Abel Torres-Espin, University California, San Francisco, USA

Big-data and machine learning for ultra-acute ‘physiome’ biomarker discovery

Funded in: 2020, 2021, 2022


Back to overview

Problem: Outcome and recovery is hard to predict
Target: Investigate physiological signals to understand the complexity of acute spinal cord injury
Goal: Big-data and machine learning algorithms help to design better personalized treatments

Recent developments in science are changing how we predict the future of patients: biomarkers, the constant increase in computer power, and the progress in artificial intelligence and machine learning. Biomarkers are signals produced by the body in response to a disease, injury or physiological process. These signals can be used to understand what is happening within a patient and how the patient is likely going to progress. In this project, the researchers want to investigate physiological signals such as the arterial pressure and heart rate to understand the complexity of acute SCI patients.
Nowadays, a lot of that information is regularly collected in hospitals at a growing pace, creating what is known as big-data. Making sense of the amount of data available requires modern methods of analysis. The scientists will use the power of big-data and machine learning algorithms to find patterns in physiological signals that allow them to be more precise about the current and future state of an SCI patient. Thus, these patterns could be measured during the early time the patient is in the hospital and be used as biomarkers to predict the progression of the patient with more precision then what it can be done with current tools. Better predictions from easy to use measures would help medical providers and caregivers to design better personalized treatments that maximize recovery after injury.

In addition, the researchers’ work will generate valuable information on the use of clinical big-data for advancing research in SCI, accelerating new discoveries for medical practice.