Towards closed-loop stimulation for stroke rehabilitation

DSpace Repositorium (Manakin basiert)


Dateien:

Zitierfähiger Link (URI): http://nbn-resolving.de/urn:nbn:de:bsz:21-opus-73964
http://hdl.handle.net/10900/50120
Dokumentart: Dissertation
Erscheinungsdatum: 2013
Sprache: Englisch
Fakultät: 7 Mathematisch-Naturwissenschaftliche Fakultät
Fachbereich: Informatik
Gutachter: Bogdan, Martin (Prof. Dr.)
Tag der mündl. Prüfung: 2014-01-27
DDC-Klassifikation: 004 - Informatik
Schlagworte: Schlaganfall , Rehabilitation , Hirnstimulation
Freie Schlagwörter: Gehirn-Computer-Schnittstelle , Adaptive Stimulation
Stroke rehabilitation , brain stimulation , brain-computer interface , closed-loop stimulation , brain-state-dependent stimulation
Lizenz: http://tobias-lib.uni-tuebingen.de/doku/lic_mit_pod.php?la=de http://tobias-lib.uni-tuebingen.de/doku/lic_mit_pod.php?la=en
Gedruckte Kopie bestellen: Print-on-Demand
Zur Langanzeige

Inhaltszusammenfassung:

Da für Schlaganfallpatienten mit Lähmungen der Hand momentan keine generell effektive Therapie existiert, ist die Entwicklung neuer Rehabilitationsverfahren ein wichtiges Ziel um die Effektivität der Physiotherapie zu verbessern. Relevante neue Entwicklungen für diese Arbeit sind kortikale Stimulation zur Erregung des Gehirngewebes und Gehirn-Computer-Schnittstellen (BCIs), die den Willen zur Bewegung der gelähmten Gliedmaße in Steuersignale für Orthesen umsetzen. In dieser Arbeit werden zwei Ansätze für die Kombination dieser Konzepte zu einem closed-loop System untersucht, das Stimulation in Abhängigkeit der aufgenommenen Hirnsignale und der daraus detektierten Intention des Patienten ermöglicht: (i) Zeitliche Kopplung der Stimulation an die Intention des Patienten, seine gelähmte Hand zu bewegen und (ii) Adaption der Stimulationsparameter zur Kontrolle der stimulationsevozierten Potentiale. Der Fokus dieser Arbeit liegt auf 3 Schlaganfallpatienten mit einer Lähmung der linken Hand, denen epidurale Elektroden zur Aufnahme des Elektrokortikogramms und zur elektrischen Stimulation implantiert wurden. Für den ersten Ansatz muss sichergestellt werden, dass die Stimulationseffekte die Merkmalsextraktion zur Dekodierung der Bewegungsintention möglichst wenig stören. Daher werden in dieser Arbeit Methoden zur Spektralschätzung während Stimulation vorgeschlagen und verglichen. Dabei wird gezeigt, dass der MEMgap-Algorithmus für die Bestimmung autoregressiver Modelle bei lückenhaften Daten, verzerrungsfrei, und damit am besten für die Anwendung geeignet ist. Aus diesen Ergebnissen entstand das erste bidirektionale BCI, das zur Rehabilitation von chronischen Schlaganfallpatienten eingesetzt wurde. Der zweite Ansatz, inspiriert durch Arbeiten zu visuellen Neuroprothesen, zielt darauf ab, die Variabilität der evozierten Potentialen durch die Anpassung von Stimulationsparametern an das gemessene Gehirnsignal zu verringern. Anhand von Experimenten mit Schlaganfallpatienten und gesunden Probanden wird gezeigt, dass die Hirnaktivität vor einem Stimulationspuls zwar einen signifikanten Einfluß auf die Form der evozierten Potentiale hat, dass aber die Interaktion zwischen der Stimulationsintensität, der Prä- und der Poststimulusaktivität nicht stark genug ausgeprägt ist, um ein closed-loop Stimulationssystem auf dieser Basis vielversprechend erscheinen zu lassen. Durch die Analyse der Abhängigkeit evozierter Potentiale von Stimulationsparametern und der Gehirnaktivität im Moment der Stimulation konnten aber neue elektrophysiologische Marker für neuronale Konnektivität und Rehabilitation vorgeschlagen werden. Dies ist eine der ersten Arbeiten zur Anwendung von closed-loop Stimulation bei Menschen und das erste Mal, dass die Ableitung von Hirnsignalen und Hirnstimulation zur Rehabilitation gleichzeitig durchgeführt wurde. Sie legt das Fundament für weitere Studien zu bidirektionalen BCIs zur Schlaganfallrehabilitation und eröffnet durch die Analyse evozierter Potentiale neue Möglichkeiten um den Rehabilitationsprozess zu überwachen und neue Stimulationssysteme zu optimieren.

Abstract:

The restoration of motor function to chronic stroke patients with a severe hand impairment is challenging, because no effective rehabilitation option exists. In recent years, brain-computer interfaces (BCIs) have shown potential to remedy this problem by decoding the patient’s movement intention from the brain signal and translating it into movements of an orthosis. In addition, brain stimulation has been investigated for the treatment of a variety of neurological conditions, including stroke. Combining both approaches into a closed-loop stimulation system has been proposed to increase the effectiveness of this treatment approach. However, this has not been put into practice before, therefore it is unknown, if closed-loop paradigms can be realized for stroke patients. In this thesis, closed-loop stimulation for stroke rehabilitation is studied for the first time. Two approaches for a closed-loop system are investigated: (i) Stimulation coupled to the movement intention of the patient and (ii) adaptation of stimulation parameters to control the shape of the evoked activity. The focus lies on 3 chronic stroke patients with a paralyzed left hand that had been implanted with epidural electrodes for electrocorticogram (ECoG) recording and electrical stimulation. To implement intention-dependent stimulation for the first approach, one has to compensate for the distortions introduced by the stimulation after-effects into the brain signal in order to ensure reliable decoding of the movement intentions. Methods to solve this problem are presented and compared in this work. The MEMgap algorithm for spectral estimation in the presence of gaps in the signal is shown to allow an unbiased decoding of the movement intention. With these algorithms, the first bidirectional BCI system used for stroke rehabilitation was implemented. The second approach, inspired by earlier work on visual neuroprosthetics, aims to reduce the variability of the evoked brain activity through adaptation of stimulation parameters to the ongoing brain activity. Experiments with stroke patients and healthy participants show that the brain activity preceding a stimulation pulse has a significant effect on the shape of the evoked activity but that the interaction of the stimulation intensity with the prestimulus brain activity in the generation of the evoked activity is too weak to make closed-loop stimulation for the stabilization of evoked activity promising. However, based on analysis of the dependency of stimulation-evoked neural activity on stimulation parameters and the brain activity at the moment of stimulation, novel electrophysiological markers for neural connectivity and for rehabilitation are proposed. This work is one of the first studies on closed-loop stimulation in humans and in particular the first study where such a stimulation paradigm had been applied for stroke rehabilitation. It sets the basis for the use of brain-state-dependent stimulation for stroke rehabilitation and introduces analysis methods for stimulation-evoked potentials to monitor the rehabilitation process and to optimize future stimulation systems.

Das Dokument erscheint in: