Sayılgan, Ebru
Loading...
Profile URL
Name Variants
Sayilgan, Ebru
Durmus, Ebru
Sayilgan, E.
Durmus, Ebru
Sayilgan, E.
Job Title
Email Address
ebru.sayilgan@ieu.edu.tr
Main Affiliation
05.11. Mechatronics Engineering
Status
Current Staff
Website
ORCID ID
Scopus Author ID
Turkish CoHE Profile ID
Google Scholar ID
WoS Researcher ID
Sustainable Development Goals
Documents
17
Citations
109
h-index
6
Documents
15
Citations
66
Scholarly Output
16
Articles
8
Views / Downloads
21/526
Supervised MSc Theses
2
Supervised PhD Theses
0
WoS Citation Count
48
Scopus Citation Count
81
WoS h-index
3
Scopus h-index
6
Patents
0
Projects
1
WoS Citations per Publication
3.00
Scopus Citations per Publication
5.06
Open Access Source
4
Supervised Theses
2
| Journal | Count |
|---|
Current Page: 1 / NaN
Scopus Quartile Distribution
Competency Cloud
16 results
Scholarly Output Search Results
Now showing 1 - 10 of 16
Conference Object Citation - WoS: 2Citation - Scopus: 2Analysing Sci Patients' Eeg Signal Using Manifold Learning Methods for Triple Command Bci Design(Institute of Electrical and Electronics Engineers Inc., 2024) Kucukselbes H.; Sayilgan E.; Kucukselbes, Hezzal; Sayilgan, EbruThis study analyzed EEG signals from patients with spinal cord injuries by examining five different hand-wrist movements. The signal processing steps designed for an automatic and robust EEG-based BCI system were applied sequentially. Initially, 37 different features in the time, frequency, and time-frequency domains were extracted from the preprocessed signal. After trying widely used Manifold Learning (ML) methods in the literature, including t-Distributed Stochastic Neighbor Embedding (t-SNE), Local Linear Embedding (LLE), Multi-Dimensional Scaling (MDS), and ISOmetric Mapping (ISOMAP), we attempted the Spectral Embedding method, which has not yet been utilized in EEG signal analysis. The signals were then classified using three different machine-learning algorithms. The study compared classification performance using the accuracy metric. A multi-class classification method was employed specifically the triple classification method. The most successful performance was achieved by using the ISOMAP machine learning method and kNN classifier for the Pronation-Palmar Grasp-Hand Open combination, with an accuracy of 0.993 ± 0.016. Other methods used were t-SNE, MDS, LLE, and Spectral Embedding, respectively. Regarding classifiers, the kNN, SVM, and Naive Bayes algorithms were found to be successful in that order. Based on these results, we propose a suitable methodology for designing a robust BCI system. © 2024 IEEE.Article Interactive and Deep Learning-Powered EEG-BCI for Wrist Rehabilitation: A Game-Based Prototype Study(LIDSEN Publishing Inc, 2025) Sayilgan, E.Motor deficits induced by neurological disorders impose a severe impact on activities of daily life. Conventional rehabilitation practices necessitate ongoing clinical supervision, which is costly and inaccessible. EEG-based brain-computer interface (BCI) systems offer a viable solution by facilitating neurorehabilitation through the direct interpretation of brain signals. Nonetheless, current systems are confronted with issues of real-time control, portability, and classification precision. The paper describes a novel EEG-controlled wrist rehabilitation robot with deep learning-based real-time motor intention classification. EEG signals were recorded with OpenBCI, preprocessed with noise filtering, and converted into time-frequency representations. A GoogLeNet-inspired convolutional neural network (CNN) was trained for the classification of wrist movement intentions. SolidWorks was utilized for designing the mechanical structure, which was verified using finite element analysis (FEA). An Nvidia-based microcontroller was employed for controlling servo motors, while an inertial measurement unit (IMU) was incorporated into the system for enabling precise and agile movement using feedback control. The system proposed in this work attained an EEG classification accuracy of 90.24%, which was well above conventional feature-based classifiers. The 2-degree-of-freedom (2-DoF) robotic system with a lightweight structure enabled controlled wrist flexion, extension, and radial/ulnar deviation movements. Structural validation by the FEA assured mechanical stability against operational loads. The