Scientific Telekinesis? Paralyzed Man Controls Computer with Thoughts

Information · 08-14-2006, 09:37 PM

Joshua Dalsimer
Matthew Nagle, a 25 year-old quadriplegic paralyzed by a knife wound to the neck, is
the first subject in the clinical trial of BrainGate. The computer interface, built in 2003
by Massachusetts-based Cyberkinetics, is being tested at Brown University by the
neuroscience department chair and Cyberkinetics founder, John Donoghue.

Imagine this: you sit at one end of a room, gaze at a computer screen at the other end, and merely by focusing your thoughts, you open an e-mail, play a computer game, or make a robotic arm pick up a small ball. Surely this is the stuff of science fiction? Until recently it was, but now a team of scientists has implanted electrodes in a paralyzed man’s brain, and designed a program that translates his brain waves into instructions for a computer. More important than the sci-fi possibilities is the fact that the new research provides hope for victims of paralysis and neuromuscular conditions such as Lou Gehrig’s disease; moreover, it increases our understanding of how the brain functions. A paper describing the remarkable accomplishments was published in the July 13 issue of Nature.

From Voltage to Information

Neurons, the cells that make up our nervous system, register signals by creating a voltage, called an "action potential," across their membranes. Action potentials are all-or-nothing: they are either turned on or turned off. So what distinguishes one signal in the brain from another are which neurons are firing, and when. The team of neuroscientists built a 4x4 millimeter (0.16 inches to a side) sensor containing one hundred electrodes; these electrodes could potentially sense and record the electrical activity of up to 96 neurons. By combining the data from the sensor, the computer could recognize the neural firing patterns from a small region of the subject’s brain.

Next they needed a subject. They chose 25-year-old Matthew Nagle, who was stabbed in the neck and became quadriplegic, or paralyzed in all four limbs, in July 2001. After using magnetic resonance imaging (MRI) to identify which part of his brain he had used to move his arms, the scientists surgically implanted the 100-electrode sensor, and began recording a small part of the activity in his brain.

To train the computer to accurately and consistently understand Nagle’s intentions, the scientists used some more complicated exercises, in which Nagle was told to pretend to use his hand to follow a cursor controlled by a technician. The program the computer used was modeled on similar programs meant to interpret the activity in monkeys’ brains. The main difference is that the monkeys weren't paralyzed, so the original software compared neural patterns to actual movement. In contrast, the new program compared Nagle’s brain activity to what Nagle had been instructed to try to do.

But Does it Work?

Now the fun part: all the cards are in place, and it is time to test the new system. The scientists had Nagel perform seven different tasks, starting with simply moving a cursor to random targets or drawing a circle on a screen, and progressing through opening and reading an e-mail, playing the computer game "Pong," adjusting a TV set’s volume, opening and closing a robotic hand, and controlling a simple robotic arm. For each task, Nagle had to pretend to move his arm, in order to move a cursor to a desired position on a screen. So, for instance, he would have to position a cursor to the spot on the screen that says "open hand," and the robotic hand would open. The technology has not progressed sufficiently for Nagle to directly control the hand, but the fact that he can move anything by sheer willpower is pretty impressive stuff.

Joshua Dalsimer
LEFT: Matthew Nagle prepares to test the surgically implanted 100-electrode
sensor interface. RIGHT: Nagle opens his e-mail on a computer screen (top) and
manipulates the robotic hand (bottom) by using sheer willpower to move the
screen cursor.

Of course, nothing is perfect, and Nagle had a much harder time moving a cursor around than an able-bodied control-subject with a mouse. In the first exercise, in which he had to move a cursor to random targets, he was able to hit 73-95% of the targets, but it took him on average over 2 times as long as an able-bodied person, and sometimes much longer. In addition, Nagel had a hard time keeping the cursor still for any length of time. Perhaps more significantly, a second subject, with whom the team has begun to experiment, performed less accurately than Nagel and the scientists are trying to figure out why. Tests are being done on two more subjects, but no results have yet been reported.

