mathematical process models. Similar to the approach in Figure 31, the approach in Figure 32 alternates between detecting and receiving signals from electrodes in periods t-3 and t-1, etc., and sending stimulus signals for sensory perception in periods t-2 and t, etc. Furthermore, the implant device 20 does not read or detect any signals from electrodes while performing stimulation, for example, during periods t-2 and t-5. The implantable device 20, however, does not pause the generation of assistive device movement commands during the period in which stimulation is performed, but rather provides a smoother series of commands over time, for example, X t-2 and X t It is possible to generate estimated assistive device movement commands such as the above. Thus, the approach and algorithm in Figure 32 can be used, for example, X t-3 and X t-1 This allows for mixing between two consecutively generated commands, avoiding the command jumps that may occur with the approach and algorithm in Figure 31 when command generation is paused. Thus, the approach and algorithm in Figure 32 can accommodate intermittent measurements and provide smooth command updates during the period in which stimuli for sensory perception are being performed.

[0118] Referring to Figure 33, another algorithm for generating orthotic movement commands and performing stimulation is illustrated. In the approach and algorithm in Figure 33, orthotic movement commands are generated throughout the entire period. However, during periods when stimulation is also performed, such as t-2 and t, the implant device 20 performs artifact estimation to estimate artifacts introduced into the EMG signal Y from the electrodes by the stimulation, and then subtracts or filters out the estimated artifacts from the EMG signal Y and generates orthotic movement commands X based on the resulting / filtered EMG signal after artifact subtraction. In this way, the implant device 20 can remove artifacts that may have been introduced into the EMG signal as a result of performing stimulation while simultaneously reading nerve signals from electrodes attached to the free tissue graft 10. In other words, the implant device performs noise reduction, filtering, or cancellation on the EMG signal Y to remove artifacts that may have been introduced into the signal by performing stimulation for sensory perception.

[0119] For example, a template subtraction algorithm can be used to generate expected artifacts based on a given stimulus signal or set of stimulus signals. The template can be used to subtract artifacts from the EMG signal. In other words, in this approach, the EMG signal Y may be corrupted by the stimulus for sensory perception, but the algorithm or approach can filter the EMG signal Y to subtract artifacts, and as a result, an orthotic movement command can be generated based on the filtered EMG signal Y. For example, the template subtraction algorithm can be implemented by averaging artifacts immediately after the stimulus pulse using multiple exponential filters with filter learning rates. In this way, a representative template can be constructed, which is then subtracted from the original signal to produce an estimated artifact-free signal.

[0120] As an addition or alternative, the ε-normalized least squares mean algorithm can be used to remove unwanted artifacts from the EMG signal Y. For example, the ε-normalized least squares mean algorithm can be implemented as an improved version of the standard least squares mean algorithm and produces better performance for signals with larger and lower signal energy intervals. In the ε-normalized least squares mean algorithm, the adaptive filter relies on a reference signal that is highly correlated with stimulus artifacts. The algorithm can be adapted to various artifact waveforms without requiring complete retraining of the weights used by the algorithm. For example, the ε-normalized least squares mean adaptive filter algorithm convolves the reference signal with filter weights to predict the artifact waveform. The predicted artifacts are subtracted from the original signal to produce an estimated artifact-free signal, which is then further used to update the filter weights after each sample.

[0121] The example in Figure 33 illustrates that stimulation is performed within alternating periods, but stimulation can also be performed continuously over consecutive periods while reading the EMG signal Y and filtering / subtracting any artifacts introduced into the EMG signal Y. Nevertheless, stimulation can be performed within alternating periods, as shown in the example in Figure 33, in order to conserve battery power for the implant device 20.

[0122] Any single algorithm or approach shown in Figures 31 to 33 can be used; however, the implant device 20 can be configured to implement all three approaches and / or switch between them in order to determine and select the best approach or algorithm for a particular environment or scenario.

[0123] Figures 31-33 illustrate an approach in which stimuli for sensory perception are delivered within alternating periods, but other sequences can be used as alternatives. For example, stimuli can be delivered in two or more consecutive periods preceding one or more periods in which EMG signal Y is read to generate assistive device motor commands.

[0124] Multiple sensing electrodes Referring to Figure 34, another embodiment of the present disclosure is shown. In particular, a neural interface system 4a is shown, which is similar to the neural interface system discussed above with reference to Figure 1, and includes a free tissue graft 10 and an implant device 20 comprising a processing circuit 22. Nevertheless, the embodiment of Figure 34 includes a plurality of sensing electrodes 14a, 14b, 14c on a single free tissue graft 10. The sensing electrodes 14a, 14b, 14c sense and receive neural signals amplified by the free tissue graft 10 at different locations on the free tissue graft 10. The sensed and received signals are communicated from the sensing electrodes 14a, 14b, 14c to the implant device 20 via wires 18d, 18e, 18f. The wires 18d, 18e, 18f can be configured as a single bundle of wires, with individual wire ends individually connected to electrodes 14a, 14b, 14c and the implant device 20. The detected and received signals can be amplified by amplifiers 24a, 24b, and 24c and transmitted to a processing circuit 22. Thus, multiple electrodes 14a, 14b, and 14c can be placed at multiple different locations on a single free tissue graft 10 of the RPNI. As discussed above, one or more parts of the nerve 6 (nerve bundle 8) can regenerate within the free tissue graft 10, allowing for nerve regeneration of the tissue. Nerve regeneration can include growing sprouting nerve fibers 12, allowing for nerve regeneration throughout the muscle tissue. Multiple electrodes 14a, 14b, and 14c can be used at different locations on the free tissue graft 10 to detect and receive different motor control signals from the RPNI. The different signals can be received by the implant device 20 as different motor control inputs and used for different purposes, such as controlling different parts of the orthotic device. For example, multiple electrodes 14a, 14b, and 14c located at different positions on a single free tissue graft 10 of the RPNI can be used to receive different motor control signals, which are used to control different fingers of the prosthetic hand 110. While a prosthetic hand 110 is provided as one example, any other type of assistive device can be used in the systems and methods of the present disclosure.

[0125] Thus, a multi-electrode array, for example, electrodes 14a, 14b, and 14c, can detect independent motor signals from a single RPNI structure for high-fidelity orthotic control. When a nerve with multiple functions neurogenesis develops a free tissue graft 10, functionally different motor endplates are neurogenesis across the entire region of the free tissue graft 10. As discussed above with reference to Figure 1, a single electrode can be used for each free tissue graft 10 to detect and record the sum of nearby motor activity. This approach may be sufficient when the nerve requires a single or limited function. Nevertheless, using a single electrode may result in incomplete and lower resolution readings of signals from the nerve and nerve capabilities compared to the neural interface system 4a in Figure 34, which includes multiple sensing electrodes 14a, 14b, and 14c on a single free tissue graft 10. This embodiment may include an array of electrodes 14a, 14b, 14c comprising a single wire bundle, i.e., bundled wires 18d, 18e, 18f, and multiple electrode contacts, for example, a high-density electrode device containing many contacts, or a system in which multiple lead wires and electrodes are independently implanted within the RPNI. This device makes it possible to record independent signals from a single RPNI free tissue graft 10. For example, in humeral amputation surgery, the severed nerve is responsible for the functions of multiple fingers and thumb. Even if it is possible to divide the nerve to create a few RPNIs, each RPNI is likely to contain multiple functions. The multi-channel electrodes of this embodiment can record independent signals from the RPNI, significantly improving the fidelity of assistive device control.

