by Anqi Zhang
figures by Daniel Utter
Have you ever met anyone with only one leg or arm? I bet you have. An estimated 185,000 people undergo amputation procedures in the US every year, with the leading cause being vascular diseases. Thanks to the advances in medical devices, some of the functionalities of the lost limbs can be restored by artificial arms or legs, or prostheses. People with prosthetic legs can walk, run or dance with ease, and those with prosthetic hands can control each finger and grip in a natural, coordinated way. However, current prostheses lack one important aspect of natural limbs — the tactile senses of human skin. Scientists from Stanford University and Seoul National University have developed an artificial nerve that can ‘feel’ how hard it is being pressed and transmit the signal to control biological muscles. For people living with prostheses, this means that, one day, they may be able to restore the lost sense of touch or gain control of the disabled limbs.
Biological sensory nerves
Sensory nerves carry information from the outside world to our spinal cord and brain. In particular, our ability to perceive touch sensation is achieved by a type of sensory nerve ending called mechanoreceptors which are located in our skin. When pressure is applied to the skin, the mechanoreceptors respond by changing their electric voltage (i.e., a measure of electrical energy). The voltages from multiple mechanoreceptors are combined and transmitted to a single neuron, or nerve cell. At a certain voltage threshold, the neuron generates repetitive electrical pulses that are forwarded to other neurons via junctions called synapses, eventually reaching the neurons in the brain to register the touch sensation. The frequency at which the electric pulses are generated (measured in hertz, i.e., number per second) is determined by the applied pressure. Higher pressures produce electrical pulses at higher frequencies, while lower pressures produce lower frequency pulses (Figure 1). These electrical pulses are eventually transmitted to and processed by the brain to feel the pressure of the external stimulus, according to the pulse frequencies.
Artificial sensory nerve
Artificial sensory nerves are at a very early stage in their development and have not yet been tested in humans. However, these artificial nerves have been designed in the hopes that, one day, they will be safe and effective for use in people. To mimic its biological counterpart, the artificial sensory nerve is constructed using three components: resistive pressure sensors, ring oscillators, and a synaptic transistor, corresponding to the biological mechanoreceptors, neurons, and synapses, respectively (Figure 2).
- The resistive pressure sensors are composed of pyramid structures of rubber filled with carbon nanotubes that conduct electrical current to gold electrodes. These pressure sensors exhibit piezoelectric properties – that is, the ability to convert mechanical stress into electricity. An increase in pressure applied on the rubber pushes more carbon nanotubes onto the gold electrodes, causing a larger electrical current and larger voltage input to the ring oscillator.
- The ring oscillators are devices that can generate electrical pulses at frequencies determined by the voltage input. A larger voltage input, as a result of larger pressure on the resistive sensors, results in a higher frequency, similar to the way that sensory neurons generate electrical pulses.
- The synaptic transistor receives and combines the electrical pulses from multiple ring oscillators. Signals from the synaptic transistor can be recorded in a computer (analogous to the brain) and/or used to drive biological muscle movements (as discussed in the next section).
In analogy to the biological sensory nerve, the artificial sensory nerve receives pressure information in the range of human pressure-sensing capability (1 to 80 kilopascals, i.e., a light touch to a firm press) from clusters of pressure sensors. The pressure information is then converted into electrical pulses that match the frequencies of sensory neurons (0 to 100 hertz) by using ring oscillators, and a synaptic transistor integrates the electrical pulses from multiple ring oscillators.
Developing real-life applications for the artificial sensory nerve
En route to developing real-life applications for the artificial sensory nerve, researchers conducted two experiments to test and evaluate this novel technology: reading braille letters and inducing the movement of a detached cockroach leg. The former test showcased the technology’s ability to precisely interpret complex tactile sensations. The latter demonstrated that it is possible to interface the artificial nerve with biological muscles and mimic a realistic nervous response circuitry similar to a ‘reflex’. Both feats take us closer to futuristic notions such as sensing artificial limbs, sensitive robots, and animal-machine hybrids.
Braille alphabet interpretation
The braille alphabet is designed for the blind. Constructed from 6 dots in a 2×3 grid, the 26 letters are represented by 26 different combinations of dots. For example, the braille letter ‘E’ is the combination of dot 1 and dot 5. In the first test, the authors fabricated 6 regions containing resistive pressure sensors, located as 6 dots similar to the layout of the braille grid. Each pressure sensor region was connected to a ring oscillator. The output from different combinations of ring oscillators was used as input for synaptic transistors. When the dots representing certain braille letters were pressed, the corresponding synaptic transistors interpreted the electric pulses from the ring oscillators and generated distinct signals for different letters. These sensory nerves can thus be implemented to improve the tactile abilities of robots. Further, if the output signals can be transmitted into a patient’s brain rather than into a computer, these artificial sensory nerves could dramatically improve the sensing capabilities of prosthetic limbs.
Hybrid bio-electronic reflex
In the biological nervous system, the spinal cord and/or brain receives and processes sensation from sensory nerves. After receiving inputs from sensory nerves, the spinal cord and/or brain can generate an output response via motor nerves (i.e., nerves that send motor signals from the spinal cord and/or brain to the muscles of the body) to generate a mechanical response to the sensation. The mode of communication of motor nerves is thus opposite compared to that of sensory nerves. For that response to be voluntary, the brain must be involved to interface between the sensory and motor nerves, like when you sense something tickling your leg and actively decide to move your arm and fingers to scratch it. On the other hand, when the body wants to react to the stimulus as quickly as possible, the signal and response are coordinated directly by the spinal cord, bypassing the brain altogether. One example of such an ‘involuntary’ response is the knee-jerk reflex, the sudden extension of the leg in response to a tap below the kneecap. In this case, the extension movement is controlled by signals sent from the spinal cord to the leg muscle, skipping the brain all-together.
While true brain interfaces are currently out of reach, the artificial sensory nerve can already emulate a spinal cord reflex. In the second test of their technology, the authors connected the artificial nerve’s synaptic transistor to a stimulating electrode (i.e., metal wire) inserted into the motor nerve of a detached cockroach leg (Figure 3). The electrical pulse output from the synaptic transistor, analogous to the output from the spinal cord, caused the leg to move, much like a reflex. An increase or decrease in the intensity of pressure applied on the resistive sensors increased or decreased the force of leg extension. This test demonstrated that artificial sensory nerves are able to directly communicate with the biological motor nerve and effect a muscle response to an external stimulus.
Promises and challenges
Artificial sensory nerves that can feel pressure and interface with biological muscles represent an important step towards helping people living with prosthetic limbs regain lost sensations. They may even one day give robots the artificial skin that can sense and respond to the environment. In its present state, however, the artificial nerve is still far from having all the sensations of real skin. Specifically, the current artificial nerve can only generate neuron-like signals in response to pressure, while real skin can also feel vibration, texture, temperature, pain and itch. In addition, it remains to be seen whether the human brain can process the signals from the artificial nerves, which is the key to restoring the lost sensations of touch. Future development of this technique aims to incorporate different types of sensors and explore human brain interface, perhaps eventually bringing cyborgs to real life.
Anqi Zhang is a Ph.D. student in the Department of Chemistry & Chemical Biology at Harvard University.