kognitive neurophysiologie des menschen human cognitive ... · illustrations: all ﬁgures should...

KOGNITIVE

NEUROPHYSIOLOGIE DES

MENSCHEN

HUMAN COGNITIVE

NEUROPHYSIOLOGY

c© 2009 W. Skrandies, Aulweg 129, D-35392 Giessenhttp://geb.uni-giessen.de/geb/volltexte/2008/6504/

ImpressumHerausgeber / Editor: Wolfgang Skrandies

c© 2009 W. Skrandies, Aulweg 129, D-35392 [email protected]

Editorial Board:M. Doppelmayr, SalzburgA. Fallgatter, WürzburgT. Koenig, BernH. Witte, Jena

ISSN 1867-576X

ii Human Cognitive Neurophysiology 2009, 2 (1)

Kognitive Neurophysiologie des Menschen wurde im Jahr 2008 gegründet. Hier sollenwissenschaftliche Artikel zu Themen der kognitiven Neurophysiologie des Menschen er-scheinen Sowohl Beiträge über Methoden als auch Ergebnisse der Grundlagen- und klinischenForschung werden akzeptiert. Jedes Manuskript wird von 3 unabhängigen Gutachtern beurteiltund so rasch wie möglich publiziert werden.Die Zeitschrift ist ein elektronisches ”Open Access”-Journal, ohne kommerzielle Interessen;http://geb.uni-giessen.de/geb/volltexte/2008/6504/.

Eine dauerhafte Präsenz der Zeitschrift im Internet wird durch die Universität Giessengewährleistet.

Human Cognitive Neurophysiology was founded in 2008. This journal will publish contribu-tions on methodological advances as well as results from basic and applied research on cogni-tive neurophysiology. Both German and English manuscripts will be accepted. Each manuscriptwill be reviewed by three independent referees.This is an electronic ”Open Access”-Journal with no commercial interest, published athttp://geb.uni-giessen.de/geb/volltexte/2008/6504/.

Online presence is guaranteed by the University of Giessen.

2009, 2 (1) Kognitive Neurophysiologie des Menschen iii

Instructions for Authors

Only original and unpublished work will be considered for publication unless it is explicitly statedthat the topic is a review. All manuscripts will be peer-reviewed. Both German and Englishversions are acceptable. After publication, the copyright will be with the editor of the journal.Usage of published material for review papers will be granted. Manuscripts (as WORD or TEXfiles) should be sent to [email protected].

Organization of manuscripts: The title page with a concise title should give the authors’ names,address(es), and e-mail address of the corresponding author. The manuscript should includean abstract in English (maximum 300 words). Organize your work in the sections Introduction,Methods, Results, Discussion, and Literature.

Illustrations: All figures should be submitted as jpeg or Coreldraw files. Please supplyfigure legends that explain the content of the figures in detail. Since this is an electronic journalcolor figures will be published free-of-charge.

The Literature should only include papers that have been published or accepted for publication.The reference list should be in alphabetical order by author. In the text, references should becited by author(s) and year (e.g. Johnson, Hsiao, & Twombly, 1995; Pascual-Marqui, Michel, &Lehmann, 1994; Zani & Proverbio, 2002).

Examples of reference formatJohnson, K., Hsiao, S., & Twombly, L. (1995). Neural mechanisms of tactile form recognition. In

M. Gazzaniga (Ed.), The Cognitive Neurosciences (p. 253-267). Cambridge, Mass.: MITPress.

Pascual-Marqui, R., Michel, C., & Lehmann, D. (1994). Low resolution electromagnetic tomog-raphy: a new method for localizing electrical activity in the brain. International Journal ofPsychophysiology , 18, 49-65.

Zani, A., & Proverbio, A. (Eds.). (2002). The Cognitive Electrophysiology of Mind and Brain.Academic Press, San Diego: Elsevier USA: Academic Press San Diego.

iv Human Cognitive Neurophysiology 2009, 2 (1)

Inhalt — Contents

Inhalt — Contents

W. Skrandies — Abstracts of the 17th German EEG/EP Mapping Meeting . . . . . . . 1M. Doppelmayr, E. Weber, K. Hoedlmoser & W. Klimesch — Effects of SMR Feedback

on the EEG Amplitude . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21A. J. Fallgatter, Ch. G. Baehne, A.-Ch. Ehlis, M. M. Plichta, M. M. Richter, M. J.

Herrmann, A. Conzelmann, P. Pauli, Ch. Jacob, & K.-P. Lesch — Impact of COMTGenotype on Prefrontal Brain Functioning in Adult ADHD . . . . . . . . . . . . . 33

Ankündigungen – Announcements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

0 Human Cognitive Neurophysiology 2009, 2 (1)

W. Skrandies — Abstracts of the 17th German EEG/EP Mapping Meeting

Abstracts of the 17 thGermanEEG/EP Mapping Meeting,

Giessen, October 31 -November 2, 2008

”Imaging Genetics” with endophenotypicmeasures of attention-deficit hyperactivitydisorder. A. J. Fallgatter, M. M. Plichta, C.Baehne, M. M. Richter, M. M. Schecklmann,K.-P. Lesch & A.-C. Ehlis, Department of Psy-chiatry, Psychosomatics and Psychotherapy,University of Würzburg, Würzburg, Germany.Deficits in response inhibition are, amongstothers, considered as candidate endopheno-types of altered prefrontal brain function inADHD. Based on their superior time resolu-tion, electrophysiological methods like Event-Related Potentials (ERPs) are adequate forthe measurement of such endophenotypes,i.e. abnormalities in brain functions underly-ing psychiatric diseases like ADHD. Moreover,ERPs seem to be particular suited to measureeffects of functionally relevant genetic variantsdirectly affecting neurotransmission systemsand brain function. This principle of imaginggenetics with ERPs has been demonstrated asearly as 1999 for the serotonin transporter pro-moter polymorphism affecting prefrontal brainfunction (Fallgatter et al., International Jour-nal of Neuropsychopharmacology, 1999). Weemployed a multi-channel EEG during perfor-mance of a Go-NoGo task to assess the elec-trophysiological basis of the endophenotyperesponse inhibition in healthy subjects as wellas in patients with ADHD. The ERP-measurederived from this protocol was termed NoGo-Anteriorisation (NGA) and is characterized by

a high interindividual stability, high short- andlong-term test-retest reliability and, moreover,is independent from age and gender. In pa-tients with ADHD the NGA was diminishedas compared to age- and sexmatched healthycontrols. Furthermore, a three-dimensionalsource location analysis with Low ResolutionElectromagnetic Tomography (LORETA) indi-cated an electrical dysfunction of the medialprefrontal cortex comprising the anterior cin-gulate cortex (ACC) in ADHD patients in child-hood as well as in adulthood. Recent stud-ies showed a significant influence of variantsof dopaminergic as well as serotonergic geneson this measure of prefrontal brain function.These results exemplify the imaging geneticsapproach by measurement of disease relateddisturbances in brain function with ERPs. Fu-ture studies will show whether such electro-physiological endophenotypes may contributeto the diagnosis of subgroups of ADHD andwhether they may serve as endophenotypes tofurther clarify genetic contributions to the dis-ease.

