Tag Archives: university of pittsburgh

Brain-machine interface success allows paralysed woman to feed herself for first time in a decade.

Today the Guardian newspaper has a fascinating report on how a woman named Jan Scheuermann, quadraplegic for over a decade due to a spinal  degenerative disease, was able to feed herself with the help of two intracortical microelectrode arrays that monitored her motor neuron activity and allowed her to manipulate a robotic arm and hand with unprecedented fluency and accuracy.

Commenting on her performance Professor Andrew Schwartz of the University of Pittsburgh said:

“We were blown away by how fast she was able to acquire her skill, that was completely unexpected, at the end of a good day, when she was making these beautiful movements, she was ecstatic.”

Professor Schwartz’s name may well be familiar to readers of this blog, and with good reason. Back in 2008 Professor John Stein wrote a post for this blog on a landmark study in which Professor Schwartz and his colleagues implanted identical microelectrody arrays into the brains of two  macaque monkeys, enabling them to manipulate a robotic arm feed themselves marshmallows. This successful study – itself the product of more than two decades of – led directly to the clinical study reported today.

It’s worth remembering that this technology will require further refinement before it is ready for wider use outside of a laboratory setting, but this study shows what can be achieved and provides the proof-of-principle necessary to encourage further investment in this approach.  It is also worth noting that it is far from the only game in town, last year we reported on Prof. Schwartz’s teams success using another method known as electrocorticography to enable a man named Tim Hemmes to control a robotic arm, though with somewhat less dexterity than that reported today, and this year we have also taken a look at the very promising results obtained through the use of spinal electrostimulation and olfactory ensheathing cell transplant to overcome paralysis. Not all these approaches will be appropriate for all patients, and ultimately they may be combined in some cases, but they do provide strong evidence that after decades of hard work and important discoveries in laboratories around the world, neuroscience is now poised to transform the treatment of spinal injuries.

Speaking of Research





ERV blogs on GMO Herpes vs severe cancer pain

As gene therapy emerges as one of the hottest areas of medical research, one thing that is striking is how it employs viruses – sometimes very nasty viruses – to deliver the gene to where it is needed in the human body.

Yesterday virologist Abbie Smith discussed another excellent example of this on the ERV blog in a post entitled “GMO Herpes vs. severs cancer pain”, describing how scientists at the Universities of Michigen and Pittsburgh have used a genetically modified herpes virus to deliver the preproenkephalin gene – which produced a precursor to pain-killing opiates – to the nerve cells of terminal cancer patients who were suffering from severe pain.

Abbie remarks that “This was one of the most depressing, yet hopeful, papers I have ever read.”. It’s difficult to disagree, after all most of the patients participating in the trial died within 3 months of it starting. But to focus on this sobering statistic would miss the reason for this study, namely that the pain-relief available to patients with severe chronic pain is often inadequate, as the drugs are not specific enough and cause unacceptable side effects when used at the high doses often required for prolonged periods of time. By targeting the opiate molecules to the nerve ccells themselves these side effects can be avoided, and more effective pain relief provided.

The paper “Gene Therapy for Pain: Results of a Phase I Clinical Trial” is available for anyone to read in PubMed Central and makes it very clear that this is a therapy that was discovered, evaluated and refined in animal models of different types of pain before entering this first clinical trial. The first two paragraphs of the introduction noting that:

A significant limitation to the development of analgesic drugs is that off-target effects at doses below the maximal analgesic threshold restrict the ability to selectively interrupt nociceptive neurotransmission1. To address this limitation, we developed a series of replication defective HSV-based vectors to deliver gene expression cassettes directly to DRG neurons from skin inoculation 2, 3. The anatomically defined projection of DRG axons allows targeting of specific ganglia by injection into selected dermatomes. In preclinical studies, the release of anti-nociceptive peptides or inhibitory neurotransmitters in spinal dorsal horn from the central terminals of transduced DRG neurons effectively reduced pain-related behaviors in rodent models of inflammatory pain, neuropathic pain, and pain caused by cancer4-9.

The human PENK gene encodes for preproenkephalin, a precursor protein proteolytically cleaved to produce the endogenous opioid peptides met- and leu-enkephalin. In the spinal cord, enkephalin peptides inhibit pain signaling through actions at presynaptic opioid receptors located on central terminals of primary afferent nociceptors and postsynaptic opioid receptors on second order neurons involved in nociceptive neurotransmission10. HSV vectors expressing opioid peptides appear to be particularly effective in animal models of inflammatory and cancer pain4, 5, 8.”

