A patient under observation at home with the pilot
version of the vision system.

It strikes in infancy and steals children’s muscle strength, breathing ability, and even their lives. In Spinal Muscular Atrophy (SMA), motor neurons gradually lose their ability to carry messages from the spinal cord to the muscles. Although this degenerative disease doesn’t yet have a cure, several promising therapies are being developed. However, the high cost of clinical trials is a barrier to their approval by the FDA. Babies in clinical trials for SMA therapies must be admitted to a hospital or monitored at home by a clinician, both of which are expensive.

Enter Columbia University Medical Center's Petra Kaufmann and NREC’s Dave LaRose. They are investigating whether intelligent vision systems can be used in clinical trials for SMA. Because it affects movement, SMA is a good candidate for computer vision. Babies would be observed in their homes during clinical trials by a vision system that monitors their breathing and other muscular movements. The data it gathers would show what effect a therapy has on muscle tone, respiration, and other markers for SMA. This home-based monitoring would be much less costly than hospital admissions or clinician visits, speeding the approval of new therapies. It would also be more comfortable and less disruptive for patients and their families. And the unblinking eye of the camera would record every move that a patient made, giving researchers insight into the disease’s progression.

The pilot study is looking at whether respiration data gathered by a computer vision system correlates with respiration data recorded by body sensors and clinical observation. Patients in the study wear a sensor vest (or life shirt) that records their respiration. At the same time, they are monitored in their cribs by stereo cameras. The vision system tracks the three-dimensional movement of markers on the life shirt. Additionally, it tracks the movement of the patient’s wrists and ankles, which also can indicate the progress of the disease. The cameras use infrared light that’s invisible to humans, allowing the babies to be monitored day and night. A laptop computer stores the image data for processing.

The system correlates the computer vision estimates of the patient’s respiration with the respiration data from the life shirt. It also correlates limb motion with observations gathered by human clinicians. If the vision system data and life shirt data correlate well, the computer vision system can be further developed and adopted for use in clinical trials.

Graph of abdominal and chest movements
detected by the vision system. They track the
patient’s respiration, which is affected by SMA.