Monitoring Babies’ Lives

| June 27, 2017

Researchers at the Division of Biokinesiology and Physical Therapy and the Viterbi School of Engineering team up to examine poor motor development in infants.

Photo courtesy of Kidsana

Modern technology helps doctors make quicker, more accurate diagnoses. An Ebola test, which once took 12 hours, now takes 30 minutes with 99 percent accuracy. However, some medical diagnoses still come from broad, subjective procedures that sometimes fall short.

According to First 5 California, a program created under Proposition 10 to promote children’s health in their first five years, “more than 10 percent of kids under 5 in California have a disability or special need that may impact their ability to play and learn.”

Currently, healthcare professionals identify and evaluate poor motor development in infants using an observational checklist. Limitations to this approach include subjective ratings by the observer and the variability of infant behavior. Infants may or may not demonstrate a specific skill during a specific observation period.

Delays in establishing movements, such as crawling and walking, may indicate slower brain development, or in more severe cases, neurological disorders such as cerebral palsy.

Additionally, the interval given for a “normal age” to reach motor milestones, such as walking, can span from nine months to 18 months.

With the current standard diagnosis procedure and the broad “normal age” intervals, it can be a challenge for healthcare professionals to accurately identify early motor impairment. Delayed identification typically results in later intervention in a situation where early intervention is the goal.

Beth Smith, an assistant research professor in the USC Division of Biokinesiology and Physical Therapy, began research five years ago to provide a method that will give more accurate early identification of impaired neuromotor development. She collaborates with USC Viterbi’s Integrated Media Systems Center (IMSC) on the research project.

Five years ago, Smith employed the use of a wearable Opal sensor on infants to collect continuous data on their movements across a full day. She said this approach overcomes some challenges that come with evaluating an infant’s motor development in a single, short, “snapshot” assessment.

“The variability in healthy development makes it more challenging to identify the [motor] delay,” said Smith, adding that she expects her new approach to provide more accuracy than current diagnostic techniques.

Smith and her team are working with IMSC researchers to analyze the continuous information from the wearable sensors. They look for patterns in the movement that reveal a correlation between motor behavior and development.

The research is supported by grants from the American Physical Therapy Association’s Academy of Pediatric Physical Therapy awarded to Smith.

After enough data is collected and processed, numerical measurements for typical infant motor development will emerge. The goal is to provide a more straightforward and precise diagnostic method.

“Current assessment methodology is very subjective,” said Luciano Nocera, a senior research associate at IMSC. “It is up to the interpretation of one medical expert and their evaluation could be incorrect.”

Added Smith: “We think a more quantitative approach will provide greater accuracy in predicting an infant’s neuromotor status. If intervention occurs earlier, then we can positively influence the infant’s motor development.”

Early results have been positive, Smith said. The study collected data from 12 typically developing infants and 24 infants at risk for developmental delay between birth and walking onset.

Infants wore a sensor on each leg at home for eight to 10 hours a day. The sensor collects details of the infant’s movements. At USC Viterbi’s IMSC, analysts use the information to evaluate the activity.

The data are revealing how the patterns of typically developing infants differ from infants with atypical motor development.

Once Smith and the IMSC collect enough data from infants with typical and atypical motor development, big data analytics will process it all and generate a numerical method to identify patterns of typical and atypical motor development.

“If you identify the infants that are at risk for not developing correctly [earlier], then you can help them out with physical therapy interventions,” said Nocera, adding that this predictive approach could help thousands of infants receive the treatment they need to improve brain and motor progress.