And in the conclusion:

In preclinical animal studies, skin inoculation of HSV vectors expressing PENK reduce acute hyperalgesic responses27, and reduce pain-related behaviors in models of arthritis28, formalin injection4, peripheral nerve damage6 and bone cancer5. Because this was the first human trial employing HSV vectors to achieve gene transfer, we elected to carry out the phase 1 clinical trial for safety and dose-finding in patients with pain caused by cancer…This Phase I clinical trial primarily addressed the question of whether intradermal delivery of NP2 to skin would prove to be safe and well tolerated by subjects. The small number of patients and the absence of placebo controls warrant circumspect interpretation of the secondary outcome measures. But the observation that subjects in the low dose cohort had little change in the NRS or SF-MPQ while subjects in the higher dose cohorts reported substantial reduction in NRS and improvement in SF-MPQ is encouraging.”

Encouraging is possibly an understatement, seeing clear evidence of therapeutic benefits in a Phase I trial like this is very promising, or as Abbie puts it “A trial turning out this successful is a great starting point for optimizing this kind of therapy.”.

Paul Browne

p.s. Those interested in a more detailed account of the research that led to this clinical trial can find it in this review published in 2008 and available to read online for free.

Tom talks nerdy to Cara Santa Maria about monkeys, prosthetic hands and brain machine interfaces.

Speaking of Research founder Tom Holder was  recently interviewed by the Huffington Post’s new science correspondent Cara Santa Maria for her blog “Talk Nerdy To Me” .

In her latest post Cara examines whether research performed on monkeys by a Chinese group with the aim of developing improved brain-machine interface technology to control a prosthetic hand is justifiable.

Vodpod videos no longer available.

It is worth noting that in addition to preventing the monkey from pulling the wires out of the electrodes by accident, the restraint chairs – in which the monkeys are only kept for short periods – also prevent the monkey from simply reaching out and grabbing the juice, obliging it to use its brain instead.

This is field of research we have discussed on several occasions since Speaking of Research was founded, most recently in a post last October when we took a look at a successful early clinical trial of a brain machine interface developed through research in monkeys by scientists at the University of Pittsburgh, which allowed a paralyzed man to control a robotic arm.

We also discussed research being undertaken at Duke University , where scientists are developing a system that they hope will allow patients to feel what their prosthetic limb is touching, allowing for much finer control and dexterity. The electrodes implanted in the brains of the human patients are essentially the same as those used in the monkey studies, and they are painless once implanted, and are implanted under anesthesia – general anesthesia for monkeys but usually local anesthetic  for humans (so the patient can help position the implant).

A paralysed man touches his girlfriend’s hand…thanks to animal research.

Earlier this year we reported that scientists at the University of Pittsburgh had launched clinical trials of two different brain implant systems, known as brain machine interfaces,  that aim to give quadriplegic patients control over a prosthetic limb. At the time we noted that this technology was built on years of basic and translational research in animals, with research on monkeys playing an especially important role.

Now the Pittsburgh Post-Gazette reports on the first success of these trials. An Electrocorticographic implant enabled Tim Hemmes, who has been paralyzed below the neck since a motorcycle accident 7 years ago, to control a robotic arm with great precision, just as had been predicted from the studies in monkeys.  Being able to gently touch his girlfriends hand was a very emotional moment for Tim Hemmes, but it could hardly have been much less emotional for the team of scientists and physicians who developed the implant.

Tim Hemmes reaches out. Associated Press.

It is exciting news, and one that will spur further research in this fast-moving area of research. One drawback in the brain machine interface technology that is being evaluated in the current series of trials at the University of Pittsburgh is that the system does not include sensory feedback – the person using it cannot feel what they are touching and must rely on sight alone to guide their movements.

A solution to this problem may be close, earlier this month a team led by Professor Miguel Nicholelis of Duke University reported that they have developed a brain machine interface that uses electrical signals sent directly to the brain to enable monkeys to “feel” what a virtual arm is touching, and then control the movement of that virtual arm in response to the sensation. It is an important advance, and again one that depended on years of careful animal research to identify the correct parameters for the electrical signals used, and the optimal location for implanting the brain machine interface.

