Bench Press

The “2.0” moniker is often abused to hype up connections with the new dynamic web applications of today. This is why I’ve co-opted it to describe a new application which is to WebMD’s 1.0 as Gmail/Twitter are to old-school webmail/message boards.

As many of you know, WebMD is a leading health information internet portal, providing a wide range of medical information for both casual patients who browse the site and medical professionals. But, while the information WebMD provides is rich and valuable, it is still a static website, valuable, but not dynamic or intuitive or mobile.

Wolfram Research (maker of popular computational algebra system/scientific computing software Mathematica)’s new WolframAlpha search engine attacks the first two challenges. One of the fundamental problems with traditional portals like WebMD is that to find specific information, the user needs to have some idea of where the information is or how it’s stored/used in conjunction with other information – in other words, it requires contextual knowledge. For example, to figure out how to determine if someone is overweight, the user needs to know:

1. That BMI is the relevant metric
2. Where to find how to calculate BMI and where/how to get the information that goes into the BMI formula
3. Compare BMI with relevant comparisons
4. Understand the limitations and implications of the BMI metric

WolframAlpha tries to simplify this by reducing the dependence of the quality of the search results on contextual knowledge. In the BMI example above, WolframAlpha hasn’t quite solved how to easily and quickly answer steps 1 and 4, but it has made it much easier to do steps 2 and 3. For instance, the BMI of a person who is 5’ 10” and weighs 165 lbs and the relevant comparisons (to the US population as a whole, and to clinical definitions of “overweight”, etc.) What’s incredible is that WolframAlpha won’t only provide you with the data, if it’s feasible, it will even provide you with charts and graphs to illustrate them. While this is a far cry from Jeopardy-playing supercomputers, it is a much welcomed change to the current heavily context-dependent search paradigm.

And, WolframAlpha does a lot more than that. On their sample page of examples of health/medical requests, WolframAlpha breaks out several other capabilities:

• find growth charts/body statistics
• compute calories burned
• figure out Blood Alcohol Content levels
• estimate risks/mortality rates for diseases
• look up diagnoses by ICD-9 code
• understand medical test results
• look up information about drugs, hospitals, and public health information

Very cool, and, despite the more technical bent to information, much more usable than a standard search engine query for finding relevant health information. There is great promise in this technology, especially if it’s natural language processing algorithms improve to the point where it can provide useful information for healthcare professionals or curious/nervous patients (perhaps by combing through/organizing the information in WebMD’s healthcare-professional-oriented).

(Image Credits)

A few weeks ago I wrote a post about the development of the Standoff Patient Triage Tool, an impressive use of lasers in order to make health critical readings of patients from a distance. Well one of the best things about science is that many people can utilize the same tools to come up with unique methods and solutions for any given problem. In this case, researchers have used lasers to develop a technology that could someday revolutionize imaging procedures in medicine.

Photoacoustic imaging of melanoma in vivo.

Photoacoustic tomography is the basis behind a new imaging technology being developed in hopes of providing more flexible and cost effective devices for physicians. The technique takes advantage of ultrasonic emissions produced when a non-ionizing laser pulse is directed towards a tissue. The emissions, resulting from transient thermoelastic expansion of the target tissue due to absorption of the laser energy, are detected and analyzed with various algorithms to construct an image (2D or 3D) of the targeted area. This differs from the reliance on the doppler shift produced by the reflected laser beam in the SPTT.

Images of vasculature like the one seen on the right can be produced by using photoacoustic tomography without the injection of contrast as differences between the molecular composition of the target can be used instead. In the example to the right, the difference between oxygenated and deoxygenated blood is an effective natural contrast. Photoacoustic tomography also presents other benefits over traditional imaging techniques as explained by The Economist:

CT scans also involve potentially harmful ionising radiation. And MRI and CT scans are very expensive, using machines that cost millions of dollars and require dedicated staff to operate them. Photoacoustic tomography, by contrast, could eventually be performed using portable hand-held devices, similar to those used for ultrasound scanning. This would allow doctors to diagnose and monitor patients in clinics, and reduce the need to refer them to consultants.

