Article
Computer Science, Artificial Intelligence
Cigdem Guluzar Altintop, Fatma Latifoglu, Aynur Karayol Akin, Adnan Bayram, Murat Ciftci
Summary: In this study, a nonlinear analysis of EEG signals was performed to classify comatose patients with different GCSs. Nonlinear features were successfully extracted from EEG signals evoked by tactile and auditory stimuli, and achieved a high accuracy in classifying different levels of consciousness.
INTERNATIONAL JOURNAL OF NEURAL SYSTEMS
(2022)
Article
Computer Science, Cybernetics
Shivam Sharma, Aakash Shedsale, Rishi Raj Sharma
Summary: This article proposes a grasp motor imagery identification approach based on multivariate fast iterative filtering (MFIF), which can be used for the development of a low-cost EEG based grasp identification system.
INTERNATIONAL JOURNAL OF HUMAN-COMPUTER INTERACTION
(2023)
Article
Engineering, Biomedical
Jian Shen, Yanan Zhang, Huajian Liang, Zeguang Zhao, Qunxi Dong, Kun Qian, Xiaowei Zhang, Bin Hu
Summary: Depression, a severe psychiatric illness, has a significant impact on patients' thoughts, behaviors, feelings, and well-being. However, current clinical practice lacks effective methods for recognizing and treating depression. Electroencephalogram (EEG) signals, which reflect the internal workings of the brain, show promise as an objective tool for depression recognition and diagnosis. In this study, we propose a regularization parameter-based improved intrinsic feature extraction method using empirical mode decomposition (EMD) to enhance depression recognition performance.
IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
(2023)
Article
Biology
Nesma E. ElSayed, A. S. Tolba, M. Z. Rashad, Tamer Belal, Shahenda Sarhan
Summary: A method to improve the efficiency of correcting EOG artifacts in raw EEG recordings is introduced in this study, utilizing feature engineering and machine learning to extract brain information and remove artifacts, with different classification models and regression learners to enhance accuracy in classification and prediction.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Medical Informatics
Hesam Akbari, Muhammad Tariq Sadiq, Siuly Siuly, Yan Li, Paul Wen
Summary: This paper proposes a novel method for depression detection using EEG signals, which includes preprocessing, mode selection, feature extraction, and classification. A new diagnostic index for depression is also proposed, aiding in faster and more objective identification of depression.
HEALTH INFORMATION SCIENCE AND SYSTEMS
(2022)
Article
Neurosciences
Maryam Sadeghijam, Saeed Talebian, Samer Mohsen, Mehdi Akbari, Akram Pourbakht
Summary: The study found that tinnitus patients exhibit higher entropy in EEG, reflecting chaotic brain behavior. Nonlinear methods (Entropy) in understanding tinnitus neurophysiology could be of significant importance and could potentially serve as a suitable criterion for clinical practice.
NEUROSCIENCE LETTERS
(2021)
Article
Automation & Control Systems
Manish Sharma, Divyansh Anand, Sarv Verma, U. Rajendra Acharya
Summary: Sleep is crucial for human well-being, and insomnia is a common sleep disorder that affects both physical and mental health. This study proposes a method that uses single-channel EEG signals to automatically identify insomnia, extracting features using a deep convolutional network and developing a model for sleep stages classification.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Chemistry, Analytical
Taraneh Aminosharieh Najafi, Antonio Affanni, Roberto Rinaldo, Pamela Zontone
Summary: This paper investigates the evaluation of drivers' mental attention state in a simulated environment. By analyzing physiological signals, such as Electrodermal activity, Electrocardiogram, and Electroencephalogram, the study finds that manual driving requires more mental effort and attention compared to autonomous driving. This approach proves to be an appropriate way to monitor driver attention.
