Detection of mild cognitive impairment using a counterpropagation network based system. An e-health solution

Abstract

Dementia is considered to be the emergent and uncontrolled epidemic of nowadays, and one of the major socio-sanitary challenges our societies face. Alzheimer's disease (AD) is the most prevalent cause of dementia, and a progressive disorder. Thus, the detection of mild cognitive impairment (MCI), syndrome equivalent to the prodromal stage of AD, is important. It improves the early diagnosis of AD and allows treatment in the initial stages to begin, which can stabilize and delay the evolution of the pathology. MCI diagnosis is a difficult task, essentially in primary care (PC) setting and is underdiagnosed. We propose a simple solution to assist in MCI diagnosis, a CPN based system using a combination of neuropsychological scales and some modifying factors. We present our preliminary results exploiting the public ADNI dataset. This intelligent system is embodied in a clinical virtual environment, EDEVITAL ZH, providing an e-health solution for universal diagnosis of MCI, AD and other dementias. EDEVITALZH guides clinicians through their workflow and aids in diagnosis, prognosis and follow-up tasks.