Tag Archives: Neuroscience


Human Brain Project Needs Artificial Brains to Understand Real Ones

mail_image_preview-1-1383232402792If neuroscientist Henry Markram had a dollar for every neuron he wants to map, he still wouldn’t have enough money.

As it happens, the Swiss Federal Institute of Technology in Lausanne (EPFL) researcher has a billion euros, or $1.38 billion, from the European Union to spend over the next ten years, but the normal means of determining a neuron’s activity can cost $1 million and take a year. By the time he got through the 3000-odd pathways shown in the photograph of a pinhead-sized slice of brain behind him in a conference room last month, he’d be flat broke, decades older, and he’d still have to map countless more pinheads’ worth of neurons to understand the brain.

As Markram has been telling everyone since he got the €1 billion nod to lead the Human Brain Project, the way researchers study the brain needs to change. His approach—and it’s not the only one—stands on an emerging type of computing that he and others claim will let machines learn more like humans do. They could then offer generalizations from what’s known about a handful of neural pathways and find shortcuts to understanding the rest of the brain, he argues. The concept will rely as much on predictions of neural behavior as on experimental observations.

Yet such predictions will have to come from people until they can better train their computers to do it. So-called cognitive computing, which relies on recognizing elements of a familiar thing in new settings, is difficult to achieve through the kind of raw calculation to which most supercomputers are suited. It’s not like winning at chess or even “Jeopardy!”, two tasks IBM machines have mastered. But IBM researchers are already turning Watson, the supercomputer that beat “Jeopardy!”, into a recipe-remixing machine, and they are sure to program it for other tasks that require massive data sifting and some level of semantic analysis.

That’s the direction Markram expects computing to go for biologists, who need their computers to think more like people do. Human intelligence seems to rely on the art of the analogy, as Douglas Hofstadter writes in his new book on artificial intelligence, which James Somers explores at length in The Atlantic this month. That’s why CAPTCHAS have been so hard to defeat: the letters are easy for a computer to learn but difficult to recognize out of context. Yet we can quickly hypothesize what’s important enough about a letter to recognize it when it is distorted.

Markram is counting on those computing capabilities to improve over the course of the project, he says. He’s also counting on being able to persuade his colleagues that such computer-generated hypotheses about neural behavior will be good enough to start making higher-level hypotheses about the brain’s emergent structures. The computer they will use, an updated version of the Blue Brain project’s Swiss-owned IBM Blue Gene, will use a hybrid memory approach, which is handy for keeping massive datasets close to the processors, but it does not yet offer any special artificial intelligence.

In the meantime Markram is focusing his efforts on another kind of computing: cultural. A major element of the Human Brain Project is the ability to unite researchers around common problems and share questions, findings, and interpretations: “we’re basically layering social networking on top of neuroscience,” Markram says.

But he will have a lot of persuading to do. One colleague told The Observer; “whatever your take is on Big Neuro, do not expect them to make good on all their promises to find causes, let alone cures, for any of the big neuro diseases they list in 10 years, and as for new computing technologies? They are pulling your leg.”

To his credit, Markram did not oversell cures to diseases during a conversation with journalists earlier this month shortly after the project’s formal launch. He also gave a realistic reply to a question about whether cognitive findings from the brain project could change the way supercomputing is done: the short version is “not yet.”

Instead, the remarkable thing about the Human Brain Project may not be its computing power so much as its convening power. The social networking layer, Markram says, “is designed for tens of thousands of scientists to be able to collaboratively work on unifying all the knowledge that we have about the brain.”

First published in IEEE Spectrum’s Tech Talk blog: [html] [pdf]

Train of Thought Derailed: How an Accident Can Affect Your Brain

My cousin Guillermo Cassinello Toscano was on the train that derailed in Santiago de Compostela, Spain, last week when it went around a bend at twice the speed limit. Cassinello heard a loud vibration and then a powerful bump and then found himself surrounded by bloody bodies in wagon number nine. Shaking, he escaped the wreckage through either a door or a hole in the train—he cannot recall—then sat amid the smoke and debris next to the track and began to cry. Seventy-nine passengers died.

Cassinello doesn’t remember everything that happened to him. The same mechanisms that kept his brain sharp enough to escape immediate danger may also make it harder for him both to recall the accident, and to put the trauma behind him. “The normal thing is that the person doesn’t remember the moment of the accident or right after,” says clinical psychologist Javier Rodriguez Escobar of trauma therapy team Grupo Isis in Seville, who helped treat and study victims of the 2004 Madrid train bombings. That’s because the mind and the body enter a more alert but also more stressed state, with trade-offs that can save your life, but harm your mind’s memory-making abilities.

