A Novel Approach for the Early Diagnosis of Neurocognitive Disorders (NCDs): Integration of EEG Parameters, Clinical and Radiological Data with Advanced Deep Learning Models

Neurocognitive disorder represents a significant global health challenge, ranking as the seventh leading cause of death worldwide. It is estimated that the number of people with dementia will increase from 57.4 million cases globally in 2019 to 152.8 million cases by 2050.

With the global population aging, the early and accurate diagnosis of neurocognitive disorders has become increasingly difficult. This challenge is driven by multiple factors, including the subtle and non-specific nature of early symptoms, limited public awareness of early dementia signs, and the high cost and limited accessibility of current diagnostic methods. As a result, there is a critical need for affordable, non-invasive, and widely accessible diagnostic tools that can facilitate early and accurate identification of neurocognitive disorders.

In recent years, deep learning and artificial intelligence (AI) methods have gained attention for their potential in the early diagnosis of neurocognitive disorders. According to a growing body of literature, electroencephalography (EEG)—especially quantitative EEG (qEEG)—can detect brain changes associated with neurocognitive disorders, even in the very early (mild cognitive impairment) stages. In some cases, EEG has been found to be more sensitive than MRI or CT scans. However, current EEG data remains limited and carries its own challenges.

Therefore, integrating EEG data with clinical tools and advanced deep learning algorithms holds great promise for improving the accuracy, efficiency, and accessibility of neurocognitive disorder diagnosis. This approach could also prove more cost-effective and scalable, making it suitable for widespread application.

The purpose of this study is to develop and evaluate deep learning and AI models for diagnosing minor and major neurocognitive disorders (Alzheimer’s Disease, Vascular Dementia, Frontotemporal Dementia, and Dementia of Lewy body type) using quantitative EEG (qEEG) features, alongside clinical data and known risk factors of NCD.

The objective is to leverage the power of deep learning and AI to uncover complex relationships between input variables (EEG and clinical features) and clinical outcomes. This includes identifying latent computational EEG biomarkers within high-dimensional datasets that are not easily discernible by traditional methods.

The study aims to support more accurate clinical decision-making, particularly for the early diagnosis of NCD, and to differentiate affected individuals from healthy age-matched controls. Early diagnosis will allow timely treatment, potentially transforming patient outcomes and alleviating the rising burden of neurocognitive disorders on society.