**1. Introduction**

Neural Interfaces are data links between human nervous system and an external device, and allow the transmission of information to and from the human nervous system to the external device. Bioelectric signals can be obtained from the nervous system, and/or transmitted to the nervous system via implanted or surface electrodes. One beneficial application of such neural interfaces is to obtain control signals from undamaged sensory‐motor areas of patientʹs brain for controlling neuroprosthetic devices such as artificial limbs or wheelchairs. Paralyzed or amputated people can reconstruct certain motor functions by using such neuroprosthetic devices. Brain‐machine interfaces, which are direct data links between human brain and machines, are one kind of neural interfaces [1]. Such interfaces have been proposed during this decade to control prosthetic limbs, or to control machines such as wheelchairs or robotic manipulators [2]. However, our present knowledge on brain functions is so limited that we do not fully understand the coding of information expressing the behavior in motions; specially, we do not know the variation of the coding in individual differences or in related thoughts or emotions.

In order to design practical neuroprosthetic devices, we focus on neural interfaces which link peripheral nervous system and external devices in this paper (PNI, Peripheral Neural Inter‐ face). Since peripheral nervous system is much simpler than central nervous system, we may avoid the difficult problems which come from ourlimited knowledge on brain functions while the neuroprosthetic devices interact just with the peripheral nerves. In Section 2, we describe interfaces to peripheral motor neurons by using reinnervation type electrodes. The electrodes are constructed by implanting embryonic neurons into peripheral motor units. The endplates of the implanted neurons which grow into muscles make biological interfaces for motor

© 2013 Obinata et al.; licensee InTech. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2013 Obinata et al.; licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

commands to the muscles. The other ends of the implanted neurons are connected by special types of electrodes to communicate with electric devices of motor controllers. The diameter of mammalian motor neurons is from 0.5 μm to 20 μm. There is a reason why our research requires bio‐compatible micro‐nano technologies for achieving such interfaces of new reinnervation type electrodes.

precursor cells derived from hESCs in patients with thoracic spinal cord injuries [7]. The company Advanced Cell Technology has reported promising results from the clinical trials on Stargardtʹs Macular Dystrophy with retinal pigment epithelium derived from hESCs [8]. However, neuronal replacement in central nervous system is still limited due to a number of cells required to reconstruct complex structures and narrow therapeutic time window [9]. There are still several hurdles to be overcome in order to establish clinical application of

Neural Interfaces: Bilateral Communication Between Peripheral Nerves and Electrical Control Devices 15

One experimental approach to rescue denervated muscles from damage to lower motor neurons or axonal disconnection is transplantation of motoneurons into a peripheral nerve as a source of neurons for muscle reinnervation. Since Erb reported the reinnervation of dener‐ vated muscle by embryonic motoneurons transplanted into peripheral nerve in 1993 [10], several studies have investigated the factors that improve motoneuron survival in peripheral nerve [11‐13]. Considering the simplicity of neural network and wide windows of opportunity forthe treatment [14], peripheral nerve system can be an ideal target for neuronalreplacement therapy. This transplantation strategy may provide the potential to excite these muscles artificially with electrical stimulation. The aim of this study was to evaluate whether trans‐ planted rat embryonic motorneurons into adult rat peripheral nerve would survive, and

The sciatic nerves of adult Fischer 344 rats were transected; 1 week later, dissociated embryonic spinal neurons were transplanted into the distal stump of the tibial and peroneal nerves. Surgical controls underwent the same surgeries but had only medium injected into the peripheral nerves. Tissue analysis and measurement of ankle angles were performed twelve

Ventral spinal cord cell were obtained from Fischer 344 rat embryos (Japan SLC,Inc., Shizuoka, Japan). After Fischer rats on day 14 of pregnancy were anesthetized with isoflurane, ventral spinal cords were resected from the fetuses using a surgical microscope and were cut into small pieces in ice‐cold Hanksʹ balanced salt solution. Ventral spinal neurons were dissociated using papain‐containing separation solution (MB‐X9901; Sumitomo Bakelite Co. Ltd, Tokyo, Japan). For implantation, dissociated neurons were suspended in Neurobasal medium (Gibco)

All procedures were performed on 8‐week‐old male Fischer 344 rats (Japan SLC, Inc.) under isoflurane anesthesia.The sciatic nerves of Fischerrats were completely transectedat midthigh. The nerves were ligated on both ends and the proximal nerve stump was sutured into hip muscles to prevent reinnervation. Approximately 1 x 106 neurons contained within 10 μl of medium were injected into the distal stumps of the tibial and peroneal nerves after 1 week of

containing B27 supplement (Gibco), Glutamax (Gibco), and N‐2 supplement (Gibco).

neuronal replacement therapies in central nervous systems.

whether transplanted motorneurons would form functional motor units.

**2.2. Methods**

*2.2.1. Cell preparation*

weeks after neural transplantation.

*2.2.2. Surgical procedures and transplantation*

Neuroprosthetic devices with interfaces detecting electromyogram (EMG) are in practical use [3]. Such EMG interfaces require to place the electrodes at the end plate of a motor neuron; thus, the paralyzed or amputated users can not control the paralyzed or amputated muscles with the same passes of motor neurons as before paralyzed or amputated. Moreover, EMG interfaces pickup the signals at end plate of a motor neuron; sensory signals from receptors are not obtainable via EMG. This means that two‐way communication by EMG interfaces is impossible. We can assemble sensors and stimulators into a neuroprosthetic device. If we have a way to send the signals from the assembled stimulators to sensory nervous system, ideal neuroprosthetic devices with two‐way communication will be achievable. The combination of EMG pickup and electrical stimulation with surface electrodes was proposed for two‐way communication [4]. However, there is no clear result on the amount of transmissive informa‐ tion through the afferent passes. In Section 3, we show a preliminary study on the possibility of sensory feedback via axial fibers of peripheral sensory neurons. In other words, we try to achieve an artificial afferent pass forfeeding signals back to the brain. A method forimproving on the amount of transmissive information has been proposed by using a configuration with multi‐channel electrodes.

One application of PNIs is to achieve unconscious muscular movements such as walking, writing, dancing, playing musical instruments, and so on. For an example, in walking motions some neural networks generate the pattern of walking, and the steady behavioris closed within the peripheral nervous system and the pattern generators in spinal cord. The upper central nervous system provides the triggers of walking such as start/stop, speedup/slowdown, turn‐ right/turn‐left. The understanding on pattern generators for walking is now enough to simulate human walking or some animalsʹ walking. In Section 4, we give a walking simulation of ʺratʺ to show the possibility of practical usages of PNIs.