[0126] Multiple sensing electrodes for regenerative peripheral nerve interface (RPNI). The basic functional unit of skeletal muscle is the motor unit, which refers to a motor neuron and all the muscle fibers it neurogenesis. Each individual muscle contains hundreds of motor units, each with a different number of fibers that produce different levels of contractile force. In RPNI (Restorative Neuromuscular Nipple Occlusion), motor units are formed when motor neurons within a severed nerve regenerate existing fibers within a muscle tissue graft. Histological images 350, illustrated in Figure 35, show different motor units forming a neuromuscular junction within a single RPNI.

[0127] A single RPNI can be created on a larger nerve branch, or alternatively, a larger nerve may be divided into several groups of fiber bundles, each group of fiber bundles having a different RPNI structure. The hand and fingers are controlled by 34 muscles. The movement of the thumb itself is controlled by 9 muscles, half of which are located in the hand. For example, if the median nerve is divided into 2-3 RPNIs, each RPNI will have nerve regeneration territories from fiber bundles with multiple functions. In other words, a median RPNI may be regenerated by fiber bundles that control thumb opposition and fiber bundles that control index finger flexion. Thus, motor units within an RPNI can be activated independently for multiple distinct movements, depending on the somatotopic localization of the nerve. When RPNIs are formed on nerves that control multiple hand functions, the muscle tissue can be regenerated by motor unit topography, which creates distinct contraction regions for different movements. Figure 36 illustrates different regions of an RPNI on the median nerve contracting in response to different movements of individual fingers and thumbs as seen under ultrasound. Similarly, Figure 37 illustrates different regions of the RPNI on the ulnar nerve that contract in response to different movements of individual fingers and thumbs as observed under ultrasound. In Figures 36 and 37, arrows illustrate which regions of the RPNI in the ultrasound images map to which joints of the fingers.

[0128] Within the RPNI, the activity of individual motor units can be recorded by a single implanted electrode. In the examples illustrated in Figures 38A and 38B, different motor units are activated for different movements, such as flexion and adduction of the little finger, controlled by the ulnar nerve. Existing control strategies rely on software to analyze these two distinct movements from a single RPNI.

[0129] Electromyograms from a single electrode are processed into features, and algorithms interpret the differences in the processed signals, or individual motor units are inferred through signal decomposition. However, these methods cannot reliably scale as the complexity and number of movements increase proportionally to the number of RPNIs. Furthermore, exponential attenuation of signal amplitude occurs with distance from the electrode recording surface. Simulations estimated that the majority of motor unit activity recorded from intramuscular electrodes likely originates from muscle fibers within 1.5 mm to 2 mm of the electrode, depending on the electrode size and contact spacing. Motor units further from the electrode are recorded with a lower signal-to-noise ratio and are not captured as reliably as closer motor units.

[0130] The system and method of this disclosure allows multiple sensing electrodes to capture activity from different motor units on independent hardware channels. For example, Figure 39 illustrates how two motor units 1 and 2 are connected to two channels 1 and 2, respectively, and input to a processing circuit 22 for processing. This allows for more sophisticated control of more motion. Importantly, this is a fundamentally more reliable method for achieving higher signal resolution than software decomposition. The recorded signals from the individual contacts of the multiple sensing electrodes are dominated by the nearest motor unit. Continuing with the two motion examples above, motor unit 1 can be directly mapped to motion 1, and motor unit 2 can be directly mapped to motion 2. For example, the processing circuit 22 can use the input received via channel 1 to generate a control signal 391 corresponding to motion 1, and use the input received via channel 2 to generate a control signal 392 corresponding to motion 2. This mapping may be performed manually, or, for convenience, an algorithm may be used to learn the relationships. The improved signal independence ensures that this method is performed with higher reliability than conventional methods using a single channel and software decomposition.

[0131] Implanting multiple sensing electrodes can also expand the sampling area, potentially providing complete coverage of the RPNI (Range-Protective Neural Inspection). Both of these factors allow for more effective utilization of the unique information contained within a single RPNI structure for assistive device control.

[0132] In one embodiment, instead of implanting a single electrode within the RPNI, multiple instances of the same electrode can be implanted in opposing regions or along the length of the RPNI. For example, Figure 34 illustrates sensing electrodes 14a, 14b, and 14c along the length of a single free tissue graft 10. These electrodes can have geometric shapes similar to those of currently used intramuscular EMG electrodes. Each electrode can have either a single unipolar recording contact or two bipolar (differential) contacts and is typically implanted along the length of the RPNI. In one example, to optimize the placement of multiple electrodes in the RPNI, each electrode can have a diameter of less than 1.5 mm, a recording contact length of 2 mm to 1 cm, and a spacing of less than 2 cm between bipolar contacts. These dimensions and placements are provided as examples, and other dimensions and placements can also be used. Implanting two or more of these electrodes in this way allows them to spread sufficiently across a larger RPNI and record independent signals from different functional regions.

[0133] In another embodiment, as shown in Figure 40, the recording resolution can be further improved by increasing the number of smaller recording contacts 400 on each electrode. For example, a high-resolution electrode may have two or more contacts that are smaller, e.g., less than 2 mm. This type of exemplary electrode may have eight contacts, each less than 0.2 mm. These dimensions are provided as examples, but other dimensions may also be used. Each contact of the electrode selectively records a small number of distinguishable and independent motor units. The high-resolution electrode may be implanted across the width of the RPNI so as to traverse multiple muscle fibers along the length of the RPNI, as shown in Figure 40. Alternatively, matrix array contacts spanning multiple dimensions of the RPNI may be used. One or more high-resolution electrodes may be inserted into the RPNI.

[0134] Referring to Figure 41, the processing circuit 20 can include multiple separate input channels for each RPNI. For example, RPNI 410 generates outputs for eight channels input to the processing circuit 20. RPNI 412 generates outputs for two channels input to the processing circuit. RPNI 414 generates an output for one channel input to the processing circuit 20. In the example in Figure 41, the processing circuit, along with amplifiers and circuits for receiving inputs from the various channels from RPNI 410, 412, and 414 as described above, can be contained within a single implantable device 20.

[0135] Referring to Figure 42, in another embodiment, when multiple high-resolution electrodes are used, a network of implant devices comprising multiple modules can be used. For example, a wireless body area network (BAN) can be used for communication between multiple separate implant devices / modules. Alternatively, a wired network can be used to connect multiple separate implant devices / modules and enable communication between them. Each module may include the hardware, circuitry, memory, and processor described above as being included in the processing circuit 22. Furthermore, the modules may include sufficient communication hardware and software stored in memory to enable communication between modules. In this way, the modules can receive inputs from various input channels of the RPNI, work together to generate system outputs, and control the assistive device. For example, Figure 42 shows three modules: Module 1, Module 2, and Module 3. Modules 1 and 2 are each connected to eight channels that receive inputs generated by the corresponding associated RPNI. Module 3 is connected to a first RPNI that generates inputs to Module 3 on two channels, and a second RPNI that generates inputs to Module 3 on a single channel.

