Systematic and Integrative review on Artificial intelligence in the Diagnosis of Dementia, Brain Neuroplasticity, and Healthy Aging

Full text

Text extracted from the article PDF for reading and search. Tables, figures and equations are in the PDF, which remains the authoritative version.

Introduction

According to Walhovd, Fjell, and Espeseth (2014) [39], brain aging is a multifactorial process characterized by progressive structural and functional changes that can result in cognitive decline and increased vulnerability to neurodegenerative diseases. This process occurs heterogeneously among individuals and is influenced by genetic, environmental, behavioral, and social factors, which reinforces the need for integrated and personalized approaches to understanding healthy aging.

In this context, the neuroscience of aging has advanced significantly by demonstrating that the brain maintains adaptive capacity throughout life, a phenomenon known as neuroplasticity (ALVES, 2024) [3]; (OLIVEIRA et al., 2025) [25]. Evidence suggests that cognitive stimuli, physical activity, multimodal interventions, and psychosocial support can modulate neural circuits, delay cognitive decline, and reduce the risk of dementia, even in advanced ages (STYLIADIS et al., 2015) [37]; (CUI et al., 2018) [8]. At the same time, the exponential growth of digital technologies and artificial intelligence-based systems has transformed the way brain aging and neurodegenerative diseases are studied, diagnosed, and monitored (LYALL et al., 2023) [19]; (GUTMAN et al., 2024) [13]. Machine learning and deep learning algorithms have shown high potential in the analysis of large volumes of clinical data, neuroimaging, electroencephalography, and cognitive profiles, contributing to the early detection of mild cognitive impairment and Alzheimer's disease (DE OLIVEIRA, 2025) [10]; (WU et al., 2024) [40].

In addition to diagnosis, artificial intelligence has been applied in the development of personalized therapeutic strategies, including cognitive rehabilitation, neuromodulation, neurofeedback, and brain-computer interfaces, expanding the possibilities for intervention in elderly populations (RUTKOWSKI et al., 2018) [31]; (CALDERONE et al., 2025) [7]. These technologies have the potential to optimize the stimulation of neuroplasticity and promote greater functional autonomy, especially in individuals at risk for dementia (KORADE et al., 2024) [16]; (XIONG, 2024) [41].

However, despite the growing volume of publications, there are still gaps in the systematization of available scientific evidence, especially regarding the integration of artificial intelligence, neuroplasticity, and brain aging in different clinical and preventiv y contexts (ONCIUL et al., 2025) [26]; (GKINTONI et al., 2025) [12]. The methodological heterogeneity of the studies, as well as the ethical and technological challenges associated with the use of artificial intelligence in health, reinforce the need for critical and rigorous syntheses of the literature.

Thus, it is essential to conduct systematic reviews that consolidate existing knowledge, identify trends, assess limitations, and guide future research and clinical applications in the field of artificial intelligence-assisted brain aging (BERNAL et al., 2024) [4]; (QANSUWA et al., 2025) [27].

In this scenario, the present systematic review seeks to contribute to the advancement of knowledge by analyzing, in a structured and evidence-based manner, the role of artificial intelligence in promoting healthy aging and preventing neurodegenerative diseases.

Objectives

General Objective To analyze, through a systematic review of the literature, the available scientific evidence on the application of artificial intelligence (AI) in the context of brain aging, neuroplasticity, and the prevention, diagnosis, and management of dementia, with an emphasis on evidence-based approaches and innovative strategies for promoting healthy aging.

Specific Objectives • Identify the main artificial intelligence methods and algorithms applied to the study of brain aging and neurodegenerative diseases; • To assess the role of artificial intelligence in the early detection of cognitive decline, mild cognitive impairment, and dementias, especially Alzheimer's disease; • Analyze the relationship between artificial intelligence, neuroplasticity, and cognitive, physical, or combined interventions in elderly populations; • Describe emerging applications of digital technologies, wearable devices, neurotechnologies, and brain-computer interfaces in healthy aging; • Synthesize evidence regarding the benefits, limitations, and ethical challenges of using artificial intelligence in clinical and preventive contexts related to brain aging.