The adaptability of this nascent technology is also impressive as researchers are already looking at using it to detect specific ailments such as brain lesions and cancer. In the case of cancer, the ability to accurately image vasculature could allow doctors to monitor patients for the development of new blood vessels (angiogenesis) a hallmark of cancer development.

While there are some issues to work out with this new technique, such as the lack of imaging depth (ultrasound signal emitted is reduced the deeper the tissue lies) and ultrasound distortion from varying tissue types within the human body (e.g. bone vs muscle), photoacoustic imaging is a very promising new technology.

(Source)

A lot of the “hot” research occurs at the edges of disciplines. In my humble opinion, this is a natural effect stemming from the fact that the research that occurs at those edges typically involves the use of new equipment/techniques (e.g. applying physical techniques to biology) or because of the need for experts in different fields to come together and exchange thoughts and perspectives (e.g. astrophysics requires an appreciation of the very large [traditional astronomy], the very small [quantum mechanics, statistical mechanics], and the very weird [relativity, ok so it’s not that weird, but from the perspective of someone who’s never moved at relativistic velocities, I think it’s weird]).

But, of course, interdisciplinary research, has its limits (HT: Abstruse Goose):

Anyone else have any equally bad ideas for “interdisciplinary research”?

(Image Credit – Abstruse Goose)

Computers can do some pretty incredible stuff. We’ve seen computers analyze evolutionary trees, master the game of Go, and even take on the challenge of Jeopardy!. However, as amazing as computers are, the one glaring deficit of computers is its inability to extrapolate and make connections, something that comes easy to humans. The culprit is how computers are designed. Computers are built to follow algorithms: a strict set of guidelines which, when executed faithfully, will yield the correct answer. However, ask a computer to build upon these algorithms is a whole different issue. Humans, on the other hand, quickly learn to develop schemas and a general model of understanding. From these basic concepts, we can use deduction to answer a question we’ve never been asked before by combining different skills together to achieve this goal. For example, humans will quickly identify that a certain object on a street is a car, even if we’ve never seen that certain model of car before.

Enter Cognitive-Level Annotation Using Latent Statistical Structure (CLASS), a project that is trying to push computers to recognize specific classes of objects the way humans can. Luc Van Gool of Belgium’s Leuven University (KUL), a member of the CLASS team, sees great promise in CLASS and has already found a marketable use for it in cell phones. Through a mobile service, CLASS will be able to let its users take a picture of an object, say a famous monument, identify what it is, and provide further information to the user via the internet.

“It’s like the object itself becomes the link to further information,” observes Van Gool. He expects the application of this technology to expand rapidly. For instance, cities and museums may offer interactive guided tours or guide books through kooaba.

While CLASS is still a long way off from being complete, it is clearly a giant step towards extending the capabilities of computers and furthering the field of artificial intelligence.

Like my fellow Bench Press blogger Ben, I’m a fairly avid Star Trek fan. Having spent many hours during my formative years watching syndicated episodes I always wondered if we’d ever have some of the amazing devices in the show. One of the devices that captured my imagination was the Tricorder (pictured right). It amazed me that such a tiny device could provide so much utility throughout the show. The Tricorder’s versatility allowed it to do pretty much anything in the show, but what I always remembered was it’s medical utility. It allowed characters like “the Doctor” to analyze any number of ailments quickly and accurately.

While we don’t have the technology to make a Tricorder as effective as the ones used in Star Trek, the U.S. Department of Homeland Security’s Science and Technology Directorate (S&T) is developing a tool called the Standoff Patient Triage Tool (SPTT). The purpose of the SPTT is to aid first responders at a disaster triage patients quickly and accurately. Triaging patients with traditional methods can take 3-5 minutes per person. This can become an extremely difficult and time intensive task during a disaster, exactly the opposite of what we’d like. Therefore, the goal of the SPTT is to reduce triage time to 30 seconds per patient by providing accurate readings of pulse, body temperature, and respiration from up to twenty paces away all in a portable package about the size of a legal notebook.