[0136] Figures 43A and 43B are photographs of RPNIs equipped with multiple sensing electrodes. Figure 43A shows an RPNI with a patch electrode on the back. Figure 43B shows an RPNI with an RPNI on the front, along with a patch electrode on the back.

[0137] Referring to Figure 44, two bipolar patch electrodes were implanted on the same RPNI, equidistant from the stimulated nerve cuff. In Figure 44, patch 1 corresponds to the solid line graph and patch 2 corresponds to the dashed line graph, showing recordings obtained from the two patch electrodes in the same recording window. Stimulation of the nerve cuff generated independent motor signal transmissions with different waveforms and maximum amplitudes, demonstrating that selective signals need to be recorded on independent hardware channels to capture different signals across the RPNI.

[0138] Referring to Figures 45 and 46, two sets of thin wire electrodes were implanted on the same RPNI, equidistant from the nerve stimulating cuff. In Figure 45, the dashed line illustrates the location of the RPNI. It can be seen that the sets of thin wire electrodes are attached to the RPNI. Referring to Figure 46, different waveforms were recorded from the two sets of electrodes, similar to the results with patch electrodes. In the graph of Figure 46, the solid line corresponds to the first set of thin wire electrodes, and the dashed line corresponds to the second set of thin wire electrodes. Similar to the RPNI, the various configurations described further below have independent motor units within the muscle tissue. Some of the following embodiments include a single electrode in the muscle and a single electrode in the skin, but embodiments can also be extended to include multiple sensing electrodes instead of each single electrode to capture more independent signals.

[0139] Composite regenerative peripheral nerve interface (C-RPNI) Referring to Figure 47, another embodiment is shown that utilizes a composite regenerative peripheral nerve interface (C-RPNI) structure 500, which includes a combination of a free skin or dermal graft 502 and a muscle graft 504 fixed around a targeted mixed sensorimotor nerve 506 implanted between the dermal graft 502 and the muscle graft 504. Once implanted, the nerve 506 demonstrates preferentially targeted nerve regeneration, such that sensory nerve fibers regenerate the dermal graft 502 and motor nerve fibers regenerate the muscle graft 504. Thus, the dermal graft is regenerated by sensory feedback nerve fibers connected to the sensorimotor nerve 506, and motor control nerve fibers, also connected to the sensorimotor nerve 506, regenerate the muscle graft 504.

[0140] As shown in Figure 47, stimulating electrodes 30a, 30b, and 30c are attached to a dermal graft 502 and provide stimulation signals transmitted from the implant device 20a via wires 34a, 34b, and 34c to sensory nerve fibers within the dermal graft 502. Similarly, sensing electrodes 14a, 14b, and 14c are attached to a muscle graft 504 and detect motor control signals from motor control nerve fibers within the muscle graft 504. The sensing electrodes 14a, 14b, and 14c transmit the detected motor control signals to the implant device 20a via wires 18d, 18e, and 18f. Although the embodiment in Figure 47 illustrates three stimulating electrodes 30a, 30b, and 30c and three sensing electrodes 14a, 14b, and 14c, any number of sensing electrodes or stimulating electrodes can be used according to this disclosure. For example, according to this disclosure, a structure having one sensing electrode and one stimulating electrode can be used. Furthermore, according to this disclosure, a structure having multiple sensing electrodes and only one stimulating electrode can be used. Furthermore, according to this disclosure, a structure having one sensing electrode and multiple stimulating electrodes can be used. Furthermore, according to this disclosure, a structure having two or more, for example, two, three, or four sensing electrodes and two or more, for example, two, three, or four stimulating electrodes can be used.

[0141] The implant device 20a is similar to the implant device 20 discussed above, except that it includes separate processing circuits for stimulation and detection. For example, the implant device 20a includes a stimulation processing circuit 22a connected to wires 34a, 34b, and 34c, which transmits stimulation signals from the implant device 20a to the stimulation electrodes 30a, 30b, and 30c. The implant device 20a also includes a detection processing circuit 22b connected to wires 18d, 18e, and 18f, which receives detected motor control signals from electrodes 14a, 14b, and 14c. The motor control signals received by electrodes 14a, 14b, and 14c can be used by the detection processing circuit 22b and the implant device 20a to perform motor control of the prosthetic device. Similarly, the implant device 20a and the stimulation processing circuit 22a can receive sensory feedback information from the prosthetic device and generate and transmit stimulation signals to electrodes 30a, 30b, and 30c, thereby providing sensory feedback signals to the sensorimotor nerve 506. In this way, the C-RPNI structure 500 can simultaneously amplify motor signals via muscle grafts 504 while simultaneously providing sensory feedback via dermal grafts 502.

[0142] According to this disclosure, a device comprising multiple sensing electrodes can be used with multi-tissue RPNI grafts, i.e., C-RPNI having both dermal and muscle tissue. For example, a single electrode implanted in muscle and a single electrode implanted in skin can be used independently to provide simultaneous or real-time bidirectional interfacing. To further enhance the selectivity of this interface, a multi-channel electrode device can also be used, in which multiple sensing electrodes can record independent signals from muscle, while multi-contact stimulating electrodes (as discussed above) may be used to activate different fibers for selective sensory feedback or to perform current manipulation on dermal tissue.

[0143] Muscle cuff regenerative peripheral nerve interface (MC-RPNI) Furthermore, while the above embodiments outline nerve endings placed within an RPNI or C-RPNI to regenerate muscle tissue within the RPNI or C-RPNI, any of the above embodiments can also be used in conjunction with an RPNI or C-RPNI attached to surround a healthy, original nerve. For example, any of the above embodiments can be used in a muscle cuff regeneration peripheral nerve interface (MC-RPNI) having a structure comprising a free muscle graft and / or free dermal graft circumferentially fixed to a healthy, original peripheral nerve. The free muscle graft and / or free dermal graft then become regenerated by the nerve it contains, amplifying nerve signals to facilitate reliable and accurate detection of motor intent and transmit sensory feedback stimuli to the healthy, original nerve. MC-RPNI devices can be used in conjunction with powered exoskeleton devices to restore function in individuals with reduced limb muscle strength. Thus, RPNI or C-RPNI structures can be attached circumferentially around healthy, intact nerves to create MC-RPNIs, where the healthy, intact nerves regenerate around the neurodermized muscle grafts. The RPNI muscle grafts amplify signals, which can be detected by electrodes and used, for example, to control an exoskeleton. Furthermore, dermal grafts can amplify stimulus signals to stimulate healthy, intact nerves, providing sensory feedback to those nerves.

[0144] According to this disclosure, a device comprising multiple sensing electrodes can also be used to record independent signals from MC-RPNIs circumferentially attached around a portion of a nerve, rather than at the nerve terminal. The MC-RPNI can be a single-tissue structure (e.g., having only muscle grafts or dermal grafts) or a multi-tissue structure (e.g., having both muscle graft portions and dermal graft portions), and the device can provide selective interfacing for controlling an exoskeleton or other prosthetic or medical device. In the multi-tissue embodiment, the MC-RPNI structure can have a different geometric shape from the C-RPNI structure for nerve terminals. For example, compared to the “sandwich” configuration in the nerve terminal embodiment, the MC-RPNI may use concentric wrapping of tissue. Furthermore, different geometric shapes or implantation procedures can be used for multiple electrodes, such as multiple sensing electrodes and / or multiple stimulating electrodes.

