Current Alzheimer Research - Volume 22, Issue 8, 2025

Alzheimer's disease (AD) is the most common form of dementia, and it is important to diagnose the disease at an early stage to help people with the condition and their families. Recently, artificial intelligence, especially deep learning approaches applied to medical imaging, has shown potential in enhancing AD diagnosis. This comprehensive review investigates the current state of the art in multimodal deep learning for the early diagnosis of Alzheimer's disease using image processing.

The research underpinning this review spanned several months. Numerous deep learning architectures are examined, including CNNs, transfer learning methods, and combined models that use different imaging modalities, such as structural MRI, functional MRI, and amyloid PET. The latest work on explainable AI (XAI) is also reviewed to improve the understandability of the models and identify the particular regions of the brain related to AD pathology.

The results indicate that multimodal approaches generally outperform single-modality methods, and three-dimensional (volumetric) data provides a better form of representation compared to two-dimensional images.

Current challenges are also discussed, including insufficient and/or poorly prepared datasets, computational expense, and the lack of integration with clinical practice. The findings highlight the potential of applying deep learning approaches for early AD diagnosis and for directing future research pathways.

The integration of multimodal imaging with deep learning techniques presents an exciting direction for developing improved AD diagnostic tools. However, significant challenges remain in achieving accurate, reliable, and understandable clinical applications.

Alzheimer’s disease is a progressive neurodegenerative condition characterized by the gradual deterioration of cognitive functions. Early identification of functional brain changes is crucial for timely diagnosis and effective intervention. This study employs multiplex network analysis to examine alterations in brain connectivity topology associated with Alzheimer's Disease, to identify early biomarkers and uncover potential therapeutic targets.

This study presents a secondary cross-sectional analysis based on a publicly available EEG dataset comprising spectral coherence measurements from 25 patients with clinically diagnosed Alzheimer's Disease (AD) and 25 age- and gender-matched Healthy Controls (HC). Functional connectivity matrices were generated across seven distinct frequency bands, with each brain region modeled as a network node and inter-regional coherence values represented as weighted edges. These matrices were then used to construct multiplex brain networks, which were rigorously analyzed using graph-theoretical approaches. The analysis encompassed key metrics, including modularity, centrality measures (Betweenness and MultiRank), motif distribution, and network controllability, to characterize and compare the underlying patterns of functional brain organization in AD and healthy aging.

Networks associated with AD exhibited significantly reduced modularity, disrupted centrality patterns, and a higher occurrence of 2 and 3-node motifs, indicating local reorganization of connectivity. Additionally, the spatial distribution of driver nodes was markedly altered in AD. Centrality analyses revealed a pronounced shift in network hubs toward the temporal and insular cortices, suggesting compensatory or pathological reallocation of influence. Controllability assessments demonstrated a lower energy requirement for network control in AD, accompanied by increased inter-layer fragmentation, reflecting compromised integrative function across frequency bands.

The findings revealed specific topological alterations, including reduced modularity, altered centrality, and decreased controllability, all of which are closely linked to AD-related network degeneration. By leveraging multi-frequency EEG data, the multiplex approach shows significant clinical potential for monitoring disease progression and supporting personalized treatments, with the ability to detect subtle connectivity disruptions before cognitive symptoms manifest.

Multiplex network analysis reveals distinct and robust alterations in the functional brain architecture of individuals with Alzheimer’s Disease. These network-level disruptions offer valuable insights into the pathophysiology of AD and highlight potential avenues for early diagnosis and targeted therapeutic strategies aimed at preserving cognitive function.

Much emphasis has been given to the biological activities of Fenugreek against various diseased conditions. This study investigated the effect of fenugreek leaf extract on behavioural and cognitive function of transgenic Drosophila having human Aβ-42 expression in the neurons, herein referred as Alzheimer’s disease model flies (AD flies).

AD flies were exposed to four different doses of fenugreek leaf extract (FE) containing i.e., 0.005, 0.010, 0.015 and 0.02 g/ml for 30 days. Thereafter, behavioural and cognitive assessment was done using climbing ability, activity pattern, aversive phototaxis and odour choice indexes. The life span of different groups of flies was also recorded. The effect of FE on the oxidative stress markers, acetylcholinesterase, monoamine oxidase (MAO) and caspase 3 and 9 activities were determined. The deposition of Aβ-42 aggregates in the brain tissue of the flies was studied by performing immunostaining. Also, the metabolic profile of different groups of flies was studied by performing LC-MS/MS. Compared with control flies, 22 selected metabolites were found to be upregulated and downregulated among transgenic AD flies and FE exposed AD flies compared to control.

The findings of this study showed the neuroprotective role of fenugreek extract, which could be employed for the treatment of Alzheimer’s disease. The AD flies exposed to FE showed a dose-dependent postponement in the decline of climbing ability, activity and cognitive impairments. A significant dose dependent increase in the life span was also noticed in the AD flies exposed to FE. A significant reduction in the oxidative stress, acetylcholinesterase, monoamine oxidase, and caspase-3&9 activities was also observed in a dose dependent manner. The results obtained from the immunostaining suggest the reduction in the deposition of Aβ-42 fibril, which was also confirmed by the docking studies showed the energetically favoured interaction useful for inhibiting the acetylcholinesterase and Aβ-42 aggregates.

This study demonstrates the neurological potency of fenugreek leaf extract (FE) in a Drosophila model of AD due to its antioxidantive, anti-cholinesterase, and neuroprotective properties. Using a combination of behavioral, biochemical, histological, and metabolomic approaches, we evaluated the therapeutic potential of FE in mitigating AD-like symptoms in transgenic flies expressing Aβ-42.

Fenugreek leaf extract may serve as a potential natural remedy for slowing down or alleviating the progression of AD.

Cerebral Amyloid Angiopathy (CAA) is a common form of cerebral small vessel disease (CSVD), characterized by the accumulation of amyloid-β (Aβ) protein in the walls of cortical and leptomeningeal arteries and arterioles. The sporadic form primarily affects the elderly and is closely associated with Alzheimer’s disease (AD). Despite previous studies on cognition, the specific neuropsychological profile of CAA remains unclear. This study aims to describe the cognitive profile of CAA patients and characterize their neuropsychological aspects in the absence of a clinical diagnosis of AD.

We present a case series of six patients with probable CAA, without clinical evidence of AD, who underwent extensive neuropsychological assessment. Additionally, a narrative review was conducted to synthesize current knowledge of the cognitive and neuropsychological aspects of sporadic CAA.

The narrative review indicates that CAA predominantly affects executive functioning, processing speed, episodic memory, global cognition, and visuospatial functions. In our case series, all patients exhibited impairments in these domains, except for global cognition. Notably, a specific dissociation was observed in the Rey Auditory Verbal Learning Test (RAVLT), with impaired delayed recall but preserved recognition.

Sporadic CAA in patients without AD contributes to cognitive impairment, particularly affecting executive functioning, processing speed, visuospatial functions, and episodic memory. In our sample, memory impairment in CAA follows a dysexecutive pattern, characterized by retrieval deficits with preserved storage. This contrasts with the amnestic profile seen in AD and amnestic mild cognitive impairment (aMCI), where both retrieval and storage are compromised.

This distinct memory profile may represent a useful neuropsychological marker for differentiating CAA-related cognitive impairment from that associated with AD and its prodromal forms. This differentiation has potential implications for diagnosis, prognosis, and the development of tailored therapeutic strategies.