A drawing of the proposed 15 inch by 8.5 inch x 6 inch Standoff Patient Triage Tool (SPTT) with the following features. 1) 4 x 6 display window; 2) Control button; 3) Infrared camera window; 4) Visible camera window; 5) Ranging subassembly window; 6) Shock bumpers. (Source: DHS S&T)

The SPTT takes advantage of Laser Doppler Vibrometry (LDV) in conjunction with a visual and infrared camera to make it’s readings. LDV takes advantage of the doppler shift produced when a laser bounces off a moving target. The shift in frequency upon return of the laser beam is measured and then analyzed in order to determine the velocity over time of the target. In this case, the SPTT’s vibrometer detects the movement of blood vessels and utilizes algorithms to extrapolate relevant data. So far researchers have found that the SPTT can produce strong readings from the head, chest, abdomen and foot. Currently, taking readings from the cartoid artery region of the neck appears to be the best option. Further testing needs to be done on patients in awkward positions, as well as with differing layers of clothing.

While the SPTT can’t do everything a tricorder can do, it appears to be taking a great first step in providing a portable device capable of providing first responders with accurate job critical data that will help them save lives. Maybe it’s only a matter of time before doctors start waving a small device around the patients while asking them what brings them into the office that day.

(Image Source)(DHS S&T Press Release)

On Bench Press, we usually discuss technologies to empower scientists to do research or to better communicate with each other and with the public. But, technologies can also help the “casual” fan appreciate science as well. Case in point: Google recently announced a new application available for Android phones called Google Sky Map. The concept is quite ingenious. If ancient navigators used starcharts and compasses to determine their location, why can’t a phone that knows its location (via built-in GPS) and its direction (via magnetometer and accelerometer) be used to figure out what stars are up above?

Or in other words, Google took this physical setup (very cute setup from Google for how Google Sky Map works) of what is basically a sextant, compass, and calendar:

And replaced it with a smartphone:

I haven’t had a chance to play with it (as I don’t own a G1), but the application looks very nicely packaged. You can use it to figure out what stars are right above you. But, the coolest feature is the ability to use the application to point you in the direction of something you’re interested in seeing (e.g. the planet Saturn).

More details are in the video below. You can get Sky Map for Android through the Android Market. For anyone who’s tried it, let us know what you think in the comments!

(Image credit - Google)

The N-body problem is a classic question in physics and mathematics. It is interesting, because while it is very simple to state, it results in a wide array of interesting behavior and very deep insights into how the universe works.

The setup is as follows: if we have N objects in space, with N masses, initial positions, and speeds, where and how fast will each of those N objects be moving if they are subject only to each other’s gravity (and the laws of classical physics – e.g. Newton’s Laws of Motion, etc)?

For a situation with 2 objects (N=2) and some of the cases with 3 objects (N=3), the problem has been well-discussed (the former leads to elliptical orbits, the latter results in the dynamics of the moon orbiting around the earth while subject to the sun’s gravity). But for problems more complex, the solution is anything but pretty. Behavior can emerge which may seem regular and predictable but degenerate into pure chaos. In fact, Poincare’s study of the N=3 problem eventually served as the foundation for the study of chaos theory, or the rise of chaotic, seemingly random behavior (like turbulence) out of “orderly” equations and behavior.

But, the challenge of studying these problems and formulating/testing hypotheses becomes increasingly more challenging as N increases. After all, how do you test an idea for how the N body problem plays out when the problem is that its so difficult to figure out how the N body problem plays out?

Enter the age of the computer simulation. Mathematicians/physicists/astronomers now have the tools to test their ideas (or just pass the time watching simulations) using computers to simulate the behavior of complex systems of N bodies.

But, the fun doesn’t stop with just a handful, or even hundreds, of particles. For an N-body simulation to be sufficient for as astronomer trying to study the structure of the universe, it would have to be scaled up to model billions of particles over distances in the billions of light years.

These super simulations are staggering to comprehend. The Millennium Simulation, run in 2005 to test our understanding of quasars and dark matter, simulated ~10 billion particles (with each “particle” representing a mass about a billion times larger than our sun) across a mind-boggling 8 trillion quadrillion cubic light years expanse of space containing over 20 million galaxies.