[0145] C-RPNI and MC-RPNI hardware and signal processing Referring to Figure 48, the implantable device 20a, the stimulation processing circuit 22a, and the detection processing circuit 22b can be used in communication with multiple C-RPNI structures 500. Although Figure 48 shows C-RPNI structures 500, MC-RPNI structures may be used as an alternative. The stimulation processing circuit 22a and the detection processing circuit 22b may each include a stimulation front end and a detection front end. The stimulation front end and the detection front end may include modules, processors, programming, memory, circuits, and / or chips sufficient to perform the functions described in this disclosure as being performed by the stimulation processing circuit 22a and the detection processing circuit 22b. Therefore, the terms stimulation front end and detection front end can be used interchangeably with the terms stimulation processing circuit 22a and detection processing circuit 22b, respectively.

[0146] Thus, each C-RPNI structure, having one or more stimulating electrodes and one or more sensing electrodes, can utilize and perform dedicated hardware and signal processing for its stimulating and sensing electrodes. Having separate stimulating and sensing front ends in the implantable system allows for a wider selection of ASICs that are fully optimized for either low-noise sensing or enhanced stimulating capability, unlike systems with a single processing circuit that provides both sensing and stimulating functions, such as the Intan RHS2116 chip mentioned above.

[0147] Thus, according to this disclosure, the implant device 20a is capable of simultaneously stimulating and recording signals from the same nerve.

[0148] Referring to Figure 49, the data illustrated were collected by alternating stimulation and recording to the same nerve. This process was repeated 100 times to verify consistent signal transmission. The constant compound muscle action potential (CMAP) in the graph on the right of Figure 49 illustrates that the muscle graft was not fatigued by simultaneous stimulation of the dermal side of C-RPNI, as shown by the compound sensory nerve action potential (CSNAP) in the graph on the left of Figure 49.

[0149] The approaches and algorithms discussed above with reference to Figures 31-33 can also be used in the detection and stimulation embodiments shown in Figures 47 and 48. For example, as described above with reference to Figures 31 and 32, the detection signal from the muscle graft 504 can be ignored while the stimulation signal is being transmitted to the dermal graft 502. Nevertheless, the embodiments in Figures 47 and 48 use separate, independent processing circuits / front-ends (i.e., separate stimulation processing circuits 22a and detection processing circuits 22b) and separate electrodes (i.e., separate detection electrodes 14a, 14b, 14c and stimulation electrodes 30a, 30b, 30c).

[0150] The use of independent electrodes and processing circuits / front-ends allows for simultaneous interfacing, but stimulation artifacts can still cause interference. Nevertheless, the embodiments in Figures 47 and 48 offer greater flexibility in mitigating artifacts. The embodiments in Figures 47 and 48 can also be used in conjunction with the approach discussed above with reference to Figure 33 to remove stimulation-induced artifacts by template / artifact extraction. The use of the C-RPNI structure 500 allows the implant device 20 and the sensing / processing circuit 22b to continuously sense and record signals from the same nerve being stimulated by the stimulation / processing circuit 22a. In a single-electrode and RPNI configuration, the RPNI needs to be excluded from sensing / recording while signals from other RPNIs / channels are being sensed and recorded.

[0151] By using processing circuits separate from detection and stimulation (22a, 22b), the implant device 20a can simultaneously detect and record signals from all channels, e.g., all C-RPNI or MC-RPNI, while independently transmitting stimulation signals to the same C-RPNI or MC-RPNI. According to this disclosure, the implant device, stimulation processing circuit 22a, and / or detection processing circuit 22b can be configured to switch control parameters during operation depending on whether stimulation is being performed simultaneously with detection / recording. In this approach, the detection processing circuit 22b can be programmed to implement a machine learning model using specific control parameters / weights, configured and trained to receive motor control nerve signals from detection electrodes 14a, 14b, 14c and generate control signals for, for example, an orthotic device. The machine learning model may include, for example, a neural network configured and trained with initial control parameters and weights, in which case the initial control parameters and weights are adjusted as the machine learning model and neural network are trained.

[0152] In this approach, the detection processing circuit 22b can be configured to utilize different control parameters to generate control signals depending on whether the stimulation processing circuit 22a is simultaneously transmitting sensory stimulation signals to the stimulation electrodes 30a, 30b, and 30c. For example, the detection processing circuit 22b can be initialized with initial default control parameters. Once implanted in a patient and connected to the orthotic device, the patient can be commanded to perform specific actions using the orthotic device. In this first phase of the process, the detection processing circuit 22b can detect and record motor signals generated from nerves in the muscle graft 504 via the detection electrodes 14a, 14b, and 14c, and generate control signals to control the orthotic device. The control parameters used and stored by the detection processing circuit 22b can then be updated based on feedback regarding the control operation of the orthotic device. Thus, the first phase of the process generates a first set of control parameters for use during periods when the stimulation processing circuit 22a is not simultaneously providing stimulation signals to the skin graft 502 portion of the C-RPNI structure 500.

[0153] In the second phase, the patient can be instructed to perform the same action using the orthotic device. However, in this second phase of the process, the stimulation processing circuit 22a is configured to simultaneously provide stimulation signals to the skin graft 502 via stimulation electrodes 30a, 30b, and 30c, while the detection processing circuit 22b detects and records motor signals generated from nerves in the muscle graft 504 via detection electrodes 14a, 14b, and 14c, and generates control signals to control the orthotic device. The control parameters used and stored by the detection processing circuit 22b can then be updated based on feedback regarding the control operation of the orthotic device to generate a second set of control parameters. Thus, the second phase of the process generates a second set of control parameters for use during the period when the stimulation processing circuit 22a is simultaneously providing stimulation signals to the skin graft 502 portion of the C-RPNI structure 500. This second set of control parameters beneficially and automatically takes into account any artifacts generated within the C-RPNI structure 500 by the stimulation signals transmitted to the stimulation electrodes by the stimulation processing circuit 22a.

[0154] As a result of the first and second phases of the training process described above, the detection processing circuit 22b is capable of storing and utilizing two separate sets of control parameters, including a first set of control parameters to be used when the stimulation processing circuit 22a is not providing a stimulation signal to the skin graft 502 of the C-RPNI structure 500, and a second set of control parameters to be used when the stimulation processing circuit 22a is providing a stimulation signal to the skin graft 502 of the C-RPNI structure 500.

[0155] Thus, the detection processing circuit 22b can utilize a basic set of "stimulus off" control parameters for receiving signals from the detection electrodes 14a, 14b, and 14c while no stimulation is being performed, and a set of "stimulus on" control parameters for receiving signals from the detection electrodes 14a, 14b, and 14c while stimulation is being performed. Furthermore, multiple sets of "stimulus on" control parameters can be generated and utilized to accommodate different combinations of stimuli. The "stimulus on" control parameters ensure that all implant devices 20a maintain control of the prosthetic device even while stimulation artifacts caused by the stimulation processing circuit 22a are present. This approach, like the other approaches described above with reference to Figures 31 and 32, for example, can beneficially avoid the use of alternating time windows, thereby reducing the update / detection rate of the detection processing circuit 22b by half. This approach can also beneficially avoid the use of template matching algorithms and precise timing adjustments required for the approaches described above with reference to Figure 33.

