The Effects of endogenous and exogenous risk factors in patients with Alzheimer's and Parkinson's diseases using clinical indexes and endophenotypes (biomarkers) as inputs to artificial intelligence

The Effects of endogenous and exogenous risk factors in patients with Alzheimer’s and Parkinson’s diseases using clinical indexes and endophenotypes (biomarkers) as inputs to artificial intelligence (PREDICT-NEURODEGEN, PNRR-MAD-2022-12376415) project investigated the hypothesis that statistically modeling (1) the endogenous and exogenous risk factors, (2) the presence of the amyloid and alpha-synuclein pathological aggregates in CSF, (3) brain connectivity, and (4) clinical and biological indices such as vigilance, general motility, the sleep-wake cycle, and genomic instability and DNA damage MAY correlate with ("predict") the clinical condition of mild cognitive impairment (MCI) and dementia in patients with Alzheimer's disease (AD) and Parkinson's disease (PD). Cognitively intact healthy (Nold) participants served to define "abnormalities" in the markers derived from the hospital setting. AD and PD neuropathological hallmarks in the above diseases were measured from cerebrospinal fluid (CSF), peripheral blood lymphocytes, and by SPECT mapping. Brain connectivity was assessed using MRI biomarkers and resting-state EEG. Vigilance, motility, and the sleep-wake cycle were investigated using scales administered in the hospital setting and wearable devices and tablets used in patients' homes for 1 week. Endogenous risk factors included recurrent genotypes associated with sporadic forms of AD and PD (e.g., APOE, GBA, etc.), age, sex, and the level of general "frailty" in the geriatric sense. Among the exogenous risk factors, level of education and lifetime intellectual activity (as indicators of cognitive reserve), diet, degree of sedentary lifestyle, socioeconomic status, and social and family inclusion were considered. The relationships among the above variables, with particular interest in the influence of endogenous and exogenous disease risk factors on the clinical manifestations of AD and PD, were studied using both linear statistical modeling and machine learning tools. The main design hypotheses were as follows: (1) MRI and EEG biomarkers might be different and specific for AD patients with MCI and dementia compared with PD patients with MCI and dementia, contributing to the understanding of the types of brain dis-connectivity related to cognitive deficits in the two neurodegenerative diseases and to the level of oxidative DNA damage and genomic instability. Such biomarkers could serve as reference endpoints for tertiary prevention strategies and for monitoring disease progression in hospital settings. In addition, they were correlated with measures of vigilance, motor and cognitive activity, and sleep-wake rhythms detected through telemonitoring in patients' homes. These telemonitoring measures could also serve as baseline indices for tertiary prevention strategies and for monitoring the clinical manifestations of the two neurodegenerative diseases. (2) MRI and EEG biomarkers specific to the two neurodegenerative diseases might be influenced by the aforementioned endogenous and exogenous risk factors. The results indicate how these factors should be appropriately accounted for in prevention and monitoring strategies based on MRI and EEG biomarkers. All ethical principles for biomedical research set out in the Declaration of Helsinki were followed. The raw project data and articles reporting the project results, published in peer-reviewed international journals, were archived with open access. In accordance with national regulations, all individual data were anonymized, and participants' identities were always protected.