Phase synchronization analysis of event-related brain potentials in language processing [Elektronische Ressource] / von Carsten Allefeld
93 pages
Deutsch

Phase synchronization analysis of event-related brain potentials in language processing [Elektronische Ressource] / von Carsten Allefeld

Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres
93 pages
Deutsch
Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres

Informations

Publié par
Publié le 01 janvier 2004
Nombre de lectures 21
Langue Deutsch
Poids de l'ouvrage 1 Mo

Extrait

Institut fur¨ Physik
Arbeitsgruppe Nichtlineare Dynamik
Phase Synchronization Analysis
of Event-Related Brain Potentials
in Language Processing
Dissertation
zur Erlangung des akademischen Grades
doctor rerum naturalium (Dr. rer. nat.)
im Fach Physik / Nichtlineare Dynamik
eingereicht an der
Mathematisch-Naturwissenschaftlichen Fakult¨at
der Universit¨at Potsdam
von
Carsten Allefeld
Potsdam, M¨arz 2004Zusammenfassung
Das Forschungsthema Synchronisation bildet einen Schnittpunkt von Nichtlinearer Dyna
mik und Neurowissenschaft. So hat zum einen neurobiologische Forschung gezeigt, daß
die Synchronisation neuronaler Aktivitat¨ einen wesentlichen Aspekt der Funktionsweise
des Gehirns darstellt. Zum anderen haben Fortschritte in der physikalischen Theorie zur
Entdeckung des Phanomens¨ der Phasensynchronisation gefuhrt.¨ Eine dadurch motivierte
Datenanalysemethode, dieonisations Analyse, ist bereits mit Erfolg auf em
pirischeDatenangewandtworden.
DievorliegendeDissertationknupft¨ andiesekonvergierendenForschungslinienan.Ih
ren Gegenstand bilden methodische Beitrage¨ zur Fortentwicklung der Phasensynchronisa
tions Analyse, sowie deren Anwendung auf ereigniskorrelierte Potentiale, eine besonders
indenKognitionswissenschaftenwichtigeFormvonEEG Daten.
Die methodischen Beitrage¨ dieser Arbeit bestehen zum ersten in einer Reihe spezia
lisierter statistischer Tests auf einen Unterschied der Synchronisationsstarke¨ in zwei ver-
schiedenen Zustanden¨ eines Systems zweier Oszillatoren. Zweitens wird im Hinblick auf
denviel kanaligenCharaktervonEEG DateneinAnsatzzurmultivariatenPhasensynchro
nisations Analysevorgestellt.
Zur empirischen Untersuchung neuronaler Synchronisation wurde ein klassisches Ex
periment zur Sprachverarbeitung repliziert, in dem der Effekt einer semantischen Verlet
zungimSatzkontextmitdemjenigenderManipulationphysischerReizeigenschaften(Schrift
farbe) verglichen wird. Hier zeigt die Phasensynchronisations Analyse eine Verringerung
derglobalenSynchronisationsstarke¨ fur¨ diesemantischeVerletzungsowieeineVerstarkung¨
fur¨ die physische Manipulation. Im zweiten Fall laßt¨ sich der global beobachtete Synchro
nisationseffekt mittels der multivariaten Analyse auf die Interaktion zweier symmetrisch
gelegenerGehirnarealezuruckf¨ uhr¨ en.
Die vorgelegten Befunde zeigen, daß die physikalisch motivierte Methode der Phasen
synchronisations Analyse einen wesentlichen Beitrag zur Untersuchung ereigniskorrelier-
terPotentialeindenKognitionswissenschaftenzuleistenvermag.
Abstract
The topic of synchronization forms a link between nonlinear dynamics and neuroscience.
Ontheonehand,neurobiologicalresearchhasshownthatthesynchronizationofneuronal
activity is an essential aspect of the working principle of the brain. On the other hand,
recent advances in the physical theory have led to the discovery of the phenomenon of
phasesynchronization. Amethodofdataanalysisthatismotivatedbythisfinding—phase
synchronizationanalysis—hasalreadybeensuccessfullyappliedtoempiricaldata.
The present doctoral thesis ties up to these converging lines of research. Its subject
aremethodicalcontributionstothefurtherdevelopmentofphasesynchronizationanalysis,
as well as its application to event related potentials, a form of EEG data that is especially
importantinthecognitivesciences.
Themethodicalcontributionsofthisworkconsistfirstlyinanumberofspecializedsta
tistical tests for a difference in the synchronization strength in two different states of a sys
tem of two oscillators. Secondly, in regard of the many channel character of EEG data an
approachtomultivariatephasesynchronizationanalysisispresented.
Fortheempiricalinvestigationofneuronalsynchronizationaclassicexperimentonlan
guageprocessingwasreplicated,comparingtheeffectofasemanticviolationinasentence
context with that of the manipulation of physical stimulus properties (font color). Here
phasesynchronizationanalysisdetectsadecreaseofglobalsynchronizationfortheseman
ticviolationaswellasanincreaseforthephysicalmanipulation. Inthelattercase,bymeans
ofthemultivariateanalysistheglobalsynchronizationeffectcanbetracedbacktoaninter-
actionofsymmetricallylocatedbrainareas.
The findings presented show that the method of phase synchronization analysis mo