Frontal brain function in schizophrenias:Topographical ERP measures in the diag-nosis and treatment of psychotic disorders.A.-C. Ehlis (1), P. Pauli (2), M. J. Herrmann(1) & A. J. Fallgatter (1), (1) Department ofPsychiatry, Psychosomatics and Psychother-apy, University of Würzburg, Würzburg, Ger-many (2) Department of Psychology I, Univer-sity of Würzburg, Würzburg, Germany.Schizophrenic illnesses have been associatedwith a wide range of neurophysiological abnor-malities, one of the most consistent findingsconcerning a reduced metabolism and dimin-ished functional activation of the frontal cor-tex (”cerebral hypofrontality”). We conducted

2009, 2 (1) Kognitive Neurophysiologie des Menschen 1


a series of experiments on prefrontal dysfunc-tion in different subgroups of psychotic disor-ders, investigating its diagnostic and prognos-tic meaning as well as the influence of an-tipsychotic drugs on measures of frontal lobefunction. We used the NoGo-Anteriorization(NGA) - that quantifies the frontalization of thebrain electrical field during processes of in-hibitory control and has been associated withincreased prefrontal activation - as a topo-graphical ERP measure of frontal lobe func-tion. We observed a) diminished prefrontal ac-tivation in schizophrenic patients during tasksof executive control; b) more pronounced ab-normalities in patients with a chronic progres-sion than in subgroups with a phasic course ofthe disease; c) a more positive impact of atyp-ical than typical antipsychotics on prefrontalbrain function; and d) prognostic properties ofthe NGA in predicting the treatment responseto different kinds of antipsychotic medication.The results are in line with previous findingsand with assumed antipsychotic drug actions,and confirm the usefulness of topographicalERP measures in neuro-psychiatric research.

Neural correlates of executive functionsin Attention-Deficit/Hyperactivity Disorder(ADHD). N. Wild-Wall (2), R. D. Oades (1),M. Schmidt-Wessels (1), H. Christiansen (1) &M. Falkenstein (2), (1) Biopsychology Group,University Clinic for Child and Adolescent Psy-chiatry and Psychotherapy, Essen, Germany(2) Leibniz Research Centre for Working En-vironment and Human Factors, Institute of Oc-cupational Physiology, University of Dortmund,Dortmund, Germany.The present study is aimed at assessingchanges of different executive functions andtheir neural correlates in children with ADHD

and their unaffected siblings compared tohealthy control children. Specifically, the pro-cessing of irrelevant stimuli, the control overinappropriate responses, and the detection oferrors, is analysed, using a modified flankertask. Event-related potentials are used todirectly measure these processes. While thebehavioral data showed no group differences,the ERPs did. The post-stimulus ERPs mainlyshowed an attenuation of flanker processing(P1) and of late inhibitory processes (N2, P3a)in ADHD. The post-response ERPs showeda general enhancement of the post-responsenegativity in the siblings. In particular therewas no significant attenuation of error pro-cessing, as reflected in Ne/ERN and Pe, inthe patients. Finally, preparatory processeswere attenuated in siblings and more so inthe patients. The pattern of results revealsspecific changes of various cognitive controlprocesses in ADHD which are not reflectedin overt behaviour. The larger post-responsepotentials in the unaffected siblings may re-flect compensatory enhancement of responsemonitoring. The study shows that ERPs havean additive value for assessing subtle cognitivechanges in ADHD.

Monoamine challenge tests in psychiatryresearch: results of neurophysiologicalstudies. C. Norra, Department of Psychia-try, LWL University Hospital, Ruhr University,Bochum, Germany.Monoaminergic challenge tests allow inves-tigating central nervous changes in humansunder acute depletion of specific neurotrans-mitters (5-HT, DA, NE). In the design of thesestudies, various biochemical and methodolog-ical aspects have to be taken into account.Here, the tryptophan depletion test (TDT) rep-



resents the currently most established humanchallenge tool for the assessment of brainserotonin functioning and alterations in vari-ous psychiatric disorders. With respect to oneproposed electrophysiological marker of theserotonergic system, the stimulus intensity ofauditory evoked potentials, TDT in animal andclinical tudies suggests an inverse influence onthe 5-HT neurotransmission. However, similareffects of TDT in healthy humans remainedmostly unconfirmed including own studiesshowing only a slight and non-significantincrease of the loudness dependence. Asopposed to other selective challenges withe.g. SSRI, interactions of TDT with furthertransmitter systems are worth discussing.Regarding auditory sensory gating and pro-cessing, TDT led to reduced acoustic startleamplitudes, but no change of prepulse inhibi-tion. Further results of monoamine depletionin electrophysiological studies (i.e. electroen-cephalography, magnetoencephalography,polysomnography, evoked potentials) will bepresented. All in all, depletion techniques offera potential non-invasive and reliable biologicalmarker of human monoaminergic vulnerabil-ity or dysfunction to investigate changes ofneurophysiological parameters.

Electrophysiological correlates of execu-tive functions in psychiatric patients. M.Ruchsow (1), M. Spitzer (2), M. Kiefer (2), G.Grön (2), M. Falkenstein (3) & L. Hermle (1),(1) Dept. of Psychiatry Christophsbad, Göp-pingen, Germany (2) Dept. of Psychiatry, Uni-versity of Ulm, Germany (3) Leibniz ResearchCentre for Working Environment and HumanFactors (IfADo), Dortmund, Germany.Dysfunctions of prefrontal and anterior cin-gulate cortex have been found in a variety

of psychiatric diseases like major depression(MDD), obsessive-compulsive disorder (OCD),and borderline personality disorder (BPD) re-sulting in impaired executive functions whichencompass - among others - error monitor-ing and response inhibition. The Ne/ERN isa negative going ERP component associatedwith erroneous responses, whereas Nogo-N2and Nogo-P3 are closely related to responseinhibition processes. Several ERP studieswere performed in order to investigate whetherNe/ERN, Nogo-N2 and Nogo-P3 are usefultools for the assessment of depressiveness,impulsivity, and compulsiveness. Participantsperformed a hybrid flanker-Go/Nogo paradigmwhile a 64-channel EEG was recorded. BPDpatients had reduced and OCD patients hadenhanced Ne/ERN amplitudes in relation totheir respective control groups. With regardto response inhibition patients with MDD andBPD had reduced Nogo-P3 amplitudes andpatients with OCD had enhanced Nogo-N2amplitudes compared to their respective con-trol groups. Three ERP components werefound to mirror the psychopathological state ofpsychiatric patients. It remains an open ques-tion whether these ERP parameters are ableto predict the clinical outcome, as well.

