Multi-dimensional electrical impedance dataanalysis for amyotrophic lateral sclerosisdisease diagnosis
Başlık çevirisi mevcut değil.
- Tez No: 761311
- Danışmanlar: PROF. VİSAKAN KADİRKAMANATHAN
- Tez Türü: Doktora
- Konular: Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Computer Engineering and Computer Science and Control
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2020
- Dil: İngilizce
- Üniversite: The Unıversıty Of Sheffıeld
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: Belirtilmemiş.
Özet
Özet yok.
Özet (Çeviri)
Amyotrophic Lateral Sclerosis (ALS) is the most common type of motor neurone disease (MND), which affects the motor neurones that control the voluntary skeletal muscles. ALS patients with bulbar involvement experience weaknesses in muscles that control speech, swallowing, and breathing. Recent attempts in using Electrical Impedance (EI) techniques have shown promising results for early diagnosis of the bulbar involvement of the ALS disease. A non-invasive, handheld EIM device is being developed as a potential biomarker for ALS bulbar involvement by a group in Sheffield. The device has a novel electrode configuration setting that employs a combination of eight electrodes used on both surfaces of the tongue to obtain the spectral EI measurements. This thesis proposes analysis of the acquired data using machine learning and system identification methods. The aim is to find patterns in data to assess bulbar dysfunction for diagnosis, and longitudinal assessment of the disease. A data specific outlier removal algorithm was implemented as part of preliminary data analysis, and electrode configurations that did not comply with the EI characteristics of the tongue muscle along the full spectrum are eliminated. Classification of the patients and healthy volunteers demonstrated that the novel electrode configurations were capable of differentiating the two groups. Classification using the feature selection method not only gave better accuracy, but also revealed the most discriminatory frequency combination for disease diagnosis. Impedance data was modeled with an Integrated Fractional-Order Transfer Function. A model selection scheme over the model space was employed to avoid underestimation, overestimation, and over-parametrization. Estimated parameters were used to identify disease related patterns. Analysis shows the device's capacity to diagnose the disease and the non-parametric data classification has shown better accuracies than the parametric data classification.
Benzer Tezler
- Kısmi yüklü dalgakılavuzları ile mikrodalga cihazı tasarımının ters saçılma ve optimizasyon problemi olarak incelenmesi
Analysis of partially filled waveguides for the synthesis of microwave devices as an inverse and optimization problem
AHMET AYDOĞAN
Doktora
Türkçe
2017
İstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
Yükleniyor...
- Derin öğrenme destekli dielektrik rezonatör anten tasarımı
Deep learning assisted dielectric resonator antenna design
FİDAN GAMZE KIZILÇAY
- Ultra geniş bant uygulamaları için esnek papyon anten tasarımı
Design of flexible bow-tie antenna for ultrawide band applications
AYKUT CAN DEMİRBAŞ
Yüksek Lisans
Türkçe
2026
Sakarya ÜniversitesiElektrik-Elektronik Mühendisliği Ana Bilim Dalı
Yükleniyor...
- Batı Anadolu'nun özdirenç yapısının elektrik-elektromanyetik yöntemler ile araştırılması
The investigation of resistivity structure of Western Anatolia using electrical and electromagnetic methods
MURAT BAYRAK