Date of Graduation

Summer 2021

Degree

Master of Natural and Applied Science in Computer Science

Department

Computer Science

Committee Chair

Tayo Obafemi-Ajayi

Keywords

deep learning, traumatic brain injury, residual learning, magnetic resonance imaging, transfer learning

Subject Categories

Artificial Intelligence and Robotics | Data Science | Diseases | Statistics and Probability

Abstract

One of the most significant frontiers for computational scientists is the engineering of human healthcare delivery based on intelligent analysis of health data. In a variety of neurological disorders such as Traumatic Brain Injury (TBI), neuro-imaging information plays a crucial role in the decision-making regarding patient care and as a potential prognostic marker for outcome. TBI is a heterogeneous neurological disorder. Due to the economic burdens of the disorder, sorting out this heterogeneity could provide more insights and better understanding of TBI recovery trajectories, thus improving overall diagnosis and treatment options. Magnetic Resonance Imaging (MRI) is a non-invasive technique that for examining the anatomy and pathology of the brain. This work examines a residual convolutional neural network to build a predictive model for TBI severity using varied MR images and three different combinations of TBI assessment scores obtained from Federal Interagency Traumatic Brain Injury Research. To address the challenges of insufficient data and increase efficiency. The framework consists of five components which include data curation, data augmentation, residual learning model, model validation and clinical relevance assessment. The data curation phase pre-processes the images into a format reliable for use by the model. To address the problem of insufficient, unbalanced and highly skewed data, the data augmentation generates different forms of the images to improve the generalization capability of the model. The residual learning model integrates transfer learning by utilizing a network that has been pre-trained on general data and then fine-tuned for MR images to improve the model performance and reduce training time. Model validation consists of both quantitative and qualitative means. The clinical relevance assessment phase includes the identification of meaningful subgroups to better understand the how the results correlate with the MRI data. A mixed effects Analysis of Variance (ANOVA) model is performed using varied TBI outcome measures to assess the clinical significance of the results. The experimental results showed that our model achieve a high precision on the test sample.

Copyright

© Dacosta Yeboah

Open Access

Share

COinS