[0156] Bidirectional interface with electrodes in different tissues for composite regenerative peripheral nerve interface (C-RPNI). Stimulation of the RPNI enhances kinesthetic (e.g., movement) and cutaneous sensations. Furthermore, Figure 50 shows the patient's stable sensory detection threshold for RPNI over time. Figure 51 shows a hand map illustrating the accuracy of somatotopic localization of the referenced sensation. However, the evoked cutaneous sensations were not reported as spontaneously felt, and were often described as tingling or vibration. This may be due to the different composition of sensory end organs compared to skin, and the inability to separate sensory fibers from muscle fibers in RPNI.

[0157] C-RPNI consists of severed nerves implanted between a muscle graft and a dermal graft. As discussed above, this allows sensory nerves to regenerate on the dermal side and motor nerves to regenerate on the muscle side. Therefore, motor signal transmission can be facilitated not only through electromyographic (EMG) recording from the muscle component, but also independently through stimulation of the dermal component to facilitate sensory feedback. C-RPNI can be used to record efferent motor signals or to provide afferent sensory feedback. However, many nerves in the body, especially in high-level transection surgeries, are a mixture of motor and sensory nerves. In such cases, using multiple electrodes in C-RPNI provides an interface for completely independent, near-simultaneous sensory feedback and motor control.

[0158] Referring to Figure 52, the bidirectional interface consists of a sensing electrode 520 implanted in muscle tissue 504. The sensing electrode 520 can be a single unipolar or bipolar electrode, as shown in Figure 52. For example, the sensing electrode 520 can include recording contacts 524, such as the two recording contacts 524 shown in Figure 52. Alternatively, the sensing electrode 520 can include multiple electrodes or a high-resolution sensing electrode. One or more stimulating electrodes can also be implanted. For example, the stimulating electrode 522 can be a multi-contact electrode with multiple stimulating contacts 526, as shown in Figure 52. A multi-contact electrode with three or more small electrode contacts is described in U.S. Patent No. 1,0779963, which is incorporated entirely herein by reference. Multiple sensing contacts selectively activate fibers near the dermal graft, thereby enabling finer resolution control of perceived location and quality of sensation. The stimulating electrode 522 can also be a unipolar or bipolar electrode with larger contacts. Nevertheless, the stimulation electrode 522, with its larger contact area, may unintentionally activate muscle tissue.

[0159] Referring to Figure 53, the implantable device 20a can use a dedicated sensing ASIC optimized for low-noise sensing and a dedicated stimulating ASIC. For example, the implantable device 20a can include the stimulating processing circuit 22a and the sensing processing circuit 22b on separate ASICs. Alternatively, referring to Figure 54, a combined processing circuit 540 for sensing and stimulating can be used and integrated on a single ASIC. Typically, a combined ASIC minimizes size at the expense of reduced performance and capability compared to implementations using separate ASICs for sensing and stimulating.

[0160] Referring to Figure 55, the processing circuit can be contained within a single implant device 550. For example, as shown in Figure 55, the single implant device 550 is a sensing and stimulating implant device with a high number of channels. Referring to Figure 56, the processing circuit can be distributed across a network of multiple implanted devices / modules in the system. Existing implementations of networked implants can be utilized. For example, a wireless body area network (BAN) or a wired networked implant can be used. The network architecture can be used to increase the total number of channels in the system and may be organized geographically as shown in Figure 56 or functionally as shown in Figure 57. Each module can include the hardware, circuitry, memory, and processor described above as being included in the processing circuit 22. Furthermore, modules can include sufficient communication hardware and software stored in memory to enable communication between modules. In this way, modules can receive inputs from various input channels of the C-RPNI500, work together to generate system outputs, and control the assistive device. For example, Figure 56 shows three modules, Module 1, Module 2, and Module 3, each module containing a detection and stimulation processing circuit for performing detection and stimulation functions via multiple detection and stimulation channels connected to each C-RPNI500. On the other hand, Figure 57 includes detection module 1 and stimulation module 1. Detection module 1 is connected to a bus connected to the detection channels of each C-RPNI500, and stimulation module 1 is connected to a separate bus connected to the stimulation channels of each C-RPNI500. Alternatively, a single bus can be used for both modules and all C-RPNIs.

[0161] In each of the embodiments, various techniques can be applied to mitigate stimulus artifacts for a true bidirectional interface. For example, as discussed above, one technique involves successively alternating between a stimulus window and a recording window. This technique enables near-simultaneous afferent and centrifugal signal transmission. To minimize significant delays in the assistive device control system, the sum of each block of the stimulus window and recording window should be less than 200 ms (ideally less than 100 ms), although other durations may be used. The stimulus window and recording window do not need to be the same length. Rather, the stimulus window can be minimized as much as possible while eliminating degradation of centrifugal signal measurements due to artifacts. If other signal filtering or artifact subtraction techniques successfully remove stimulus artifacts, the stimulus window can be completely eliminated for true simultaneous control.

[0162] As shown in Figure 58, a photograph of the C-RPNI structure is shown, which includes a severed nerve implanted between a dermal graft 580 and a muscle graft 582.

[0163] Figure 59 shows a photograph of the electrodiagnostic setup, in which afferent signaling was generated and recorded by stimulating a dermal graft with a patch electrode and recording afferent activity using a nerve cuff on the proximal CP nerve. Muscle signaling was generated and recorded by stimulating the proximal CP nerve with a nerve cuff and recording EMG signals with bipolar intramuscular electrodes in the muscle graft.

[0164] Within each C-RPNI, dermal grafts spontaneously attract sensory fibers, which are interfaced using dedicated stimulating electrodes and stimulating front-ends, i.e., stimulating processing circuits. Muscle grafts spontaneously regenerate nerves through motor fibers interfaced using dedicated recording electrodes and sensing front-ends, i.e., sensing processing circuits.

[0165] C-RPNI allows for simultaneous stimulation and recording from the same nerve. Referring to Figure 60, the graph illustrates data collected by alternating stimulation and recording to the same nerve. This was repeated 1,000 times to verify consistent signal transmission. The constant CMAP illustrated in Figure 60 demonstrates that the muscle did not fatigue by stimulating the dermal side with C-RPNI.

[0166] Composite cuff-regenerative peripheral nerve interface (CC-RPNI) An ideal control system for a reinforced exoskeleton responds to user intent by directly utilizing signals from the nervous system. Because the reinforced exoskeleton must generate movement with minimal latency in conjunction with functional muscles providing partial mobility, a direct interface for capturing movement intent from the nerves is key. Sensory feedback is a crucial component of control for both the upper and lower limbs. Patients with healthy, intact nerves may only retain partial motor function and partial sensory feedback in their affected limb. Therefore, in these cases, a dual interface is required to fully restore function. The Composite Cuff-Regenerative Peripheral Nerve Interface (CC-RPNI) is an ideal interface for this dual function, as it includes a dermal component capable of nerve regeneration by sensory fibers for sensory feedback, and a muscular component capable of amplifying motor signals to detect movement intent.

