Akademik Veri Yönetim Sistemi
Tezin Türü: Yüksek Lisans
Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Elektrik ve Elektronik Mühendisliği Bölümü, Türkiye
Tezin Onay Tarihi: 2016
Öğrenci: Vesal Rasoulzadeh
Danışman: İLKAY ULUSOY
Özet:Dyslexia is a learning disability that makes reading a challenge, despite normal level of intelligence and receiving adequate instructions. The core deficit in dyslexia is attributed to phonological processing. It’s been suggested that dyslexia is a disconnection syndrome. In this sense, the major sites of phonological processing in the brain are intact and the interconnection between these areas are disturbed. In this study, the disturbance in dyslectic brains based on effective connectivity models in “pre-reading” and “while reading” stages is investigated, which explains the causal interactions between different regions of the brain. Dynamic Bayesian Networks were constructed for the EEG data in theta, alpha and beta frequency bands to model the effective connectivity patterns of the brain in dyslectic and normal subjects in these bands. Analysis was performed based on the data obtained from two independent experiments, reading a word and a non-word by each subject. As the main objective of the thesis, dyslexic and normal children were classified based on the information obtained from the underlying effective connectivity models of their brains which reveal the abnormal patterns in the brain that may lead to detection and diagnosis of the condition. Dyslectic subjects were found to have a different effective connectivity patterns in “pre-reading” period, regardless of the reading task and theta frequency vi band is reported to be the most informative one about the disturbance in the casual influence between two groups in this period. The classification rate of 86.21% were obtained based on “pre-reading” models. the classification rates of 86.21% in reading a word experiment and 81.03% in reading a non-word experiment were obtained in alpha band. Features used to classify two groups are the connectvity weights (obtained fron DBN models) that are significantly different between dyslectics and controls. The connection include the ones from both dorsal (which is more activated while reading a word) and ventral (which is more activated while reading a non-word) pathways. This indicates the distruption of them both in dyslectic brains.