Classification of fNIRS Data Using Deep Learning for Bipolar Disorder Detection
| Autor: | Haluk Barkin Evgin, Yasemin Hosgoren, Oguzhan Babacan, Damla Sayar, Ilkay Ulusoy, Adnan Kusman, Halise Devrimci Özgüven, Bora Baskak |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Artificial neural network
Computer science business.industry Deep learning 05 social sciences Pattern recognition medicine.disease Convolutional neural network 050105 experimental psychology 03 medical and health sciences 0302 clinical medicine Mood medicine 0501 psychology and cognitive sciences Artificial intelligence Bipolar disorder Time domain business 030217 neurology & neurosurgery |
| Zdroj: | SIU |
| Popis: | With the use of ecologically validated tools more applicable measurements can be obtained, especially of individuals who have psychological disorders. Functional Near- Infrared Spectroscopy (fNIRS) is a neural imaging method that comes into prominence for imaging patients who have psychological disorders. It is a desired method because of its feasibility, high resolution in time and its partial resistance to head movements. Following the developments in the artificial intelligence, individuals' medical data obtained from various methods are started to be used in neural networks to classify various health conditions. In this research, 1 dimensional time domain data of fNIRS, which is acquired during prepared tasks, are used to train a neural network for the diagnosis of a common mood disorder, the Bipolar Disorder. With the classification of this data, the distinguishability of ill subjects from healthy subjects is investigated by using a 1 dimensional Convolutional Neural Network (CNN), which is a feed-forward deep neural network. By means of the obtained results, it is observed that the Bipolar Disorder can be classified even during the remission period. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|