[0167] The electrode interface, implantable device, and processing circuitry of the CC-RPNI structure are similar to those described above for bidirectional interfacing of electrodes in different tissues in the composite regenerative peripheral nerve interface (C-RPNI). However, in the case of the CC-RPNI structure, the multi-tissue graft is created on the healthy, original nerve to maintain residual downstream function.

[0168] As shown in Figure 61, the CC-RPNI structure includes a free muscle graft 504 and a free dermal graft 502 circumferentially fixed to a healthy, original peripheral nerve 610. The free muscle graft 504 and free dermal graft 502 are subsequently regenerated by the nerves they contain, amplifying nerve signals to facilitate reliable and accurate detection of motor intention and enable the communication of sensory feedback.

[0169] Similar to the discussion of the C-RPNI structure discussed above, the bidirectional interface consists of a sensing electrode 520 implanted in muscle tissue 504. The sensing electrode 520 can be a single unipolar or bipolar electrode, as shown in Figure 61. For example, the sensing electrode 520 can include recording contacts 524, such as the two recording contacts 524 shown in Figure 61. Alternatively, the sensing electrode 520 can include multiple electrodes or a high-resolution sensing electrode. One or more stimulating electrodes can also be implanted. For example, the stimulating electrode 522 can be a multi-contact electrode with multiple stimulating contacts 526, as shown in Figure 61. A multi-contact electrode with three or more small electrode contacts is described in U.S. Patent No. 10779963, which is incorporated entirely herein by reference. Multiple sensing contacts selectively activate fibers near the dermal graft, thereby enabling finer resolution control of perceived location and quality of sensation. The stimulating electrode 522 can also be a unipolar or bipolar electrode with larger contacts. Nevertheless, the stimulation electrode 522, with its larger contact area, may unintentionally activate muscle tissue.

[0170] Referring to Figure 62, a photograph of the CC-RPNI structure 626 is shown. The CC-RPNI structure 626 includes a muscle graft 620 wrapped circumferentially around a healthy, unaltered nerve 622. A dermal graft 624 is fixed to the muscle graft 620 around the healthy, unaltered nerve 622.

[0171] Referring to Figure 63, afferent signal transmission was generated and recorded by stimulating a dermal graft with a patch electrode and recording afferent activity with a nerve cuff on the proximal CP nerve. Muscle signal transmission was generated and recorded by stimulating the proximal CP nerve with a nerve cuff and recording EMG signals with a thin wire electrode in the muscle graft.

[0172] Referring to Figure 64, the graph illustrates the afferent and efferent signaling of CC-RPNI. The left-hand side of the graph in Figure 64 illustrates successful nerve regeneration in the dermal portion of the graft for afferent sensory signaling. Electrical stimulation of the dermal graft produces CSNAP recorded upstream from the nerve cuff. The right-hand side of the graph in Figure 64 illustrates successful nerve regeneration in the muscular portion of the graft for efferent motor signaling. Stimulation of the nerve cuff produces downstream CMAP recorded from the muscle.

[0173] Muscle cuff and skin regeneration peripheral nerve interface (RPNI) equipped with electrodes in muscle and skin. Typical implanted electrodes have large contact points relative to the size of RPNI and C-RPNI structures. Stimulating the structure with large electrodes can create a large activation field that unintentionally activates muscle and nerve fibers. Referring to Figure 65, an example setup for an electrophysiological experiment on a C-RPNI 650 structure is shown, including a dermal graft 654 and a muscle graft 656 attached to a nerve 658. In the example setup, a stimulating dermal electrode 651 is configured to transmit a stimulating electrical signal to the dermal graft 654, and a recording muscle electrode 652 is configured to record electrical signals from the muscle graft 656. A stimulating and recording nerve cuff electrode 653 is configured to detect recording sensor-feedback electrical signals from the nerve 658, such as signals transmitted through the dermal graft 654. The stimulating and recording nerve cuff electrode 653 is also configured to transmit direct electrical signals indicating motor intention to the nerve 658. The recording muscle electrode 652 is then capable of recording the electrical signals transmitted from the stimulation and recording nerve cuff electrode 653 to the nerve 658, and then to the muscle graft 656.

[0174] Referring to Figure 66, signal recordings using the setup illustrated in Figure 65 are shown. The graph on the left side of Figure 66 illustrates nerve cuff and muscle graft recordings when a dermal graft 654 is stimulated using a stimulating dermal electrode 651. For example, the afferent nerve action potential generated by stimulating the dermal graft 654 is shown in the upper left graph of Figure 66. Nevertheless, the muscle graft 656 is also activated by electrical stimulation, as evidenced by the large amplitude compound motor action potential shown in the lower left graph of Figure 66. The magnitude of this activation, shown in the lower left graph of Figure 66, is comparable to the activity generated from efferent nerve signaling, as shown in the lower right graph of Figure 66. The graph on the right side of Figure 66 illustrates recordings when nerve 658 is stimulated by a stimulating and recording nerve cuff electrode 653. As shown in the lower left and lower right graphs of Figure 66, the muscle graft recordings are similar when stimulation is provided by nerve 658 and when stimulation is provided by dermal graft 654. As discussed above, near-simultaneous interfacing can be achieved without fatigue of muscle tissue. Nevertheless, if the intention is to produce only controlled somatosensory feedback, such as tactile feedback, this unintended activation can lead to undesirable clinical effects, such as uncontrolled motor sensation. Uncontrolled sensation can disrupt and degrade sensory feedback. Furthermore, undesirable clinical effects may include muscle activation caused by electrical stimulation, which can disrupt voluntary motor commands. As discussed above, multi-point sensing electrodes may be used to mitigate this problem.

[0175] To address these issues, the sensing and stimulating interfaces are geometrically separated, allowing for the use of larger, simpler electrodes that can stimulate afferent sensory fibers without interfering with muscle activation.

[0176] Referring to Figure 67, a structure is shown having a single tissue cuff circumferentially surrounding a portion of the nerve 670 and an RPNI with a different tissue located at the end of the nerve 670. For example, a free muscle graft cuff 672 is attached circumferentially around the nerve 670, and a free dermal graft 674 is attached at the end of the nerve 670. A sensing electrode 520 is implanted within the muscle tissue 672. The sensing electrode 520 can be a single unipolar or bipolar electrode, as shown in Figure 67. For example, the sensing electrode 520 can include recording contacts 524, such as the two recording contacts 524 shown in Figure 67. Alternatively, the sensing electrode 520 can include multiple electrodes or a high-resolution sensing electrode. One or more stimulating electrodes can also be implanted. For example, the stimulating electrode 522 can be a multi-contact electrode implanted within the dermal graft 674, having multiple stimulating contacts 526, as shown in Figure 67. In U.S. Patent No. 1,0779963, which is incorporated in its entirety herein by reference, a multi-contact electrode having three or more small electrode contacts is described. Multiple sensing contacts selectively activate fibers near the dermal graft, thereby enabling finer resolution control of perceived location and sensory quality. The stimulating electrode 522 may also be a unipolar or bipolar electrode with larger contacts. However, a stimulating electrode 522 with larger contacts may unintentionally activate muscle tissue.