The Destination, not the Journey

The same issue of Nature that carried the study involving Nagle also published a paper by a second team of scientists in which they proposed a different approach to interpreting neural activity. Instead of working with a paralyzed human, they conducted tests on monkeys, and achieved some promising results. Their basic idea was that instead of focusing on tracing movement towards a target, the program could focus on the target itself.

These researchers, led by Dr. Gopal Santhanam of Stanford University, made use of the fact that before a monkey actually begins to move its arm, its brain has already determined the arm’s destination. The scientists would first show a monkey a target, or many targets, and only afterwards instruct the monkey to grab a target with its hand. They then used the monkey’s brain activity from the time the monkey saw the target until it began to move towards the target, called the "delay period," to predict where it was going to grab. In the same way that the first team analyzed Nagle’s neural patterns to determine how he was pretending to move his arm, the second team analyzed the monkey’s neural patterns to determine the final destination of its arm.

There are two principal advantages to the second approach. The first is that since the program focuses on only one point, as opposed to an entire trajectory, there’s less estimation and hence more accuracy. The second is that the system uses only data recorded before the monkey moves its arm, so it doesn't have to wait for actual movement, and can progress much faster.

Santhanam’s team, which measured brain activity with a device similar to the one used by the first team, divided the delay period into two portions: in the first, which they referred to as Tskip, the monkey processes what it sees and decides which target to hit. In the second, Tint, the brain actually prepares to move towards the target. The scientists chose to ignore Tskip, and focus on Tint to predict where the monkey would grasp. After deciding the length of Tskip to be 150 milliseconds, the issue became how long Tint should be. The longer it is, the more accurate the prediction; with a long enough time, they were able to make correct predictions up to 90% of the time. On the other hand, longer Tint’s slow down the process.

To reach a happy medium between speed and accuracy, the scientists used a statistic called information transfer rate capacity, or ITRC, which measures the speed at which the computer gathers information from the monkey’s brain. They found that the ITRC is the highest, about 6.5 bits per second, when Tint is short—about 60-130 ms. An ITRC of 6.5 is more than 4 times higher than what other methods of brain-wave interpretation were able to achieve, and corresponds to typing at about 15 words per minute. Although the results are preliminary, and involve monkeys and not humans, they are surely encouraging.

Down the Road

Certainly, a lot more work needs to be done before either team’s findings lead to practical medical therapies. For one thing, at present all parts of the device are connected with wires; so, for instance, there were actually wires coming out of Nagle’s head. A practical device would almost certainly be wireless, both to make it more convenient and to lessen the chance of an infection. Another problem that has to be surmounted is that the electrodes start to lose their ability to measure neural signals after about six months. Neither team of scientists has been able to figure out why this is so, though Santhanam’s team believes they may have a partial solution.

ShowbizIreland/Getty Images
Physicist Stephen Hawking suffers from Lou Gehrig's Disease, a disease that attacks
the central nervous system. With the help of a computer and voice synthesizer,
Hawking is able to communicate.

Currently, many paralysis victims use systems that interpret either their voice or their eye movements to perform tasks. At present, these techniques are far more accurate and practical than interpreting brain waves, and in addition do not require surgery. On the other hand, the advantage of communicating directly with the brain is that it allows the patient to look wherever he wants, or talk at the same time as he performs the task. Given these pluses, the brain-wave approach could become standard if it is further refined.

Laboriously and uncertainly pushing a cursor around on a computer screen is certainly a far cry from what science fiction writers have imagined, such as a super-human bionic man who uses brain waves to completely control robotic limbs. On the other hand, the advances described in the papers mark a major step forward—and even if they never produce a super-human, they offer hope of greatly improving the lives of the paralyzed.

