Classification of Schizophrenia patients based on Stimuli of Speech Perception using Deep Learning and Machine Learning

Vetal, Darshan (2024) Classification of Schizophrenia patients based on Stimuli of Speech Perception using Deep Learning and Machine Learning. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

This research focuses on classifying individuals as schizophrenic or healthy, and further distinguishes schizophrenic patients with and without auditory hallucinations using MRI scans stimulated by speech perception tasks. Schizophrenia, a mental disorder characterized by symptoms such as delusions, memory loss, disorganized thinking, and auditory hallucinations, was studied using deep learning and machine learning models. The analysis targeted Broca’s area, associated with auditory processing, to understand neural differences. For Research Question 1, distinguishing hallucinating from non-hallucinating schizophrenic patients, a Convolutional Neural Network (CNN) achieved 48% accuracy initially. Applying ADASYN for oversampling improved accuracy to 65%, demonstrating its efficacy in handling class imbalance. For Research Question 2, A Convolutional Neural Network (CNN) model was implemented, achieving 32% accuracy initially. However, the dataset was imbalanced, necessitating oversampling techniques like SMOTE and ADASYN, which improved model robustness and yielded accuracies of 50% and 57%, respectively. Lazy Predict, a Python library, was employed to benchmark multiple traditional models, with Linear Discriminant Analysis, Linear SVC, and Logistic Regression achieving accuracies of 68%, 62%, and 62%. Grad-CAM visualizations enhanced interpretability by highlighting key regions influencing model predictions. This study provides insights into the neural correlates of schizophrenia and auditory hallucinations, aiming to improve diagnostic accuracy, treatment personalization, and patient outcomes. The findings highlight the potential of combining MRI data with advanced computational techniques for enhancing clinical decision-making in schizophrenia diagnosis.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Menghwar, Teerath Kumar UNSPECIFIED
Uncontrolled Keywords:	Schizophrenia; Auditory Hallucinations; MRI; Deep Learning; CNN; SMOTE; ADASYN; Machine Learning; Lazy Predict
Subjects:	R Medicine > RC Internal medicine > RC0321 Neuroscience. Biological psychiatry. Neuropsychiatry > Neurology. Diseases of the Nervous System. > Psychiatry Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning R Medicine > RA Public aspects of medicine > RA790 Mental Health
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	08 Sep 2025 08:58
Last Modified:	08 Sep 2025 08:58
URI:	https://norma.ncirl.ie/id/eprint/8834

Actions (login required)

View Item