Tomographic neurofeedback in healthyadults - implications for ADHD treatment.M. Liechti (1,2), P. Ianiro (1,2), M. Mächler (1),M. Döhnert (1,3), L. Jäncke (2), D. Brandeis(1) & R. Drechsler (1), (1) Department ofChild and Adolescent Psychiatry, Universityof Zürich, Zürich,Switzerland (2) Departmentof Neuropsychology, Institute for Psychology,University of Zürich, Zürich, Switzerland (3)Department of Child and Adolescent Psychia-try, University of Leipzig, Leipzig, Germany.



Tomographic neurofeedback (NFB) based onmultichannel scalp electroencephalographic(EEG) aims to control intracerebral activityin specific brain regions. In the context ofimproving NFB for treatment of Attention-Deficit-/Hyperactivity-Disorder (ADHD), tenhealthy adults (age: 23-33 years) were trainedto increase their beta/theta ratios in two con-secutive counterbalanced blocks targetingdifferent prefrontal regions (Anterior cingulatecortex = ACC and right dorsolateral prefrontalcortex = DLPFC). Pre- and post-tests includedresting EEG, and tests of attention (TAP).EEG power in different frequency bands wasanalyzed. We hypothesized that successfulregion specific NFB has corresponding effectson resting EEG, neuropsychological tests andsubjective ratings. Subjects partly learned tocontrol their prefrontal activity, given subjec-tive feelings of control, but significant objectiveimprovement was limited to the ACC training.Objective control of ACC and DLPFC activitywas uncorrelated. ACC training improved TAPmeasures of flexibility, DLPFC training mea-sures of sustained attention and inhibition.Resting EEG with eyes closed was unaf-fected. The results provide first evidence thattomographic ACC NFB has specific cognitiveeffects. Its clinical relevance remains to beevaluated with subjects suffering from ADHD.(Supported by the SBF COST B27 ENOC andby a Grant to the GD of the Kanton of Zurich.)

Diagnostic EEG mapping in children withADHD - a failure to replicate. D. Brandeis (1),M. Liechti (1,2), L. Valko (1), U. Müller (1), S.Maurizio (1), H.-C. Steinhausen (1), R. Drech-sler (1) & M. Döhnert (1,3), (1) Departmentof Child and Adolescent Psychiatry, Universityof Zürich, Zürich, Switzerland (2) Department

of Neuropsychology, Institute for Psychology,University of Zürich, Zürich, Switzerland (3)Department of Child and Adolescent Psychi-atry, University of Leipzig, Leipzig, Germany.Recent work suggests that resting EEG thetaand theta/beta ratios at Cz are so consistentlyelevated in ADHD that they can serve as di-agnostic markers. We attempted a replica-tion in a typical child psychiatric EEG researchsetting and explored the topography of thesepresumed markers of ADHD. Resting EEG(46 EEG, 2 EOG electrodes) was mappedfor 3 min each with eyes closed and eyesopen. Children (8.5-13y) with ADHD com-bined (N=44) and without ADHD symptoms(N=20 normal controls) were compared. Stan-dard FFT of the EEG (256Hz/ 2s epochs) us-ing average and mean ear reference followedmanual plus automatic artefact rejection. In-creases of absolute theta, relative theta, andtheta/beta ratio in the ADHD group were testedon 12 measures (2 conditions x 2 references x3 markers at Cz). Topographies of group differ-ences were explored with t-maps. For none ofthese 12 measures, increases in ADHD couldbe replicated (at p<.05 uncorrected), and noconsistent increases were found at other elec-trodes. Instead, some measures even tendedto reduced (e.g., relative theta /eyes closed/average reference). These results demon-strate that ”classical” quantitative EEG mark-ers can not be used for ADHD diagnostics intypical child psychiatry settings. (Supported bythe SNF project ”MFAA” (32-109591))

Clinical neuro-monitoring by visual evokedpotentials - New ways for useful perfor-mance. P. Christophis, M. Preuss, F. Wanis& V. Schreiber, Department of Neurosurgery,University of Giessen, Giessen, Germany.



Intraoperative monitoring via acoustic andsomato-sensory evoked potentials is suitableto protect brain stem functions during surgeryand is routinely used. Conflicting reports ofthe usefulness of intra-operative monitoringvia visual evoked potentials (VEP) and also ofthe VEP derivation in patients of the intensivecare unit, were the cause for this study. Mostinteresting for the neurosurgeon is the moni-toring via flash-VEP (f-VEP) during operationsalong the visual pathways for detect of immi-nent lesions and minimise the risk of lesionsof visual system respectively. Own pilot studyin 6 cases and last reports show a possibilityto realise an intraoperative f-VEP-monitoringin intravenous anaesthesia in about 50% ofthe cases. In the present follow-study we de-rive the steady-stay f-VEP and use frequencyanalytic methods for clear identification of theP100 and of previous f-VEP-waves in healthypersons additionally. In a second step we willderive f-VEP in brain-healthy persons underintravenous narcosis to find the best combi-nation of the used drugs, which allow a clearidentification of P100 and also of the previousf-VEP-waves. The third and last step will bethe use of f-VEP-monitoring in neurosurgicaloperations along the visual pathways.

