Powering Discoveries
Faith Singer
Good for Researchers, Good for Science
Texas cyberinfrastructure helps scientists across UT System make meaningful contributions to society
Researchers and students at 14 universities and medical schools across the state of Texas use advanced computing resources at TACC that are among the most powerful in the world.
They rely on these resources to conduct groundbreaking research in all domains of science from natural hazards to novel nanomaterials to medical research. The resources and staff expertise are under an umbrella called The University of Texas System Research Cyberinfrastructure.
Following are three current collaborations that illustrate how TACC and research teams work together to employ the hardware and software necessary to foster innovation. TACC also provides the consulting, training, and intellectual collaboration required for researchers to make meaningful contributions to their field and to society.
Mapping the Brain
UT Health San Antonio
In the 1970s and early 1980s, the first non-invasive brain imaging techniques were developed—computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET). These imaging advances were informed by massive government investments in signal-detection and nuclear technology during World War II. Knowledge gained through the Manhattan Project, for
example, was used to launch the field of nuclear medicine.
Peter Fox, MD, and director of the Research Imaging Institute at UT Health San Antonio, has used PET and MRI imaging for more than three decades to map brain funcion and organization.
To document the precise locations of task-induced neuronal activity, Fox introduced a Cartesian coordinate system, assigning a unique x-y-z address to each image voxel. Coordinate-based addressing rapidly became the most widely adopted analysis and publishing standard in human neuroscience.
Building upon his new standard, Fox built an online database of coordinate-based neuroimaging results called BrainMap.
“I developed a PET task-mapping strategy and software because nothing like this was available at the time,” Fox said. “As the brain-mapping field matured, the paradigms advanced from simple visual stimuli (flashing checkerboards) and motor performances (finger tapping) to mapping language, memory, and attention. Now the mental operations being mapped are very sophisticated, including social cognition, problem solving, risk-reward balance, and lie detection.”
This type of research is computationally intensive. Brain activation patterns are not analyzed by visual inspection, the way a radiologist does, i.e., by describing it in words and assigning a score. Rather, everything Fox and the brain-mapping community do is objective and rigorously quantitative, using computational pipelines composed of mathematical algorithms.
Fox began working with TACC in 2018. In the first collaboration, Fox and Joe Allen, a research associate at TACC, co-wrote a proposal and won funding to develop the BrainMap Community Portal (BCP).
The BCP is a TACC-hosted digital ecosystem providing the brain-mapping community with datasets and well-standardized meta-analysis pipelines in the setting of HPC and storage. This combination of resources is particulalry powerful for discovering network architecture in the brain using coordinate-based meta-analysis.
“TACC has a clever way of implementing portals,” Fox said. “Joe has mirrored the three BrainMap databases, which contain study results from more than 250,000 subjects and detail disease alterations, structure, function, and task performance. He has implemented all our meta-analytic applications as containerized code and reproducible pipelines. Now researchers around the world use the BCP to create computational models of brain disorders.”
Going forward, Fox and Allen plan to extend the BCP to include primary (per-subject) MRI data and symptom scores from neurologic, psychiatric, infectious, and metabolic disorders to determine whether they have biomarker potential.
“TACC was the springboard to make the software functionality available and the computing capability open broadly to people who wanted to extend this work,” Fox said. “TACC also makes our code work within the portal. They have been rapid at learning what we do—they dive right in to understand the complicated biology. I’m impressed at the tremendous support and quality of my interactions with TACC.”
Helping Children Overcome Childhood Trauma in Texas
Childhood Trauma Research Network
Traumatic experiences during childhood are widespread. In 2018, Texas reported more than 198,000 confirmed cases of child maltreatment in the form of physical, sexual, or emotional abuse; most incidents of maltreatment are unreported.
Moreover, childhood traumas are not limited to maltreatment. Other traumatic events include witnessing violence; loss of a loved one to an accidental death; online victimization; medical procedures; natural disasters; or growing up in a household affected by addiction, mental illness, domestic violence, or incarceration.
When a child experiences a traumatic event, are there characteristics that indicate at-risk behaviors versus positive resilience as their life continues? The answers to this question are important to the Texas Legislature, which funds the Childhood Trauma Research Network (CTRN). This network connects all 12 medical schools in Texas and observes the impact in the years following the traumatic events of children.
“We’re learning that mental health care is much more effective if its personalized.”
“We’re learning that mental health care is much more effective if its personalized,” said Jeffrey Newport, a psychiatrist at the Dell Medical School and associate director of the CTRN.
“Sometimes intervention does not help, or some children get better but still are not well. My hope is that we get a clearer sense of the issues requiring our attention so we can give parents specific guidance as to what the needs of their child might be rather than using a one-size-fits-all approach.”
Currently, 2,500 children between the ages of 8 to 20 are participating in the study. They have access to virtual mental health visits at school, peer-to-peer support, and cognitive behavioral therapy.
“Retention has been amazing,” Newport said. “Forty percent of the children are completing the two year mark.”
CTRN is working with TACC to develop a public facing web portal and a secure portal login for members to query the database directly.
“TACC was the obvious place to turn,” Newport said. “They are doing groundbreaking work leading automated data transfer and data curation. We are grateful to have TACC as a partner.”
Parrot Babbling
The University of Texas Rio Grande Valley
With their beautiful and powerful wings, long feathery tails, and many beak shapes, parrots have the most complex language abilities of any non-human.
Parrot “babbling” can be summarized as a type of complex communication for baby parrots where they string together a large diversity of adult sounds, but out of context. Babbling has huge implications for understanding language and cognitive development in children.
“In Venezuela, our work with parrots is helping us understand important precursors to spoken language, including providing implications for those who study and predict autism spectrum disorder long before it happens in people,” said Karl Berg, associate professor in the School of Integrative Biological and Chemical Sciences at UT Rio Grande Valley.
Berg and his team are in their 37th consecutive field season—the longest ongoing study of any wild parrot population in the world. They have recorded more than 14,000 hours of vocal babbling inside specially designed
nest cavities of 650 nestlings over a period of 14 years.
“It was fortuitous that I found TACC,” Berg said. “Language models are computationally expensive to run. At every turn in the road supercomputing has critical applications to what we are doing.”
Berg and TACC archived, stored, and are protecting the 200 terabyte data collection. The next goal is to automate it online so people from anywhere in the world can access it.
The intellectual collaboration is the most important aspect of the relationship with TACC, according to Berg. Preliminary data suggest that parrots have the potential to produce more than 20,000 expressions they display in their babbling repertoire.
“We’re trying to understand if there are grammatical rules to the babbling sequence that these nestlings do—is it random when they string sounds together or are there rules? Human language and computer science are connected at the hip. Together, we make a powerful team.”
Learn more: UT System Research Cyberinfrastructure (UTRC)