This review paper explores recent advances in machine learning approaches with emphasis on deep learning techniques for non-invasive cognitive impairment detection.We examine various non-invasive indicators of cognitive decline, including speech and language, facial, and motoric mobility. The paper provides an overview of relevant datasets, feature-extracting techniques, and deep-learning architectures applied to this domain. The paper comprises 60 peer-reviewed papers that mainly utilize deep learning models to detect cognitive impairment conditions. We have analyzed the performance of different methods across modalities and observed that speech and language-based methods generally achieved the highest detection performance. Studies combining acoustic and linguistic features tended to outperform those using a single modality. Facial analysis methods showed promise for visual modalities but were less extensively studied. Most papers focused on binary classification (impaired vs. non-impaired), with fewer addressing multi-class or regression tasks. Transfer learning and pre-trained language models emerged as popular and effective techniques, especially for linguistic analysis. Despite significant progress, several challenges remain, including data standardization and accessibility, model explainability, longitudinal analysis limitations, and clinical adaptation. Lastly, we propose future research directions, such as investigating language-agnostic speech analysis methods, developing multi-modal diagnostic systems, and addressing ethical considerations in AI-assisted healthcare. By synthesizing current trends and identifying key obstacles, this review aims to guide further development of deep learning-based cognitive impairment detection systems to improve early diagnosis and ultimately patient outcomes.

Deep machine learning models including Convolutional Neural Networks (CNN) have been successful in the detection of Mild Cognitive Impairment (MCI) using medical images, questionnaires, and videos. This paper proposes a novel Multi-branch Classifier-Video Vision Transformer (MC-ViViT) model to distinguish MCI from those with normal cognition by analyzing facial features. The data comes from the I-CONECT, a behavioral intervention trial aimed at improving cognitive function by providing frequent video chats. MC-ViViT extracts spatiotemporal features of videos in one branch and augments representations by the MC module. The I-CONECT dataset is challenging as the dataset is imbalanced containing Hard-Easy and Positive-Negative samples, which impedes the performance of MC-ViViT. We propose a loss function for Hard-Easy and Positive-Negative Samples (HP Loss) by combining Focal loss and AD-CORRE loss to address the imbalanced problem. Our experimental results on the I-CONECT dataset show the great potential of MC-ViViT in predicting MCI with a high accuracy of 90.63% accuracy on some of the interview videos.

Background: Biological aging varies substantially among individuals of the same chronological age, yet biological aging biomarkers require biospecimen collection and laboratory processing, limiting scalability for repeated monitoring. Objective: This study evaluated whether consumer smartwatch-derived digital biomarkers are associated with the pace of biological aging, quantified using DunedinPACE, a biomarker that measures biological aging rates. Methods: Baseline data were analyzed from the Healthy Minds for Life study within the Precision Aging Network. Among 1305 enrolled participants, 621 community-dwelling adults aged 50–79 years had both DunedinPACE and smartwatch data available. Smartwatch-derived biomarkers were grouped into mobility, cardiorespiratory physiology, sleep, and physiological stress domains. Associations with DunedinPACE were examined using Spearman correlations and adjusted linear regression models. Results: Significant associations were observed across all four domains. Daily step count showed the strongest mobility-domain association with DunedinPACE (r = −0.25, p < 0.01). Mean sleep peripheral oxygen saturation showed the strongest sleep-domain association (r = −0.22, p < 0.01). Resting heart rate showed the strongest physiological stress-domain association (r = 0.24, p < 0.01). Adjusted regression models supported these findings, with lower daily mobility, poorer sleep oxygenation and efficiency, greater wake after sleep onset, and higher resting heart rate showing the most consistent associations with higher DunedinPACE values. Conclusions: Smartwatch-derived biomarkers captured a multidomain profile of accelerated biological aging characterized by lower mobility, poorer sleep health, and elevated physiological stress. Consumer wearables can provide low-burden, real-world indicators of biological vulnerability, although longitudinal studies are needed before clinical use.