Alzheimer’s disease (AD) is the most prevalent neurodegenerative disorder worldwide, affecting approximately 56.9 million people in 2021 and projected to reach 152 million by 2050. Its defining pathological features, amyloid-beta plaques and neurofibrillary tangles, accumulate for up to two decades before cognitive symptoms emerge, placing the preclinical and mild cognitive impairment (MCI) stages at the centre of the early detection problem. Despite this, current diagnostic practice in routine clinical settings remains unreliable, with post-mortem studies placing the specificity of clinical AD diagnosis between 44.3 and 70.8% even in specialist memory clinics. Machine learning (ML) and deep learning (DL) applied to neuroimaging and electrophysiological data have emerged as candidate tools for closing this diagnostic gap, yet whether the accuracy figures reported in published studies translate into clinically useful performance on independent data remains unresolved. This study presents a structured comparative review of machine learning and deep learning methods reported across four publicly available Alzheimer’s disease datasets, namely the Alzheimer’s Disease Neuroimaging Initiative (ADNI), the Open Access Series of Imaging Studies (OASIS), the OpenNeuro ds004504 electroencephalography (EEG) dataset, and the Kaggle Alzheimer’s magnetic resonance imaging (MRI) dataset. Thirteen model families are examined through the published literature rather than through new experiments, and for each model and dataset combination, the best accuracy reported in the source study is recorded alongside the model’s mathematical formulation. All performance figures reported in this abstract and throughout the paper are taken from the published studies reviewed, not from new experiments conducted by the authors. Across the reviewed studies, deep learning architectures pre-trained on ImageNet and fine-tuned on neuroimaging data are reported to produce the highest accuracy on MRI classification tasks. Residual Network (ResNet)-101 is reported at 98.21 percent on ADNI and 97.45 percent on OASIS, while the IncepRes fusion architecture reaches 98.35% on OASIS by combining multi-scale feature extraction from InceptionV3 with residual connectivity from ResNet152V2. Traditional machine learning classifiers remain competitive on tabular clinical and biomarker data, with Extreme Gradient Boosting (XGBoost) reaching 91% on ADNI multiclass features. Logistic Regression achieves 82 to 85% on binary ADNI classification and is the only classifier in this review that provides explicit per-feature prediction contributions without post hoc tooling. Gaussian Naïve Bayes achieves 80 to 83% on the same task. On the OpenNeuro EEG dataset, K-nearest neighbours (KNN) with singular value decomposition (SVD) entropy features achieves 91% binary accuracy, with feature engineering quality determining performance more reliably than classifier architecture. Eight principal findings emerge from the cross-dataset analysis. Binary classification consistently outperforms multiclass by 10 to 30% across all datasets, reflecting the genuine biological ambiguity of the mild cognitive impairment category. Dataset size and augmentation predict reported accuracy more reliably than model architecture. Ensemble methods outperform individual classifiers by 5 to 8% in nearly every imaging study. Deeper architectures can overfit small clinical cohorts. EEG models trail MRI models by approximately 10 to 15% on comparable binary tasks. Cross-dataset generalisation has not been systematically evaluated in most studies, and the few that have tested it report accuracy drops of 5 to 10% or more when models encounter data from different scanners or cohorts. Eight recurring limitations constrain the clinical utility of these findings. Small sample sizes and limited demographic diversity, severe class imbalance inflating raw accuracy metrics, poor cross-dataset generalisation driven by scanner heterogeneity, limited deep learning interpretability, the dominance of binary over multiclass tasks, the absence of longitudinal modelling despite available datasets, inadequate standardisation of preprocessing and evaluation protocols, and the signal-to-noise ratio constraints specific to EEG recordings of elderly patients collectively define the gap between benchmark performance and clinical readiness. Future work must prioritise multi-centre training cohorts, multimodal fusion architectures, longitudinal progression modelling, and standardised interpretability evaluation as non-optional requirements for any system intended for clinical deployment.

Alzheimer’s disease (AD) is the most common neurodegenerative disorder, and its pathology accumulates for years before symptoms appear, which places the mild cognitive impairment (MCI) stage at the centre of early detection. Machine learning (ML) and deep learning (DL) applied to structural magnetic resonance imaging (MRI) report high classification accuracy, yet these figures are usually measured within a single cohort and are rarely tested for transfer to an independent cohort, so whether they translate into clinical usefulness remains unresolved. The objective of this study is to address that gap directly. Six classifiers, three classical learners (Extra Trees, LightGBM and CatBoost) and three deep networks (ConvNeXt, CoAtNet and an introduced bidirectional Vision Mamba model, MambAD) were trained and tested on identical hippocampal slices from three publicly available datasets, namely ADNI, OASIS and a public Kaggle collection, for the three-class staging of cognitively normal (CN), MCI and AD. Every model was evaluated within cohorts and, more importantly, across cohorts, with subject-level scoring on the two clinical datasets. Within the primary ADNI cohort, the introduced model reached the highest balanced accuracy (BAC) at 0.578, although this lead was not statistically significant; across the six models, none dominated on every task, and no pairwise difference between models reached significance. BAC fell from about 0.55 within-cohort to between 0.38 and 0.45 across cohorts for every model, and the deep models transferred better on the ranking metric than the classical learners. The results show that a lightweight state-space model can match far larger pretrained backbones under matched conditions, and they confirm that single cohort accuracy overstates readiness, with cross-dataset generalisation remaining the binding constraint on clinical use.