Methodology

This is a systematic review of the literature conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines, with the aim of ensuring transparency, methodological rigor, and reproducibility of the study selection and analysis process.

Search Strategy The bibliographic search was conducted in national and international electronic databases, including indexed databases relevant to the fields of neuroscience, geriatrics, artificial intelligence, and digital health. Additionally, other sources were consulted, such as reference lists of included articles and complementary literature, in order to broaden the identification of potentially eligible studies. Controlled descriptors and free terms related to artificial intelligence, aging, neuroplasticity, cognitive decline, dementia, and Alzheimer's disease were used, combined by Boolean operators.

Eligibility Criteria Original studies, systematic reviews, narrative reviews, and observational or experimental studies addressing the application of artificial intelligence in brain aging, neuroplasticity, or dementia, published without language restrictions, were included. Duplicate articles, studies that were not directly related to the proposed topic, incomplete works, conference abstracts without full text available, and publications that did not meet the established methodological criteria were excluded.

Study Selection The selection of studies was carried out in stages. Initially, the titles and abstracts of the identified records were evaluated independently, and those that were clearly irrelevant were excluded. Subsequently, potentially eligible articles were submitted for full-text reading to verify compliance with the inclusion criteria. The selection process is described in detail in the PRISMA flowchart shown in Figure 1.

Data Extraction and Synthesis From the studies included in the qualitative synthesis, information was extracted regarding the authors, year of publication, type of study, population investigated, applications of artificial intelligence, main outcomes, and conclusions. The data were organized descriptively and analyzed qualitatively, allowing the identification of patterns, convergences, and gaps in the scientific literature on the subject.

Ethical Considerations

As this is a systematic review based exclusively on secondary data available in the literature, it was not necessary to submit the study to a research ethics committee.

Results

Figure 1 presents the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flowchart, which systematically and transparently describes the process of identification, selection, eligibility, and inclusion of the studies analyzed in this review on the application of artificial intelligence in brain aging, neuroplasticity, and dementia prevention. The use of the PRISMA flowchart allows us to detail each methodological step adopted, ensuring reproducibility, scientific rigor, and clarity regarding the criteria used for the selection of articles included in the qualitative synthesis.

Fig 1: PRISMA flowchart of the selection of studies on artificial intelligence, neuroplasticity, and brain aging. Source: Authors Initially, records were identified in electronic databases and other complementary sources, totaling 404 studies. After removing duplicates, 356 records remained for analysis, of which 304 were excluded after reading the titles and abstracts because they did not meet the previously established eligibility criteria. In the eligibility stage, 52 articles were evaluated in full, resulting in the final inclusion of 25 studies in the qualitative synthesis. This process highlights the careful application of methodological criteria and reinforces the consistency of the findings discussed throughout this work.

Table 1 summarizes the main artificial intelligence methods and algorithms identified in the scientific literature applied to the study of brain aging and neurodegenerative diseases. The included studies highlight the diversity of computational approaches employed, ranging from classic machine learning algorithms to advanced deep learning techniques and adaptive systems, applied to different types of neurological, cognitive, and clinical data (LYALL et al., 2023) [19]; (WU et al., 2024) [40].