EEG correlates of fMRI resting state net-works. T. Koenig (1), K. Jann (1), M. Kottlow(1), C. Boesch (2) & T. Dierks (1), (1) Dept. ofPsychiatric Neurophysiology, University Hos-pital of Psychiatry, University of Bern, Bern,Switzerland (2) Dept. of Clinical Research(AMSM), University of Bern, Bern, Switzer-land.Functional synchronization of brain areas is amain concept of how information processingin the brain is performed. Synchronous oscil-

lations have been observed and investigatedin EEG during the past decades and more re-cently such oscillations, however in a differenttimescale, have been discussed in fMRI. Os-cillations occur over a wide frequency range.In EEG, one prominent oscillation is the al-pha rhythm, present when a subject is re-laxed keeping the eyes closed. The spectralpower of the alpha rhythm has recently beeninvestigated in simultaneous EEG/fMRI con-sidering the relationship between neuronal ac-tivity and BOLD fluctuations. Even more re-cently, spontaneous fluctuations of the BOLDsignal in fMRI measurements have been ob-served establishing the notion of so called rest-ing state networks (RSN). The similarity of thenetworks delineated by either method is sur-prising considering their different temporal dy-namics. Hence, their relationship is presentlyinvestigated. One hypothesis is that the sub-units assembling a specific RSN might be co-ordinated by different EEG rhythms. Simulta-neous EEG/fMRI recordings were performedin 10 healthy subjects. Time-frequency anal-ysis on the EEG data and Independent Com-ponent Analysis (ICA) on the fMRI BOLD sig-nal were performed to reveal relationships be-tween EEG features and RSNs. Correlationscould be found between different RSNs andEEG rhythms on individual data sets, evidenc-ing the hypothesis that BOLD oscillations arelinked to EEG oscillations. The understandingof relations between BOLD signal fluctuationsand associated EEG rhythms may also play arole in pathology, since several psychiatric dis-orders show changes in BOLD signals and/orEEG oscillations.

Simultaneous EEG-fMRI during a workingmemory task. L. Michels (1,4), K. Bucher



(2,4), E. Martin (2,4) , D. Jeanmonod (1,4) &D. Brandeis (3,4), (1) Functional Neurosurgery,University Hospital Zürich, Zürich, Switzer-land (2) MR-Center University Children’s Hos-pital, Zürich, Switzerland (3) Department ofChild and Adolescent Psychiatry, University ofZürich, Zürich, Switzerland (4) Zürich Cen-ter for Integrative Human Physiology (ZIHP),Zürich, Switzerland.

Increased EEG alpha power is often reportedduring resting states. However, the role ofalpha during cognitive tasks is still unclear.For theta it has been reported that the spectralpower increases with working memory load atfrontal electrodes. Furthermore, negative cor-relations between frontal theta and the fMRI(BOLD) signal have been found within the de-fault network (DFN) during rest. In this study,we used time-frequency analyses to inves-tigate correlations between task-dependentfluctuations in theta and alpha power and theBOLD signal. Brain activity was measured bysimultaneous EEG-fMRI in 16 healthy adults.During the task, sets of either 2 or 5 letterswere presented for 2s. Subjects had to retainthem in memory to deciding whether a probeletter shown after 3.5s was part of the set.Occipital alpha power negatively correlatedwith BOLD activity in parietal and lateral pre-frontal areas. The same cortical network wasobserved by the fMRI contrast ’task vs. rest’.Our findings demonstrate that alpha is not onlya marker for resting state activity but furthermodulates cognitive processing. Frontal thetapower at AFz increased with memory load andthese theta load effects were correlated withBOLD load effects in prefrontal regions andthe angular gyrus. Frontal theta activity wasnegatively correlated with the BOLD signal in

fronto-parietal areas. A comparison of theseregions with the DFN revealed a strong over-lap. We propose that resting state activity inthe DFN is at least partially controlled by adesynchronization of frontal theta EEG.

Dimensions of meaning in multi-sensoryperception: Comparison of early and latecomponents of evoked potentials. A. Hiessl,S. Hörner & W. Skrandies, Institute of Physi-ology, Justus Liebig University, Giessen, Ger-many.The semantic differential is used to define con-notative dimensions of meaning, and the pro-cessing of words by the brain depends on suchdimensions. Earlier studies demonstrated thatstimuli of the different semantic classes lead todifferences in neuronal processing. We investi-gated the influence of connotative meaning onmulti-sensory processing (words strongly re-lated to odour, taste, vision or texture). A groupof 795 subjects rated 197 food words on thebasis of 11 pairs of opposed adjectives. Fac-tor analysis revealed three dimensions (Evalu-ation, Potency and Texture). Words with highpositive or negative scores and low scoreson the other dimensions were used as stim-uli in an ERP experiment with 40 adults. EEGwas recorded from 30 channels, and aver-aged according to semantic stimulus class.Component latency, global field power and to-pography were influenced by semantic mean-ing. We found differences between early (be-tween 80 and 200 ms) and late (around 500ms) ERP components. Major differences wereseen when the polarity of stimulus classeswas compared: GFP evoked by positive wordclasses was smaller than that evoked by nega-tive classes with early components. This effectwas reversed with late components. Similar ef-



fects were observed with component latency.In summary, semantic dimensions influenceneuronal processing of words related to multi-sensory perception. The semantic dimensionsfound were similar to those described earlier.Instead of ”activity” we found ”texture” whichreflects the ”mouthfeel” of food items. In addi-tion, we describe early effects on ERP compo-nents, and the influence of semantic meaningwas different with late cognitive components.

Dimensions of meaning in multi-sensoryperception: Gender-specific differences ofevoked potentials. S. Hörner, A. Hiessl &W. Skrandies, Institute of Physiology, JustusLiebig University, Giessen, Germany.According of Osgood every word is placed ina 3D space of connotative meaning that isdefined by the factors ”evaluation”, ”potency”,and ”activity”. One aim of our study was to in-vestigate whether there are additional factorswhich represent sensory modality. A total of197 food words were rated by 795 subjects,and in addition to ”evaluation” and ”potency”the dimension ”texture” emerged which is re-lated to the consistency and surface of fooditems. Females had a larger awareness of”healthy” and ”unhealthy” food, and they dif-ferentiated more exactly among different foodcategories. These results provided the ba-sis for an ERP study in which unambiguousfood words were presented to 40 healthy sub-jects (20 males, 20 females). Each stimu-lus belonged to one of six classes (E+, E-, P+, P-, T+, T-.). EEG was recorded from30 channels, and averaged according to se-mantic stimulus class. Peaks of Global FieldPower (GFP) defined 11 components whichwere then compared between males and fe-males. With all components GFP was signifi-

cantly higher in females than in males. In ad-dition, there were gender-specific differencesin scalp topography revealed by the scalp lo-cation of centroids: as early as at 107 ms thelocation of the positive centroids was close tothe midline with females while in male sub-jects there was a significant lateralization to-wards the right hemisphere. A number of othereffects with late components support the no-tion that there were differences between malesand females in electrophysiological parame-ters. In summary, we could show that thereare gender-specific differences in the evalua-tion of food items as well as in neurophysio-logical processing of multi-sensory stimuli be-longing to different semantic classes.