Article
Engineering, Electrical & Electronic
Smith K. Khare, Varun Bajaj, U. Rajendra Acharya
Summary: The study proposes an automated identification method for schizophrenia using time-frequency analysis and convolutional neural network, achieving an accuracy of 93.36%. Compared to traditional methods, this approach shows superior efficiency in feature extraction and has the potential to aid in diagnosing schizophrenia.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Engineering, Biomedical
Puja A. Chavan, Sharmishta Desai
Summary: In this research, a train-optimized hidden Markov model (HMM) based on human learning optimization is proposed to automatically detect epileptic seizures by distinguishing focal and non-focal epileptic EEG signals. The model is trained and optimized using frequency band-based features with the help of the human learning optimization algorithm. The effectiveness of the model is proved by measuring evaluation metrics, showing higher accuracy and sensitivity compared to state-of-the-art methods.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Chemistry, Analytical
Felix Nieto-del-Amor, Raja Beskhani, Yiyao Ye-Lin, Javier Garcia-Casado, Alba Diaz-Martinez, Rogelio Monfort-Ortiz, Vicente Jose Diago-Almela, Dongmei Hao, Gema Prats-Boluda
Summary: This study aimed to assess the clinical value of various entropy metrics for predicting preterm birth, with dispersion and bubble entropy performing better in discriminating between preterm and term births. Entropy metrics provided complementary information to linear features, while the inclusion of other nonlinear features did not significantly improve model performance. The best model achieved an F1-score of 90.1% in testing dataset, indicating potential for real-time applications and clinical practice.
Article
Engineering, Biomedical
Nilima Salankar, Pratikshya Mishra, Lalit Garg
Summary: Emotion recognition from EEG signals is a challenging task due to the non-stationarity of the signals. Extracting relevant features and classifying patterns in EEG signals is difficult. The study used the DEAP dataset with EEG signals from 32 participants, categorizing them into quadrants related to different emotions.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2021)
Article
Engineering, Biomedical
Jianliang Min, Chen Xiong, Yonggang Zhang, Ming Cai
Summary: This study discussed a fatigue identification framework based on forehead EEG signals, achieving better outcomes in fatigue detection by enhancing signal quality with a hybrid model and multiple common entropy measures.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2021)
Article
Engineering, Biomedical
Miguel Angel Lujan, Jorge Mateo Sotos, Ana Torres, Jose L. Santos, Oscar Quevedo, Alejandro L. Borja
Summary: This paper presents a new automated procedure based on deep learning methods for schizophrenia diagnosis. By analyzing electroencephalogram signals, the proposed method effectively distinguishes between schizophrenia patients and healthy volunteers with high accuracy. The experimental results show that this method outperforms other machine learning methods in schizophrenia detection.
JOURNAL OF MEDICAL AND BIOLOGICAL ENGINEERING
(2022)
Article
Computer Science, Information Systems
Alexander Voznesensky, Denis Butusov, Vyacheslav Rybin, Dmitry Kaplun, Timur Karimov, Erivelton Nepomuceno
Summary: In this paper, a novel denoising algorithm called EITD is proposed to process chaotic signals using specific chaotic noise generators. The developed approach is compared with other filtration algorithms and shows better denoising quality.
Article
Computer Science, Artificial Intelligence
Meghana Karri, Chandra Sekhara Rao Annvarapu, U. Rajendra Acharya
Summary: Automated brain tumor segmentation using MRI is crucial for clinical decision-making and surgical planning. This study proposes a method called SGC-ARANet, which utilizes deep learning models and multiple modules to improve segmentation performance. The results demonstrate that this approach outperforms several state-of-the-art algorithms in various evaluation measures.
APPLIED INTELLIGENCE
(2023)
Review
Engineering, Biomedical
V. Jahmunah, Joel En Wei Koh, Vidya K. Sudarshan, U. Raghavendra, Anjan Gudigar, Shu Lih Oh, Hui Wen Loh, Oliver Faust, Prabal Datta Barua, Edward J. Ciaccio, U. Rajendra Acharya
Summary: In this study, information extraction methods were used to differentiate Celiac Disease (CD) from non-CD. Statistical and non-linear methods were found to be the most important for information extraction. These tools can reduce variability in clinical workflows, but bias introduced during the design of diagnostic support systems may limit the general validity of the results. Understanding the limitations of information extraction tools can improve the process.
