MIT Department: Media Arts and Sciences
Undergraduate Institution: Rensselaer Polytechnic Institute
Faculty Mentor: Rosalind Picard
Research Supervisor: Daniel Lopez Martinez
My name is Allison Rojas, I am a Peruvian American raised in Miami, FL. I am an Architecture undergraduate student at Rensselaer Polytechnic Institute, involved in research spanning art direction, wearable design, painting analysis, and architectural design integration of virtual reality. I am intent on understanding the relationship between public health and design and harnessing it in order to innovate the designed environment to become a tool for healing ourselves. Specifically in the future by improving the quality of life of citizens through redesigning systems which affect communities and families’ lives directly such as in the context of a hospital, health care system, or urban design system. I also enjoy reading and spending time with my family.
2017 Research Abstract
Data Visualization to Detect Infections
Allison Rojas, Department of Architecture, Rensselaer Polytechnic Institute,
Daniel Martinez, Harvard-MIT Division of Health Sciences and Technology, Affective Computing, Media Lab, Massachusetts Institute of Technology
Rosalind Picard, Affective Computing, Media Lab, Massachusetts Institute of Technology
The development of wearable biosensors has provided us with the opportunity to monitor physiological health parameters non-invasively throughout a continuous duration of time. Through this explicit record, we have acquired the ability to analyze and find correlating temporal relationships between skin conductance, temperature, heart rate variability, and motion. By recording over 9,00,000 daily measurements for up to 19 individuals by documenting first baseline data followed by 3-4 days after inoculation with Rhinovirus, a common cold virus, we are in the process of investigating correlations between symptoms and physiological data. We utilized accelerometer data to create criteria in which to hypothesize a circadian rhythm and fluctuation in sleep duration throughout the period of infection. The fundamental question we are trying to answer is if it is possible to detect physiological and behavioral responses due to the common cold and flu solely using wearable sensor data. Studying the data sets has also illustrated to us that using the still posture of sleep duration is effective in collecting more subtle but accurate personal data due to the reduction of noise caused by daily activities. Utilizing this data, we categorized before and after inoculation mean, standard deviation, maximum, minimum and range of the previously mentioned physiological signals in order to create future classifiers to define which parameters are most efficient in identifying the innate shift from health to infection. This can lead to greater implications in public health and epidemiology such as understanding how individuals are becoming sick daily, monthly, yearly and accessing information about the dissemination rate of a pandemic.