Eye dominance and nasal-temporal retinaldifferences are reflected by the scalp to-pography of visual evoked brain activity.W. Skrandies, Institute of Physiology, JustusLiebig University, Giessen, Germany.There are anatomical and neurophysiologicaldifferences between nasal and temproal areasof the human retina. Thus, we expect thateye dominance affects cortical activity. In agroup of 23 healthy adults with different eyedominance checkerboard reversal activity wasmeasured from 30 channels (binocular stimu-lation; 2 reversals / sec; 21 x 15 test field;50’ check size; 97% contrast). Stimulation oc-curred either in the center, or in the left or righthalf-field. Latency, global field power (GFP),and topography were compared between ex-perimental conditions. Mean latency (110.75ms) and GFP were not affected while topog-raphy of the P100 component displayed thewell known paradoxical lateralization. Analy-sis of the scalp location of the positive cen-troids revealed small but significant differences



between left or right eye dominant subjects:with central fixation, the centroids of subjectswith right eye dominance showed a strongerlateralization over the right hemisphere whilesubjects with left eye dominance displayed theopposite pattern (more extreme lateralizationover the left hemisphere). This was signifi-cant (F(1,21) = 6.62; p = 0.017). Similar ef-fects were seen with lateralized stimuli. Sincethe human pattern ERG demonstrates differ-ences between nasal and temporal retinal ar-eas (Skrandies & Leipert, Int. J. Neuroscie.,1988, 39, 137-146), our present results mightreflect basic functional differences extendingfrom the retina to the visual cortex of man.

Resting EEG-based intracortical func-tional connectivity in Alzheimer’s diseaseand frontotemporal degeneration. F. M.Dahinden (1,2), L. R. R. Gianotti (1), G.Künig (3), R. D. Pascual-Marqui (1), P. L.Faber, (1), D. Lehmann (1), K. Kochi (1) & U.Schreiter-Gasser (4), (1) The KEY Institutefor Brain-Mind Research, University Hospitalof Psychiatry, Zürich, Switzerland (2) De-partment of Neuropsychology, University ofZürich, Zürich, Switzerland (3) Departmentof Geriatric Psychiatry, University Hospital ofPsychiatry, Zürich, Switzerland (4) Praxis fürPsychiatrie Rehalp, Zürich, Switzerland.We recorded resting-state EEG (19 chan-nels) from two groups of mild-to-moderateAD (n=21) and FTD (n=16) patients and agroup of age-matched healthy controls (n=17).sLORETA (Pascual-Marqui, Methods FindExp Clin Pharmacol, 2002, 24, Suppl. D,5-12 ) was applied to calculate the sourcestrength of the intracortical generators innine regions of interest (ROI, frontal left/right,temporal left/right, central, parietal left/right,

occipital left/right) and in the seven classi-cal frequency bands. A 3-way ANOVA withgroup as between-subject factor and ROI andfrequency bands as within-subject factors re-vealed a significant interaction between groupx frequency bands x ROI (F=1.7, p<0.0001).Post-hoc analyses (t-statistics) confirmed asignificantly decreased connectivity in AD rel-ative to healthy controls in the alpha1 band(8.5-10Hz). Also, AD showed a significantlydecreased connectivity compared to FTD inthe alpha1 and the theta (6.5-8Hz) frequencybands. The decreased connectivity in ADrelative to the two groups concerned thewhole brain, suggesting a generalized loss offunctional interactions between different brainregions in AD. In line with previous reports(e.g., Pijnenburg et al., Clin Neurophysiol,2008) which found no detectable deviations ofEEG parameters in FTD in spite of widespreadneuronal degeneration, our study’s EEG in-tracerebral connectivity remained in the nor-mal range during mild-to-moderate stages ofthe disease.

Scalp EEG connectivity and intracerebralelectrical connectivity (sLORETA laggedcoherence) during resting and five medi-tation traditions. P. L. Faber (1), S. Tei (2),R. D. Pascual-Marqui (1), L. R. R. Gianotti (1),H. Kumano (2), K. Kochi (1) & D. Lehmann(1), (1) The KEY Institute for Brain-MindResearch, University Hospital of Psychiatry,Zürich, Switzerland (2) Department of StressScience and Psychosomatic Medicine, Grad-uate School of Medicine, The University ofTokyo, Tokyo, Japan.Brain electric functional connectivity was stud-ied in experienced meditators of five traditions(13 Tibetan Buddhists ’TB’, 15 QiGong ’QG’,



14 Sahaja Yoga ’SY’, 14 Ananda Marga Yoga’AY’, 15 Soto Zen ’Zen’) during tradition-specific meditation (self-dissolution, QiGong,Samadhi, Satori) and during wakeful restingbefore (’rest1’) and after (’rest2’) meditation.EEG (19-56 electrodes) was computed (viasLORETA, current density in 6239 voxels)into intracerebral waveshapes of 19 intrac-erebral regions (ROIs) that correspond tocortex underlying the 10/20 electrode posi-tions. Functional connectivity was computedfrom scalp-recorded data as conventionalcoherence between 19 locations, and fromsLORETA waveshapes as ’lagged coherence’between 19 ROI’s; lagged coherence onlymeasures connections with time delay; theseare interpretable as true functional connectiv-ity. - For each meditator group, t tests identi-fied significant coherence differences betweenrest1 vs meditation and rest2 vs meditationin each of 8 EEG frequency bands (delta togamma). Between 19 locations or ROIs, thereare 171 connections. For each subject and fre-quency band, the percentage of connectionswere counted that reached significant differentcoherence between rest1 vs meditation andrest2 vs meditation; from these two values,mean was computed, and averaged across all8 bands for each tradition separately. For scalpcoherences, in the 5 traditions, between 1% to4% of the connections were significant higherin meditation than rest, between 6% to 36%lower; for intracerebral lagged coherence, 0%were higher, between 26% to 68% were lower.On average across the 5 traditions, scalp co-herence decreased most strongly in alpha1&2and beta1&2, while intracerebral lagged co-herence decreased most strongly in delta,theta, beta1&2. For the gamma frequency

band alone, scalp coherences were higher be-tween 1% to 13%, lower between 1% to 27%;intracerebral lagged coherences were higherin 0%, lower between 2% to 75% of cases.In sum, all 5 traditions clearly showed moresignificant decreases than increases in scalpcoherence, and only significant decreases, noincreases in intracerebral lagged coherencethat avoids distorting volume conduction. Con-trary to published reports of strongly increasedgamma band coherence in meditation, our5 traditions on average in scalp coherenceincreased significantly only 4% of the gammaband coherences while 9% decreased; in in-tracerebral lagged coherence none increasedbut 44% decreased. (Partial support by BialGrant No. 44 2006/2007 ).

Active intracerebral areas (EEG LORETA)in non-meditators and experienced medi-tators differ during resting. P. L. Faber (1),S. Tei (2), D. Lehmann (1), L. R. R. Gian-otti (1), T. Tsujiuchi (3), H. Kumano (2) & K.Kochi (1), (1) The KEY Institute for Brain-MindResearch, University Hospital of Psychiatry,Zürich, Switzerland (2) Department of StressScience and Psychosomatic Medicine, Grad-uate School of Medicine, The University ofTokyo, Tokyo, Japan (3) Department of HealthScience and Social Welfare, Faculty of HumanSciences, Waseda University, Tokorozawa,Saitama, Japan.Do experienced meditators and meditation-naïve people have different brain functionalstates during no-task resting 19 channelEEG was recorded (versus average refer-ence) from 8 QiGong meditators with 3 30years experience (mean 11.5+/-8.8) and 9meditation-naïve controls (mean ages 41+/-10 years (3 males), and 37+/-6 years (3



males), respectively) during eyes closed rest(sitting; meditators did not meditate). Allartifact-free 2-second EEG epochs (mean33.9+/-8.5/subject) were recomputed intointracortical 3 dimensional generator distribu-tions using LORETA (2394 voxels) for eachsubject and each of the 8 EEG frequencybands. Results were normalized per fre-quency band and subject (total current densityacross all LORETA voxels scaled to 1). Cur-rent density of all voxels was tested (t tests)for differences between groups for each fre-quency band. An exceedance proportiontest correcting for multiple testing identifiedvoxels at p<0.05. Only differences in deltafrequency band (1.5-6 Hz) were significant(355 voxels): 229 were stronger, 126 weakerin meditators than controls. All but 4 strongervoxels were in anterior areas (BA 9, 10, 11,44, 45, 46, 47), 81 of them left, 144 right;all weaker voxels were in central-posteriorareas (BA 4, 6, 7, 18, 19, 22, 30, 31, 32, 37,39, 40), 107 of them left, 19 right. - In sum:during task-free resting, experienced medi-tators had different brain states compared tonon-meditators. Meditators had stronger deltaEEG activity than non-meditators in frontalcortex (64% right hemisphere voxels), andweaker delta activity in central-posterior cortex(85% left hemisphere voxels). In view of thegeneral assumption that EEG delta activityrepresents inhibition, experienced meditatorshave stronger inhibitory activity than controlsanterior right-preponderant, and less inhibitoryactivity central-posterior predominantly left.These results suggest that meditators re-duce internal information processing whileenhancing input and output processing, aninterpretation that agrees with the meditators’

subjective experience of disengaging fromperceived information. (Partial support by BialGrant No. 44 2006/2007.)

Unconscious processing of words andnon-words. M. Hollenstein (1), T. Koenig (2),W. J. Perrig (1) & M. Kubat (1), (1) Institute ofPsychology, University of Bern, Bern, Switzer-land (2) Dept. of Psychiatric Neurophysiology,University Hospital of Psychiatry, University ofBern, Bern, Switzerland.A well debated issue in cognitive psychologyand neuropsychology is the question to whichlevel an unconsciously perceived linguistic in-put is processed. This has often been inves-tigated with subliminal pattern masking tech-niques that reduce the stimuli’s sensory ac-tivation strength, which has implied problemsto define an objective threshold of awareness.In this event-related-potential-(ERP-) study, anew mirror-masking technique was used thatdoes not reduce sensory activation, but as-sumingly reduces top-down attentional acti-vation. Stimuli were words and non-wordswritten in a quadratic font that were mergedwith their inverted counterpart mirrored at thebaseline of the letters (mirror-masking). Thevisual result of such mirrored letter stringswere unfamiliar, abstract patterns that pre-vented subjective awareness of the linguisticinformation hidden in the seemingly meaning-less patterns. Significant differences in ERPsbetween mirror-masked words and non-wordswere found in a time range from 112-160ms af-ter stimulus-onset above right posterior areas.A Low Resolution Electromagnetic Tomogra-phy (LORETA) model localised neuronal gen-erators of these differences between maskedwords and non-words in right supramarginalgyrus (BA 40) and posterior temporal lobe,



suggesting that at least an automatic differen-tiation occurred at a sublexical, phonologicallevel of processing. Our results strongly indi-cate that subjects unconsciously and automati-cally extracted some linguistic information frommasked words.

Localization by norm. E. Januts (1), A.Klein (2) & T. Sauer (2), (1) Medical Op-tics, Institute of Medical Physics, Friedrich-Alexander University, Erlangen, Germany (2)Numerical Mathematics, Justus-Liebig Univer-sity, Gießen, Germany.Underdetermined inverse problems which usu-ally appear in the context of source localizationare usually regularized by choosing a solutionthat is minimized with respect to a quadraticenergy norm. However, this approach is moreof traditional nature and motivated by the factthat equality constrained quadratic minimiza-tion problems are easy to solve. Unfortunately,the quality of such a location often leaves a lotto be desired, which, nevertheless, is not nec-essarily a measurement artifact but a purelymathematical one stemming from the underly-ing algorithm. Already the simple change ofpassing to the 1-norm which has a smaller,better localized unit ball, often leads to sig-nificantly better results that can be computedquickly and efficiently due to the wide avail-ability of good software for linear programming.We will illustrate this phenomenon with someexamples from optical tomography.

Continuous wavelet transformation andEEG analysis. A. Klein & T. Sauer, Numer-ical Mathematics, Justus-Liebig University,Gießen, Germany.Wavelets transforms have become a standardtool for the decomposition of signals in the

time-frequency-domain. Much effort has beenput into algorithms for discrete wavelet trans-forms (DWT) which allow for very fast imple-mentations. Along with their inverse trans-forms, DWTs serve very well as bandpass-filters and as tools for data compression, butdespite their representing all information con-tained in the signal, the results of DWTs areoften lacking in terms of interpretability, es-pecially when it comes to giving visuals cuesabout the data being analysed. On the otherhand, continuous wavelet transforms (CWT)have been used for a long time as a means ofdata analysis, because they give very many vi-sual cues, but due to their cumbersome inver-sion their use for any kind of real data process-ing beyond sheer analysis has been very lim-ited. In order to overcome these problems, wehave developed a set of tools allowing for anal-ysis as well as synthesis of data at a somewhatslower but still competitive speed as comparedto DWTs.