BIOCYBERNETICS AND BIOMEDICAL ENGINEERING
(2023)
Article
Automation & Control Systems
Moloud Abdar, Mohammad Amin Fahami, Leonardo Rundo, Petia Radeva, Alejandro F. Frangi, U. Rajendra Acharya, Abbas Khosravi, Hak-Keung Lam, Alexander Jung, Saeid Nahavandi
Summary: This article proposes a simple and novel hierarchical attentive multilevel feature fusion model for medical image classification, combined with an uncertainty-aware module for uncertainty quantification. The model is tested on various medical imaging datasets and achieves excellent classification accuracy.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Moloud Abdar, Soorena Salari, Sina Qahremani, Hak-Keung Lam, Fakhri Karray, Sadiq Hussain, Abbas Khosravi, U. Rajendra Acharya, Vladimir Makarenkov, Saeid Nahavandi
Summary: The COVID-19 pandemic poses a major threat to human health, making the development of computer-aided detection systems a priority. This study introduces a new deep learning feature fusion model called UncertaintyFuseNet, which accurately classifies CT scan and X-ray images. The results demonstrate the efficiency and robustness of the model.
INFORMATION FUSION
(2023)
Article
Computer Science, Artificial Intelligence
Prabal Datta Barua, Emrah Aydemir, Sengul Dogan, Mehmet Ali Kobat, Fahrettin Burak Demir, Mehmet Baygin, Turker Tuncer, Shu Lih Oh, Ru-San Tan, U. Rajendra Acharya
Summary: In this study, a novel multilevel hybrid feature extraction-based classification model was developed for the classification of myocardial infarction (MI). The model achieved accurate MI classification with low time complexity by utilizing pooling functions and feature selection algorithms to extract features from ECG signals.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2023)
Article
Medicine, General & Internal
Faruk Oztekin, Oguzhan Katar, Ferhat Sadak, Muhammed Yildirim, Hakan Cakar, Murat Aydogan, Zeynep Ozpolat, Tuba Talo Yildirim, Ozal Yildirim, Oliver Faust, U. Rajendra Acharya
Summary: Dental caries is a common dental health issue that can cause pain and infections, reducing quality of life. Applying machine learning models for caries detection can lead to early treatment, but lack of explainability may hinder their acceptance. In this study, an explainable deep learning model for detecting dental caries was developed and evaluated. The ResNet-50 model showed slightly better performance compared to EfficientNet-B0 and DenseNet-121, achieving an accuracy of 92.00% and a sensitivity of 87.33%. The heat maps provided by the model helped explain the classification results, enabling dentists to validate and reduce misclassification.
Article
Computer Science, Interdisciplinary Applications
Oliver Faust, Simona De Michele, Joel E. W. Koh, V Jahmunah, Oh Shu Lih, Aditya P. Kamath, Prabal Datta Barua, Edward J. Ciaccio, Suzanne K. Lewis, Peter H. Green, Govind Bhagat, U. Rajendra Acharya
Summary: This study aimed to utilize AI models to distinguish between normal individuals, individuals with celiac disease (CD), and individuals with non-celiac duodenitis (NCD) based on characteristics of the small intestinal lamina propria. By preprocessing high magnification biopsy images, the study found that AI models can effectively differentiate between different types of intestinal disorders and can potentially be applied in clinical settings.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2023)
Article
Computer Science, Interdisciplinary Applications
Meghana Karri, Chandra Sekhara Rao Annavarapu, U. Rajendra Acharya
Summary: This study proposes a new two-phase cross-domain transfer learning system for effective skin lesion segmentation from dermoscopic images. By using model-level and data-level transfer learning, combined with a high-performing DL network called nSknRSUNet, the system achieves excellent segmentation performance. The results show that the system outperforms existing methods in both qualitative and quantitative evaluations.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2023)
Article
Engineering, Biomedical
Sueleyman Yaman, Elif Isilay Unlu, Hasan Guler, Abdulkadir Sengur, U. Rajendra Acharya