As the train fell over, several changes would have swept through Cassinello’s body. His adrenal glands, near his kidneys, would have released adrenaline (also known as epinephrine) into his bloodstream. The adrenaline would have directed blood to the powerful muscles of his arms and legs, where it would help him escape the wreckage faster. The hormone would have raised his heart and breathing rates. It also would have stimulated his vagus nerve, which runs from his spine to his brain. Although adrenaline cannot cross the blood–brain barrier, the vagus can promote noradrenaline production in the brain. That hormone activates the amygdala, which helps form memories.

Just the right amount of noradrenaline, researchers have found, can boost memory storage; too much can destroy it. Figuring out the balance could allow researchers to harness the hormone. Neuroscientist Christa McIntyre at the University of Texas at Dallas and colleagues have been studying how the chemical shapes memory-making in rats (her team is planning a human trial). When the team stimulated rats’ vagus nerves the animals’ memories improved. McIntyre has to keep the dose low, however, because other experiments have shown that too much noradrenaline appears to impede memory-making. Researchers are still trying to determine whether the excess noradrenaline directly causes the memory lapses or if the hormone is associated with high stress levels that cause some other chemical system to interfere. “That’s the part we don’t really understand: if there’s too much [noradrenaline] or if there’s another system that kicks in and puts a brake on it,” McIntyre says.

Cassinello’s memory lapses may be due to a noradrenaline overflow. But there may be other explanations for the gaps in his memory. His brain may have narrowed his attention at the time of the crash to only those things that matter for survival, such as escaping the train, leading him to ignore things that do not, such as whether the path out of the train passed through a door or a hole. Researchers have shown that humans report selective hearing during stressful events and that stressed people pay attention to different things than do unstressed people (pdf).

Cassinello’s uncle picked him up from the accident scene and drove him to a hospital for a checkup. Apart from a few minor scratches, he was fine. But Cassinello says he has flashbacks to the disaster. “The images of shattered people in my cabin and outside are in my head,” he says. Flashbacks are a normal part of the stress response. If Cassinello is lucky, the flashbacks will fade within weeks as he learns to suppress the bad memories cued up by triggers such as the sound of a train.

That process is called fear extinction. McIntyre and colleagues want to be able to influence it, so as to better help victims of post-traumatic stress disorder (PTSD). Scientists could activate a trauma victim’s vagus nerve, amplifying the memory-writing process while the patient practiced healthy responses to a fear-inducing stimulus. If the process works, it could speed up recovery. Other researchers are working on drug-enhanced fear extinction using chemicals such as zeta inhibitory peptide (ZIP) or D-cycloserine. Another approach, called fear reversal, aims to provoke fear-inducing memories into a malleable state, such as all memories enter when we access them, and then changing them with the help of a different drug, propranolol, which interferes with protein formation, or even with precisely timed talk therapy aimed at blocking the reconsolidation of bad memories.

One thing that is almost certain is that his memories of the event will change with time. Studies after the September 11, 2001, attacks on the World Trade Center found that New Yorkers’ reports of their experience of the attack changed over the years.

For now, survivors of traumatic experiences such as Cassinello can lean on the trauma therapists who rushed to Santiago after the crash. Some 70 to 80 percent of car accident survivors get away without PTSD, McIntyre reckons. As Rodriguez points out, however, most of those therapists are volunteers in town for a few days. It may take a few weeks or even months of therapy for patients to get past the worst of their experiences.

Read the rest of this story where it first appeared, in Scientific American: [html] [pdf]
Yahoo News also syndicated it: [html]


Avatar’s gaze illuminates social brain

gazingThey may seem a little unsettling but the staring eyes of this female avatar were designed to grab your gaze and hold it, and also to obligingly follow where you look. By performing these actions with people placed inside a brain scanner, she has helped to demonstrate that guiding the gazes of others activates different brain areas than following.

This could help unravel the brain activity underlying the process of “joint attention”, thought to be key to complex, human social interactions. It could also offer insights into why social interactions can break down for people with autism.

