Publication Date

Fall 2023

Degree Type

Thesis

Degree Name

Master of Science (MS)

Department

Biomedical Engineering

Advisor

Folarin Erogbogbo; Srikantan Nagarajan; Carly Demopoulos

Abstract

Auditory evoked fields (AEF) paradigm is a common research tool to study human auditory responses by using Magnetoencephalography (MEG), a neuroimaging tool. AEF paradigm requires repetition over many trials to achieve adequate signal-to-noise ratio (SNR), but Autistic children and some other populations cannot tolerate prolonged exam time due to various reasons. To address these challenges, this project uses a novel machine learning algorithm, Champagne with baseline noise learning, to reconstruct AEF data with fewer trials (80%, 60%, 40%, 20%) but produce the same results as using all AEF trials (100%). The results show that this novel machine learning algorithm can produce reliable latency results with only 60% of all AEF trials.

Download

Share

COinS