Locating and Isolating Neurological Disorders in MRI Datasets Using Hardware-Accelerated Delaunay Triangulation

Parkinson's disease and stroke are among the deadliest progressive neurological disorders affecting movement. Researchers have attributed the increasing prevalence of Parkinson's disease to the aging global population, making it the fastest-growing neurodegenerative disease. Interdisciplinary research in computer science and bioengineering frequently finds alternative and improved techniques that can greatly assist in the diagnosis and management of Parkinson's disease. Studies to date still confirm that Parkinson's disease is challenging to diagnose, as there are no specific tests to confirm the disease, and many symptoms overlap with those of other neurodegenerative disorders, such as essential tremor, a condition that causes involuntary and rhythmic shaking similar to that observed in Parkinson's disease. Connectomes and artificial intelligence have recently been examined for computer-assisted disease diagnosis and therapy management. Connectomes map and analyze neural linkages within the human nervous system, providing a viable approach for diagnosing brain diseases requiring detailed assessment of vascular structures. This study proposes a hardware-accelerated brain connectomic framework for locating and isolating possible neurological disorders in Magnetic Resonance Imaging (MRI) datasets using the Delaunay triangulation algorithm. The framework was implemented using the Microsoft .NET environment, primarily in C# and Visual Basic (VB), and integrated with hardware acceleration for enhanced processing. The connectomic framework was evaluated using Magnetic Resonance (MR) brain images from the Department of Surgery, University of North Carolina, United States, which mostly contained healthy datasets and certain brain abnormalities. Notably, the framework was not only able to successfully depict the connectomes of the datasets but also achieved processing times of less than 17 s for datasets containing more than 300 images. Moreover, the connectomic framework is considered resourceful in revealing the vascular structures in MR brain images, irrespective of the captured neurological disorders, with all features optimized for web and browser compatibility.