Background and Purpose:Gait dysfunctions have been frequently observed in patients with Alzheimer's disease (AD). Previous studies have used various methods to assess the gait impairment in AD. Here, we developed a wearable gait sensor, a Smart-insole, which is embedded in shoe insoles. We evaluated whether gait parameters measured using the Smart-insole were associated with cognitive performance in patients with AD. Methods:Participants aged 45-90 years, who were either cognitively unimpaired (CU) or had Alzheimer's disease dementia (ADD) were recruited from a hospital-based outpatient clinic, between January and December 2023. Participants performed three gait tasks (walk straight and turn test, timed up and go test, and ramp and stair test) while wearing the Smart-insole. The association of gait parameters with individual's cognitive status (CU vs. ADD) and cognitive test scores (Mini-Mental State Examination [MMSE] and various domain-specific cognitive tests from the Seoul Neuropsychological Screening Battery, 2nd Edition) was assessed. Results:Patients with ADD demonstrated decreased gait pace, rhythm, and stability as evidenced by longer task, cycle, and stance times, a higher number of steps, greater variability in swing time, and lower cadence during the gait task. Furthermore, gait parameters showed nominal associations with individual's MMSE score and each cognitive domain score (fronto-executive, memory, and language function). Conclusions:Smart-insole wearable sensors showed exploratory gait-related alterations in patients with ADD. These hypothesis-generating findings suggest that wearable insole-based gait assessment may provide complementary information for characterizing gait and cognitive dysfunction in ADD.
Alzheimer disease (AD) is no longer viewed only as a clinical syndrome of memory impairment but as a biological disease continuum driven by interacting amyloid-β, tau, glial, vascular, metabolic, and synaptic mechanisms. This shift has practical therapeutic consequences. Symptomatic agents such as cholinesterase inhibitors and memantine remain useful for selected patients, but they do not directly modify the core neurodegenerative process. In contrast, amyloid-β monoclonal antibodies have provided proof that biomarker-confirmed early AD can be slowed, although the magnitude of clinical benefit is modest, treatment is restricted to carefully selected patients, and amyloid-related imaging abnormalities require structured monitoring. Tau pathology, neuroinflammation, blood-brain barrier dysfunction, mitochondrial and oxidative stress, and impaired proteostasis are therefore increasingly important druggable axes rather than secondary background phenomena. A staged, biomarker-guided druggability framework can evaluate each target by biological proximity to clinical decline, biomarker measurability, therapeutic modifiability, implementation feasibility, and suitability for sequencing or combination therapy. Validated and emerging targets are summarized with emphasis on how biomarkers reshape patient selection and outcome assessment and how anti-amyloid therapy can be integrated with stage-specific interventions targeting tau propagation, glial activation, vascular injury, clearance failure, and metabolic vulnerability. The central conclusion is that AD drug development should move from a single-target rescue model toward biomarker-guided, mechanism-matched, and combination-ready strategies that can be tested earlier in the disease course while maintaining realistic safety and implementation standards.
Magnetic resonance imaging (MRI)-based artificial intelligence (AI) models are increasingly applied to brain MRI for diagnosing Alzheimer's disease (AD) and mild cognitive impairment (MCI), but their overall diagnostic performance remains unclear. We systematically searched PubMed/MEDLINE, Embase, Web of Science, Scopus, and IEEE Xplore until February 15, 2026, for diagnostic accuracy studies of machine-learning or deep-learning models using structural brain MRI to distinguish AD vs. cognitively normal (CN) controls and MCI or late mild cognitive impairment (LMCI) vs. CN controls. Seven studies met inclusion criteria, contributing five AD vs. CN and two MCI/LMCI vs. CN tasks, predominantly using deep-learning architectures applied to Alzheimer's Disease Neuroimaging Initiative cohorts. For AD vs. CN (five studies), pooled sensitivity was 0.96 (95% confidence interval [CI], 0.93-0.97) and pooled specificity was 0.95 (95% CI, 0.92-0.97), indicating excellent discrimination. For MCI/LMCI vs. CN (two studies), sensitivity was consistently high (0.91-0.93), whereas specificity varied widely (0.54-0.98), limiting the interpretability of pooled estimates. MRI-based AI models therefore show strong performance for established AD but heterogeneous specificity for MCI, underscoring the need for larger, externally validated studies in diverse populations.