system proved to be feasible for real-time, user-intended motion control. The proposed study offers a cost-effective, portable, and deep learning-based EEG-BCI rehabilitation robot, rendering a possible solution to neurorehabilitation. The high classification accuracy and real-time control features of the system highlight the potential for personalized rehabilitation. Future endeavors will focus on the development of deeper learning frameworks, the advancement of motor control strategies, and the implementation of extended clinical trials. © 2025 Elsevier B.V., All rights reserved.Article Citation - WoS: 12Citation - Scopus: 17Investigating the Effect of Flickering Frequency Pair and Mother Wavelet Selection in Steady-State Visually-Evoked Potentials on Two-Command Brain-Computer Interfaces(Elsevier Science Inc, 2022) Sayilgan, E.; Yuce, Y. K.; Isler, Y.Introduction: Steady-state visually evoked potentials (SSVEPs) have become popular in brain-computer interface (BCI) applications in addition to many other applications on clinical neuroscience (neurodegene-rative disorders, schizophrenia, epilepsy, etc.), cognitive (visual attention, working memory, brain rhythms, etc.), and use of engineering researches. Among available methods to measure brain activities, SSVEPs have advantages like higher information transfer rate, simplicity in structure, and short training time. SSVEP-based BCIs use flickering stimuli at different frequencies to discriminate distinct commands in real life. Some features are extracted from the SSVEP signals before these commands are classified. The wavelet transform (WT) has attracted researchers among feature extraction methods since it utilizes the non-stationary signals well. In the WT, a sample function (named mother wavelet) represents the SSVEP signal in both time and frequency domains. Unfortunately, there is no universal mother wavelet function that fits all the signals. Therefore, choosing an appropriate mother wavelet function may be a challenge in WT-related studies. Although there are such studies in three-and seven-command SSVEP-based studies, there is no study for two-command systems in our knowledge.Materials and Methods: In this study, two user commands flickered at the combinations of seven different frequencies were tested to determine which frequency pairs give the highest performance. For this purpose, three well-known wavelet features (energy, entropy, and variance) were calculated for each of derived EEG frequency bands from the discrete WT coefficients of SSVEP signals. The WT was repeated for six different mother wavelet functions (Haar, Db4, Sym4, Coif1, Bior3.5, and Rbior2.8). Then, four feature sets (every three features, and all together) were applied to seven commonly-used machine learning algorithms (Decision Tree, Discriminant Analysis, Logistic Regression, Naive Bayes, Support Vector Machines, Nearest Neighbors, and Ensemble Classifiers).Results and Discussion: We achieved 100% accuracies among these 3,528 runs (7 classifiers x 4 feature sets x 6 mother wavelets x 21 flickering frequency pairs) using the mother wavelet function of Haar and the Ensemble Learner classifier. The highest classifier performances are 100% when two commands have the flickering frequency pairs of (6.0 and 10 Hz), (6.5 and 8.2 Hz), or (6.5 and 10.0 Hz).Conclusion: We obtained three main outcomes from this study. First, the most representative mother wavelet function was Haar, while the worst one was Symlet 4. Second, the Ensemble Learner classifier gave the maximum classifier performance in a two-command SSVEP-based BCI system. Besides, two user commands from SSVEP should be one of the frequency pairs of (6.0 and 10.0 Hz), (6.5 and 8.2 Hz), and (6.5 and 10.0 Hz) to achieve the maximum accuracy.