Summary: This study presents a new deep learning-based ensemble model and a novel double iterative ReliefF (DIRF) feature selection algorithm to enhance the automatic detection of atrophy, ischemia, and WMI using MRI images.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Engineering, Biomedical
Serkan Kirik, Sengul Dogan, Mehmet Baygin, Prabal Datta Barua, Caner Feyzi Demir, Tugce Keles, Arif Metehan Yildiz, Nursena Baygin, Ilknur Tuncer, Turker Tuncer, Ru-San Tan, U. Rajendra Acharya
Summary: We developed an efficient handcrafted feature engineering model for EEG-based language identification using four directed graphs modeled on Feynman graph patterns (FGPat). We collected a dataset of 3252 EEG signals from native English-speaking Nigerian-born and Turkish subjects. In our FGPat18 model, input EEG signals and their Q wavelet transform-decomposed wavelet bands were used to extract textural and statistical features. The obtained feature vectors were input to a classification algorithm to achieve high accuracy rates. The FGPat18 model achieved a classification accuracy rate of 99.38% with 10-fold cross-validation.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Engineering, Biomedical
Samiul Based Shuvo, Syed Samiul Alam, Syeda Umme Ayman, Arbil Chakma, Prabal Datta Barua, U. Rajendra Acharya
Summary: This study aims to discover the optimal transformation method for detecting cardiovascular diseases using noisy heart sound signals and propose a noise-robust network to improve classification performance. The results indicate that Continuous Wavelet Transform (CWT) is the optimal transformation method for noisy heart sound signals. The proposed Convolutional Recurrent Neural Network (CRNN) with CWT demonstrates higher accuracy in detecting cardiovascular diseases.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Review
Computer Science, Information Systems
S. Janifer Jabin Jui, Ravinesh C. Deo, Prabal Datta Barua, Aruna Devi, Jeffrey Soar, U. Rajendra Acharya
Summary: An automated Neurological Disorder detection system is a cost-effective and resource efficient tool for medical and healthcare applications. The concept of entropy is a promising approach for processing electroencephalogram signals. Support Vector Machines and sample entropy are the most commonly used machine learning model and entropy measure respectively.
Article
Computer Science, Information Systems
Manish Sharma, Harsh Lodhi, Rishita Yadav, Niranjana Sampathila, K. S. Swathi, U. Rajendra Acharya
Summary: Sleep is important for health and well-being, covering one-third of life. The traditional evaluation of sleep quality does not consider important transient phenomena, such as K-complexes and transient fluctuations, which are crucial for diagnosing sleep disorders. This study proposes an automated, computerized approach for classifying sleep phases using a machine learning model with explainable artificial intelligence capabilities based on wavelet-based Hjorth parameters.
Article
Remote Sensing
Suat Gokhan Ozkaya, Nursena Baygin, Prabal D. Barua, Arvind R. Singh, Mohit Bajaj, Mehmet Baygin, Sengul Dogan, Turker Tuncer, Ru-San Tan, U. Rajendra Acharya
Summary: This study aims to develop a computationally lightweight, accurate, and explainable machine learning model for the automated detection of seismogram signals that could serve as an effective warning system for earthquake prediction. The model achieved a classification accuracy of 96.82% using a balanced dataset of 5001 earthquakes and 5001 non-earthquake signals. It utilized multilevel feature extraction, selector-based feature selection, and a greedy algorithm for model selection.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Computer Science, Artificial Intelligence
Sinan Tatli, Gulay Macin, Irem Tasci, Burak Tasci, Prabal Datta Barua, Mehmet Baygin, Turker Tuncer, Sengul Dogan, Edward J. Ciaccio, U. Rajendra Acharya
Summary: This study aims to propose a new algorithm for early diagnosis of multiple sclerosis (MS) using machine learning. The algorithm utilizes transfer learning and hybrid feature engineering, and calculates feature vectors through multiple layers of neural networks, resulting in high classification accuracy.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)