Table 1: Main artificial intelligence methods and algorithms applied to the study of brain aging and neurodegenerative diseases Author/Year Type of Study AI Method/Algorithm Main Application Disease/Context Lyall et al., 2023 [19] Narrative Review Supervised Machine Learning Predictive modeling and digital diagnosis Dementia and brain aging Wu et al., 2024 [40] Systematic review Deep learning, convolutional neural networks (CNNs) Prediction of brain age Healthy and pathological aging Rutkowski et al., 2018 [31] Methodological study Machine learning based on information geometry Digital EEG biomarkers Dementia Buscema et al., 2015 [6] Experimental study Adaptive AI systems (I-FAST) EEG-based diagnosis Alzheimer's disease Gutman et al., 2024 [13] Historical review Machine learning and predictive models Predictive medicine Neurology and aging De Oliveira, 2025 [10] Narrative review ML and DL algorithms Early detection Alzheimer's and Parkinson's Calderone et al., 2025 [7] Narrative review AI applied to neuromodulation Personalized therapies Neurodegenerative diseases Korade et al., 2024 [16] Interdisciplinary review AI integrated with neuroplasticity Cognitive optimization Mental health and aging Xiong, 2024 [41] Applied review Brain-Computer Interfaces (BCI) Communication and Alzheimer's and Parkinson's Fuentes-Mendoza et al., 2025 [11] Narrative review AI applied to connectome analysis Brain plasticity Brain tumors and aging Onciul et al., 2025 [26] Narrative review Multimodal machine learning Translational clinical research Neuroscience Gkintoni et al., 2025 [12] Systematic review EEG, AI, Virtual Reality Cognitive rehabilitation Mild Cognitive Impairment Malarkhodi et al., 2024 [20] Computational study Supervised Machine Learning Prediction of brain age Aging Natalizi et al., 2025 [24] Observational study ML algorithms Brain volumetric analysis Alzheimer Zhang et al., 2025 [42] Narrative review AI applied to cognitive exercise Cognition improvement Elderly population Sisubalan et al., 2025 [36] Narrative review AI in wearable devices Continuous monitoring Healthy aging It is observed that supervised machine learning and deep learning algorithms, especially convolutional neural networks, have been widely used for brain age prediction, identification of digital biomarkers, and early diagnosis of dementias, with emphasis on Alzheimer's disease (BUSCEMA et al., 2015) [6]; (RUTKOWSKI et al., 2018) [31]; (NATALIZI et al., 2025) [24]. In addition, emerging approaches such as brain-computer interfaces, AI-assisted neuromodulation, and connectome analysis broaden the scope of applications, reinforcing the role of artificial intelligence as a central tool in understanding, preventing, and managing neurodegenerative changes associated with aging (CALDERONE et al., 2025) [7]; (XIONG, 2024) [41]; (ONCIUL et al., 2025) [26].

De Oliveira (2025) [10] and Lyall et al. (2023) [19] point out that artificial intelligence has expanded the ability to identify subtle cognitive changes early on by integrating large volumes of clinical, digital, and neurophysiological biomarker data into predictive models. In this sense, Table 2 brings together the main evidence from the selected literature on the role of AI in the early detection of cognitive decline, mild cognitive impairment (MCI), and dementia, with an emphasis on strategies applied to Alzheimer's disease.

Table 2: Evidence on the role of artificial intelligence in the early detection of cognitive decline, mild cognitive impairment, and dementia (emphasis on Alzheimer's)

Author/Year Type of study Data/Main source AI approach Contribution to early detection Target condition Lyall et al., 2023 [19] Narrative review Digital health and clinical/multimodal data ML applied to predictive models Support for screening and risk stratification with digital tools Dementia (general)

De Oliveira, 2025 [10] Narrative review Clinical data, imaging, and digital signals ML and DL Summary of applications for early detection and assisted diagnosis Alzheimer's and Parkinson's Wu et al., 2024 [40] Systematic review Neuroimaging and population databases DL (incl. CNN) and ML Prediction of “brain age” as an early marker of risk and vulnerability Pathological aging/risk of dementia Rutkowski et al., 2018 [31] Methodological study EEG (digital biomarker) ML with information geometry Development of EEG biomarkers associated with task load and dementia Dementia Buscema et al., 2015 [6] Experimental/methodological study Raw (unprocessed) EEG AI system (I-FAST) Classification/diagnostic support for Alzheimer's disease based on EEG Alzheimer Rossini et al., 2020 [29] Biomarker study (review/neurophysiological approach)

Brain connectivity/biomarkers Integration with analytical models (AI support)