(c) 2022 AGBM. Published by Elsevier Masson SAS. All rights reserved.Article Citation - WoS: 11Citation - Scopus: 14Evaluation of Mother Wavelets on Steady-State Visually-Evoked Potentials for Triple-Command Brain-Computer Interfaces(Tubitak Scientific & Technical Research Council Turkey, 2021) Sayilgan, Ebru; Yuce, Yilmaz Kemal; Isler, YalcinWavelet transform (WT) is an important tool to analyze the time-frequency structure of a signal. The WT relies on a prototype signal that is called the mother wavelet. However, there is no single universal wavelet that fits all signals. Thus, the selection of mother wavelet function might be challenging to represent the signal to achieve the optimum performance. There are some studies to determine the optimal mother wavelet for other biomedical signals; however, there exists no evaluation for steady-state visually-evoked potentials (SSVEP) signals that becomes very popular among signals manipulated for brain-computer interfaces (BCIs) recently. This study aims to explore, if any, the mother wavelet that suits best to represent SSVEP signals for classification purposes in BCIs. In this study, three common wavelet-based features (variance, energy, and entropy) extracted from SSVEP signals for five distinct EEG frequency bands (delta, theta, alpha, beta, and gamma) were classified to determine three different user commands using six fundamental classifier algorithms. The study was repeated for six different commonly-used mother wavelet functions (haar, daubechies, symlet, coiflet, biorthogonal, and reverse biorthogonal). The best discrimination was obtained with an accuracy of 100% and the average of 75.85%. Besides, ensemble learner gives the highest accuracies for half of the trials. Haar wavelet had the best performance in representing SSVEP signals among other all mother wavelets adopted in this study. Concomitantly, all three features of energy, variance, and entropy should be used together since none of these features had superior classifier performance alone.Article Citation - WoS: 16Citation - Scopus: 20Principal Component Analysis and Manifold Learning Techniques for the Design of Brain-Computer Interfaces Based on Steady-State Visually Evoked Potentials(Elsevier, 2023) Yeşilkaya, Bartu; Sayilgan, Ebru; Yuce, Yilmaz Kemal; Perc, Matjaz; Isler, YalcinSteady-state visually evoked potentials (SSVEP) are stochastic and nonstationary bioelectric signals. Because of these properties, it is difficult to achieve high classification accuracy, especially when many considered features lead to a complex structure. We therefore propose a manifold learning framework to decrease the number of features and to classify SSVEP data by comparing lower dimensional matrices with well-known machine learning algorithms. We use the AVI-SSVEP Dataset, which includes stimuli at seven different frequencies and 15360 samples per person. The SSVEP features are extracted from relevant and distinctive frequency -domain, time-domain, and time-frequency domain properties, creating a total of 55 feature vectors. We then analyze and compare five divergent manifold learning methods with respect to their performance on nine different machine-learning algorithms. Among all considered manifold learning methods, we show that the Principal Component Analysis has the best classifier performance with an average of 22 components. Moreover, the Naive Bayes classifier with the Principal Component Analysis achieves the maximum accuracy of 50.0%-80.95% for a 7-class classification problem. Our research thus shows that the proposed analytical framework can significantly improve the decoding accuracy of 7-class SSVEP problems, and that it exhibits notable robustness and efficiency for small group datasets.Article Citation - Scopus: 4Classifying Eeg Data From Spinal Cord Injured Patients Using Manifold Learning Methods for Brain-Computer Interface-Based Rehabilitation(Springer Science and Business Media Deutschland GmbH, 2025) Sayilgan, E.Spinal cord injuries (SCI) occur when the spinal cord is damaged due to any trauma. Treatment of this condition typically involves a long and challenging rehabilitation process. Brain-computer interface (BCI) controlled rehabilitation systems show promise for people with SCI, as they offer repetitive and controlled treatment in the home environment without requiring a specialist. There is a growing demand for electroencephalography (EEG)–based BCI rehabilitation systems, particularly for patients with SCI. In this study, EEG signals from ten SCI patients were analyzed. At the same time, they imagined performing rehabilitation movements for five different hand and arm actions (pronation, supination, palmar grasp, lateral grasp, and hand opening). The study tested manifold learning algorithms, feature extraction, and classification methods in EEG analysis to improve usability in real-time applications. Manifold learning algorithms were used to represent complex