Project

Improved Anesthesiologist Situational Awareness Via Neuromorphic Computing

Project Status: Active

The purpose of this project is to evaluate the feasibility of using neuromorphic computing to improve battlefield anesthesia capabilities via arecommender-type of system.  This effort consists of the following research objectives:1) Perform a detailed literature search from existing research to understand prior work in this area.2) Conduct a series of computational experiments incorporating a government provided data base using neuromorphic spiking neural networks (SNN)algorithms on the provided labelled patient vitals and laboratory data (e.g., standard physiologies such as respiration, heart rate and bloodpressure as a function of ventilator adjustments and anesthesia adjustments ) to determine the feasibility of training a SNN model for this data.3) Create a training and evaluation set from the provided data to compare the accuracy of the SNN approach to standard machine learning and thehumanperformance provided in the data. This comparison will be most likely based on intervention recommendations to an anesthesiologist such asventilation settings, fluids, inhalational agents, and medication administration.4) Demonstrate the feasibility of a Field-Programmable Gate Array (FPGA) based neuromorphic computing system running a SNN trained onprovided datatomake recommendations displayed on a laptop in near real-time.5) Conduct testing of an FPGA neuromorphic computing system at the Uniformed Services University of the Health Sciences (USUHS) simulation andtraining laboratory. The purpose of the testing program is to understand the ability of the software to maintain patient vitals within the required rangeovera range of combinations of variables (ex. patient vitals, ventilator adjustment, medication, fluid and anesthesia adjustments) on a “simulant.”Initially the system will be operated open loop.  Potential feasibility of closed loop operation will be explored.6) Provide an assessment of the feasibility of using neuromorphic computing to improve battlefield anesthesia situational awareness.7) Conduct final demonstration of neuromorphic computing system at the USUHS simulator.

The purpose of this project is to evaluate the feasibility of using neuromorphic computing to improve battlefield anesthesia capabilities via arecommender-type of system.  This effort consists of the following research objectives:1) Perform a detailed literature search from existing research to understand prior work in this area.2) Conduct a series of computational experiments incorporating a government provided data base using neuromorphic spiking neural networks (SNN)algorithms on the provided labelled patient vitals and laboratory data (e.g., standard physiologies such as respiration, heart rate and bloodpressure as a function of ventilator adjustments and anesthesia adjustments ) to determine the feasibility of training a SNN model for this data.3) Create a training and evaluation set from the provided data to compare the accuracy of the SNN approach to standard machine learning and thehumanperformance provided in the data. This comparison will be most likely based on intervention recommendations to an anesthesiologist such asventilation settings, fluids, inhalational agents, and medication administration.4) Demonstrate the feasibility of a Field-Programmable Gate Array (FPGA) based neuromorphic computing system running a SNN trained onprovided datatomake recommendations displayed on a laptop in near real-time.5) Conduct testing of an FPGA neuromorphic computing system at the Uniformed Services University of the Health Sciences (USUHS) simulation andtraining laboratory. The purpose of the testing program is to understand the ability of the software to maintain patient vitals within the required rangeovera range of combinations of variables (ex. patient vitals, ventilator adjustment, medication, fluid and anesthesia adjustments) on a “simulant.”Initially the system will be operated open loop.  Potential feasibility of closed loop operation will be explored.6) Provide an assessment of the feasibility of using neuromorphic computing to improve battlefield anesthesia situational awareness.7) Conduct final demonstration of neuromorphic computing system at the USUHS simulator.

Last Updated: September 1, 2020 - 4:09 pm