Alzheimer's disease is a progressive neurodegenerative disorder that impairs memory, cognition, and behaviour. Advances in AI technology, particularly in deep learning and medical imaging, offer powerful tools for early detection and classification of Alzheimer's disease, improving diagnosis and treatment outcomes. In this research, the authors highlighted the impact of selected digital data on various pre-processing techniques for Alzheimer’s disease classification. They further highlight the potential challenges of implementing FreeSurfer’s "recon-all" module in healthcare applications. Data collection plays a crucial role in standardizing images for consistent analysis. The authors developed and evaluated three different pre-processing strategies using a custom EfficientNetV2S architecture. Results indicate that more complex pre-processing steps, such as skull-stripping, lead to improved classification precision. However, practical challenges like long processing times and FreeSurfer's closed code environment limit its practicality in fast-paced healthcare settings. The hypothesis suggests that skull-stripped MRI sequences processed through FreeSurfer offer a more accurate method for detecting Alzheimer’s than alternative strategies. The findings of this research support this hypothesis, with skull-stripping showing better accuracy. However, the differences between methods are minimal given the current dataset and model setup. Future research should focus on expanding the dataset and enhancing model robustness to uncover more significant distinctions, highlighting the potential advantages of skull-stripping for Alzheimer's detection. |
*** Title, author list and abstract as submitted during Camera-Ready version delivery. Small changes that may have occurred during processing by Springer may not appear in this window.