Interpretable Machine Learning for the classification of Mild Cognitive Impairment patients using actigraphy data - MSc Thesis

Aug 5, 2026

Upload by Eden

Early and accurate diagnosis of Mild Cognitive Impairment (MCI) is crucial because it often precedes Alzheimer’s disease, enabling timely intervention that may slow cognitive decline. This thesis develops a non-invasive digital medicine approach using actigraphy, which captures activity and rest-activity cycles via wearable devices, to support earlier detection and better management. A study pipeline is built to extract parametric cosinor features and train interpretable machine learning models for MCI vs. Normal classification, analyzed with SHAP. Results on the ALBION dataset indicate lower Mesor, Acrophase, and Amplitude align with MCI, with model performance reaching expected specificity 0.88 at sensitivity 0.5 on unseen data.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages67

Reading Time2 h 49 min

AccessFree to download

Abstract

Background and motivation

Data source and feature extraction

Machine learning pipeline and interpretation

Results and implications

NATIONAL AND KAPODISTRIAN UNIVERSITY OF ATHENS SCHOOL OF SCIENCE DEPARTMENT OF INFORMATICS AND TELECOMMUNICATIONS INTERDISCIPLINARY MSc PROGRAM "DATA SCIENCE AND INFORMATION TECHNOLOGIES" SPECIALIZATION “BIOINFORMATICS – BIOMEDICAL DATA SCIENCE” MSc THESIS Interpretable Machine Learning for the classification of Mild Cognitive Impairment patients using actigraphy data Marios A. Gavrielatos Supervisor: Elias S. Manolakos, PhD, Professor, Department of Informatics and Telecommunication, National and Kapodistrian University of Athens ATHENS ΕΘΝΙΚΟ ΚΑΙ ΚΑΠΟΔΙΣΤΡΙΑΚΟ ΠΑΝΕΠΙΣΤΗΜΙΟ ΑΘΗΝΩΝ ΣΧΟΛΗ ΘΕΤΙΚΩΝ ΕΠΙΣΤΗΜΩΝ ΤΜΗΜΑ ΠΛΗΡΟΦΟΡΙΚΗΣ ΚΑΙ ΤΗΛΕΠΙΚΟΙΝΩΝΙΩΝ ΔΙΑΤΜΗΜΑΤΙΚΟ ΠΡΟΓΡΑΜΜΑ ΜΕΤΑΠΤΥΧΙΑΚΩΝ ΣΠΟΥΔΩΝ" ΠΙΣΤΗΜΗ ΔΕΔΟΜΕΝΩΝ ΚΑΙ ΤΕΧΝΟΛΟΓΙΕΣ ΠΛΗΡΟΦΟΡΙΑΣ" ΕΙΔΙΚΕΥΣΗ “ΒΙΟΠΛΗΡΟΦΟΡΙΚΗ – ΕΠΙΣΤΗΜΗ ΒΙΟΙΑΤΡΙΚΩΝ ΔΕΔΟΜΕΝΩΝ” ΔΙΠΛΩΜΑΤΙΚΗ ΕΡΓΑΣΙΑ ΜΕΤΑΠΤΥΧΙΑΚΟΥ Ερμηνεύσιμη μηχανική μάθηση για την ταξινόμησηασθενών με ήπια γνωσιακή διαταραχήχρησιμοποιώντας δεδομένα ακτιγραφίας Μάριος Α . Γαβριελάτος Επιβλέπων : Ηλίας Σ . Μανωλάκος, Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικό και Καποδιστριακό ΠανεπιστήμιοΑθηνών ΑΘΗΝΑ ΙΟΥΝΙΟΣ 2024 MSc THESIS Interpretable Machine Learning for the classification of Mild Cognitive Impairment patients using actigraphy data Marios A. Gavrielatos SRN: 7115152100023 Supervisor: Elias S. Manolakos, Professor, Department of Informatics and Telecommunications, National and Kapodistrian University of Athens EXAMINATION COMMITTEE: Elias S. Manolakos Nikolaos Skarmeas Stavros Perantonis Professor, Department of Informatics and Telecommunications, National and Kapodistrian University of Athens Professor of Neurology, Department of Medicine, National and Kapodistrian University of Athens Research Director, National Center for Scientific Research“Demokritos” ATHENS ΔΙΠΛΩΜΑΤΙΚΗ ΕΡΓΑΣΙΑ ΜΕΤΑΠΤΥΧΙΑΚΟΥ Ερμηνεύσιμη μ ηχανική μάθηση για την ταξινόμηση ασθενών με ήπιαγνωσιακή διαταραχή χρησιμοποιώντας δεδομένα ακτιγραφίας Μάριος Α . ΓαβριελάτοςΑ . Μ .: 7115152100023 ΕΠΙΒΛΕΠΩΝ: Ηλίας Σ . Μανωλάκος, Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών ΕΞΕΤΑΣΤΙΚΗ ΕΠΙΤΡΟΠΗ: Ηλίας Σ . ΜανωλάκοςΝικόλαος ΣκαρμέαςΣταύρος Περαντώνης Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών,Εθνικό και Καποδιστριακό Πανεπιστήμιο ΑθηνώνΚαθηγητής Νευρολογίας, Ιατρική Σχολή, Εθνικό καιΚαποδιστριακό Πανεπιστήμιο ΑθηνώνΔιευθυντής Ερευνών, ΕΚΕΦΕ, Δημόκριτος ΑΘΗΝΑ ΙΟΥΝΙΟΣ 2024 ABSTRACT Early and accurate diagnosis of Mild Cognitive Impairment (MCI) is crucial as it is often a precursor stage to Alzheimer’s disease, allowing for timely intervention and treatment that can potentially slow the progression of cognitive decline. A non-invasive diagnostic approach would be highly beneficial, as it would be more comfortable and accessible for patients, increasing the likelihood of early detection and enabling prompt action to preserve cognitive function and quality of life. Developing easy and non-invasive digital medicine methods for diagnosing MCI could lead to earlier intervention and better management of this condition, which is a major step towards preventing or delaying the onset of Alzheimer’s disease and its devastating effects on individuals and their families. Actigraphy, which involves the continuous monitoring of physical activity and rest-activity cycles using wearable devices, can offer valuable insights into the presence of MCI. This graduate thesis project was conducted in collaboration with the Aiginition Longitudinal Biomarker Investigation of Neurodegeneration (ALBION) study at the University of Athens, Greece. 7-day actigraphy time series data from individuals with normal cognitive function and those with MCI were collected via an actigraph device. We extracted parametric features, including Mesor (average activity), Amplitude (highest magnitude of activity), and Acrophase (the timing of the largest pe

FAQs

Why is early diagnosis of Mild Cognitive Impairment important?

Mild Cognitive Impairment (MCI) is often a precursor to Alzheimer’s disease. Early detection enables timely intervention that can potentially slow cognitive decline and improve quality of life.

How does the thesis use actigraphy data for classification?

The work uses 7-day actigraphy time series from normal and MCI participants and extracts parametric features via multi-cosinor analysis. These features feed a machine-learning pipeline to classify MCI vs. Normal based on clinician-assigned diagnostic labels.

Which actigraphy features and model interpretation findings are reported?

For the ALBION dataset, lower Mesor, Acrophase, and Amplitude values are associated with the MCI class. SHAP-based analysis identifies the most informative features and their effects on classification outcomes.

Aug 5, 2026

Upload by Eden