Development of EEG-based technologies for the characterization and treatment of neurological diseases affecting the motor function

Ibáñez Pereda, Jaime
Mínguez Zafra, Javier (dir.) ; Castillo Sobrino, María Dolores del (dir.) ; Serrano Moreno, José Ignacio (dir.)

Universidad de Zaragoza, 2014
(Informática e Ingeniería de Sistemas)


Abstract: This thesis presents a set of studies applying signal processing and data mining techniques in real-time working systems to register, characterize and condition the movement-related cortical activity of healthy subjects and of patients with neurological disorders affecting the motor function. Patients with two of the most widespread neurological affections impairing the motor function are considered here: patients with essential tremor and patients who have suffered a cerebro-vascular accident. The different chapters in the presented thesis show results regarding the normal cortical activity associated with the planning and execution of motor actions with the upper-limb, and the pathological activity related to the patients' motor dysfunction (measurable with muscle electrodes or movement sensors). The initial chapters of the book present i) a revision of the basic concepts regarding the role of the cerebral cortex in the motor control and the way in which the electroencephalographic activity allows its analysis and conditioning, ii) a study on the cortico-muscular interaction at the tremor frequency in patients with essential tremor under the effects of a drug reducing their tremor, and finally iii) a study based on evolutionary algorithms that aims to identify cortical patterns related to the planning of a number of motor tasks performed with a single arm. In the second half of the thesis book, two brain-computer interface systems to be used in rehabilitation scenarios with essential tremor patients and with patients with a stroke are proposed. In the first system, the electroencephalographic activity is used to anticipate voluntary movement actions, and this information is integrated in a multimodal platform estimating and suppressing the pathological tremors. In the second case, a conditioning paradigm for stroke patients based on the identification of the motor intention with temporal precision is presented and tested with a cohort of four patients along a month during which the patients undergo eight intervention sessions. The presented thesis has yielded advances from both the technological and the scientific points of view in all studies proposed. The main contributions from the technological point of view are: ¿ The design of an integrated upper-limb platform working in real-time. The platform was designed to acquire information from different types of noninvasive sensors (EEG, EMG and gyroscopic sensors) characterizing the planning and execution of voluntary movements. The platform was also capable of processing online the acquired data and generating an electrical feedback. ¿ The development of signal processing and classifying techniques adapted to the kind of signal recorded in the two kinds of patients considered in this thesis (patients with essential tremor and patients with a stroke) and to the requirements of online processing and real-time single-trial function desired for BCI applications. Especially in this regard, an original methodology to detect onsets of voluntary movements using slow cortical potentials and cortical rhythms has been presented. ¿ The design and validation in real-time of asynchronous BCI systems using motor planning EEG segments to anticipate or detect when patients begin a voluntary movement with the upper-limb. ¿ The proof of concept of the advantages of an EEG system integrated in a multimodal human-robot interface architecture that constitutes the first multimodal interface using the combined acquisition of EEG, EMG and gyroscopic data, which allows the concurrent characterization of different parts of the body associated with the execution of a movement. The main scientific contributions of this thesis are: ¿ The study of the EEG-based anticipation of voluntary movements presented in Chapter 5 of the thesis was the first demonstration (to the author's knowledge) of the capacity of the EEG signal to provide reliable movement predictions based on single-trial classification of online data of healthy subjects and ET patients. This study also provides, for the first time, the results of a BCI system tested in ET patients and it represents an original approach to BCI applications for this group of patients. ¿ It has been presented the first neurophysiological study using EEG and EMG data to analyze the effects of a drug on cortical activity and tremors of patients with ET. In addition, the obtained results have shown for the first time that a significant correlation exists between the dynamics of specific cortical oscillations and pathological tremor manifestation as a consequence of the drug effects. ¿ It has been proposed for the first time an experiment to inspect whether the EEG signal carries enough information to classify up to seven different tasks performed with a single limb. Both the methodology applied and the validation procedure are also innovative in this sort of studies. ¿ It has been demonstrated for the first time the relevance of combining different cortical sources of information (such as BP and ERD) to estimate the initiation of voluntary movements with the upper-limb. In this line, special relevance may be given to the positive results achieved with stroke patients, improving the results presented by similar previous EEG-based studies by other research groups. It has also been proposed for the first time an upper-limb intervention protocol for stroke patients using BP and ERD patterns to provide proprioceptive feedback tightly associated with the patients' expectations of movement. The effects of the proposed intervention have been studied with a small group of patients.

Pal. clave: actividad cerebral ; inteligencia artificial ; tratamiento de señales ; fisiología del movimiento

Knowledge area: Ciencia de la computación e inteligencia artificial

Department: Informática e Ingeniería de Sistemas

Nota: Presentado: 30 10 2014
Nota: Tesis-Univ. Zaragoza, Informática e Ingeniería de Sistemas, 2014

Creative Commons License



 Record created 2014-11-20, last modified 2019-02-19


Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)