Information · 08-14-2006, 09:39 PM

Bibliography

Biello, David. "Tiny Chip Converts Paraplegic's Thought into Action." Scientific American.com, (July 13, 2006) [accessed July 18, 2006]: www.sciam.com/ article.cfm? chanID= sa003& articleID= 0004DD3A- 6E37- 14B5- AE3783414B7F0000.

Hochberg, Leigh R. et al. "Neuronal ensemble control of prosthetic devices by a human with tetraplegia." Nature, (July 13, 2006) [accessed July 17, 2006]: www.nature.com/ nature/ journal/ v442/ n7099/ full/ nature04970.html.

"Is this the Bionic Man?" Nature, (July 13, 2006) [accessed July 17, 2006]: www.nature.com/ nature/ journal/ v442/ n7099/ full/ 442109a.html.

Pollack, Andrew. "Paralyzed Man Uses Thoughts to Move a Cursor." New York Times, July 13, 2006, page A1.

Santhanam, Gopal, et al. "A high-performance brain-computer interface." Nature. (July 13, 2006) [accessed July 17, 2006]: www.nature.com/ nature/ journal/ v442/n7099/ full/ nature04968.html.

Luka · 08-25-2006, 08:29 PM

I always wanted to have Telekenisis after watching Carrie and after watching Charmed. If only... -crap emote-

Nebacudnezar · 08-26-2006, 12:16 PM

Ye i seen this on dicovery or some where like that,

He has a very small micro chip on his brain transmitting signals to the damaged part of his brain and it makes him able to control him movement again,

But i dont think we will need machiens to have telekentic abilities i think that a good few generations down the human evelutionary line we will evolve sutch abilities natureally that,s what i think my self. !

йіĝђŧмąяε · 08-26-2006, 02:42 PM

Quote:

Originally Posted by Luka

I always wanted to have Telekenisis after watching Carrie and after watching Charmed. If only... -crap emote-

