skip to main content

Title: Development of a Short-Term to Long-Term Supervised Spiking Neural Network Processor
; ; ; ;
Award ID(s):
1718428 1556294 1556301
Publication Date:
Journal Name:
IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Page Range or eLocation-ID:
2410 to 2423
Sponsoring Org:
National Science Foundation
More Like this
  1. OBJECTIVES/GOALS: Spinal cord stimulation (SCS) is an intervention for patients with chronic back pain. Technological advances have led to renewed optimism in the field, but mechanisms of action in the brain remain poorly understood. We hypothesize that SCS outcomes will be associated with changes in neural oscillations. METHODS/STUDY POPULATION: The goal of our team project is to test patients who receive SCS at 3 times points: baseline, at day 7 during the trial period, and day 180 after a permanent system has been implanted. At each time point participants will complete 10 minutes of eyes closed, resting electroencephalography (EEG). EEG will be collected with the ActiveTwo system, a 128-electrode cap, and a 256 channel AD box from BioSemi. Traditional machine learning methods such as support vector machine and more complex models including deep learning will be used to generate interpretable features within resting EEG signals. RESULTS/ANTICIPATED RESULTS: Through machine learning, we anticipate that SCS will have a significant effect on resting alpha and beta power in sensorimotor cortex. DISCUSSION/SIGNIFICANCE OF IMPACT: This collaborative project will further the application of machine learning in cognitive neuroscience and allow us to better understand how therapies for chronic pain alter resting brain activity.