E-Locus - Institutional Repository of the University of Crete

Home Search

Results - Details

Search command : Author="Ζαγανάς" And Author="Ιωάννης"

Current Record: 15 of 36

[Add to Basket]

Identifier

000447756

Title

Ιατρικοί και ψυχολογικοί παράγοντες που επηρεάζουν την έναρξη και πρόοδο της ήπιας γνωστικής διαταραχής : Συγχρονικά και διαχρονικά δεδομένα

Alternative Title

Medical and psychosocial factors affecting the onset and progression of mild cognitive impairment

Author

Ζιώγα, Μαρία

Thesis advisor

Σίμος, Παναγιώτης

Reviewer

Ζαγανάς, Ιωάννης
Μπάστα, Μαρία

Abstract

A variety of processes associated with brain aging are responsible for cognitive difficulties or deficits, as people reach advanced age. Clinical significance of acquired cognitive deficits results when (a) they manifest to an unexpected degree depending on the age and educational level of the individual, (b) they complicate daily functioning or cause significant subjective disturbances that affect the quality of life, and / or (c) are considered risk factors for developing dementia. The present study aimed to examine, through machine learning techniques, sociodemographic, behavioral, neuropsychological and medical factors associated with diagnosis of such deficits in a large cross-sectional sample. To this end, a database comprised of 812 participants (mean age 72 years) was created by integrating and homogenizing data from 3 existing cohorts, from rural and urban areas of Heraklion in Crete was created. The sample included non-cognitively impaired persons (NCI, n = 358), individuals diagnosed with Mild Cognitive Impairment (MCI, n = 294) and patients with dementia (n = 151). A total of 4 machine learning models were developed, using the Random Forest algorithm for effective cognitive impairment classification. For classification of MCI and NCI, 60 neuropsychological, socio-demographic and medical variables, determined algorithmically, yielded a balanced accuracy of 87%. Finally, concerning the classification of MCI and Dementia, 20 algorithmically determined neuropsychological variables yielded a balanced accuracy of 80%.

Language

Greek, English

Subject

Dementia

Machine learning

Άνοια

Γνωστικά ελλείμματα

Μηχανική μάθηση

Issue date

2022-03-30

Collection

School/Department--School of Medicine--Department of Medicine--Post-graduate theses