Congruent music and affective attitudeinfluence the anticipation and perceptionof emotional events - a sLORETA studyon healthy subjects with low depressionscores. M. Kottlow (1), R. Willi (2), T. Baum-gartner (3), G. Tanner (4), T. Koenig (1) & L.Jaencke (2), (1) Department of PsychiatricNeurophysiology, University Hospital of Psy-chiatry, University of Bern, Bern, Switzerland(2) Department of Neuropsychology, Instituteof Psychology, University of Zürich, , Zürch,Switzerland (3) Institute for Empirical Re-search in Economics, University of Zürich,Zürich, Switzerland (4) Institute of HumanMovement Sciences, ETH, Zürich, Switzer-land.Recent studies have shown the influence of



music on the experience of emotion inten-sity. In this study, we replicated the results,additionally analyzing the cued expectationof the emotional stimuli. 17 female subjectswith low depressiveness scores participatedin this EEG study expecting and afterwardsprocessing blocks of sad or happy pictureswith or without congruent music. Peripheralpsychophysiology and psychometrical rat-ings were collected. Alpha frequency bandpower, inversely related to neural activity, wasanalyzed with sLORETA (standardized lowresolution brain electromagnetic tomography).A significant increase in brain activity was visi-ble during the announced expectation of happyemotional pictures with music vs. without mu-sic in anterior cingulate and inferior parietalregions. During the presentation of conditionswith vs. without music, we found significantlyhigher activations in emotion relevant brainregions, supported by increased empathyratings and HR during happy conditions withmusic. Happy compared with sad conditionsresulted in activity within emotion relevant ar-eas, whereas sad conditions activated rathervisual and auditory perception areas. Ourresults show intensity related anticipation andprocessing of emotional events, presumablyinfluenced by the positive affective attitude ofsubjects. Definitive argumentations require theoutstanding comparison with healthy subjectswith high depressiveness scores.

Influence of COMT-genotype on sensorygating in patients with ADHD and healthycontrols. J. Langer, A.-C. Ehlis, T. Daubitz, C.Bähne & A. J. Fallgatter, Department of Psy-chiatry, Psychosomatics and Psychotherapy,University of Würzburg, Würzburg, Germany.Catechol-O-Methyltransferase (COMT) is

an enzyme involved in the degradation ofdopamine (DA) in the synaptic cleft mainlyin the prefrontal cortex. In humans theCOMT gene shows a functional polymorphism(AA158/AA108, Val158Met) resulting in threephenotypes (Met/Met, Val/Met, Val/Val) differ-ing in the availability of DA in the brain (highestMet/Met, lowest Val/Val). Mainly enervated bydopaminergic neurons, the dorso-lateral pre-frontal cortex (DLPFC) is especially affectedby this polymorphism. The DLPFC is cruciallyinvolved in working memory and attention andtherefore often investigated in ADHD. Oneelectrophysiological construct associated withlong term attention and the DLPFC is sensorygating (SG). It is supposed to represent a pro-tective neuronal mechanism filtering irrelevantstimuli to prevent informational ”overload” inthe brain. We assessed genotype and SG viathe P50 double-click paradigm in 23 ADHDpatients (18 to 35 years) and 25 age andgender matched healthy controls. Using a21-channel EEG event-related potentials wererecorded at Cz. ADHD patients overall showeddeficient SG. In controls subjects with Val/Valgenotype showed the worse SG as dopamineis degraded most quickly in this subgroup.No influence of the COMT-Genotype on SGwas found in ADHD-patients which could beconfounded by medication influencing theDA-level.

EEG in Tibetan Buddhist meditators duringrest and meditation, and effects of medi-tation experience. D. Lehmann (1), S. Tei(2), P. L. Faber (1), H. Kumano (2), L. R.R. Gianotti (1) & K. Kochi (1), (1) The KEYInstitute for Brain-Mind Research, UniversityHospital of Psychiatry, Zürich, Switzerland (2)Department of Stress Science and Psycho-



somatic Medicine, The University of Tokyo,Tokyo, Japan.Nineteen-channel EEG from 13 Tibetan Bud-dhist meditators (27-58 years old, mean 39+/-8) with 2 to 25 years experience in meditationwas recorded during task-free resting whilenot meditating, and during meditation aimingat the condition of self-dissolution. The se-quence ’rest - meditation - rest’ was recordedtwice. 286 artifact-free seconds EEG duringrest (1084 seconds during meditation) wereavailable on average/subject. EEG was ana-lyzed into power spectra (1.5 to 44 Hz, versusaverage reference). Spectra were averaged foreach subject, separately for the 2 meditationand the 4 rest conditions. From mean spectraacross the 19 channels we computed 1) cen-troid spectral frequency for the full 1.5-44 Hzband, and 2) integrated power for each of the8 EEG frequency sub-bands (delta to gamma).From the 19 channel spectra, we computed3) the location of the gravity center of bandpower distribution over the 19 scalp record-ing locations. Results were A) compared be-tween meditation and rest, and B) correlatedwith years of meditation experience controlledfor age. Significant results were: A) Com-paring meditation versus rest, full-band spec-tral centroid frequency was higher during med-itation, power in theta and alpha 1 frequencybands was lower during meditation, and loca-tion of gravity center of band power distributionon the scalp was more left for delta, and moreposterior for beta bands during meditation.B) With increasing meditation experience, full-band spectral centroid frequency decreased inmeditation, theta and alpha1 frequency bandpower increased but beta2 and beta3 bandpower decreased in meditation (theta power in-

creased also in rest), and gravity center loca-tion of band power distribution of alpha1 andof all higher frequency bands moved posteriorduring rest and meditation. - These resultsare in agreement with concentration-relatedchanges from rest to meditation, and suggestlessening of these differences with increasingexperience in meditation. (Supported by BialGrant No. 44 2006/2007 ).

The neurophysiological signature of habit-ual prospective memory processes. S. Mat-ter (1), T. Koenig (2) & B. Meier (1), (1) Insti-tute of Psychology, University of Bern, Bern,Switzerland (2) Dept. of Psychiatric Neuro-physiology, University Hospital of Psychiatry,University of Bern, Bern, Switzerland.Tasks that require remembering to realize de-layed intentions at the appropriate occasionare defined as prospective memory tasks. Re-alizing the same prospective memory task re-peatedly transfers an episodic into a habitualprospective memory task. Using event-relatedpotentials (ERPs) the goal of the present studywas to investigate whether characteristic neu-rophysiological episodic to habitual transitioneffects emerge when a prospective memorytask is repeated. ERPs evoked by correctlyanswered prospective memory cues in thefirst part of the experiment were contrastedwith those of the second part of the exper-iment under the assumption that the formerrather represents episodic and the latter ha-bitual prospective memory. Results revealeda significant episodic to habitual transition ef-fect expressed by enhanced frontal negativityand parietal positivity in the ERP 400 - 600ms post-stimulus, a time-window interpretedas critical for retrieval of the intention and post-retrieval processes. Additionally, source local-



ization using LORETA attributed the transitionto a significantly lower activation in the anteriorfrontal regions while posterior parietal and oc-cipital regions showed higher activation in thesecond compared to the first part of the exper-iment. The findings indicate that episodic andhabitual prospective memory tasks require dif-ferent processes. While episodic prospectivememory is supported by controlled retrievalprocesses guided by frontal structures, habit-ual prospective memory was found to be sup-ported by automatic retrieval processes. Thepresent study adverts to the importance of aconceptual distinction between episodic andhabitual prospective memory tasks.