i did tooooo NOT really

08-14-2006, 09:37 PM	#1
Information Not even a newbie yet... Join Date: Aug 2006 Posts: 29	Scientific Telekinesis? Paralyzed Man Controls Computer with Thoughts Joshua Dalsimer Matthew Nagle, a 25 year-old quadriplegic paralyzed by a knife wound to the neck, is the first subject in the clinical trial of BrainGate. The computer interface, built in 2003 by Massachusetts-based Cyberkinetics, is being tested at Brown University by the neuroscience department chair and Cyberkinetics founder, John Donoghue. Imagine this: you sit at one end of a room, gaze at a computer screen at the other end, and merely by focusing your thoughts, you open an e-mail, play a computer game, or make a robotic arm pick up a small ball. Surely this is the stuff of science fiction? Until recently it was, but now a team of scientists has implanted electrodes in a paralyzed man’s brain, and designed a program that translates his brain waves into instructions for a computer. More important than the sci-fi possibilities is the fact that the new research provides hope for victims of paralysis and neuromuscular conditions such as Lou Gehrig’s disease; moreover, it increases our understanding of how the brain functions. A paper describing the remarkable accomplishments was published in the July 13 issue of Nature. From Voltage to Information Neurons, the cells that make up our nervous system, register signals by creating a voltage, called an "action potential," across their membranes. Action potentials are all-or-nothing: they are either turned on or turned off. So what distinguishes one signal in the brain from another are which neurons are firing, and when. The team of neuroscientists built a 4x4 millimeter (0.16 inches to a side) sensor containing one hundred electrodes; these electrodes could potentially sense and record the electrical activity of up to 96 neurons. By combining the data from the sensor, the computer could recognize the neural firing patterns from a small region of the subject’s brain. Next they needed a subject. They chose 25-year-old Matthew Nagle, who was stabbed in the neck and became quadriplegic, or paralyzed in all four limbs, in July 2001. After using magnetic resonance imaging (MRI) to identify which part of his brain he had used to move his arms, the scientists surgically implanted the 100-electrode sensor, and began recording a small part of the activity in his brain. To train the computer to accurately and consistently understand Nagle’s intentions, the scientists used some more complicated exercises, in which Nagle was told to pretend to use his hand to follow a cursor controlled by a technician. The program the computer used was modeled on similar programs meant to interpret the activity in monkeys’ brains. The main difference is that the monkeys weren't paralyzed, so the original software compared neural patterns to actual movement. In contrast, the new program compared Nagle’s brain activity to what Nagle had been instructed to try to do. But Does it Work? Now the fun part: all the cards are in place, and it is time to test the new system. The scientists had Nagel perform seven different tasks, starting with simply moving a cursor to random targets or drawing a circle on a screen, and progressing through opening and reading an e-mail, playing the computer game "Pong," adjusting a TV set’s volume, opening and closing a robotic hand, and controlling a simple robotic arm. For each task, Nagle had to pretend to move his arm, in order to move a cursor to a desired position on a screen. So, for instance, he would have to position a cursor to the spot on the screen that says "open hand," and the robotic hand would open. The technology has not progressed sufficiently for Nagle to directly control the hand, but the fact that he can move anything by sheer willpower is pretty impressive stuff. Joshua Dalsimer LEFT: Matthew Nagle prepares to test the surgically implanted 100-electrode sensor interface. RIGHT: Nagle opens his e-mail on a computer screen (top) and manipulates the robotic hand (bottom) by using sheer willpower to move the screen cursor. Of course, nothing is perfect, and Nagle had a much harder time moving a cursor around than an able-bodied control-subject with a mouse. In the first exercise, in which he had to move a cursor to random targets, he was able to hit 73-95% of the targets, but it took him on average over 2 times as long as an able-bodied person, and sometimes much longer. In addition, Nagel had a hard time keeping the cursor still for any length of time. Perhaps more significantly, a second subject, with whom the team has begun to experiment, performed less accurately than Nagel and the scientists are trying to figure out why. Tests are being done on two more subjects, but no results have yet been reported. The Destination, not the Journey The same issue of Nature that carried the study involving Nagle also published a paper by a second team of scientists in which they proposed a different approach to interpreting neural activity. Instead of working with a paralyzed human, they conducted tests on monkeys, and achieved some promising results. Their basic idea was that instead of focusing on tracing movement towards a target, the program could focus on the target itself. These researchers, led by Dr. Gopal Santhanam of Stanford University, made use of the fact that before a monkey actually begins to move its arm, its brain has already determined the arm’s destination. The scientists would first show a monkey a target, or many targets, and only afterwards instruct the monkey to grab a target with its hand. They then used the monkey’s brain activity from the time the monkey saw the target until it began to move towards the target, called the "delay period," to predict where it was going to grab. In the same way that the first team analyzed Nagle’s neural patterns to determine how he was pretending to move his arm, the second team analyzed the monkey’s neural patterns to determine the final destination of its arm. There are two principal advantages to the second approach. The first is that since the program focuses on only one point, as opposed to an entire trajectory, there’s less estimation and hence more accuracy. The second is that the system uses only data recorded before the monkey moves its arm, so it doesn't have to wait for actual movement, and can progress much faster. Santhanam’s team, which measured brain activity with a device similar to the one used by the first team, divided the delay period into two portions: in the first, which they referred to as Tskip, the monkey processes what it sees and decides which target to hit. In the second, Tint, the brain actually prepares to move towards the target. The scientists chose to ignore Tskip, and focus on Tint to predict where the monkey would grasp. After deciding the length of Tskip to be 150 milliseconds, the issue became how long Tint should be. The longer it is, the more accurate the prediction; with a long enough time, they were able to make correct predictions up to 90% of the time. On the other hand, longer Tint’s slow down the process. To reach a happy medium between speed and accuracy, the scientists used a statistic called information transfer rate capacity, or ITRC, which measures the speed at which the computer gathers information from the monkey’s brain. They found that the ITRC is the highest, about 6.5 bits per second, when Tint is short—about 60-130 ms. An ITRC of 6.5 is more than 4 times higher than what other methods of brain-wave interpretation were able to achieve, and corresponds to typing at about 15 words per minute. Although the results are preliminary, and involve monkeys and not humans, they are surely encouraging. Down the Road Certainly, a lot more work needs to be done before either team’s findings lead to practical medical therapies. For one thing, at present all parts of the device are connected with wires; so, for instance, there were actually wires coming out of Nagle’s head. A practical device would almost certainly be wireless, both to make it more convenient and to lessen the chance of an infection. Another problem that has to be surmounted is that the electrodes start to lose their ability to measure neural signals after about six months. Neither team of scientists has been able to figure out why this is so, though Santhanam’s team believes they may have a partial solution. ShowbizIreland/Getty Images Physicist Stephen Hawking suffers from Lou Gehrig's Disease, a disease that attacks the central nervous system. With the help of a computer and voice synthesizer, Hawking is able to communicate. Currently, many paralysis victims use systems that interpret either their voice or their eye movements to perform tasks. At present, these techniques are far more accurate and practical than interpreting brain waves, and in addition do not require surgery. On the other hand, the advantage of communicating directly with the brain is that it allows the patient to look wherever he wants, or talk at the same time as he performs the task. Given these pluses, the brain-wave approach could become standard if it is further refined. Laboriously and uncertainly pushing a cursor around on a computer screen is certainly a far cry from what science fiction writers have imagined, such as a super-human bionic man who uses brain waves to completely control robotic limbs. On the other hand, the advances described in the papers mark a major step forward—and even if they never produce a super-human, they offer hope of greatly improving the lives of the paralyzed.

