12 Therapies and Prosthetics

This lecture is about how technology can help people with disabilities regain abilities they have lost. It has two learning objectives:

1. Describe how technology can be used to help restore ability to people with disabilities.
2. Explain the neuroscience mechanisms that make neural prosthetics possible.

Exam 2 came on the following Wednesday, and the section of the course after it covers mental health and drugs. The lecture opened with a memory test on the previous lecture, machine learning (see Check yourself).

Sensory substitution (review)

Sensory substitution was first proposed by Paul Bach-y-Rita, who developed the first device to use brain plasticity to give visual information to blind patients. How can sensory substitution allow one sense to be detected through another? It is made possible by re-wiring the brain through experience.

The slide shows Bach-y-Rita’s early apparatus: a camera on a tripod, a cabinet of electronics, and a chair whose backrest holds a grid of points, beside two images of faces built out of dots on that grid. Below it is a modern device: a woman wears sunglasses fitted with a small camera, wired to a handheld controller and to a small pad that she holds to her mouth (compare the electrical stimulation of the tongue from the lecture on hearing and the vestibular system).

A new era of prosthetics

The next part of the lecture, “Putting the ‘science’ back in sci-fi!”, introduces a new era of prosthetics.

Neural prosthetics

Neural prosthetics are distinct from sensory substitution, though the two categories overlap:

• all sensory substitution devices are neural prosthetics,
• but not all neural prosthetics use sensory substitution.

The slide shows a robotic hand picking up a small ball.

How it works

A neural prosthetic does signal transduction like a set of synthetic receptors. It converts environmental information into electrical signals and sends them to the brain. The brain then strengthens the new pathway so that it can use it more effectively.

Replacing the receptors is the key. In a sensory disability, people have lost the ability to detect a sense, while often keeping the ability to process and encode it. Once a signal makes it past a receptor, it is all action potentials, so if you can replace the receptors, you’re in business.

The figure on this slide comes from a brain-imaging study. Panels A to C show the brain viewed from behind, with clusters of activation at the back of the brain, over the occipital lobe; the colors give the z-score, a measure of how strong the activation is, and small inset slices are labeled y = −76 and y = −48. Panel D is a scatter plot of the difference in reaction time (“Reversal − Repetition”, in ms) against the number of activated voxels (the small cubes a brain image is divided into). The points are spread widely, and the fitted line slopes downward, from about 230 ms at 500 voxels to about 10 ms at 3,800 voxels.

Visual prosthetics: the bionic eye

There are two kinds of bionic eye.

• Mounted camera. A camera maps the image across an array of implanted electrodes. Depending on the range of damage, the electrodes can stimulate different regions: retinal neurons, the optic nerve, or the visual cortex.
• Retinal implant. Electronic photoreceptors are implanted under the retina, where they modulate pulses of incoming light.

The photo shows an older man wearing glasses with a small camera mounted on the bridge.

The bionic eye has been in development since 2000, and $50 million is currently allocated to its research and development. Resolution and detail are still very low, but it allows basic visual perception. Three images on the slide show what this means: a camera photo of a person standing by a window, a scene reduced to a few flat dark and light regions with a rectangle marked on it, and a blurry, glowing, low-resolution image of a few objects.

Auditory prosthetics

An auditory prosthetic detects sound with a microphone and converts the sound waves into electrical signals, much as the hair cells in the ear do. There are three kinds:

• cochlear implants,
• auditory brainstem implants,
• auditory midbrain implants.

The slide’s diagram of a cochlear implant shows a sound processor worn behind the ear, an internal implant placed under the skin behind the ear, and a thin electrode running from it into the cochlea, next to the hearing nerve.

Inside the implant, the signal is processed, digitized, and filtered into frequency bands. Each band is routed to the region of the cochlea that responds to its frequency. The cochlea and the auditory cortex are both tonotopically organized (the lecturer’s word is “tonographical”): each region responds best to particular frequencies, and this sound map is maintained from the cochlea to the cortex.

