Πλοήγηση ανά Συγγραφέας "Tsichlaki, Styliani"

Τώρα δείχνει 1 - 2 of 2
  • Μικρογραφία εικόνας
    Τεκμήριο
    Development of a web-based platform for cognitive training.
    (Τ.Ε.Ι. Κρήτης, Τεχνολογικών Εφαρμογών (Σ.Τ.Εφ), Τμήμα Μηχανικών Πληροφορικής Τ.Ε., 2019-05-21) Tsichlaki, Styliani; Τσιχλάκη, Στυλιανή
    In this thesis, a web-based platform for cognitive training was developed, as a tool for a) the training of cognitive functionality of high-risk individuals or individual facing early Mild Cognitive Impairment (eMCI), and b) the support of clinical professionals’ examining eMCI individuals. The cognitive training platform was developed on Moodle, a Learning Management System (LMS), which is based on Social Constructivist Pedagogy. Towards this scope, a series of interactive, specialized exercises were designed as described by Dr Foteini Kounti, cognitive psychologist and co-founder of The Greek Association of Alzheimer’s Disease and Related Disorders1, [1][2][3][4] 2. Mild Cognitive Impairment (MCI) is described as the intermediate stage of cognitive impairment [2], which concerns a group of individuals who may or may not be at high risk for developing dementia[5][6]. Early Mild Cognitive Impairment (eMCI) describes the earliest and mildest symptomatic phase of Mild Cognitive Impairment (MCI). In these early stages, abnormal brain functional capabilities have already occurred in the form of subtle changes in the life of the individual [7][8]. The early and accurate detection and intervention of eMCI is of high importance for the individual’s health[8]. However, the pathology of eMCI remains mostly unknown[8], which can be challenging to the clinical diagnosis[8]. To this end, more work needs to be done on applications related to eMCI. The cognitive training platform include training exercises which are focusing on activities of daily living, with each one to be consist of three difficulty levels. Each level consists of nine specialized exercises focused on eMCI individuals, some of which include some sub-exercises, all of which are developed using vivid interactive contents. The platform also offers consistency in workflow among the exercises, as well as detailed descriptions to avoid misunderstanding. At the end of each exercise and while users have acted on what they were asked to do, they receive the appropriate feedback. All the exercises, of each level, are evaluated concluding to an overall level score. Users also have the ability to access their grades and have a detailed view of their scores on each exercise. Finally, the platform has been developed with a simple look and feel, while it contains a wide range of user-friendly interactive content.
  • Μικρογραφία εικόνας
    Τεκμήριο
    Type 1 diabetes hypoglycemia prediction based on continuous glucose monitoring and heart rate.
    (ΕΛ.ΜΕ.ΠΑ., Σχολή Μηχανικών (ΣΜΗΧ), ΠΜΣ Πληροφορική και Πολυμέσα, 2022-01-14) Tsichlaki, Styliani; Τσιχλάκη, Στυλιανή
    Hypoglycemia is a condition that arises when blood glucose levels decrease below 60 mg/dL. This incident can occur due to a variety of causes, such as taking additional doses of insulin, skipping meals, or over-exercising. In this thesis, we examined the use of biosignals and other measurements provided by a wearable device along with self-assessment parameters, for the development of a hypoglycemia predictive model. Glucose measurements were captured by a clinically certified continuous glucose monitoring sensor, while the predictive model was trained using machine learning techniques. In addition, a diabetes management mobile application was developed and used for the required data collection from the patient, i.e. finger-stick glucose measurements, insulin doses, food and exercise, mood and a diabetes distress level questionnaire. Hypoglycemia threshold was defined as a blood glucose value below 70 mg/dL. The results of the hypoglycemia prediction model that was developed revealed, for patient with ID 575, that the 30-minute prediction curve held an RMSE score of 20.25 mg/dL and a MAE score of 13.26 mg/dL. Finally, we sincerely consider that the proposed model produces useful and applicable outcomes for T1D patients, and we suggest that a 30-minute RMSE of 20.25 mg/dL can provide a basis for avoiding a potentially critical, for the patient’s health, hypoglycemic episode.