See the entire story and accompanying video on NewScientist.com [html] [video]


A Memorable Device

science_cover090313It was over drinks at a local pub in the spring of 2006 that cognitive psychologist Martin Conway of the University of Leeds in the United Kingdom first told his colleague Chris Moulin about using a wearable camera for memory research. But it took more than a few pints of beer to convince Moulin that SenseCam, a camera that periodically takes still photos while worn on the user’s chest, might be a game-changer in the study of what psychologists call autobiographical memory. Although skeptical of the small device’s usefulness, Moulin did finally agree to take one for a test drive.

Or rather, he took it on a test walk. Moulin regularly wore a SenseCam on a series of walks. When he reviewed the images 6 months later, to see how well his memories matched the camera’s visual record, Moulin says he experienced an unexpected feeling of “mental time travel.” One of the images triggered the memory of the song–Thom Yorke’s “Black Swan”–that was playing on his iPod when the image was taken.

Conway says that many SenseCam users likewise report a sudden flood of memories of thoughts and sensations, what he calls “Proustian moments,” when they review images taken by the device. SenseCam’s images “correspond to the nature of human memory—they’re fragmentary, they’re formed outside your conscious control, they’re visual in nature, they’re from your perspective. All these features are very like what we call episodic memory,” says Conway.

That’s why he, Moulin, and dozens of other researchers have begun to test whether the images can help resolve how the brain handles personal memories. Cognitive experiments, however, represent just one line of inquiry supported by Microsoft Research, the scientific arm of the software giant and the inventor of SenseCam. Medical researchers are also evaluating whether the device can help people with memory problems due to illness or injuries.

In 2004, Narinder Kapur and Emma Berry, neuropsychologists at Addenbrooke’s Hospital in Cambridge, U.K., were the first to use a SenseCam for memory rehabilitation work. They found that the device significantly helped Mrs. B, an elderly woman with memory problems due to brain damage from an infection. Mrs. B normally forgot events after 3 to 5 days, and even keeping a diary that she periodically reviewed helped her remember events for only about 2 weeks. But when she regularly reviewed SenseCam images of events, she could recall more details—and her memories persisted for months after she ceased reviewing the past images. Encouraged by that data, Kapur says he and Berry grew hopeful that “periodic, regular review of visual images of personal events … really does help long-term [memory] consolidation.”

They and others are getting a chance to test that hypothesis. After the pair reported the results from Mrs. B, Microsoft Research decided to provide more than $550,000 in funding to seven research groups, most of them focusing on people with memory problems, and to loan hundreds of cameras to other scientists. SenseCam has “very obvious applications in a whole range of clinical disorders,” says one of the grant recipients, psychologist Philip Barnard of the University of Cambridge.

Personal black boxes

SenseCam grew out of a Microsoft Research project that aimed to create a “black box for the human body” which would record data that doctors might find useful if a person were in an accident, says Ken Wood of Microsoft Research Cambridge. In 1999, computer scientist Lyndsay Williams, then also at the same lab, suggested adding a camera to the device so it could double as a memory aid for mundane tasks such as finding lost keys.

In 2002, Kapur heard then-Microsoft CEO Bill Gates mention the project in a talk. Because his hospital is just a few miles from Microsoft Research Cambridge, it was easy enough for him and Berry to suggest using SenseCam prototypes for patients with memory problems due to Alzheimer’s or brain injuries.

Clinicians who work with such people have typically focused on helping them with their prospective memory, i.e., remembering tasks to be completed in the future, such as keeping appointments. For this, the best aids are still simple tools such as checklists and alarm clocks. But for patients with difficulty recalling past events, clinicians have had little to offer beyond diary-keeping, a task many people, such as Mrs. B and her husband, complain is onerous.

In contrast, SenseCam records images passively, permitting a person to go about their day without interruption. The latest version is about the size and weight of a clunky mobile phone and appears to observe the world through two unmatched eyeballs. One is a passive infrared sensor, tuned to trigger the camera whenever another person passes by. The other is a wide-angle camera lens, set to capture most of the user’s field of view. The device is also equipped with an ambient light sensor that triggers the camera when its user moves from one room to another, or goes in or out of doors. The camera can also be set to snap an image if the sensors haven’t triggered a photo after an arbitrary number of seconds. A typical wearer might come home with 2000 to 3000 fragmentary, artless images at the end of a day.

It may be just those characteristics of the SenseCam images that make them so useful for memory rehabilitation and research, Kapur says. Like Conway, he suspects that the reason the images stimulate memory retrieval and possibly consolidation is because they mimic “some of the representations that we have” of past events in our brains.

