An AI that can instantly detect genetic diseases. A giant search engine containing all the medical research in the world. A virtual nurse on your phone that tells you exactly what treatments you need.
In the 21st century digital world, medicine must keep up with the times. Hundreds of technology and computing researchers at USC’s Information Sciences Institute (ISI) have developed innovations like these designed to streamline medical research and jumpstart life-saving treatments.
Through twenty years of collaborations with the USC Keck School of Medicine, Children’s Hospital Los Angeles, and many more medical associations and universities worldwide, ISI has used tools like machine learning, advanced data processing, and massive databases to revolutionize the ways medical researchers are finding solutions to the world’s health problems.
Robots and AI: The Next Generation of Doctors
In a world where watches and fridges are becoming smarter, making everyday tasks easier, it only makes sense that biotechnology should gain some IQ points as well.
NEURODEGENERATIVE DISEASES. The Center for Neuronal Longevity (CNL), a group at the Keck School of Medicine focused on treating neurodegenerative diseases, is working with ISI to develop smart implants that collect data directly from the patient to adjust their treatment in real time. For a technique like deep brain stimulation, where electrodes are implanted in the brain to produce impulses, a smart implant in the brain can use artificial intelligence to personalize the levels of stimulation it provides based on the body’s reaction.
ADRENAL GLAND DISEASE. Artificial intelligence is also being used to diagnose disorders using indicators undetectable by a human doctor. Researchers at ISI’s Visual Intelligence and Multimedia Analytics Laboratory (VIMAL) taught a computer to recognize congenital adrenal hyperplasia (CAH), a life-threatening genetic condition of the adrenal glands, simply by showing it pictures of faces of individuals with and without the disorder.
BRAIN CANCER. In a collaboration with Keck’s Radiation Oncology Department, ISI researchers are developing an AI-based computer tracking system that can convert miniscule head movements in a brain cancer patient to tangible feedback regarding the success of an ongoing radiation treatment.
NURSE IN YOUR PHONE. And HealthProtect, another ISI-Keck collaboration, is perhaps the smartest of all—a behavioral health program designed to intake data from smart devices, medical records, and user feedback to provide personalized health recommendations. Like a bedside nurse in your phone, it will be able to give instant medical advice whenever and wherever necessary.
Big Data Means Big Results
Medical research often means taking the microscopic and analyzing it on a massive scale. This can mean tracking the trillions of cells in the human body, or figuring out how a disease can travel through billions of individuals. ISI’s advanced data processing technology has helped researchers analyze massive amounts of data and find trends that can prevent diseases and understand the human body better.
MEMORY FORMATION. An ISI-developed method to track and archive data allowed for the imaging of synapses, the connections between neurons, in the brain of zebrafish. This enabled USC researchers Don Arnold, Scott E. Fraser, and ISI’s Carl Kesselman to discover how the memories form in the brain: the creation and destruction of synapses. Helped by ISI’s data collection technology, alongside a USC custom-designed microscope and a method of altering fish DNA so that the synapses glow, the group collected this data without harming the fish. This research paves the way for future studies of memory-altering disorders like PTSD.
KIDNEY DEVELOPMENT. Kesselman and his ISI team have also developed a program that was used to discover similarities between kidney development in mice and humans. The software program, called DERIVA (Discovery Environment for Relational Information and Versioned Assets), was used by a team of USC stem cell scientists to automate the task of recording data, saving researchers time and money. The platform displays the data in an easy-to-access manner, which makes the study not only accessible but easily reproducible, allowing it to be improved upon by any researcher interested in the subject.
MEDICARE RECORDS. Kesselman has also collaborated with Dana Goldman of the USC Price School of Public Policy, to use grid computing for the comparison of over six billion Medicare records in order to provide insights on the value of health procedures. Thanks to the innovative technology, Medicare records can be analyzed simultaneously alongside other datasets, such as Medicaid and private insurance, with relative ease.
OVARIAN CANCER & ALZHEIMER’S. ISI researcher Greg Ver Steeg teamed up with research consultant Shirley Pepke to scour through public databases and find patterns in the expressions of certain genes in ovarian cancer patients, allowing them to create personalized treatment plans based on patients’ genes. Using their discoveries Pepke was able to create a treatment plan for her own cancer diagnosis. Ver Steeg used the same machine learning techniques to find markers of Alzheimer’s disease in blood samples, opening the door for earlier diagnoses and non-invasive ways of tracking the progress of the disease in a patient. Blood markers had not been seriously considered as a method of diagnosis before, but the algorithm found the correlation that years of study could not.
Digital Collaborations Across the World
As research gets more advanced, collaboration can become difficult. Every leading research institution has its own software and practices, so it can be difficult for other institutions to interpret and adapt that data, needlessly slowing the pace of innovation. Many of ISI’s efforts over the past decade have focused around creating central databases to store data, or software that can easily interpret data from different places of origin.
