This paper presents an automated framework for early detection of Alzheimer's Disease (AD) using a combination of Otsu thresholding with the Tunicate Swarm Algorithm (TSA) and Deep Belief Networks (DBN). The framework aims to improve the accuracy and efficiency of AD detection by segmenting regions of interest (ROI) from structural Magnetic Resonance Imaging (sMRI) data, extracting texture features, and classifying the images using DBN. The TSA optimizes the threshold value for Otsu thresholding, reducing computational time and improving segmentation accuracy. Local Binary Pattern (LBP) and Local Directional Pattern variance (LDPv) descriptors are used to extract texture features from the segmented ROI. The extracted vectors are then fed into the DBN for classification. The proposed method achieves high classification accuracy (99.80% and 99.92%) on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and Australian Imaging, Biomarker and Lifestyle flagship work of ageing (AIBL) datasets, outperforming conventional detection models. The effectiveness of the proposed framework is validated through various performance measures, including Jaccard Similarity Coefficient (JSC), Dice Similarity Coefficient (DSC), Pixel Accuracy (PA), specificity, execution time, classification accuracy, and sensitivity. The results demonstrate that the proposed method not only improves diagnostic performance but also accurately identifies pathological regions in sMRI images.This paper presents an automated framework for early detection of Alzheimer's Disease (AD) using a combination of Otsu thresholding with the Tunicate Swarm Algorithm (TSA) and Deep Belief Networks (DBN). The framework aims to improve the accuracy and efficiency of AD detection by segmenting regions of interest (ROI) from structural Magnetic Resonance Imaging (sMRI) data, extracting texture features, and classifying the images using DBN. The TSA optimizes the threshold value for Otsu thresholding, reducing computational time and improving segmentation accuracy. Local Binary Pattern (LBP) and Local Directional Pattern variance (LDPv) descriptors are used to extract texture features from the segmented ROI. The extracted vectors are then fed into the DBN for classification. The proposed method achieves high classification accuracy (99.80% and 99.92%) on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and Australian Imaging, Biomarker and Lifestyle flagship work of ageing (AIBL) datasets, outperforming conventional detection models. The effectiveness of the proposed framework is validated through various performance measures, including Jaccard Similarity Coefficient (JSC), Dice Similarity Coefficient (DSC), Pixel Accuracy (PA), specificity, execution time, classification accuracy, and sensitivity. The results demonstrate that the proposed method not only improves diagnostic performance but also accurately identifies pathological regions in sMRI images.