Answers are grounded in this paper and cite their sources.
Amyotrophic lateral sclerosis (ALS) is a progressive neurological disease leading to motor function deterioration and eventually respiratory failure and death. Identifying early diagnostic biomarkers for sporadic ALS is challenging due to the undefined risk population, making large datasets difficult to obtain. In this project, we used a dataset of 190 autopsy brain images from the Gregory Laboratory at the University of Aberdeen to explore convolutional neural networks for ALS classification. We developed Miniset-DenseSENet, a model combining DenseNet121 with a Squeeze-and-Excitation (SE) attention mechanism, to analyze these images. Our model aims to distinguish ALS patients from control groups and identifycognitiveimpairmentsinALSpatients. Miniset-DenseSENetoutperformed other transfer learning models and attention mechanisms, achieving 97.37% accuracy, an average MCC of 0.84, and sensitivity and specificity of 1 and 0.95, respectively. These results highlight the potential of integrating transfer learning and attention mechanisms to improve diagnostic accuracy and provide new insights into cognitive impairment associated with ALS. This approach not only improves the ability to identify ALS features but also provides a new perspective for distinguishing ALS patients with and without cognitive impairment.
No takes yet. Share an insight, caveat, or question.
Xia et al. (2024) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: