Atıf Sayısı
1
Görüntüleme
3
İndirme
3
Bu çalışmada multipl skleroz (MS)’ te akustik ses karakteristiklerinin betimlenmesi ve ses değişimi kriterlerinin saptanmasıyla MS’te erken tanı amaçlanmaktadır. Bu amaç doğrultusunda çalışmamızda hastalığa özgü bir foniatrik reedükasyon yöntemi geliştirilerek MS’ te erken tanıda kullanılabilecek ses değişimi kriterlerini belirleyici paremetrelerin oluşturulması planlanmıştır. Çalışma Mart 2013-Mart 2014 yılları arasında Sultan Abdülhamid Eğitim ve Araştırma Hastanesi Nöroloji Kliniğine gelen McDonald 2010 tanı kriterlerine göre hastalığı ilk kez teşhis edilen veya hastalık süresi 0-10 yıl arasında olan 55 (30 kadın+25 erkek) MS hastası ile herhangi bir nörolojik ve ses patolojisi olmayan 55 (33 kadın+ 22 erkek) sağlıklı gönüllü üzerinde yürütülmüştür. Hastaların yaşları 18 ile 55 (ort: 29,86, std sapma: 8,97), sağlıklı bireylerin yaşları ise 18 ile 50 (ort: 32,55, std sapma: 10,79) arasında değişmektedir. Seslerin akustik incelenmesinde The Multi-Dimensional Voice Program (MDVP) akustik ses analizi programı, seslerin öznitelik sınıflandırmasında ise k-En Yakın Komşu (k-EYK) algoritması kullanılmıştır. k-EYK sınıflandırıcısının kullanıldığı bu çalışmada MS ve sağlıklı bireylere ait ses özniteliklerine bakılarak hasta-sağlıklı ayrımı yapmada en yüksek ses tanımlama başarımı sürdürülmüş fonasyonlar /a/,/e/ ve /u/ seslerin (%88.76) özniteliklerinden elde edilmiştir. k-EYK algoritması kullanılarak MS hastalarında görülen patolojik seslerin sağlıklı bireylerden ayrılmasında başarılı sonuçlar verdiği gözlenmiştir. Sürdürülmüş fonasyonlar, kelimeler ve cümle bazında genel bir değerlendirme yapıldığında MS hastalarının /a/ ve /e/ vokallerini belirgin derecede her üç grupta da sağlıklı bireylere göre patolojik seslettiği görülmüştür.
In this study, multiple sclerosis (MS) is aimed at early diagnosis in MS with the description of the acoustic sound characteristics and the determination of the criteria for sound change. For this purpose, our work has developed a phoniatric reeducation method that is specific to the disease and is planned to create parameters that determine the criteria for sound exchange that can be used in early diagnosis in MS. The study was conducted on 55 (30 women + 25 men) with MS patients who were diagnosed for the first time or with a duration of 0-10 years, according to the 2010 diagnostic criteria from the McDonald's Neurology Clinic of Sultan Abdulhamid Education and Research Hospital between March 2013 and March 2014 and 55 (33 women + 22 men) with a healthy volunteer without any neurological and sound pathology. The age of patients ranges from 18 to 55 (medium: 29,86, std deviation: 8,97), while the age of healthy individuals ranges from 18 to 50 (medium: 32,55, std deviation: 10,79). The Multi-Dimensional Voice Program (MDVP) acoustic sound analysis program is used in the acoustic study of sounds, while the classification of sounds is used by the k-Last Neighbor (k-EYK) algorithm. The k-EYK classificator is used in this study, taking into account the sound properties of MS and healthy individuals, not making a patient-healthy distinctionHigh sound identification success was obtained from the properties of continued fonations /a/,/e/ and /u/ sounds (88.76%). The k-EYK algorithm has been observed to give successful results in the separation of pathological sounds seen in MS patients from healthy individuals. A general assessment on the basis of continued fonations, words and phrases found that MS patients expressed their /a/ and /e/ vocals significantly in each of the three groups compared to healthy individuals.
The aim of this study is to delineate the acoustic charecteristic of Multiple Sclerosis (MS). In accordance with this goal, it was planned the affordance of phonaiatric table specific to the disease, and from this point of view, development of parameters that determine the phonetic change criterias which can be used in the diagnosis in MS. The study was conducted at Sultan Abdülhamid Trainin and Research Hospital Department of Neurology between March 2013 and March 2014. This research was carried out 55 RRMS (30 female+25 male) patients diagnosed for the first time according to McDonald's 2010 diagnostic criteria,or duration of disease between 1-10 years and EDSS≤3. Control group consisted of 55 (33 female+22 male) healthy volunteers without any neurological and voice pathology. Acoustic investigation of sound was analyzed by The Multi-Dimensional Voice Program (MDVP). Classification of sound features that are extracted from the voice samples of each subject were fed into k-nearest neighbor algorithm (k-EYK). In this study using k-EYK classifier, the highest voice recognition success was achieved sustained phonations that / a /, / e / and / u / sounds features (88.76%), in making patient-healthy discrimination based on the sound features of MS and healthy individuals. It was observed that using k-EYK algorithm has been succesful in separating pathological sounds which seen in MS patients from healthy individuals. It was found that MS patients had pathological vocalization of / a / and / e / vocals at significantly higher levels in all three groups compared to healthy subjects when a general evaluation is made based on sustained phonations,words and sentences.
Dergi Türü : Uluslararası
Makale Detay 
Atıf Yapanlar
Benzer Makaleler
PDF Görüntüle
Dergi Bilgisi
Eseri Dinleyin
Bu Sayfayı Yazdırın
Paylaş
Mendeley
