CHAOS ANALYSIS OF EEG DATA DURING REAL AND IMAGINARY MOVEMENTS

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: İstanbul Üniversitesi, Fen Bilimleri Enstitüsü, Fen Fakültesi Bölümü, Türkiye

Tezin Onay Tarihi: 2024

Tezin Dili: İngilizce

Öğrenci: RABİA TULUK

Asıl Danışman (Eş Danışmanlı Tezler İçin): Emine Şeküre Nazlı Arda

Eş Danışman: Aydın Akan

Özet:

The neuronal loss caused by an accident that end up with a paralyzation or a disorder (Amyotrophic Lateral Sclerosis (ALS), Multiple Sclerosis (MS), Stroke etc.) is one of the most fearful lifetime scenario due to its irreversible. Lately the BCI studies that aim to regain some abilities artificially has increased its interest on EEG signals during motor imagery. The aim of our study is to find out if there is any difference in chaos parameters between real and imaginary finger movements which may offer an approach for BCI applications. A wavelet-chaos methodology was used for data analysis. First, EEGs of subjects were decomposed into 5 EEG sub-bands (alpha, beta, delta, theta, gamma) by discrete wavelet transform. Then, Katz, Higuchi’s fractal dimensions (KFD and HFD), Hurst exponent, correlation dimension were calculated as complexity measures for full-band and sub-band EEGs. After signal processing, Independent T-test was used to compare the full band and sub-bands of motor imagery and execution of finger movement EEG data. In this study our findings indicate significant complexity variations in specific states of brain function during both real and motor imagery movements. Notably, the analysis revealed consistent patterns in fractal dimensions and other nonlinear features across different subbands, highlighting the distinct yet interrelated neural processes involved in real and imagined movements. According to our results, Higuchi's Fractal Dimension (HFD) is found to be more accurate compared to the other chaos analysing methods.