and high-dimensional data in a lower-dimensional space, allowing for better representation and separation of the temporal and spatial characteristics of brain activity. The spectral embedding algorithm was used in this study, which, to the best of our knowledge, is the first time this algorithm has been applied to the data of SCI patients. Additionally, we conducted a comparative analysis of commonly encountered methods in the literature, including multi-dimensional scaling (MDS), isometric feature mapping (ISOMAP), local linear embedding (LLE), and t-distributed stochastic neighbor embedding (t-SNE). Machine learning algorithms, such as k-nearest neighbor (k-NN), support vector machine (SVM), and Naive Bayes methods, were used to obtain classification results for both multi-class and binary classes. The combination of pronation and hand open was found to yield the best performance for binary classification of the movements. The study determined that the ISOMAP manifold learning method with the k-NN algorithm is the optimal method for processing times of both the train and the test. The method also demonstrated a high accuracy value of 0.967 and a short time of 0.088 units in multiple classifications, which is promising. The study utilized the spectral embedding method for the first time and achieved an accuracy rate of 0.649 in multi-class classification and 0.933 ± 0.049 in binary-class classification. © The Author(s) 2025.Article Bağımsız Bileşen Analizi ve Makine Öğrenmesi Kullanılarak Omurilik Yaralanması Olan Kişilerden Alınan Eeg Sinyallerinden El Hareketlerinin Sınıflandırılması(2024) Sayılgan, EbruBu çalışmanın temel amacı, Omurilik Yaralanması (OY) olan kişilerin, kol ve el hareketlerinin, kodu çözülebilir nöral bağıntılarını koruduğunu göstermektir. OY’li on kişiden pronasyon, süpinasyon, palmar kavrama, lateral kavrama ve el açma hareketleri düşündürülerek alınan ElektroEnsefaloGrafi (EEG) sinyallerinin ayırt edici hareket bilgisi araştırılmıştır. Bunu yaparken kullanılan yöntemlerde Bağımsız Bileşen Analizi (BBA/ICA) yöntemi hem artefakt gidermede hem de yeni bir yaklaşım olarak öznitelik vektörlerini çıkarmada kullanılmıştır. Önerilen yöntemde öznitelik vektörleri bağımsız bileşenlerde ortak bilgi matrisi çıkarılarak oluşturulmuştur. Çıkarılan ve seçimi yapılan öznitelik vektörleri dört farklı makine öğrenmesi modeli (Destek Vektör Makinesi (DVM), k-En Yakın Komşuluk (k-EYK), AdaBoost ve Karar Ağaçları (KA)) ile test edilmiştir. Model değerlendirme aşamasında aşırı öğrenmenin önüne geçmek için 5-katlamalı çapraz doğrulama ve hata matrisi yöntemleri kullanılmıştır. Sonuç olarak, incelenen beş sınıfa göre elde edilen başarım oldukça yüksek çıkmıştır. Deneklerin ortalaması alındığında elde edilen model doğruluk değerleri sırasıyla DVM’de 0.9024±0.0781, k-EYK’da 0.8582±0.0985, AdaBoost’ta 0.7924±0.0937 ve KA’da 0.8089±0.0645 olarak hesaplanmıştır. Bu sonuçlara dayanarak OY olan bireylerin kol ve el hareketlerinin ayırt edicilik performansının önerilen yöntem ile oldukça yüksek sonuçlar verdiği görülmektedir. BBA yöntemine dayalı bir öznitelik çıkarma ve DVM modeli ile sınıflandırma metodolojisinin OY’li hastaların rehabilitasyon tedavisinde EEG temelli beyin bilgisayar arayüzü uygulamalarına önemli bir katkısı olacağı düşünülmektedir.Conference Object Citation - Scopus: 1Design of a Low-Cost Wrist Rehabilitation Robot for Home Use(IEEE, 2024) Sayilgan, EbruThe primary objective of this study is to develop a lightweight, compact, and portable medical wrist rehabilitation device. The idea behind the project is that the device will be used by the patient, in the comfort of their home; so the device needs to be easy to wear and use, also safety is a key factor. The target users are patients who need assistance with two specific wrist movements - flexion-extension and radial-ulnar deviation. To achieve this objective, 3D CAD models will be developed to design the mechanical components of the robot. Kinematic and kinetic calculations will be done and simulated using tools like Solidworks Motion. Additionally, finite element analysis using software such as Solidworks and ANSYS will analyze component stresses to ensure safety factors for all parts meet or exceed industry standards. After validating the calculations with simulation, the next steps will involve selecting the electric components. This