Brain volume decrease is usually connected to neurodegeneration and aging. In this environment,an important percentage of elderly persons suffer from mild cognitive impairment (MCI), a kind ofdementia that can lead to Alzheimer’s disease (AD). Since the symptoms of cognitive impairment arescarcely discernible, developing a safe and effective method for early MCI detection has emerged asan important challenge. According to this regard, numerous cognitive training tests can be targetedto help aging people retain a good quality of life, especially in the case of fragility disorders. AConsortium to Establish a Registry for Alzheimer’s Disease (CERAD) task was initially created todetect the early stages of AD. This task specifically targets various tests related to specific cognitivedomains. However, it has since developed into a popular diagnostic tool for many kinds of dementia,such as MCI. Several low-cost equipment, such as electroencephalography (EEG) and heart ratevariability (HRV), may be useful for predicting MCI. On the other side, various machine learning(ML) models can be employed to extract/analyse relevant features from biomedical and physiologicalsignals, especially in the context of anomaly detection and classification. To this regard, we developeda new method based on ML models to categorize MCI and healthy control (HC) patients duringthe CERAD task using EEG and HRV multimodal data. Our dataset includes 15 subjects who wererandomly assigned to training and testing groups of 7 HC and 8 MCI, respectively. Our raw EEGand HRV data are analyzed to extract time, frequency, and non-linear features. A scaling step isemployed to reduce the significant disparity between features. For the classification task, five MLmodels are evaluated, including support vector machine (SVM), k-nearest neighbors (KNN), decisiontree (DT), random forest (RF), and gradient boosting (GB). To enhance accuracy, a hybrid ML modelwith a voting system is developed, combining the top ML models with the highest accuracy rates.A comparison step is performed between the use of ML and hybrid ML models. The experimentalfindings demonstrated the efficacy of our proposed technique, which included a hybrid ML model. Anaverage accuracy of 93.86%, a sensitivity of 93.87%, and a specificity of 93.53% are achieved. Theobtained results allow one to conclude that the first CERAD test plays a prominent role as a novelbiomarker with an ultra-short duration for early MCI identification through the combination of EEGand HRV signals.

The need for multi-modal data in detecting complex relationships between physiological processes to improve anomaly characterization becomes evident with the growing of medical records modalities. Robust fusion techniques are not often used for biomedical data, and current multi-modal approaches need to be employed more effectively. This makes it possible to early identify a variety of anomalies as Alzheimer's disease (AD). To possibly prevent AD, it is essential to identify mild cognitive impairment (MCI) patients who are more likely to develop this disease. Our study makes use of the Consortium to Establish a Registry for Alzheimer's Disease (CERAD). This protocol has become a prevalent clinical method for a variety of dementias, including MCI. Numerous modalities including electroencephalography (EEG) and heart rate variability (HRV) can be used as helpful tools for MCI diagnosis. In this research, we developed a new deep learning (DL) method using the fusion of EEG and HRV signals acquired during the CERAD task. Using our data, a Transformer architecture was adapted to categorize participants into MCI and healthy control (HC). Our experimental results show that the proposed method has a benefit over the state-of-the-art in terms of classification accuracy. An accuracy test of 95.02%, a sensitivity of 95.47%, an F1 score of 95.03%, and a precision of 94.57% were achieved due to the proposed method.