When it comes to classifying EEG signals, having the right tools can make all the difference in accurately interpreting brain activity. With advancements in technology, there are several tools available that can aid in this process. Below are some of the best tools to use when classifying EEG signals:
DL models cannot directly work on graph-structured input data because they consider the brain network features as a vector of one dimension [293]. The human brain connectivity represents the brain as a graph with interacting nodes in non-Euclidean space, and existing DL methods generally disregard the interaction and association of brain connectivity networks [294]. GNNs aim to learn graph representations by using a neural network and to pass information via a message-passing algorithm [295,296]. Unlike neural networks, GNNs update the representations of nodes while maintaining the graph topology.
Hence, signals such as those from an electrocardiogram (ECG) or electrooculogram (EOG) are required to separate artifacts from EEG signals. Although the regression method is based on simple mathematical intuition and therefore is widely used because of the minimal computation required, the dependence of this method on reference channels for ECG and EOG removal is considered a drawback [48]. (A) Classification accuracy of POWER dataset without ocular artifact rejection (yellow), with ICA-based component rejection method (blue) and ICA-based epoch rejection method (green). (D) Classification accuracy using phase locking value (PLV, blue) and weighted phase lag index (WPLI, purple). Of the three analytical measures, PAC (Fig. 2C, ​,3C)3C) provides for the maximal discrimination. For 1 electrode and 1 x-band (Fig. 3C top, left), the best discrimination is provided by the F8, F4, P4 and F7 electrodes in the theta-high gamma x-band (Fig. 3C middle, left).
In this diagram, the EEG data are first read and then the eye artifacts have been removed from the recorded signals by ICA technique. Then, after artifacts removing process, the EEG signals will be segmented into fixed time windows of 50 s. Next, we feed the output of the segmentation process into a band-pass filter to remove the noises. To extract the features of EEG signals, construct the feature vectors, and improve classification accuracy, LBP, SD, variance, kurtosis, and entropy are used along with DWT. Bubble plot of studies according to classification algorithms and feature extraction methods.
Each subset contains 100 EEG signals of 23.6 sec in length, and the sampling frequency is 173.6 Hz. Among them, there were two subsets of EEG recorded during epileptic seizures, which had 200 samples, and one set of EEG records in the seizure period had 100 samples. Figure 1 shows two types of signals in epilepsy patients during nonepilepsy and epilepsy. Among them, 200 samples are classified as F and 100 samples are classified as S. Class F is labeled as a nonepileptic seizure EEG signal, while class S is a seizure signal.
The size of the bubble indicates the performance of classification models for each task as marked with different colors. WT is a time-frequency transform that considers the features of the EEG signals within a frequency domain and is perfectly localized within the time domain [70]. This method has good performance in spectral analysis of irregular and nonstationary signals within different size windows [85]. One advantage of WT is that it provides accurate frequency information and time information at low and high frequencies, respectively. That is, a narrow window is typically used to evaluate high frequencies, and a wide window is applied to assess low frequencies [68,86].
1. EEG Signal Processing Software
Despite its ease of use and quick setup, EEG recording from dry electrodes yields data with more artifacts than does gel-based EEG recording, thus affecting EEG analysis. In this case, identifying the most sensitive biomarkers for the specific task is essential [43]. In addition, some experimental studies have confirmed that selecting the proper features can improve the performance of ML/DL methods. 1, with a linear discriminant using 55 variables or more, a classification accuracy of 100% is obtained using the d4 wavelet, although with 50 variables this same wavelet reaches results close to 100%, improving the results reported in [21]. Regarding the other wavelets examined, we can see how all of them, except for Haar, achieve an accuracy between 98% and 100% using 40 variables, while with 45 variables all of them exceed 99% of accuracy. On the other hand, if we compare the use of a linear discriminant versus the quadratic discriminant, we see that, for these data, the linear discriminant is more accurate.
EEG signal processing software plays a crucial role in transforming raw EEG data into usable information. These software programs typically include algorithms for filtering, artifact removal, feature extraction, and classification. Some popular options include EEGLAB, Brainstorm, and MNE-Python.
This system ensures that the naming of electrodes is consistent across laboratories. In most clinical applications, 19 recording electrodes (plus ground and system reference) are used.[51] A smaller number of electrodes are typically used when recording EEG from neonates. Additional electrodes can be added to the standard set-up when a clinical or research application demands increased spatial resolution for a particular area of the brain. High-density arrays (typically via cap or net) can contain up to 256 electrodes more-or-less evenly spaced around the scalp.
2. Machine Learning Algorithms
Machine learning algorithms have shown great promise in classifying EEG signals. By training these algorithms on labeled data, they can learn to distinguish patterns in the signals that correspond to different mental states or activities. Common machine learning algorithms used for EEG classification include Support Vector Machines (SVM), Random Forest, and Convolutional Neural Networks (CNN).
3. Feature Extraction Techniques
Feature extraction techniques are essential for identifying relevant information in EEG signals. These techniques involve extracting specific features from the signals, such as power spectral density, frequency bands, or time-frequency representations. By selecting the right features, researchers can improve the performance of their classification models.
4. Data Visualization Tools
Data visualization tools can help researchers gain insights into the patterns and trends present in EEG signals. By visualizing the data in various ways, such as spectrograms, topographical maps, or time-series plots, researchers can better understand the underlying dynamics of brain activity. Popular data visualization tools for EEG signals include MATLAB, Python’s Matplotlib, and NeuroPype.
Overall, utilizing a combination of EEG signal processing software, machine learning algorithms, feature extraction techniques, and data visualization tools can enhance the accuracy and efficiency of classifying EEG signals. By leveraging these tools effectively, researchers can uncover valuable insights into brain function and cognitive processes.