But even the Millennium simulation is mere child’s play compared to Project Horizon, which aimed to model “half the observable universe with enough resolution to describe a Milky Way-like galaxy with more than 100 dark matter particles”.

And, probably, this is merely the tip of the iceberg for what such super simulations may be capable of. How the future of this will shape out, I believe, is dependent on the following N=3 (pun intended) questions:

1. It is relatively simple (although computationally challenging) to create a toy simulation to validate or disprove one’s theory. It is much more challenging to build a simulation which can reveal testable/actionable (e.g. not simply if one’s theory is right or wrong) conclusions to further our understanding of the system of interest. For example, will future models of the human metabolome reveal genes or regulatory pathways of interest, or are they simply to validate existing models (something which may bias the model design away from revealing more interesting behavior)? Answering this question in the affirmative is essential, or else these simulations becomes vanity toys – something to use up research funds on without driving real value for the underlying science.
2. Will Moore’s Law/alternative processors/computer science keep up with demands for greater precision and computational power? Otherwise, these super simulations will hit constraints like energy consumption or the introduction of calculation error.
3. Will scientists become more adept at communicating the value of these models? This is potentially the most important question as the investment necessary to run these simulations are significant, not only in terms of cost, but in terms of manpower and energy consumption. With the current financial crisis and a potential future energy crisis, demonstrating value to the public and to more “traditional” scientists/doctors/engineers who rely on non-computational techniques will become more and more important for these future simulations to survive.

For now, I’ll leave you with a video summarizing the gorgeous results of the Millennium simulation:

During the past month’s global swine flu emergency providing health care professionals around the world with accurate information was critical to understanding and potentially containing the outbreak. The time sensitive nature of dealing with an emerging disease highlighted the importance of developing an effective communication channel that is quick, accurate, and accessible by numerous individuals. Traditional paper distribution channels, mailing notifications to primary care physicians, can delay the receipt of time sensitive materials by 72-96 hours. Thus, the question becomes how do you design a system that can be accessed quickly and easily by a maximum number of health care professionals, while still providing quality information.

A new web application developed by the Indiana Health Information Exchange (IHIE) interfacing with a service called Docs4Docs®, provided by the Regenstrief Institute, appears to have answered that question. The IHIE’s web portal allows electronic communication of public health messages to any registered health care provider. Registered users can also send messages back through the portal to be disseminated across the network. The web portal’s simplicity allows it to bridge the gap between paper-based and electronic-based medical offices thereby ensuring that even doctors in rural areas without advanced IT infrastructure can receive and contribute critical information.

Docs4Docs® also leverages the Indiana Network for Patient Care (INPC) which is a secure community health records system, providing patient data whenever needed. Dr. Shaun Grannis, a Regenstrief researcher, explains “[b]y working with our public health partners to seamlessly deliver public health alerts in precisely the same manner that physicians receive time-sensitive clinical information for patient care, we ensure that physicians have the right information at the time they need to see it”. This was exemplified by the first electronic health alert sent out across the Docs4Docs® network with regards to the emerging H1N1 crisis on April 29, 2009.

Last year Regenstrief scientists received a $10 million, 5 year contract from the Centers for Disease Control in order to continue working on developing electronic records and notification systems like those that make up the backbone of the Docs4Docs® service in Indiana. I for one believe that money is going to good use and look forward to seeing other states follow Indiana’s lead with regards to developing new electronic records and notification systems.

When most people think of laboratory/medical equipment, they think of massive machines full of sophisticated electronics and gear. But, thanks to Moore’s Law (which helps electronics get smaller, cheaper, and more power efficient), equipment that formerly required massive machinery, may be duplicated in the form of handheld devices, like this USB ultrasound gear from Laborie Medical Technologies (FDA approved):

LMT bundles software with their ultrasound probes to deliver images on PCs running Windows XP, but to use it on a phone you’ll need software from William Richard’s group from Washington University in St. Louis, who, with funding from Microsoft, have created a client allowing you to access ultrasound images on Windows Mobile devices! They’ve even released an SDK to help other enterprising researchers create other applications which can make use of these portable UltraSound devices and have them work on any Windows Mobile phone with a USB interface!