[0177] As shown in Figure 67, the free muscle graft 672 is physically and geometrically separated on the nerve 670, creating a gap 676 and physical separation between the muscle graft 672 and the dermal graft 674 along the axis of the nerve 670. The gap 676 between the two tissues geometrically isolates the interface while minimizing signal transmission interference. In this example, a muscle cuff is formed so that the muscle graft 672 is located on the healthy, original portion of the nerve proximal to the dermal graft 674 created at the nerve terminal, and the muscle graft 672 and dermal graft 674 are separated by a physical gap 676 along the axis of the nerve 670. Motor fibers regenerate the motor endplate on the muscle graft 672 of the muscle cuff, and sensory fibers regenerate the dermal graft 674 at the nerve terminal. The motor fibers at the nerve terminal do not have a nerve regeneration target within the dermal graft and are removed.

[0178] Referring to Figure 68, the structure of this embodiment geographically separates the motor interface and the sensory interface along the nerve axis. This significantly improves the isolation between the motor and sensory portions of the tissue interface, enabling independent activation of sensory and motor functions using independent electrodes. Amplified efferent motor signals are recorded from the muscle cuff portion of the structure, and afferent sensory feedback is transmitted through a skin graft at the terminal of the severed nerve 670.

[0179] Referring to Figure 69, signal recordings using the structures of Figures 67 and 68 are shown. For example, the graph in the upper left of Figure 69 shows the graph of afferent nerve action potentials generated by stimulating the dermal graft 674. Unlike the other structures, as shown in the graph in the lower left of Figure 69, no motor unit action potentials are observed in the muscle graft 672, establishing isolation. In other words, in the structures of Figures 67 and 68 of this disclosure, the muscle graft 672 is isolated from the stimulation signals supplied to the dermal graft 674. Nevertheless, as shown in the graph in the lower right of Figure 69, efferent motor signal transmission from the nerve generates large-amplitude motor unit action potentials within the muscle cuff tissue. Similar to C-RPNI, the stability of the cutaneous RPNI and muscle cuff was also verified by 1000 repetitions of afferent and efferent signal transmission, as shown in the graph in Figure 70.

[0180] Non-restrictive discussion of technical terms The foregoing description of the examples has been provided for illustrative and explanatory purposes. It is not intended to be exhaustive or to limit the disclosure. Individual elements or features of a particular example are, in general, not limited to that particular example and are interchangeable and can be used in selected examples, even if not specifically shown or described, where applicable. The same can also be varied. Such modifications should not be considered deviations from the disclosure, and all such modifications are intended to be within the scope of the disclosure. The example examples are provided so as to complete the disclosure and fully convey its scope to those skilled in the art. Numerous specific details, such as examples of particular components, devices, and methods, are described in order to provide a complete understanding of the examples of the disclosure. It will be apparent to those skilled in the art that certain details are not required to be adopted, that the example examples may be embodied in many different forms, and that neither should be taken as limiting the scope of the disclosure. In some example examples, well-known processes, well-known device structures, and well-known techniques are not described in detail.

[0181] As used herein, the term "circuit" can mean part of or include part of an application-specific integrated circuit (ASIC), an electronic circuit, a combinational logic circuit, a field-programmable gate array (FPGA), a processor (shared, dedicated, or group) that executes code, other suitable hardware components that provide the described functionality, or any combination of the above, such as a system-on-a-chip. The term "circuit" may also include memory (shared, dedicated, or group) that stores code executed by the processor.

[0182] The term "code" as used above may include software, firmware, and / or microcode, and may refer to programs, routines, functions, classes, and / or objects. The term "shared" as used above means that some or all of the code from multiple circuits may be executed using a single (shared) processor. Furthermore, some or all of the code from multiple circuits may be stored in a single (shared) memory. The term "group" as used above means that some or all of the code from a single circuit may be executed using a group of processors. Furthermore, some or all of the code from a single circuit may be stored using a group of memories.

[0183] The apparatus and methods described herein may be executed by one or more computer programs executed by one or more processors. The computer programs include processor-executable instructions stored in a non-temporary, tangible, computer-readable medium. The computer programs may also include stored data. Non-exclusive examples of non-temporary, tangible, computer-readable mediums include non-volatile memory, magnetic storage, and optical storage.

[0184] The headings and subheadings used herein (such as “Background Art” and “Summary of the Invention”) are intended solely for the general organization of the topics within this disclosure and are not intended to limit this disclosure or any aspect thereof. In particular, the subject matter disclosed in “Background Art” may include novel technology and should not be considered an enumeration of prior art. The subject matter disclosed in “Summary of the Invention” is not an exhaustive or complete disclosure of the entire scope of the art or any embodiment thereof. The classification or discussion of materials within sections of this specification that have particular utility is for convenience only, and it should not be inferred that, when a material is used in any given component, the material must necessarily or independently function according to the classification of materials herein.

[0185] All patent and patent application disclosures referenced or cited in this disclosure are incorporated herein by reference.

[0186] The descriptions and specific examples illustrate features and embodiments, but are intended for illustrative purposes only and are not intended to limit the scope of this disclosure. Furthermore, the enumeration of multiple embodiments having the described features is not intended to exclude other embodiments having additional features or incorporating different combinations of the described features. The specific examples are provided to illustrate how the described methods, systems, and components are carried out and used, and are not intended to imply, unless otherwise expressly stated, that a given embodiment has been performed or tested, or not performed or not tested.

[0187] As used herein, the words “prefer” or “preferable” describe an embodiment that provides a particular benefit under specific circumstances. Nevertheless, other embodiments may be preferred under the same or other circumstances. Furthermore, the enumeration of one or more preferred embodiments does not imply that other embodiments are unhelpful and is not intended to exclude other embodiments from the scope of this disclosure.

[0188] As used herein, the word “include” and its variations are intended to be non-limiting; therefore, the enumeration of items in the list does not exclude other similar items that may also be useful in the methods, systems, materials, components, and devices described. Similarly, the terms “can” and “may” and their variations are intended to be non-limiting; therefore, the enumeration that an embodiment can or may have a particular element or feature does not exclude other embodiments that do not have that element or feature.

[0189] The term “comprising,” which can be broadly interpreted as a synonym for non-restrictive terms such as “including,” “containing,” or “having,” is used herein to describe and assert the embodiments of this disclosure; however, embodiments may be described using more restrictive terms, such as “consisting of” or “consisting essentially of.” Accordingly, for any given embodiment that enumerates materials, components, or process steps, this disclosure also specifically includes embodiments that consist of or essentially consist of such materials, components, or processes, except for the additional materials, components, or processes (if any) and the additional materials, components, or processes (if any) that affect the essential nature of the embodiment, even if such additional materials, components, or processes are not expressly enumerated in this application. For example, an enumeration of a component or process that lists elements A, B, and C specifically assumes that the embodiment consists of A, B, and C, and that it essentially consists of A, B, and C, excluding element D which may be enumerated in the Art even if element D is not explicitly described as excluded herein.

