Dr. Rose Faghih of the Cullen College of Engineering has been spotlighted as a featured author by the Institute of Electrical and Electronics Engineers (IEEE) Xplore homepage for the month of August.
The website highlights a particular research paper of the Computational Medicine Lab, “A Marked Point Process Filtering Approach for Tracking Sympathetic Arousal From Skin Conductance,” which was authored by her doctoral student Dilranjan Wickramasuriya and Faghih.
Published in March 2020, the paper focused on how the tracking of arousal by sweat levels could lead to better health monitoring via the development of technologies like smart watches and other skin monitors. This research was supported by the National Science Foundation through the CRII: CPS: Wearable-Machine Interface Architectures under Grant 1755780.
Their research was also featured earlier this year in the journal PLOS One. Faghih was identified by the IEEE Women in Engineering Magazine with a feature in its June issue as a “Woman to Watch,” as well as one of 2020's “Innovators Under 35” by MIT Technology Review.
For his part, Wickramasuriya said this line of research interests him because it has a path to practical application, which they are currently working toward.
“There is now quite a bit of hype in the field of machine learning today, regarding its possibilities,” he said. “Unfortunately, some of these machine learning methods lack interpretability. An advantage of our method is that it affords interpretability.”
Wickramasuriya said that in the future, there is promise when it comes to expanding their method to hormones with similar profiles.
“For instance, the hormone cortisol is secreted in pulses, just like a skin conductance signal is generated by bursts of neural impulses to the sweat glands,” he said. “Similar to skin conductance, we can use the methods we have developed thus far to estimate unobserved quantities that the human body is trying to regulate internally. The general framework of methods could also find applications related to other pulsatile hormones as well.”
Insulin and growth hormone are two common examples of these pulsatile hormones. Ideally, Wickramasuriya said that further research could lead to more precise health monitoring and treatment.
“One of the advantages of our method is that figuring out what is happening inside the body based on different physiological observations is already formulated within a closed-loop control-theoretic framework,” he said. “Therefore, they are more easily deployed into situations where closed-loop controllers are needed by patients suffering from different disorders, like automated hormone infusion pumps.”
The IEEE Xplore first highlighted their paper on Aug. 8. The feature will remain on the IEEE Xplore homepage through Aug. 31. The paper was published under Creative Commons License and will remain free to read on IEEE Xplore.