Feature and Classification Analysis for Detection and Classification of Tongue Movements From Single-Trial Pre-Movement EEG

Abstract

Individuals with severe tetraplegia can benefit from brain-computer interfaces (BCIs). While most movement-related BCI systems focus on right/left hand and/or foot movements, very few studies have considered tongue movements to construct a multiclass BCI. The aim of this study was to decode four movement directions of the tongue (left, right, up, and down) from single-trial pre-movement EEG and provide a feature and classifier investigation. In offline analyses (from ten healthy participants) detection and classification were performed using temporal, spectral, entropy, and template features classified using either a linear discriminative analysis, support vector machine, random forest or multilayer perceptron classifiers. Besides the 4class classification scenario, all possible 3-, and 2-class scenarios were tested to find the most discriminable movement type. The linear discriminant analysis achieved on average, higher classification accuracies for both movement detection and classification. The right-and down tongue movements provided the highest and lowest detection accuracy (95.3±4.3% and 91.7±4.8%), respectively. The 4-class classification achieved an accuracy of 62.6±7.2%, while the best 3-class classification (using left, right, and up movements) and 2-class classification (using left and right movements) achieved an accuracy of 75.6±8.4% and 87.7±8.0%, respectively. Using only a combination of the temporal and template feature groups provided further classification accuracy improvements. Presumably, this is because these feature groups utilize the movement-related cortical potentials, which are noticeably different on the left-versus right brain hemisphere for the different movements. This study shows that the cortical representation of the tongue is useful for extracting control signals for multi-class movement detection BCIs.

Adapting to progressive paralysis: A tongue-brain hybrid robot interface for individuals with amyotrophic lateral sclerosis

“…Other BCI signals, such as movement-related potentials could be implemented to act as the movement-pedal when a TCI can no longer provide this. Movement-related potentials are also generated when imagining or performing tongue movements, and could provide the users with an intuitive replacement for the TCI [83], [84]. These signals could also be used if the individuals reach a complete locked-in stage, in which the loss of gaze functionality could prove to be impactful on the SSVEP based BCI performance (as a gaze-independent SSVEP or P300 BCI has significantly lower performance [85]- [87]).…”

A Detailed Analysis & Parametric Comparison Of Eeg Processing Models

“…In terms of complexity GA SVM [24], ICA [25], Fuzzy [37], and LDA SVM [48] are preferred, while in terms of highspeed operations LDA [13], Linear [15], and GA SVM [24] must be used so that EEG signals can be quickly classified into different categories. In terms of scalability, XWT [42], and Linear [15] showcase better performance, thus can be used for a wide variety of clinical scenarios.…”

“…On the basis of which, it can be observed that WCNN [1], Fuzzy [4], XWT [42], SSD CNN [11], DAE CNN [50], and CVMD [8] showcase high accuracy, while WCNN [1], STFT CNN [2], CNN [9], CVMD [8], and DAE CNN [50] showcase high precision, and as a result, they can be utilized for a wide variety of high-performance EEG use cases. CWT [3] In terms of complexity, the GA SVM [24], the ICA SVM [25], the Fuzzy SVM [37], and the LDA SVM [48] are favored, while in terms of high-speed operations, the LDA [13], the Linear [15], and the GA SVM [24] must be employed so that EEG data may be swiftly sorted into distinct categories. XWT [42] and Linear [15] both provide higher performance in terms of scalability, and as a result, they are suitable for a broad range of clinical applications.…”

“…In addition, it is now possible to get your hands on some portable EEG acquisition gear. For example, emotive has seen significant use in the field of brain-computer interface [11][12][13] due to its low cost, mobility, and performance that is comparable to those of medical devices. However, despite the fact that a great deal of medical gear and portable EEG collecting devices generate the EEG data, which can be used for research on epilepsy, there are no standard data formats because there are so many different sources of data.…”

Implementing Performance Accommodation Mechanisms in Online BCI for Stroke Rehabilitation: A Study on Perceived Control and Frustration

“…It should be noted though that lower recognition rates still induce plasticity [ 8 ], although a better BCI performance was suggested to improve the induction of plasticity [ 9 ]. The BCI performance may be enhanced in various ways by selecting the optimal pre-processing techniques [ 24 , 25 , 26 ], features [ 25 , 27 , 28 , 29 , 30 ], classifiers [ 29 , 31 ], or by focusing on user instructions and training [ 23 , 32 , 33 ]. By improving the BCI performance, the patients’ perceived control and frustration improve as well [ 34 , 35 , 36 , 37 , 38 , 39 , 40 ], which may help them maintain interest in the training.…”