[0190] As referred to herein, a range, unless otherwise specified, includes an endpoint and includes disclosure of all distinct values ​​within the entire range, as well as any further subdivided ranges. Thus, for example, a range "from A to B" or "about A to about B" includes A and B. Disclosure of values ​​and ranges of values ​​for specific parameters (such as temperature, molecular weight, or weight percentage) does not preclude other values ​​and ranges of values ​​that are useful herein. The use of the term "about" with respect to ranges, values, or thresholds should be considered in the context of a range, value, or threshold as understood by those skilled in the art. Where a range, value, or threshold cannot be determined from the context, the use of the term "about" may correspond to a range of 10 to 15 percent. It is assumed that two or more specific exemplary values ​​for a given parameter may define an endpoint for a range of values ​​that may be claimed for the parameter. For example, if parameter X is illustrated herein as having the value A and also illustrated as having the value Z, it is assumed that parameter X may have a range of values ​​from about A to about Z. Similarly, the disclosure of two or more ranges of values ​​for a parameter (whether such ranges are nested, overlapping, or entirely different) is intended to encompass all possible combinations of value ranges that could be claimed using the endpoint of the disclosure. For example, if parameter X is exemplified herein as having values ​​in the range of 1 to 10, or 2 to 9, or 3 to 8, it is also intended that parameter X may have other ranges of values ​​including 1 to 9, 1 to 8, 1 to 3, 1 to 2, 2 to 10, 2 to 8, 2 to 3, 3 to 10, and 3 to 9.

Claims

1. An implantable device having a detection processing circuit configured to receive electrical signals from a composite cuff tissue graft surgically attached so as to circumferentially surround a subject's healthy, original nerve, and a stimulation processing circuit configured to transmit electrical signals to the composite cuff tissue graft, wherein the composite tissue graft comprises a muscle graft portion and a dermal graft portion, the composite tissue graft is completely surrounded by the subject's non-grafted tissue and is in direct contact with the subject's non-grafted tissue, and additionally, the healthy, original nerve is located at the position of the composite tissue graft A system comprising an implant device, wherein the composite tissue graft is surgically attached to the subject to transmit electrical signals to other tissues of the subject both upstream and downstream from the location, the muscle graft portion and the dermal graft portion are excised from the subject, vascularized, nerve-removed, and then surgically attached to the subject, and the nerve is regenerated from the muscle graft portion and the dermal graft portion of the composite tissue graft after the composite tissue graft has been surgically attached to the nerve, At least one sensing electrode is attached to the muscle graft portion of the composite tissue graft and communicates electrically with the muscle graft portion, At least one stimulating electrode is attached to the dermal graft portion of the composite tissue graft and communicates electrically with the dermal graft portion, The detection processing circuit is configured to receive an electrical signal from the at least one detection electrode, process the electrical signal received from the at least one detection electrode, generate processed motion signal data corresponding to the electrical signal received from the at least one detection electrode, and transmit the processed motion signal data to the prosthetic device controller; the prosthetic device controller is configured to control the prosthetic device based on the processed motion signal data transmitted from the detection processing circuit of the implant device. A system in which the stimulation processing circuit is configured to receive sensory feedback data from the assistive device controller, generate processed sensory signal data based on the sensory feedback data, and transmit an electrical signal corresponding to the processed sensory signal data to the at least one stimulation electrode.

2. The system according to claim 1, wherein the at least one sensing electrode has a diameter of less than 1.5 millimeters, a recording contact length of 1 to 2 centimeters, and a distance between bipolar contacts of less than 2 centimeters.

3. The system according to claim 1, wherein the at least one detection electrode has a recording contact length of less than 2 millimeters.

4. The system according to claim 1, wherein the at least one detection electrode is a high-resolution electrode having eight contacts, each having a recording contact length of 0.2 millimeters or less.

5. The system according to claim 1, wherein the at least one stimulating electrode is a multi-contact electrode having three or more electrode contacts configured to transmit an electrical signal to the dermal graft portion of the composite tissue graft.

6. The system according to claim 1, wherein the detection processing circuit is included on a first application-specific integrated circuit (ASIC) of the implant device, and the stimulation processing circuit is included on a second ASIC of the implant device.

7. The system according to claim 1, wherein the detection processing circuit is included in a first module, the stimulation processing circuit is included in a second module, and the first module and the second module communicate wirelessly with each other via at least one of a wireless body area network (BAN) or a wired network implanted in the subject.

8. The system according to claim 1, wherein the electrical signal received from the muscle portion of the composite tissue graft has a voltage amplitude of approximately 150 microvolts or more.

9. The implant device's detection processing circuit receives electrical signals from a composite cuff tissue graft surgically attached to surround the subject's healthy, original nerves in a circular fashion. A step of transmitting an electrical signal to the composite cuff tissue graft by a stimulation processing circuit of the implant device, wherein the composite tissue graft comprises a muscle graft portion and a dermal graft portion, the composite tissue graft is surgically attached to the subject such that the composite tissue graft is completely surrounded by the subject's non-grafted tissue and in direct contact with the subject's non-grafted tissue, and additionally, the healthy, intact nerves communicate electrical signals to other tissues of the subject both upstream and downstream of the location of the composite tissue graft, the muscle graft portion and the dermal graft portion are respectively excised from the subject, vascularized, nerve-removed, and then surgically attached to the subject, the nerves are regenerated from the muscle graft portion and the dermal graft portion of the composite tissue graft after the composite tissue graft has been surgically attached to the nerves, A method including, At least one sensing electrode is attached to the muscle graft portion of the composite tissue graft and communicates electrically with the muscle graft portion, At least one stimulating electrode is attached to the dermal graft portion of the composite tissue graft and communicates electrically with the dermal graft portion, The detection processing circuit is configured to receive an electrical signal from the at least one detection electrode, process the electrical signal received from the at least one detection electrode, generate processed motion signal data corresponding to the electrical signal received from the at least one detection electrode, and transmit the processed motion signal data to the prosthetic device controller; the prosthetic device controller is configured to control the prosthetic device based on the processed motion signal data transmitted from the detection processing circuit of the implant device. A method comprising: a stimulation processing circuit configured to receive sensory feedback data from the assistive device controller, generate processed sensory signal data based on the sensory feedback data, and transmit an electrical signal corresponding to the processed sensory signal data to the at least one stimulation electrode.

10. The method according to claim 9, wherein the at least one sensing electrode has a diameter of less than 1.5 millimeters, a recording contact length of 1 to 2 centimeters, and a distance between bipolar contacts of less than 2 centimeters.

11. The method according to claim 9, wherein the at least one detection electrode has a recording contact length of less than 2 millimeters.

12. The method according to claim 9, wherein the at least one detection electrode is a high-resolution electrode having eight contacts, each having a recording contact length of 0.2 millimeters or less.

13. The method according to claim 9, wherein the at least one stimulating electrode is a multi-contact electrode having three or more electrode contacts configured to transmit an electrical signal to the dermal graft portion of the composite tissue graft.

14. The method according to claim 9, wherein the detection processing circuit is included on a first application-specific integrated circuit (ASIC) of the implant device, and the stimulation processing circuit is included on a second ASIC of the implant device.

15. The method according to claim 9, wherein the detection processing circuit is included in a first module, the stimulation processing circuit is included in a second module, and the first module and the second module communicate wirelessly with each other via at least one of a wireless body area network (BAN) or a wired network implanted in the subject.

16. The method according to claim 9, wherein the electrical signal received from the muscle portion of the composite tissue graft has a voltage amplitude of approximately 150 microvolts or more.

Images