To move beyond the initial case study of Mrs. B, the Addenbrooke’s team, under the direction of neuropsychologist Georgina Brown, has followed five additional people with memory problems over a nearly 3-year period, exploring the difference between the memory boost provided by visual and written diary-keeping. Establishing a baseline of how fast these people lose their memories, the team asked each about an event every other day for 2 weeks after the event, and then again after 1 month and after 3 months. Then they asked the patients to keep a diary of a separate event and review it every other day during an initial 2-week assessment, but not during subsequent months. Finally, patients reviewed their SenseCam’s images for 2 weeks following a third event.

The preliminary results suggest that SenseCam use strengthened these patients’ memories more than diary-keeping did. A full analysis of the data is in preparation, says Brown, whose team plans to submit it to the journal Memory for a special issue devoted to SenseCam research.

In a recent, separate study, Mrs. B has repeated a version of her trial, this time incorporating a brain scanner. Researchers compared the activity in her brain as she tried to remember events she had either reviewed in her written diary or with personal images from her SenseCam. Mrs. B recognized about 50% of images taken at an event she had studied using a diary, but 90% if she had studied images instead. And brain regions associated with autobiographical memory were more active when she recalled events she had studied using SenseCam images than when she recalled the diary-studied event, Berry and colleagues report online on 13 March in the Journal of Neurology, Neurosurgery and Psychiatry.

The Addenbrooke’s work represents just a few patients with varying causes of memory loss, but Berry notes that worldwide there are about 30 ongoing SenseCam studies of memory patients. Adam Zeman of the University of Exeter in the United Kingdom leads one. “I think the main interest [in SenseCam] is that it gives you an opportunity to look at memory in what you might call a more ecological fashion than laboratory stimuli generally do,” he says, and “it gives an opportunity to support and rehabilitate memory.”

Memory walks

Normally, basic research precedes clinical studies, but the history of SenseCam has been the reverse. “The initial studies had a strong pragmatic aim,” says Kapur, “but certainly once we started to collect data, [psychologists] began to look at these things from a theoretical slant.” The question for cognitive scientists is whether SenseCam, or any similar wearable, point-of-view photographic device, can illuminate how healthy autobiographical memory works. Moulin, for example, has engaged volunteers to undertake memory walks in which they read a list of words while wearing the SenseCam. His student Katalin Pauly-Takacs has tested the participants’ recall of the words on the day of their walks and then again 3 months later, with and without the help of SenseCam images. Their preliminary results suggest that volunteers remember more of the words from walks that they reviewed using SenseCam images.

Moulin’s experiment is a nod to decades of autobiographical memory research, in which volunteers were tested on their ability to recall standard images or word lists they had previously seen. Some researchers suggest that the more personal nature of SenseCam images will be key to better studying autobiographical memory storage and retrieval. “Using SenseCam we can first, have more interesting stimuli and second, test [memory] processes that can generalize more easily to real life,” explains Roberto Cabeza, a neuroscientist at Duke University in Durham, North Carolina, who is also working with the device.

Despite SenseCam’s more personal touch, there are no guarantees it will break new ground in memory research. “Whether or not it will tell us different principles or something novel is unclear,” says Larry Squire, a psychologist at the University of California, San Diego, who hasn’t yet worked with the device.

William Brewer of the University of Illinois, Urbana-Champaign, notes that nobody really knows how best to evaluate SenseCam as a memory-consolidation aid or a retrieval cue. He and his graduate student Jason Finley have tested different aspects of memory using SenseCam images as cues, asking individuals how certain they are that they’ve seen an image before, or inquiring what they did after a certain image was taken. Such baseline studies, says Brewer, should help identify the most appropriate memory tests.

In addition to the seven Microsoft Research grants handed out in 2007, dozens of groups in cognitive psychology, clinical neuropsychology, education, and computer science are conducting research with borrowed SenseCams and independent funding. But there are no current plans to commercialize the hardware or the software from the SenseCam project—a fact that puzzles some fans of the device. In fact, to keep up with the growing demand for the devices, Microsoft would like to find another manufacturer willing to mass-produce the cameras, says Wood. Microsoft currently provides the cameras to only a limited number of patients under clinical supervision.

Even though he lobbies colleagues such as Moulin to try the device, Conway remains cautious about overselling SenseCam. There is still at least a decade’s work ahead before “we can maximize its use for research and its use as an intervention scheme in helping failing memories,” says the 56-year-old investigator. “By that time, I’ll need to wear one permanently, myself.”

This feature first appeared in Science [html] [pdf]