• Blog Stats

    • 61,643 hits
  • Categories

  • Archives

  • Advertisements

Big Medical Data

MIT News (01/25/13) Larry Hardesty

Last year the Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Laboratory (CSAIL) launched bigdata@csail, a big data initiative that includes several projects designed to make medical data more accessible to physicians and patients.  For example, researchers in the Clinical Decision Making Group are developing methods for bringing artificial intelligence to the medical community.  The group participates in a large initiative to create a database system that would link genomic and clinical data so that doctors can more easily test hypotheses about connections between genetic variations and certain diseases.  The group recently presented a new approach to the problem of word-sense disambiguation, or inferring from context which of a word’s several meanings is intended.  Meanwhile, researchers in CSAIL’s Data-Driven Medicine Group are investigating techniques for detecting and predicting hospital-borne infections.  In addition, researchers in the New Media Medicine Group are developing tools to enable members of online discussion boards to gather and organize medically relevant data about their own experiences with particular diseases and courses of treatments.


Surgeons May Use Hand Gestures to Manipulate MRI Images in OR

Purdue University News (01/10/13) Emil Venere

Purdue University researchers are developing a system that recognizes hand gestures as commands to tell a computer to browse and display medical images of a patient during surgery.  The system uses depth-sensing cameras and algorithms to recognize hand gestures as commands to manipulate MRI images on a large display.  The system recognizes 10 gestures, including rotate clockwise and counterclockwise, browse left and right, browse up and down, increase and decrease brightness, and zoom in and out.  The researchers note the system’s accuracy relies on the use of contextual information in the operating room, which is achieved through cameras that observe the surgeon’s torso and head to determine what the surgeon wants to do.  “Based on the direction of the gaze and the torso position we can assess whether the surgeon wants to access medical images,” says Purdue professor Pablo Wachs.  The gesture-recognition system uses a Microsoft Kinect camera that can sense 3D space.  The researchers found that integrating context enables the algorithms to accurately distinguish image-browsing commands from unrelated gestures, reducing false positives from 20.8 percent to 2.3 percent.  The system also has an average accuracy of 93 percent in translating gestures into specific commands.

Research to Link Mobile Phones and Health

Murdoch University (12/18/12) Rob Payne

Murdoch University researchers are working with the University of Leuven’s DistriNet Research Group to develop wireless sensor network technologies that monitor people’s health.  “A lot has been done on using sensors to monitor health, but my work is the first that uses mobile phones collaboratively to detect and alert people to physical conditions like heart attacks,” says Murdoch researcher James Meneghello.  “Basically, if a person wearing a sensor has a problem with their heart, I want their phone to detect the anomaly and reach out to phones around it, using them to process the information, then pulling it back to warn the person that they’re about to have a heart attack.”  The researchers say their technology can theoretically tell if diabetic neuropathy is occurring by monitoring variations in heart rate.  Meneghello says the technology, if successful, could not only alleviate physical pain, but also spare a person from traveling very far to regularly see a physician.

Computers Can Predict Effects of HIV Policies

Brown University (07/27/12) David Orenstein

Brown University researchers have developed software that can model the spread of HIV in New York City over several years to make specific predictions about the future of the epidemic under different intervention plans.  “What we’re trying to do is identify the ideal combination of interventions to reduce HIV most dramatically in injection drug users,” says Brown University professor Brandon Marshall.  The program projects that with no change in New York City’s current HIV programs, the infection rate among injection drug users will be 2.1 percent per 1,000 by 2040.  However, strategies such as expanding HIV testing, increasing drug treatment, and providing earlier delivery of antiretroviral therapy could cut the rate by more than 60 percent, to 0.8 per 1,000.  The model creates a virtual reality of 150,000 agents who engage in drug use and sexual activity like real people.  “With this model you can really look at the microconnections between people,” Marshall says.  The researchers calibrated the program until it reproduced the infection rates among injection drug users that were known to occur in New York City between 1992 and 2002.

AI Predicts When You’re About to Get Sick

New Scientist (07/26/12) Michael Reilly

University of Rochester’s Adam Sadilek and colleagues were able to predict when individuals in New York City were about to come down with the flu up to eight days before they showed symptoms, using artificial intelligence and Twitter data.  The team analyzed 4.4 million tweets tagged with global positioning system location data from more than 630,000 users in the New York City area over one month in 2010.  The researchers trained a machine-learning algorithm to distinguish between tweets such as “I’m so sick of this traffic!” and those by people who were actually sick and showing signs of the flu.  They were able to predict when someone was about to fall ill–and then tweet about it–with about 90 percent accuracy up to eight days in the future.  Still, Sadilek says the system is limited because people do not reliably tweet about their symptoms and because getting sick is not limited to who they come in contact with.  Nonetheless, the data from the system could potentially be used for a smartphone app that warns users when they are entering a public place with a high incidence of flu.

UT Computer Science Professor Develops New Software to Aid in Disease Treatment

Daily Texan (07/18/12) David Maly

University of Texas (UT) researchers have developed software that can determine which compounds will best treat diseases using less information than was previously required.  The software relies on the 3D quasi-atomistic model of a virus’ protein cells to identify possible cures.  The researchers, led by UT professor Chandrajit Bajaj, developed algorithms to study the necessary detail on the 3D quasi-atomistic cell model needed to determine which drugs are most likely to prevent or treat a disease.  UT researcher Qin Zhang says the software will greatly help with future drug research by enabling scientists to research viruses without using the 3D atomistic model.  Although the software has been used mostly for HIV research, other scientists have approached the UT team about using it for work on other viruses.  “Each one is where they don’t have an atomistic model to begin with, and there is not any solution available other than ours,” Bajaj says.

Computer-Designed Proteins Programmed to Disarm a Variety of Flu Viruses

UW Today (06/01/12) Leila Gray

University of Washington researchers have demonstrated that proteins found in nature that do not normally bind with the flu can be engineered using computer modeling to act as broad-spectrum antiviral agents against a variety of flu virus strains.  “One of these engineered proteins has a flu-fighting potency that rivals that of several human monoclonal antibodies,” says Washington professor David Baker.  The computer-designed influenza inhibitors are constructed using a computer modeling system to fit perfectly into a specific nano-sized target on flu viruses.  The models can describe the landscapes of forces involved on the submicroscopic scale.  The researchers want to create antivirals that can react against a wide variety of H subtypes, which could lead to a comprehensive therapy for influenza.  The new methods could be “a powerful route to inhibitors or binders for any surface patch on any desired target of interest,” Baker says.  He notes that “we anticipate that our approach combining computational design followed by comprehensive energy landscape mapping will be widely useful in generating high-affinity and high-specificity binders to a broad range of targets for use in therapeutics and diagnostics.”