08-25-2006, 08:29 PM	#3
Luka John O' Scots. Join Date: Mar 2006 Location: City Of Wonder. Posts: 38,731	I always wanted to have Telekenisis after watching Carrie and after watching Charmed. If only... -crap emote- __________________ 'Cause we were both young, when I first saw you.

08-26-2006, 12:16 PM	#4
Nebacudnezar I talk too much! Join Date: Apr 2006 Location: wild west midlands :P Posts: 1,974	Ye i seen this on dicovery or some where like that, He has a very small micro chip on his brain transmitting signals to the damaged part of his brain and it makes him able to control him movement again, But i dont think we will need machiens to have telekentic abilities i think that a good few generations down the human evelutionary line we will evolve sutch abilities natureally that,s what i think my self. ! __________________ Nicked The Siggy From Blondy Uk's Thread

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

08-14-2006, 09:39 PM	#2
Information Not even a newbie yet... Join Date: Aug 2006 Posts: 29	Bibliography Biello, David. "Tiny Chip Converts Paraplegic's Thought into Action." Scientific American.com, (July 13, 2006) [accessed July 18, 2006]: www.sciam.com/ article.cfm? chanID= sa003& articleID= 0004DD3A- 6E37- 14B5- AE3783414B7F0000. Hochberg, Leigh R. et al. "Neuronal ensemble control of prosthetic devices by a human with tetraplegia." Nature, (July 13, 2006) [accessed July 17, 2006]: www.nature.com/ nature/ journal/ v442/ n7099/ full/ nature04970.html. "Is this the Bionic Man?" Nature, (July 13, 2006) [accessed July 17, 2006]: www.nature.com/ nature/ journal/ v442/ n7099/ full/ 442109a.html. Pollack, Andrew. "Paralyzed Man Uses Thoughts to Move a Cursor." New York Times, July 13, 2006, page A1. Santhanam, Gopal, et al. "A high-performance brain-computer interface." Nature. (July 13, 2006) [accessed July 17, 2006]: www.nature.com/ nature/ journal/ v442/n7099/ full/ nature04968.html.