SCHIZOPHRENIA. In 2017, ISI’s José-Luis Ambite and his team launched SchizConnect, a platform that integrates many different sources of data in the field of schizophrenia research, collectively totaling two thousand patient records. The study of schizophrenia relies heavily on MRI (magnetic resonance imaging) scans, which are incredibly costly, meaning that studies are often small in scope and stored in unique repositories. SchizConnect allows anyone to upload their data regardless of format.
FOODBORNE ILLNESSES. In collaboration with the German Federal Institute for Risk Assessment and the Kuehne Logistics University in Hamburg, ISI researcher Abigail Horn developed a mathematical model of the food supply chain, enabling scientists to predict large-scale outbreaks of foodborne diseases. Using the model, researchers can comb through massive amounts of supply chain data to pinpoint exactly where outbreaks begin and halt diseases before they proliferate.
NEUROBIOLOGY. ISI researcher Gully A.P.C. Burns’s NeuroScholar software project similarly helped alleviate information overload for researchers, developing a search system in which three-dimensional images could be searched through. In tandem with the National Institute of Health’s (NIH)’s Biomedical Information Science and Technology Initiative (BISTI) group, Burns created NeuroScholar to provide researchers “instant access to all the data sources, analysis tools, modeling tools, visualization tools, and interpretative materials necessary to do their jobs with no inefficiencies in computation or information technology.”
BIOMEDICAL DATA. ISI’s Carl Kesselman also worked with the NIH, leading a $22.2 million project to launch the Biomedical Informatics Infrastructure Research Network Coordinating Center (BIRNCC), creating a nation-wide computer network to facilitate biomedical research. This involved partnerships with research departments at UCLA, U.C. Irvine, the University of Chicago, and Massachusetts General Hospital. He also adapted grid computing, which he helped to develop, to bring together enormous amounts of medical research for the Center for Health Informatics.
GENOMICS. In another NIH project, ISI joined the Center for Genomic Studies on Mental Disorders, headquartered at Rutgers University, contributing computation infrastructure and clinical databases to the project. José-Luis Ambite and Ewa Deelman worked on web-based tools to easily query data and standardize the program’s data integration. This expanded on their previous work with Rutgers and NIH aimed at understanding the hereditary roots of diseases.
21st Century Health and Education
ISI is leading the way in adapting data and computing for real-world applications in the field of healthcare education. When training the next generation of medical professionals, from high schoolers to PhDs, the use of data tools and machine learning enables the curriculum to be precise and accurate.
BIOMEDICAL RESEARCH. In 2016, ISI and the USC School of Cinematic Arts launched Big Data U, a training coordination center that scours the internet for online courses and educational videos about the analysis of biomedical information, and centralizes them in one location for researchers from undergraduate students to even professors to access. Biomedical research as a whole is moving away from the “reductionist model,” in which one variable is tested in a controlled experiment. Instead, multiple variables are tested simultaneously, which is necessary to understand a complex system like the brain. Big Data U makes it easier for researchers to learn how to use tools designed explicitly for this type of research.
PEDIATRIC EMERGENCY RESPONSE TRAINING. ISI’s Robert Neches collaborated with Dr. Jeffrey S. Upperman of Children’s Hospital Los Angeles on a novel software system for pediatric emergency response, including the use of new telemedicine robots and tools to develop disaster drill training and regional emergency strategies. In the confusion following a large earthquake or similar disaster, the lives of children at a pediatric hospital depend on the decision makers getting accurate, real-time data about available resources and the hospital population’s needs. The software uses adaptive planning to train hospital employees before emergencies, and, in the midst of an emergency, help them follow the plan that best suits whatever situation they find themselves in.
SCIENCE CLASS AS A VIDEO GAME. A collaboration between the School of Cinematic Arts’ Interactive Media & Games Division and the USC Viterbi School of Engineering’s Department of Computer Science launched a nearly $300k effort with ISI to improve K-12 biology teaching, including an interactive video game called ImmuneAttack, designed to help high schoolers learn about the immune system. Based on a biology textbook, it presents the basic biology curriculum in a challenging and accessible manner, all while providing 20 minutes of gameplay in a virtual environment designed like the human immune system.
ISI and the Future of Medicine
As medicine evolves and researchers innovate new tests, treatments, and technology, the role of data will only grow. The ability to analyze medical records at a massive scale is crucial to unlocking new discoveries, and these revelations must be easily accessible, shareable, and reproducible. ISI’s collaborations with medical schools, hospitals, and researchers in the future will be key to ensuring sustainable, robust innovation in the medical field, leading to a healthier world.
Published on September 14th, 2022
Last updated on June 6th, 2023