Human trials of the brain machine interface system developed by Prof. Nicholelis and his colleagues are expected to begin within the next few years.

The development of brain machine interface technology to this point, where it is offering the hope of greater independence and mobility to thousands of quadriplegic patients, is a great achievement of animal research.  In order to ensure that this technology, and many others at the cutting edge of medical science, to fulfill their potential we must continue to support the use of animals in biomedical research.

Overcoming paralysis: From Monkey to Man at the University of Pittsburgh

On Friday the New York Times reported that scientists at the University of Pittsburgh are ready to start clinical trials of two different brain implant systems, known as brain machine interfaces,  that aim to give quadriplegic patients control over a prosthetic limb.

In the main project a team led by Professor Andrew Schwartz and Professor Michael Boninger will, over the next two years,  place two sensors, each of which consists of an array of 100 electrodes that record the activity of about 50 nerve cells, just beneath the skulls of three patients. The signals collected from these sensors should allow the patients to control the movement of a prosthetic arm and hand.  In 2008 Professor John Stein wrote an article for Speaking of Research on the monkey studies that Prof. Schwartz performed while developing these sensors.  In these studies the monkeys displayed a finer degree of dexterity in manipulating a robotic arm than the scientists had anticipated, and learned to use the robotic arm surprisingly quickly, suggesting that paralysis victims may also be able to learn to use the prosthetic arm in a relatively short space of time.

The smaller project uses an alternative approach called electrocorticography, which also used a sensor implanted under the skull, but measures the activity of populations of nerve cells rather than individual neurons. This technique has the advantage of being less invasive than the electrodes that need to make direct contact with neurons, as the risk of infection is reduced when the protective meninges are not penetrated during implantation of the sensor.  Although it was previously thought to be a less precise approach than the direct measurment of single neuron activity, a  recent monkey study has demonstrated that brain machine interfaces based on electrocorticogram sensors can rival the performance of sensors that measure neuron activity directly, this and other studies have prompted the clinical evaluation of this approach.

Interesting  as these trials are, they represent only a few of the technologies being developed to treat paralysis, other techniques we have examined in recent years include neuroprosthetic devices that bridge severed spinal cords, stem cells, and therapies that encourage the regrowth and repair of damaged nerve tissue. Much of this research is still at a relatively early stage, but it is exciting to see that these techniques are starting to move from the bench to the bedside.

Paul Browne

Returning control to paralyzed limbs one nerve at a time.

A few months ago we reported on a fascinating study undertaken by Andy Schwartz and colleagues at the University of Pittsburgh, who developed a brain-machine interface that when implanted into the motor cortex, the part of the brain responsible for controlling voluntary muscle movements,  of monkeys allowed then to control a robot arm with surprising precision.  This week Chet Moritz and colleagues at Washington National Primate Research Center have published another exciting paper  (1) online in the journal Nature that describes an alternative approach to the use of brain-machine interfaces to overcome paralysis.

Rather than use an implant that monitors the activity of groups of nerve cells in the motor cortex and then use complex algorithms to decode this activity and calculate appropriate control signals for external devices, the approach used by the University of Pittsburg group, Chet Moritz and colleagues used a brain implant that could detect the activity of a single nerve cell and then home in on it and measure its activity. These implants were placed in the part of the monkey motor cortex responsible for controlling the wrist muscle and used implanted wires to directly stimulate the wrist muscles using a technique known as functional electrical stimulation (FES).  Monkeys whose wrist muscles had been temporarily paralyzed by injection of anaesthetic to the nerves that control them, quickly learned to control their wrist muscles again using the brain implant-FES system. The wrist muscle movements in turn controlled the location of a cursor on a screen, and by moving the cursor to particular locations in a screen the monkeys could gain rewards in the form of a tasty snack. More surprisingly the scientists found that monkeys could also learn to use motor cortex neurons that were not normally involved in controlling the wrist muscles to control the wrist muscles.

