Experts here are building algorithms that harness AI’s power to improve cancer diagnoses in the brain, understand atypical speech and predict when hearing loss might be an early indicator of Alzheimer's disease.

Powering that work is a cadre of neuroscientists, biologists and engineers working to understand the complexities of the brain and its neurons. This study of biological intelligence is fueling Beckman’s work in this transformative frontier.

Zhi-Pei Liang is leading research at the intersection of artificial intelligence and magnetic resonance imaging of the brain. The lab's non-invasive, high-resolution metabolic imaging of the whole brain revealed differences in metabolic activity and neurotransmitter levels among brain regions. Liang's team found metabolic alterations in brain tumors. Team members mapped and characterized multiple sclerosis lesions, while patients spent only minutes in an MRI scanner.

“Our technology overcomes several long-standing technical barriers to fast high-resolution metabolic imaging by synergistically integrating ultrafast data acquisition with physics-based machine learning methods for data processing,” Liang said in a release about the work. With the new MRSI technology, the Illinois team cut the time required for a whole brain scan to 12 and a half minutes."

In subjects with multiple sclerosis, the technique detected molecular changes associated with neuroinflammatory response and reduced neuronal activity up to 70 days before changes become visible on clinical MRI images.

Zhi-Pei also supercharges collaborations that apply his AI-powered MRI expertise to new areas:

• He works closely with Dr. Dan Llano in better understanding the correlation between hearing loss and Alzheimer's disease. Dan's research pinpoints unexpected areas of brain shrinkage in Alzheimer's patients with hearing loss. The work could eventually lead to earlier intervention and new treatment strategies to delay cognitive decline in patients.

• With Yun-Sheng Chen, he's exploring how recurrent hypoglycemia impacts brain function in diabetes patients. The team uses advanced imaging technologies to investigate the brain responses tied to cognitive impairments caused by low blood sugar levels.

Cognition — and its disruptions in psychiatric and neurological disorders — emerges from the coordinated activity of brain regions organized into large-scale networks.

Caterina Gratton and members of her lab study these networks with high-resolution, individual-level brain mapping and network-based mathematics.

"These methods produce reliable, detailed maps of human brain networks to characterize how people’s brains differ from one another and change over time," she said. "Using machine learning, we can show that these individual network patterns predict important aspects of cognition measured outside the MRI, such as inhibitory control. Ongoing projects in the lab are extending these findings to study health, aging and neurodegeneration in Parkinson’s disease."

Rhanor Gillette and Ekaterina Gribkova built an artificial intelligence based on marine life, including sea slugs and octopuses. As a result, it can navigate new environments, seek rewards, map landmarks and overcome obstacles.

“This essentially makes our artificial intelligence much more animal-like than current artificial intelligences,” Ekaterina said. “We are learning how to go from something like the memory of a sea slug, which is very, very simple, to something like us.”

Machine intelligence, though advanced, is far from human. This Beckman research group is using human behavior and neuroscience to inspire artificial intelligence. Researchers in this group also hope to enhance the behavioral sciences with principles learned from computational intelligence. They’re also forging collaborations among those who study human behavior and neuroscience and those developing artificial intelligence.

The Beckman team leading the Speech Accessibility Project is making technology work for every voice. Historically, the speech recognition systems in your phone or smart speaker have been trained on audiobook recordings, so they have trouble recognizing diverse speech patterns.

Funded by Apple, Amazon, Google, Meta and Microsoft, the project recently finished recruiting people with Amyotrophic Lateral Sclerosis, cerebral palsy, Down syndrome, Parkinson's disease and those who have had a stroke. The next phase is recruiting people who stutter and those who are deaf and hard of hearing.

So far, Microsoft announced "significant improvements" in its speech recognition tech, and the research team made an automatic speech recognizer 30% more accurate in understanding people with Parkinson's. The project's data is open to outside researchers and companies; about 100 organizations are currently using it.

Studying infant behavior and emotional regulation is incredibly important — and difficult. A Beckman team, including human development expert Nancy McElwain and electrical engineer Mark Hasegawa-Johnson, have created a data collection tool perfect for the institute's youngest research participants.

LittleBeats is a wearable device that collects data via electrocardiogram, audio and motion sensors. The team then uses deep learning algorithms and the collected data to assess infant behavior in real-world settings.

“Our end goal is to give back to the community and use LittleBeats as a tool for intervention and prevention," Nancy said. "More broadly, it could be used as a way for parents to gain a detailed view of their child’s daily interactions. ... If parents and care providers can monitor a child’s development with this device, early identification of a motor or language delay, behavioral disturbance, or sleep disturbance could be possible."