Background: The brain-computer interface, also known as BCI, represents a new communication modality between the human brain and technology. This system is based on the recording, processing, decoding, and transformation of brain signals into commands that can be used to control artificial devices. The main components of the system are the human brain, the computer, and the interface. The primary application of BCI is in rehabilitation, particularly in the use of prosthetics to improve the quality of life for patients with motor disabilities or brain injuries. Additionally, it is used in research to study cognitive processes such as attention, memory, and perception. Objectives: The main objective of this thesis is to analyze a public dataset to identify and recognize instances of motor imagery (MI) in EEG recordings. The analysis focuses on applying advanced methods to precisely identify brain signals associated with the thought of movement, which are essential for the development of BCI and for the control of robotic devices. Materials and Methods: In this study, a public dataset of EEG recordings from 10 healthy subjects was used. The experimental setup included a calibration phase followed by an evaluation phase. The motor imagery (MI) classes analyzed were related to the right hand and left hand. The Common Spatial Pattern (CSP) algorithm was used for EEG signal analysis, while Linear Discriminant Analysis (LDA), a machine learning technique, was used for classification. Results: The results show a classification accuracy of 91%, classifying 50 trials, i.e., 50 movement attempts of the subjects between the left hand and the right hand. Conclusions: The automated recognition system is very effective and promising, as it does not require extensive training as is usually necessary for BCI. Therefore, it is suitable for use in rehabilitation clinics for patients with motor disabilities. Keywords: Brain-computer interface, EEG, motor imagery
Background: L'interfaccia cervello-computer, conosciuta anche come brain-computer interface (BCI), rappresenta una nuova modalità comunicativa tra il cervello umano e la tecnologia. Questo sistema si basa sulla registrazione, elaborazione, decodifica e trasformazione dei segnali cerebrali in comandi utilizzabili per il controllo di dispositivi artificiali. I principali componenti del sistema sono il cervello umano, il computer e l'interfaccia. L'applicazione principale della BCI è nella riabilitazione, particolarmente nell'uso delle protesi per migliorare la qualità di vita dei pazienti con disabilità motorie o lesioni cerebrali. Inoltre, viene utilizzata nella ricerca per studiare processi cognitivi come l'attenzione, la memoria e la percezione. Obiettivi: L'obiettivo principale di questa tesi è analizzare un dataset pubblico per individuare e riconoscere i momenti in cui è presente l'immaginazione motoria (MI) nei tracciati EEG. L'analisi si focalizza sull'applicazione di metodi avanzati per identificare con precisione i segnali cerebrali associati al pensiero del movimento, essenziali per lo sviluppo delle BCI e per il controllo di dispositivi robotici. Materiali e Metodi: In questo studio, è stato utilizzato un dataset pubblico di acquisizioni EEG registrate su 10 soggetti sani. Il set-up sperimentale comprendeva una fase di calibrazione seguita da una fase di valutazione. Le classi di immaginazione motoria (MI) analizzate erano relative alla mano destra e alla mano sinistra. Per l'analisi dei segnali EEG, è stato impiegato l'algoritmo Common Spatial Pattern (CSP), mentre per la classificazione è stata utilizzata l'Analisi Discriminante Lineare (LDA), una tecnica di machine learning. Risultati: I risultati ottenuti mostrano un'accuratezza di classificazione del 91%, classificando 50 prove, ovvero 50 tentativi di movimento dei soggetti tra mano sinistra e mano destra. Conclusioni: Il sistema automatizzato di riconoscimento è molto efficace e promettente, poiché non richiede un training estensivo come solitamente necessario per le BCI. È quindi adatto per l'uso nelle cliniche di riabilitazione in pazienti con disabilità motorie. Keywords: Brain-computer interface, EEG, immaginazione motoria
RICONOSCIMENTO AUTOMATIZZATO DEL MOTOR IMAGERY PER LE INTERFACCE BRAIN-COMPUTER
FAVERO, BEATRICE
2023/2024
Abstract
Background: The brain-computer interface, also known as BCI, represents a new communication modality between the human brain and technology. This system is based on the recording, processing, decoding, and transformation of brain signals into commands that can be used to control artificial devices. The main components of the system are the human brain, the computer, and the interface. The primary application of BCI is in rehabilitation, particularly in the use of prosthetics to improve the quality of life for patients with motor disabilities or brain injuries. Additionally, it is used in research to study cognitive processes such as attention, memory, and perception. Objectives: The main objective of this thesis is to analyze a public dataset to identify and recognize instances of motor imagery (MI) in EEG recordings. The analysis focuses on applying advanced methods to precisely identify brain signals associated with the thought of movement, which are essential for the development of BCI and for the control of robotic devices. Materials and Methods: In this study, a public dataset of EEG recordings from 10 healthy subjects was used. The experimental setup included a calibration phase followed by an evaluation phase. The motor imagery (MI) classes analyzed were related to the right hand and left hand. The Common Spatial Pattern (CSP) algorithm was used for EEG signal analysis, while Linear Discriminant Analysis (LDA), a machine learning technique, was used for classification. Results: The results show a classification accuracy of 91%, classifying 50 trials, i.e., 50 movement attempts of the subjects between the left hand and the right hand. Conclusions: The automated recognition system is very effective and promising, as it does not require extensive training as is usually necessary for BCI. Therefore, it is suitable for use in rehabilitation clinics for patients with motor disabilities. Keywords: Brain-computer interface, EEG, motor imageryFile | Dimensione | Formato | |
---|---|---|---|
Favero_Beatrice_2092736.pdf
accesso riservato
Dimensione
1.19 MB
Formato
Adobe PDF
|
1.19 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/69106