Analysis of Arm Movement Prediction by Using the Electroencephalography Signal

Abstract

Various technological approaches have been developed in order to help those people who are unfortunate enough to be afflicted with different types of paralysis which limit them in performing their daily life activities independently. One of the proposed technologies is the Brain-Computer Interface (BCI). The BCI system uses electroencephalography (EEG) which is generated by the subject's mental activity as input, and converts it into commands. Some previous experiments have shown the capability of the BCI system to predict the movement intention before the actual movement is onset. Thus research has predicted the movement by discriminating between data in the "rest" condition, where there is no movement intention, with "pre-movement" condition, where movement intention is detected before actual movement occurs. This experiment, however, was done to analyze the system for which machine learning was applied to data obtained in a continuous time interval, between 3 seconds before the movement was detected until 1 second after the actual movement was onset. This experiment shows that the system can discriminate the "pre-movement" condition and "rest" condition by using the EEG signal in 7-30 Hz where the Mu and Beta rhythm can be discovered with an average True Positive Rate (TPR) value of 0.64 ± 0.11 and an average False Positive Rate (FPR) of 0.17 ± 0.08. This experiment also shows that by using EEG signals obtained nearing the movement onset, the system has higher TPR or a detection rate in predicting the movement intention. AbstrakAnalisis Prediksi Gerakan Tangan menggunakan Sinyal Elektroensefalografi. Berbagai pendekatan teknologi telah dikembangkan untuk membantu mereka yang menderita kelumpuhan dalam melakukan aktivitas kesehariannya secara mandiri. Salah satu teknologi tersebut adalah Brain-Computer Interface (BCI). Sistem BCI menggunakan elektroensefalografi (EEG) yang dihasilkan dari aktivitas mental seorang subjek sebagai masukan, dan mengubahnya menjadi perintah. Beberapa percobaan sebelumnya telah menunjukkan kemampuan sistem BCI untuk memprediksi gerakan sebelum gerakan tubuh aktual terjadi. Penelitian tersebut memprediksi gerakan yang akan terjadi dengan membedakan data pada kondisi rest, di mana tidak ada intensi gerakan, dengan kondisi pre-movement, di mana terdapat intensi gerakan sebelum gerakan aktual terjadi. Penelitian ini dilakukan untuk melakukan analisis sistem yang dihasilkan dari pembelajaran, yang kemudian diterapkan pada data dengan interval waktu kontinu, antara 3 detik sebelum gerakan terdeteksi sampai 1 detik setelah gerakan sebenarnya terjadi. Hasil percobaan menunjukkan bahwa sistem dapat membedakan kondisi premovement dan kondisi rest dengan menggunakan sinyal EEG pada frekuensi 7-30 Hz di mana letak Mu dan ritme Beta dengan nilai rerata true positive rate (TPR) sebesar 0.64 ± 0.11 dan rerata nilai false positive rate (FPR) sebesar 0.17 ± 0.08. Hasil percobaan juga mampu menunjukkan bahwa penggunaan sinyal EEG yang dekat dengan terjadinya gerakan, membuat sistem dapat mend...

Hybrid Brain-Computer Interface: a Novel Method on the Integration of EEG and sEMG Signal for Active Prosthetic Control

“…After all sensory information are received by the brain, humans deliberates on procedural activities to response to stimuli. Some examples of this phenomenon can be observed from event-related desynchronization/event-related synchronization of brain signals if the response(s) shows correlation with body movements [5][6][7][8][9][10][11][12][13][14][15][16][17][18]. Thus, a movement starts when an individual thinks or intends to move his body until the end of movement.…”

“…All movement intentions do not always become body movements. Somatosensory motor cortex, pre-motor cortex, primary motor cortex, and the thalamus are important brain areas for producing body movements [16][17][18]. Although a command signal has been generated by primary motor cortex, the signal can reach the muscles if the thalamus continually relays this signal.…”

Data Augmentation for Brain-Computer Interfaces: Analysis on Event-Related Potentials Data

“…Note that the processing is usually resistant to different amplitudes (scaling) of the pattern due to the aforementioned data standardization. Another example of ERP classification is movement prediction based on movementrelated cortical potentials (MRCPs) (e.g., [1,5,33]). MRCPs include pre-movement components that mainly reflect preparatory processes of the brain before a movement [35].…”

Exploring the Impact of Data Continuity on EEG-Based BCI: A Comparative Study of AI Models

“…Chai et al [17] conducted a study to classify mental character writing, mental arithmetic, and mental Rubix cube rolling tasks using blind source separation techniques based on independent component analysis. Darmakusuma et al [18]predicted movements by distinguishing between the resting and the state before movement occurred. Additionally, Huang et al [19] proposed a multiscale CNN model to improve the accuracy of EEG signal classification.…”