Neural Network Analysis of MRI Scans for FND Diagnosis
Samiel Azmaien
Samiel Azmaien, Department of Computer Science, Georgia Institute of Technology, Atlanta, Georgia, United States of America (USA).
Manuscript received on 02 June 2024 | Revised Manuscript received on 12 June 2024 | Manuscript Accepted on 15 June 2024 | Manuscript published on 30 June 2024 | PP: 42-46 | Volume-4 Issue-4, June 2024 | Retrieval Number: 100.1/ijapsr.A405805011224 | DOI: 10.54105/ijapsr.A4058.04040624
Open Access | Editorial and Publishing Policies | Cite | Zenodo | OJS | Indexing and Abstracting
© The Authors. Published by Lattice Science Publication (LSP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Background Functional Neurological Disorder (FND) currently lacks a definitive method of diagnosis, leading to an extremely high rate of misdiagnosis. Methods This project aimed to address the question of improving diagnostic accuracy for FND by utilizing logistic regression models and neural networks, integrating patient MRI data and clinical history to differentiate FND from other neurological disorders. MRI scans were first pre-processed through noise reduction and feature engineering, and then used to train two types of models: logistic regression for general neurological disorder classification and a neural network specifically for FND diagnosis. The diagnostic performance was measured using the ROC AUC metric, with additional evaluation through accuracy, precision, recall, and the F1 score. Results & Conclusions By targeting the most relevant variables from the MRI data, both models demonstrated high efficacy, with the neural network showing a 92% accuracy rate in FND classification.
Keywords: Diagnosis · Functional Neurological Disorder · Logistic Regression · MRI Data Scans· Neural Networks.
Scope of the Article: Medical Physiology