The slide’s figure, titled “Tonotopic map has columnar organization”, zooms in on the primary auditory cortex in the temporal lobe. There, neurons are arranged in columns by frequency, from 500 Hz through 1000, 2000, 4000 and 8000 Hz up to 16,000 Hz; binaural columns run across these (they are not shown in the figure).

A memory check on what neural prosthetics replace came next, followed by a “memory expansion” question on which principles of synaptic plasticity underlie sensory substitution and neural prosthetics (see Check yourself).

Targeted reinnervation: motor and somatosensory prosthetics

• The electrical signals sent to the target muscle are used to drive a motorized prosthetic device.
• When the prosthetic’s skin is touched, the person perceives a tactile sensation on the missing limb.

The slide’s diagram (“Targeted Muscle Reinnervation”) shows a man’s chest and shoulder with three nerves of the arm (the musculocutaneous, median and radial nerves) ending in the chest muscle. EMG (electromyography) electrodes, which record the electrical activity of a muscle, sit over these sites and are wired to a microprocessor controller in the motorized prosthetic arm.

Training your own prosthetic arm

This prosthetic was invented by a student, Gautham Vasan. Users “teach” their own AI prosthetic arm how to move and grasp in complex patterns, using motor signals from above the elbow. It is simple for patients to use and requires minimal technical knowledge. The photos show Vasan wearing bands of sensors on his arm, and a user in a lab, with a band on his upper arm, moving his own arm while a robotic arm on the bench moves with him.

Spinal cord stimulator: the “pain pacemaker”

A spinal cord stimulator delivers mild electric stimulation to the spinal cord.

• The stimulation likely masks pain with a “tingling” sensation, altering pain processing.
• In neuropathic pain conditions (pain caused by damage to the nerves themselves), it decreases the hyperexcitability of pain neurons in the spinal cord.
• It increases the release of GABA and serotonin at the stimulation site.

The slide’s X-ray shows the stimulator’s two electrode leads, rows of small bright rectangles lying along the spine, with their wires running off to one side.

A memory check on methods of restoring ability followed (see Check yourself).

Creative expression

Assistive technologies don’t stop there: some research focuses on how to help individuals with neurological disorders express their creativity.

Machine painting with ALS

Dr. Jana Muenssinger developed the first brain–computer interface to allow creative expression, used with people who have ALS (amyotrophic lateral sclerosis).

• Brain activity is collected with an EEG (electroencephalography) cap and sent as electrical signals to a processor.
• The user works with a colored matrix to create digitized paintings using only brain activity.

The slide shows a photo of Dr. Muenssinger and three of the paintings: one made of bright blocks, circles and other shapes, and two of concentric colored rings.

Writing music

Software has also been developed to write music using a brain–computer interface. In the photo, a person wearing an EEG cap faces a monitor showing a matrix of options and a laptop showing a musical score. The matrix (a) groups its options in colored boxes: note lengths and a rest; the note names C, D, E, F, G, A and H; flat, natural and sharp signs; moving a note up or down by a semitone or an octave; stepping between notes and bars; and commands such as delete and play. Colored arrows link each group to the matching part of the score-writing program (b). (The labels are in German: H is the German name for the note B, Halbton means semitone, and Takt means bar.)

Hebb, machine learning and neural prosthetics

The principle behind all of these technologies is Hebb’s: “Neurons that fire together wire together.”

Senses and abilities can be restored by:

• using an alternative sensory modality to detect the environment (sensory substitution), or
• using sensors to electronically stimulate existing sensory nerves (neural prosthetics).

Through extended use, these new pathways become strengthened and replace the old sensory pathways. The slide shows a human hand touching fingertips with a robotic hand.

Benefits of neural prosthetics

Through these technologies, people can:

• re-experience the world through their senses;
• regain the ability to do things many take for granted;
• regain a sense of autonomy;
• express themselves creatively.

Neural prosthetics are associated with a greater quality of life and better outcomes. The slide’s photo shows a man in a wheelchair fitted with a robotic arm, with a team of doctors and researchers around him.

The lecture ended with an exit discussion reviewing the section (see Check yourself).

Check yourself