Supports early biomarkers of neurodegenerative decline Dementia/cognitive decline Natalizi et al., 2025 [24] Observational study Brain volumetry + cognitive profiles ML algorithms Prediction/stratification of cognitive impairment (amnestic) for Alzheimer's risk aMCI / Alzheimer's Gkintoni et al., 2025 [12] Systematic review EEG, VR, and neuromodulation rehabilitation and assessment Maps technologies for assessment/monitoring in MCI and rehabilitation MCI Medenica et al., 2024 [22] Review/applied study Neuropsychological assessments and neuropsychological Support for assessment, monitoring, and personalization (relevant for early identification)

Brain injury/rehabilitation (cognitive risk)

Sisubalan et al., 2025 [36] Narrative review Wearables and continuous monitoring AI in digital health Identification of risk patterns and subtle functional/cognitive changes Healthy aging/dementia risk Johnstone et al., 2025 [15] Consensus Nutrition and prevention, brain health Integration with digital models Guides prevention and risk screening (basis for applied AI)

Dementia prevention Onciul et al., 2025 [26] Narrative review Multimodal data in neuroscience AI/ML in synergy with neuroscience Translational integration for biomarkers and clinical applications Dementias and clinical neuroscience It is observed that early detection has been supported by approaches that combine machine learning and deep learning applied to multimodal data, with neuroimaging and brain age prediction standing out as risk markers, in addition to digital biomarkers derived from EEG and brain connectivity (WU et al., 2024) [40]; (RUTKOWSKI et al., 2018) [31]; (ROSSINI et al., 2020) [29]. For Alzheimer's in particular, studies and reviews describe models capable of supporting screening and diagnostic classification, as well as stratifying risk in individuals with amnestic MCI based on volumetric markers and cognitive profiles, contributing to more timely and personalized interventions (BUSCEMA et al., 2015) [6]; (NATALIZI et al., 2025) [24]; (GKINTONI et al., 2025) [12]. Complementarily, the incorporation of monitoring by wearables and digital health tends to strengthen the longitudinal surveillance of early signs, increasing sensitivity to subtle changes throughout aging (SISUBALAN et al., 2025) [36].

Styliadis et al. (2015) [37] and Cui et al. (2018) [8] demonstrate that cognitive, physical, or combined interventions can positively influence cognitive and neurophysiological outcomes in older adults, supporting the hypothesis that neuroplasticity remains modifiable throughout aging. In this scenario, Table 3 systematizes the selected evidence linking artificial intelligence, neuroplasticity, and intervention strategies (cognitive training, physical exercise, combined programs, and rehabilitation), highlighting how digital technologies and computational models can support monitoring, personalization, and therapeutic optimization.

Table 3: Relationship between artificial intelligence, neuroplasticity, and cognitive, physical, or combined interventions in elderly populations Author/Year Type of study Intervention (cognitive/physical/combined)

Role of AI/Technology Indicators linked to neuroplasticity Population/Context Oliveira et al., 2025 [25] Article (preventive approach)

Preventive and evidence-based interventions AI/technologies as support for care and prevention Neuroscientific basis for cognitive protection Elderly people / dementia prevention Styliadis et al., 2015 [37] Controlled study (eLORETA)

Computerized physical + cognitive training (combined)

Computerized training tools Changes in resting states (eLORETA) compatible with plasticity Older adults at risk for dementia Cui et al., 2018 [8] Intervention study Physical exercise (intervention associated with cognitive improvement)

Potential integration with digital monitoring Cognitive improvement associated with intervention (functional proxy)

Alzheimer's disease Zhang; Zhang; Luo, 2025 [42] Review AI-supported exercise/training physical/cognitive training Improvement in cognitive function (outcomes)

Elderly population Gkintoni et al., 2025 [12] Systematic review Cognitive rehabilitation and neuromodulation (includes VR/EEG)

Emerging AI in VR, EEG, neuromodulation Neurophysiological parameters (EEG) and response to training MCI/rehabilitation Jirayucharoensak et al., 2019 [14] Intervention study Game-based neurofeedback Digital system/feedback (basis for adaptive AI)

