Efficient Deep Neural Networks for Classification of Alzheimer’s Disease and Mild Cognitive Impairment from Scalp EEG Recordings

Fouladi, Saman; Safaei, Ali; Mammone, Nadia; Ghaderi, Foad; Ebadi, M. J.

doi:10.1007/s12559-022-10033-3

Cited by 49 publications

(22 citation statements)

References 94 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Early diagnosis and differential diagnosis of MCI have been achieved by two different deep learning architectures, including modified convolutional and convolutional autoencoder neural networks. The outcomes demonstrate a 10% increase in accuracy rate over similar research techniques, indicating that the deep learning is a promising technique for processing EEG signals ( Fouladi et al, 2022 ).…”

Section: Discussionmentioning

confidence: 94%

Bibliometric analysis of electroencephalogram research in mild cognitive impairment from 2005 to 2022

Liu

et al. 2023

Front. Neurosci.

View full text Add to dashboard Cite

BackgroundElectroencephalogram (EEG), one of the most commonly used non-invasive neurophysiological examination techniques, advanced rapidly between 2005 and 2022, particularly when it was used for the diagnosis and prognosis of mild cognitive impairment (MCI). This study used a bibliometric approach to synthesize the knowledge structure and cutting-edge hotspots of EEG application in the MCI.MethodsRelated publications in the Web of Science Core Collection (WosCC) were retrieved from inception to 30 September 2022. CiteSpace, VOSviewer, and HistCite software were employed to perform bibliographic and visualization analyses.ResultsBetween 2005 and 2022, 2,905 studies related to the application of EEG in MCI were investigated. The United States had the highest number of publications and was at the top of the list of international collaborations. In terms of total number of articles, IRCCS San Raffaele Pisana ranked first among institutions. The Clinical Neurophysiology published the greatest number of articles. The author with the highest citations was Babiloni C. In descending order of frequency, keywords with the highest frequency were “EEG,” “mild cognitive impairment,” and “Alzheimer’s disease”.ConclusionThe application of EEG in MCI was investigated using bibliographic analysis. The research emphasis has shifted from examining local brain lesions with EEG to neural network mechanisms. The paradigm of big data and intelligent analysis is becoming more relevant in EEG analytical methods. The use of EEG to link MCI to other related neurological disorders, and to evaluate new targets for diagnosis and treatment, has become a new research trend. The above-mentioned findings have implications in the future research on the application of EEG in MCI.

show abstract

Section: Discussionmentioning

confidence: 94%

Bibliometric analysis of electroencephalogram research in mild cognitive impairment from 2005 to 2022

Liu

et al. 2023

Front. Neurosci.

View full text Add to dashboard Cite

show abstract

“…Various standard features such as spectral analysis, PSD, Event-related potentials (ERPs), powers in different frequency bands, and alpha band relationships were extracted during the tasks by Hu and Zhang [59]. The early detection of Alzheimer disease (AD) by mild cognitive impairment (MCI) has been proposed by Fouladi et al [60]. The authors proposed two different deep learning architectures for the classification of subjects into AD, MCI, and healthy control using 19-channel scalp EEG data.…”

Section: Discussionmentioning

confidence: 99%

A deep learning approach to detect the electroencephalogram-based cognitive task states

2023

IJATEE

View full text Add to dashboard Cite

Brain-computer interface (BCI) is an important topic for researchers and the scientific community, as indicated by the abundance of research and study materials in the field. The purpose of the BCI is to allow interaction with any device or computer via brain signals. According to this definition, BCI strives to collect the brain signals using sensors, analyze and process these received signals, and then extract features to operate any device. Simply, it is a link between the brain and the device. The user can control the device by using the brain's neural activities. *Author for correspondence BCI was first developed for biomedical applications to enable physically impaired persons to move around by substituting for lost motor functions [1]. Nowadays, it includes non-medical applications as well [2, 3]. Newer areas of BCI research include lie detection, drowsiness detection, cognitive studies, motor imagery, virtual reality, video games, driver fatigue detection, stress detection, and many more. From these applications, cognitive ability is important to understanding brain functioning. Cognitive ability depends from person to person and is essential in controlling various mental activities [4]. BCI research has been accelerated by technological advances enabling processing and observing mental and cognitive activities [5]. Any cognitive task reveals how the person thinks, utilizes,

show abstract

“…For accurate classification, the output of the encoder layer was trained by two continuous convolution layers with 64-filter length, a 2*2-kernel function, and a 2*2-maxpooling layer with step 2. Batch-normalization and 0.1 dropout layers were also used to prevent overfitting [33].…”

Section: Convolutional Auto-encoder Neural Networkmentioning

confidence: 99%

MRI-based brain tumor detection using convolutional deep learning methods and chosen machine learning techniques

Saeedi

Rezayi

Keshavarz

et al. 2023

BMC Med Inform Decis Mak

160

View full text Add to dashboard Cite

Background Detecting brain tumors in their early stages is crucial. Brain tumors are classified by biopsy, which can only be performed through definitive brain surgery. Computational intelligence-oriented techniques can help physicians identify and classify brain tumors. Herein, we proposed two deep learning methods and several machine learning approaches for diagnosing three types of tumor, i.e., glioma, meningioma, and pituitary gland tumors, as well as healthy brains without tumors, using magnetic resonance brain images to enable physicians to detect with high accuracy tumors in early stages. Materials and Methods A dataset containing 3264 Magnetic Resonance Imaging (MRI) brain images comprising images of glioma, meningioma, pituitary gland tumors, and healthy brains were used in this study. First, preprocessing and augmentation algorithms were applied to MRI brain images. Next, we developed a new 2D Convolutional Neural Network (CNN) and a convolutional auto-encoder network, both of which were already trained by our assigned hyperparameters. Then 2D CNN includes several convolution layers; all layers in this hierarchical network have a 2*2 kernel function. This network consists of eight convolutional and four pooling layers, and after all convolution layers, batch-normalization layers were applied. The modified auto-encoder network includes a convolutional auto-encoder network and a convolutional network for classification that uses the last output encoder layer of the first part. Furthermore, six machine-learning techniques that were applied to classify brain tumors were also compared in this study. Results The training accuracy of the proposed 2D CNN and that of the proposed auto-encoder network were found to be 96.47% and 95.63%, respectively. The average recall values for the 2D CNN and auto-encoder networks were 95% and 94%, respectively. The areas under the ROC curve for both networks were 0.99 or 1. Among applied machine learning methods, Multilayer Perceptron (MLP) (28%) and K-Nearest Neighbors (KNN) (86%) achieved the lowest and highest accuracy rates, respectively. Statistical tests showed a significant difference between the means of the two methods developed in this study and several machine learning methods (p-value < 0.05). Conclusion The present study shows that the proposed 2D CNN has optimal accuracy in classifying brain tumors. Comparing the performance of various CNNs and machine learning methods in diagnosing three types of brain tumors revealed that the 2D CNN achieved exemplary performance and optimal execution time without latency. This proposed network is less complex than the auto-encoder network and can be employed by radiologists and physicians in clinical systems for brain tumor detection.

show abstract

Efficient Deep Neural Networks for Classification of Alzheimer’s Disease and Mild Cognitive Impairment from Scalp EEG Recordings

Cited by 49 publications

References 94 publications

Bibliometric analysis of electroencephalogram research in mild cognitive impairment from 2005 to 2022

Bibliometric analysis of electroencephalogram research in mild cognitive impairment from 2005 to 2022

A deep learning approach to detect the electroencephalogram-based cognitive task states

MRI-based brain tumor detection using convolutional deep learning methods and chosen machine learning techniques

Contact Info

Product

Resources

About