Such gear could bring ultrasound access to countries or regions lacking significant healthcare infrastructure, and similar devices could dramatically change how biomedical research is conducted.

For more information, read the presentation that David M. Zar gave at the Medical Records Institute’s TEPR+ (Towards the Electronic Patient Record) conference, and check out the live video demonstration as well as the UltraSoundUSB page on YouTube:

Let’s hope this is only the first in a long line of portable electronics interfacing with readily available mobile phone technology.

(Image credit – The Daily What)

Introducing today’s contestants…the IBM QA system Watson? That’s right folks. Our friends at IBM, not content with simply creating a supercomputer capable of defeating humans at Go, have taken it a step further and are currently creating a supercomputer (codenamed Watson) with the goal of it having the ability to beat humans at a game of Jeopardy!. [IBM Video on Watson at the bottom]

The interesting thing about this particular problem is, unlike with games of Go and Chess which have clearly defined rules and discrete moves/outcomes, playing a game of Jeopardy requires an understanding of semantics which has traditionally been relegated to the human domain.

Admittedly, there are natural language processing solutions out there. But at the end of the day, we’re a long way off from the computers displayed in Star Trek which can:

• Understand spoken words – This is a very challenging problem. How do you instruct a computer to not just comprehend words, but comprehend actual meaning to those words (semantics). The ability to understand that the “can” in “I can do it” is very different from the “can” in “soda can”, or that “being on pins and needles” is just an expression, or to even understand when a sentence is a question versus a statement are very deep problems. But this is only the beginning of Watson’s challenges, for Watson must also be able to…
• Search a massive database for relevant information – Merely searching a database for a list of possible results is a tractable problem that many database/search engines have already solved (e.g. searching for “Indian economy” on Google’s search engine). Searching a large database to find a particular answer behind the reams of data is much harder (e.g. understanding that “Economic Output” can be measured by a country’s GDP).
• Understand the relevant information – Just as it’s harder to understand Quantum Theory than it is to merely read the papers, IBM’s Watson must be able to parse the information that it’s found from its database. For instance, if asked to compare India’s economic output to its neighbors, a computer must not only understand that economic output is GDP, it must also understand what “neighbors” means in the context of India, understand that GDP may be “real” or “nominal” and may need to be adjusted by currency, and understand what it means to “compare” GDP’s.
• Formulate a response – This is related to the first ask, but is more challenging. Just as its harder to memorize the Bible than it is to recognize specific passages, IBM’s Watson must do more than just recognize/understand words – it must be able to create its own sentences which use the relevant information and understanding its developed.

The task is challenging, but not impossible. Already, researchers have demonstrated computers which have been able to do the scientific method (hypothesize –> experiment/test –> analyze –> formulate new hypotheses) all on their own. Granted, the scientific problem explored was more systematic in nature (and had a more well-defined solution set than a game of Jeopardy) as it was focused on finding missing pieces in metabolic networks, but the fact that a computer was capable of performing basic high level logic is very promising for fields of research (although threatening to lab techs and uncreative grad students everywhere) which were formerly intractable due to their scope (e.g. mapping out the human proteome or transcriptome).

“The essence of making decisions is recognizing patterns in vast amounts of data, sorting through choices and options, and responding quickly and accurately,” said Samuel J. Palmisano, IBM Chairman, President and Chief Executive Officer. “Watson is a compelling example of how the planet—companies, industries, cities—is becoming smarter. With advanced computing power and deep analytics, we can infuse business and societal systems with intelligence. This project is the latest example of IBM’s longstanding commitment to fundamental research and to overcoming ‘grand challenges’ in science and technology.”

Although I don’t know how well Watson would fare against Ken Jennings, Watson’s completion would be a landmark in artificial intelligence. It’ll be interesting to see if IBM’s Watson does as well as IBM promises. Although I don’t know how well Watson would fare against Ken Jennings, Watson’s completion would be a landmark in artificial intelligence and computer science. Watson may pave the way to an age where computers can actively aid doctors diagnose patients or help business executives make financial decisions (which is probably what IBM is going for here).

(Image Credit) (Video)