tivated by physics is able to provide a relevant contribution to the investigation of event
relatedpotentialsinthecognitivesciences.
3Contents
1 Introduction 7
1.1 Temporalbinding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2 Synchronization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Aimsandoutline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2 Basic Concepts 15
2.1 Electroencephalography . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Event relatedpotentials . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.3 Phasesynchronization . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3 Data Processing and Bivariate Analysis 33
3.1 Reductionofspuriouscorrelations . . . . . . . . . . . . . . . . . . . . 33
3.2 Determinationoftheinstantaneousphase . . . . . . . . . . . . . . . . 36
3.3 Quantification of bivariate phase synchronization and directional
statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4 Statistical Tests for Bivariate Phase Synchronization 47
4.1 Parametrictests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Asimplenonparametrictest . . . . . . . . . . . . . . . . . . . . . . . . 52
4.3 Bootstraptechniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.4 Datafromtimeseries . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5 Multivariate Phase Synchronization Analysis 57
5.1 Synchronizationclusteranalysis . . . . . . . . . . . . . . . . . . . . . 58
5.2 ApplicationtoERPdata . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6 A Language Processing Experiment 67
6.1 Experimentalsetupandanalysis . . . . . . . . . . . . . . . . . . . . . 68
6.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
7 Conclusion and Outlook 77
A Language Material 83
Bibliography 87
5Chapter 1
Introduction
Thetopicofsynchronizationhasrecentlyachievedanoutstandingroleinneurobi
ology and cognitive science as well as in nonlinear dynamics. In cognitive neuro
science,synchronizationismoreandmoreconsideredtobeoneofthebasicmecha
nisms of brain function, from visual perception up to highest cognitive processes.
Inphysics,thelongknownphenomenonofsynchronizationofperiodicoscillators
hasinthelastyearsbeenextendedtochaoticsystems. Theserecentadvanceshave
led to a cooperation of the two research fields which forms the background of the
presentthesis.
This chapter gives a short introduction into this converging research. It for-
mulatesthemaintheoreticalideasthathaveledtothecurrentinterestinsynchro
nizationprocessesinneuroscienceandsummarizesanumberofimportantstudies
thathavebeenpublishedinthiscontext. Thisincludesthefindingsofthephysical
theoryofsynchronizationinnonlineardynamicsandtheapplicationoftheresult
ing analysis methods to neurophysiological data. Finally, the chapter states the
1specificaimsofthisworkandgivesanoutlineofthefollowing.
1.1 Temporal binding
Neurobiologicalresearchhasshown(cf.Engeletal.,1991)thatinthevisualcortex
thereexistsahierarchyofneuronsthatdetectincreasinglycomplexfeaturesofthe
scene registered by the eyes. On the simplest level, they just copy the activation
patterns of sensory neurons in the retina, but subsequent cells react to contrasts,
movements, linear structures in a specific direction, and so on, each in a specific
areaofthevisualfield. Ifoneassumesthatthishierarchicalpatternofaccumulat
ing complexity is the functional principle of the whole brain, one has to conclude
that at the highest level for each object that is possibly relevant to the organism
thereexistsasinglededicatedneuronthatdetectsthespecialcomplexcombination
ofpropertiesthisobjectconsistsof. Sincesuchaschemeleadstoanabsurdlyhigh
number of combinations, it calls for far more neurons than are actually present in
the brain (the “combinatorial explosion”), and it implies the strange notion that
every contingent item of the world gets hard coded into the brain (presumably
during maturation), which has been caricatured by the idea of the “grandmother
neuron”—aneuronthatisactiveifandonlyifone’sgrandmotherispresent.
Ontheotherhand,forcomplexperceptionandcognitionitisnotsufficientthat
onlybasicstimulusfeaturesaredetectedinspecializedbrainareas. Ifanumberof
1In this introduction, the understanding of the concepts referred to has to be presumed. Many of
themwillbeexplainedinCh.2.
78 1 Introduction
Figure 1.1: The concept of binding, illustrated by a “bistable” image. The image
(a) has two possible interpretations: a face partially occluded by a candlestick (b),
or two opposing faces (c). Both interpretations are distinguished by the way the
edges that are detected at different places (marked by bold circles) are associated
with each other to make up a contour—that is, how they are “bound” into object
representat

  • Univers Univers
  • Ebooks Ebooks
  • Livres audio Livres audio
  • Presse Presse
  • Podcasts Podcasts
  • BD BD
  • Documents Documents