Modulation of cognitive performance (associated with learning/plasticity)

Healthy elderly and aMCI Buscema et al., 2015 [6] Experimental study — (diagnostic focus) AI applied to EEG (I-FAST)

Neurophysiological biomarkers useful for monitoring brain states Alzheimer Rutkowski et al., 2018 [31] Methodological study Cognitive tasks with ML load assessment ML for digital EEG biomarkers EEG biomarkers related to load/task Dementia/aging Fuentes-Mendoza et al., 2025 [11] Narrative review Rehabilitation/recovery associated with the connectome AI applied to the connectome Plasticity in networks (connectome)

Brain tumors (relevance to plasticity)

Reddy, 2025 [28] Book chapter Neurocognitive rehabilitation (technology)

AI for therapy personalization Optimization of strategies (functional plasticity)

Rehabilitation (applicable to older adults)

Medenica; Ivanovic; Milosevic, 2024 [22] Applied study/review Neuropsychological rehabilitation AI for rehabilitation and monitoring Therapeutic adjustment and longitudinal follow-up Brain injury/rehabilitation Sisubalan et al., 2025 [36] Review Promotion of healthy aging (activity/routine) AI + wearables Continuous monitoring of functional patterns (adaptation proxy)

Elderly/healthy aging Korade et al., 2024 [16] Interdisciplinary review Interventions for cognitive improvement and well-being AI–neuroplasticity integration Frames AI as a facilitator of cognitive enhancement Aging/mental health Qansuwa; Atalah; Salama, 2025 [27] Review Rehabilitation and equitable health AI/ML applied to Relationship between rehabilitation, plasticity, and algorithms Healthcare/rehabilitation systems Onciul et al., 2025 [26] Narrative review Clinical applications and research Multimodal AI Integration of biomarkers and interventions Clinical neuroscience Evidence converges toward a model in which AI can act as a facilitator of adaptive interventions and longitudinal follow-up, especially when integrated with neurophysiological measures (such as EEG) and digital training platforms, allowing the capture of changes associated with learning and functional reorganization (JIRAYUCHAROENSAK et al., 2019) [14]; (RUTKOWSKI et al., 2018) [31]. Combined physical-cognitive interventions and technology-supported exercises tend to be particularly relevant for elderly populations at risk for dementia, while approaches involving virtual reality, neuromodulation, and AI-assisted rehabilitation expand the possibility of therapeutic personalization and response assessment, bringing clinical practice closer to more accurate and data-driven strategies (GKINTONI et al., 2025) [12]; (ZHANG; ZHANG; LUO, 2025) [42]; (REDDY, 2025) [28].

Thus, the integration of AI and interventions aimed at neuroplasticity emerges as a promising axis for prevention and care in cognitive decline and dementia, although challenges remain in terms of methodological standardization and external validation for widespread application (ONCIUL et al., 2025) [26]; (QANSUWA; ATALAH; SALAMA, 2025) [27]. Graph 1 illustrates the distribution of the main emerging applications of artificial intelligence in the context of healthy aging, as well as the benefits, limitations, and ethical challenges discussed in the studies included in this systematic review. The categories were grouped according to the type of technology employed—digital technologies, wearable devices, neurotechnologies, and brain-computer interfaces— , in addition to clinical and preventive applications and ethical discussions associated with the use of artificial intelligence in elderly populations.

Source: Authors Fig 1: Distribution of emerging applications of artificial intelligence in brain aging and the main ethical and clinical challenges There is a greater concentration of studies related to digital technologies and the direct clinical applications of artificial intelligence, reflecting the growing use of digital health platforms, predictive algorithms, and clinical decision support systems in cognitive monitoring and early detection of cognitive decline (LYALL et al., 2023) [19]; (GUTMAN et al., 2024) [13].

Neurotechnologies, including EEG, neuromodulation, and neurofeedback, are also prominent, highlighting their potential for functional assessment and neuroplasticity-based interventions (RUTKOWSKI et al., 2018) [31]; (GKINTONI et al., 2025) [12]. To a lesser extent, but with growing relevance, brain-computer interfaces and wearable devices appear, associated with continuous monitoring and the personalization of preventive interventions (XIONG, 2024) [41]; (SISUBALAN et al., 2025) [36]. Finally, a significant number of studies address methodological limitations, ethical challenges, data privacy, and equity in access to technologies, emphasizing the need for regulation, clinical validation, and responsible approaches to the incorporation of artificial intelligence in brain aging care (ONCIUL et al., 2025) [26]; (QANSUWA et al., 2025) [27].

Discussion

Walhovd, Fjell, and Espeseth (2014) [39] emphasize that brain aging should not be understood as a linear and inevitable process of functional loss, but as a dynamic phenomenon influenced by biological, environmental, and behavioral factors throughout the course of life. In this scenario, the incorporation of artificial intelligence (AI) into neuroscience research on aging has allowed for a more refined analysis of complex patterns associated with cognitive decline and neurodegenerative diseases, expanding the capacity for prediction and early intervention (LYALL et al., 2023) [19]; (GUTMAN et al., 2024) [13].

The studies included in this review show that machine learning and deep learning algorithms have been widely used in the early detection of cognitive changes, especially through the analysis of neuroimaging, EEG, and multimodal clinical data (WU et al., 2024) [40]; (BUSCEMA et al., 2015) [6]; (NATALIZI et al., 2025) [24]. These approaches demonstrate significant potential for identifying individuals at risk in the early stages, such as mild cognitive impairment, favoring timely and personalized interventions (DE OLIVEIRA, 2025) [10].

In addition to diagnosis, the literature points out that AI has played an increasing role in supporting cognitive, physical, and combined interventions, with a direct impact on neuroplasticity in elderly populations (STYLIADIS et al., 2015) [37]; (CUI et al., 2018) [8]. Digital cognitive training platforms, technology-assisted physical exercise programs, and EEG-based neurofeedback systems show promise in stimulating functional reorganization and delaying cognitive decline (JIRAYUCHAROENSAK et al., 2019) [14]; (GKINTONI et al., 2025) [12].

Emerging applications of wearable devices and brain-computer interfaces further expand this scenario, allowing continuous monitoring, longitudinal assessment, and dynamic adaptation of interventions, especially in preventive and healthy aging contexts (SISUBALAN et al., 2025) [36]; (XIONG, 2024) [41]. However, despite technological advances, significant challenges remain related to methodological standardization, external validation of models, and effective integration of these tools into clinical practice (ONCIUL et al., 2025) [26].

Additionally, ethical aspects, such as data privacy, algorithmic bias, and equity in access to technologies, are widely discussed in the literature and represent important obstacles to the safe and responsible implementation of AI in health (QANSUWA; ATALAH; SALAMA, 2025) [27]; (BERNAL et al., 2024) [4]. Thus, there is a clear need for interdisciplinary approaches that reconcile technological innovation, scientific rigor, and ethical responsibility.

Conclusions

This systematic review shows that artificial intelligence has established itself as a central tool in research and care related to brain aging and neurodegenerative diseases. The findings demonstrate that machine learning and deep learning algorithms contribute significantly to the early detection of cognitive decline, especially in Alzheimer's disease, by integrating clinical, neurophysiological, and digital data. In addition, AI plays an important role in optimizing cognitive, physical, and combined interventions, favoring personalized strategies that enhance neuroplasticity in elderly populations. Digital technologies, wearable devices, neurotechnologies, and brain-computer interfaces expand the possibilities for continuous monitoring and prevention, bringing clinical practice closer to more accurate, data-driven models.

However, despite the promising potential, methodological, clinical, and ethical challenges persist that limit the widespread application of these technologies. Thus, future research should prioritize robust clinical validation, standardization of methods, and incorporation of ethical principles to ensure that artificial intelligence is used safely, equitably, and effectively in the context of brain aging.