includes choosing two appropriately sized motors, one for each degree of freedom, that will apply sufficient torque for joint movements, without excessive force to disturb the patient based on torque requirements for wrist rehabilitation. In addition, a controller with sufficient processing properties will be evaluated and selected to interface with the motors/drivers. Testing on people with a healthy wrist movement extent will conclude the objectives. The prototype will be tested to verify its performance through quantitative and qualitative feedback. Key metrics for testing will include a range of motion, torque output, electrical current usage, and safety. The design will also be reconsidered based on user feedback to enhance comfort, usability, and ability to smoothly assist natural wrist motions during rehabilitation exercises. Comprehensive documentation of all aspects of the design, development, and testing process will likewise be provided.Conference Object Citation - WoS: 3Citation - Scopus: 6Effective Ssvep Frequency Pair Selection Over the Googlenet Deep Convolutional Neural Network(IEEE, 2022) Avci, Meryem Beyza; Sayılgan, EbruThe acquired Electroencephalography (EEG) signal while applying a blinking image on a screen is called steady-state visually-evoked potential (SSVEP). SSVEP is a popular control signal of the EEG in real-life applications because of the advantages such as; higher information transfer rate, simplicity in structure, and short training time. Most of the studies related to the SSVEP tried to discriminate which image (frequency) is gazed at while recording and turn this frequency into control commands. In this study, we focused on the selection of the stimulating frequency pair, which has the best accuracy rate, to investigate whether there is a correlation between stimulation frequencies. To achieve this goal, first of all, recorded SSVEP signals, which include seven different frequencies (6 - 6.5 - 7 7.5 - 8.2 - 9.3 - 10 Hz) were converted into spectrogram images. After dividing the spectrogram images into folders with respect to the frequencies, they were routed to GoogLeNet deep learning algorithm for binary classification. Consequently, we obtained the best performance in 8.2 & 10 Hz frequency pairs with 91.28% accuracy.Conference Object Citation - Scopus: 3Decoding of Palmar Grasp and Hand Open Tasks From Low-Frequency Eeg From People With Spinal Cord Injury Using Machine Learning Algorithms(Institute of Electrical and Electronics Engineers Inc., 2023) Avci, M.B.; Kucukselbes, H.; Sayılgan, EbruSpinal cord injury (SCI) is a chronic disorder that is detrimental to the spinal cord and causes the loss of neuronal function, particularly sensorimotor functions. Brain-computer interface (BCI)-controlled rehabilitation systems have been proposed as a promising treatment component for people with SCI whose treatment is based on a long and tiring rehabilitation process. With respect to this, we presented a novel approach using an electroencephalography (EEG) based BCI rehabilitation system to help SCI patients. For this purpose, low-frequency EEG signals acquired from nine people with SCI were analyzed by considering attempted arm and hand movements. We used both time-domain features based on statistical changes (e.g., mean, variance, skewness, and kurtosis, etc.) and frequency-domain features based on Fast Fourier Transform in the EEG signal to decode the two intentions: hand open and palmar grasp. For binary classification, seven machine learning models (Fine KNearest Neighbour, Ensemble, Logistic Regression Kernel, Support Vector Machines Kernel, Fine Tree, Quadratic Discriminant, and Wide Neural Network) were used to classify the features. Accuracy, Precision, Recall, and F1 score criteria were used to evaluate machine learning models. In conclusion, we achieved successful results like an Accuracy of %91.70, Precision of %93, Recall of %90, and F1 Score of %91 by using frequency domain features combined with the Fine K-Nearest Neighbour model, with a prediction speed of 8848.84 obs/sec, and a training time of only 10.59 seconds. These results indicate that our methodology can decode executed hand open and palmar grasp motions from people with SCI. For this reason, it could be a critical and crucial contribution to the literature regarding the application of BCI. © 2023 IEEE.