This research has caught the attention of the mainstream press, and it’s good to see that the welcome it has received is accompanied by cautionary notes.  There’s no doubt that this is a significant advance, the Washington National Primate Research Center team have shown that a relatively simple device can be used to restore control to paralyzed muscles, but they have so far only demonstrated control of one muscle group whereas a useful limb will require the simultaneous and accurate control of many muscle groups by many nerve cells.  I’m optimistic that this won’t be as much of a problem as it may initially appear since this study, the previous work at the University of Pittsburgh, and indeed the frequently observed ability of patients with brain damage to recover lost functions, all demonstrate that the brain is surprisingly adaptable. Whether using individual nerve cells to control muscles or groups of nerve cells to control robots will prove must useful in the clinic several years down the line is impossible to say right now. It’s quite likely that elements of both techniques will be used in future systems and that he decision as to which approach should be used in an individual paralysis patient will be determined by the nature of the injury and duration of subsequent paralysis.
Several scientists involved in this work have also stressed the importance of sensory feedback, the ability of a patient to “feel” what a paralyzed or robotic limb is doing, and this is an area under investigation by several research groups that will no doubt see further advances in the coming years.  Even without the ability to feel objects, and consequently the ability to more precisely manipulate objects, I’m of the opinion that the ability to use a robotic arm, or even a patients own arm,  has the potential to greatly increase the independence of paralysis patients. For that reason I expect that we will see this technology in the clinic sooner than many people think, and will be a therapeutic advance that many paralysis patients will welcome.


Paul Browne

1) Moritz C.T., Perlmutter S.I. and Fetz E.E. “Direct control of paralyzed muscles by cortical neurons” Nature. 2008 October 15. DOI: 10.1038/nature07418 [Epub ahead of print]

Monkeys, Robots and the University of Pittsburgh – New hope for paralysis victims?

The day after Tom Holder spoke at the University of Pittsburgh about the importance of animal research, more news is coming from this academic institution.

Every 45 seconds someone in the US gets a stroke, many are left paralyzed, furthermore 14,000 people every year suffer spinal cord injuries which may also result in paralysis. There is therefore huge incentive to learn how to bypass the damaged parts of the brain by a brain-machine interface so that the patients can regain effective movements that would be a huge help in their daily lives. It is not surprising that there has been widespread press coverage of a study published online in Nature yesterday (1) that signalled a major breakthrough in this field.


Using 2 rhesus monkeys Andy Schwartz and his team at the University of Pittsburg have made a huge advance towards that aim. They trained the monkeys to use their own motor cortical activity to control a mechanized arm to feed themselves. The team extracted the control signal from recording from about 50 nerve cells in the animals’ motor cortex. This was far fewer neurons than many researchers thought would be necessary, an important discovery in itself that should make it a little easier to design electrode implants in future. Once the monkeys got used to the system they soon became astonishingly fluid, skilled and expert in moving the robot arm just by altering the firing of their motor cortical neurones. They even learnt to take advantage of the marsh mallows sticking to the robot fingers to speed its delivery to their mouths. Even though clinical use for people with disabilities is still years away because the arm requires large computers, bulky equipment and a full time technician, and the brain-implanted electrodes would not last a lifetime and lack touch feedback from the arm, Schwartz’s achievement is phenomenal and a huge leap towards helping all those people with paralysis.

It is important to emphasise that this work could not have taken place without many years of animal experiments, with monkeys playing a key role (2,3). Andy has been working with monkeys trained to make movements designed to reveal how the motor cortex works for some 20 years. Only monkeys have the kind of control over their hands that we have, so only using monkeys could he work out the kind of control signals that they use to feed themselves. 20 years of monkey experiments (only using 1 or 2 a year) allowed him to ‘take the system to pieces’ and work out how the motor cortical cells control the arm. Obviously these experiments couldn’t be done on humans, they are simply too risky at this early stage in the development of the technology, but now he’s elucidated the control circuitry it will not be long before they’ll be applied to benefiting paralysed humans.

Kind regards,

John Stein

Professor of Physiology, Magdalen College, Oxford University
1) Velliste M. et al. “Cortical control of a prosthetic arm for self-feeding” Nature. 2008 May 28. [Epub ahead of print]

2) Lebedev M.A. and Nicolelis M.A.L. “Brain–machine interfaces: past, present and future” Trends in Neurosciences Volume 29, Issue 9, Pages 536-546 (2006)

3) Schwartz A.B. et al. “Brain-Controlled Interfaces: Movement Restoration with Neural Prosthetics” Neuron Volume 52, Issue 1, Pages 205-220 (2006)