Neural activity before stimulus presenta-tion predicts the success of stimulus re-trieval. T. Padovani ( 1) , T. Koenig (1,2), D.Brandeis (3) & W. J. Perrig (1), (1) Instituteof Psychology, University of Bern, Switzerland(2) University Hospital of Psychiatry, Universityof Bern, Bern, Switzerland (3) Department ofChild and Adolescent Psychiatry, University ofZürich, Zürich, Switzerland.The brain’s information processing is state de-pendent. In this context, previous research hasshown that pre-stimulus ERP activity predictslater recollection (subsequent memory effect).In this experiment, we attempted to systemati-cally manipulate the pre-stimulus state by giv-ing the subjects a) a semantic decision taskand b) an emotional decision task and afterthat testing the incidental encoding of each sin-gle item presented in the study phase with arecognition memory test. We recorded andanalyzed 65 channel ERPs in 21 healthy vol-unteers. We found a significant interactionsof task and later memory performance usinga TANOVA in two different intervals preced-

ing the words onset (-800 to -400 ms, -400to 0 ms). When the cue indicated to executea semantic decision, ERPs of subsequentlyremembered and forgotten words differed inan interval from -800 to -400 ms. When thecue indicated to execute a emotional decisionERPs of subsequently remembered and for-gotten words differed from -400 to 0 ms. Thisindicates that the neural networks necessaryfor a successful incidental encoding depend onthe task that the subject is preparing. Whensearching for those networks, Loreta analy-ses suggested that for the semantic condi-tion, the maximal difference between remem-bered and forgotten words was in the left mid-dle frontal gyrus that is known to be relatedto semantic processes. In the emotional con-dition, the maximal difference was estimatedto be in the left middle temporal gyrus. Thisstructure has been related to memory encod-ing and processing of emotional stimuli. Theseeffects may also reflect latency differencesbetween semantic and emotional preparatoryprocesses important for efficient encoding intoepisodic memory. This study supports theidea that the formation of lasting memories in-volves the role of neural activity that both pre-cedes and follows the presentation of a stimu-lus event.

Early topographical differences elicitedby color and food words. N. Reuther &W. Skrandies, Institute of Physiology, JustusLiebig University, Giessen, GermanyWe investigated the relationship betweencolor, food, odor and taste stimuli. Subjectswere studied with questionnaires and in elec-trophysiological experiments. First, a total of144 word pairs were rated by 660 subjectswho determined whether the second stimuli



(a food or color word) matched the first stim-ulus (taste or odor word). Clearly matchingor non-matching word pairs were used in anEEG experiment. Activity was recorded from30 electrodes in 24 healthy adults while singlewords were randomly presented on a monitor.Evoked potentials were computed for differentstimulus classes (matching and non-matchingcombinations of color or food words with odoror taste words). Six components were identi-fied and topography was compared betweenconditions. Most location differences occurredin the left-right-direction. We observed sig-nificant cognitive effects as early as 100-150ms after stimulus presentation: the loca-tions of the positive centroids differed whenfood/color stimuli were paired with taste/odorstimuli (F(1,23)=4.92, p<0.05). Odor and tastewords yielded different lateralization over theright hemisphere, depending whether taste orodor words were processed by the subjects.Between 400-500 ms color and food words af-fected the locations of negative centroids (sig-nificant interaction between stimulus modalityand target type F(1,23)=4.44, p<0.05). Thesefindings suggest rapid cognitive processing ofsemantic differences between color, food, odorand taste in the visual cortex. The topograph-ical effects indicate that different underlyingneural populations are activated by differentsemantic classes.

Nogo-N2 and Nogo-P3 in psychiatric pa-tients. M. Ruchsow (1), M. Spitzer (2), M.Kiefer (2), G. Grön (2), M. Falkenstein (3) & L.Hermle (1), (1) Dept. of Psychiatry Christophs-bad, Göppingen, Germany (2) Dept. of Psy-chiatry, University of Ulm, Ulm, Germany (3)Leibniz Research Centre for Working Environ-ment and Human Factors (IfADo), Dortmund,

Germany.Nogo-N2 and Nogo-P3 were measured in pa-tients with major depressive disorder (MDD),obsessive-compulsive disorder (OCD), andborderline personality disorder (BPD) and inthree independent sex-, age-, and education-matched control groups. In a fourth study,RT residuals were calculated in a group ofhealthy subjects (n = 26) in order to split theentire group in a high impulsiveness subgroup(HI, n = 13) and a low impulsiveness sub-group (LI, n = 13) according to the method ofPailing et al. (2002). Participants performeda hybrid flanker-Go/Nogo paradigm while a64-channel EEG was recorded. MDD pa-tients, BPD patients, and HI subjects showedreduced Nogo-P3 amplitudes compared totheir respective control groups. OCD patientshad enhanced Nogo-N2 components in re-lation to control subjects. Possibly, Nogo-N2and Nogo-P3 are affected by impulsiveness,compulsivity, and depressiveness in a differentmanner. Further studies are needed to exactlydetermine the underlying neuropsychologicaland neurobiological mechanisms resulting inaltered Nogo-P3 and Nogo-N2 amplitudes inpsychiatric patients.

Ne/ERN in obsessive-compulsive spectrumdisorders. M. Ruchsow (1), M. Spitzer (2), M.Kiefer (2), G. Grön (2), M. Falkenstein (3) & L.Hermle (1), (1) Dep. of Psychiatry Christophs-bad, Göppingen, Germany (2) Dep. of Psy-chiatry, University of Ulm, Ulm, Germany (3)Leibniz Research Centre for Working Environ-ment and Human Factors (IfADo), Dortmund,Germany.The error negativity (Ne) or error-related neg-ativity (ERN) was investigated in patients withobsessive-compulsive disorder (OCD) and



borderline personality disorder (BPD) and intwo independent sex-, age-, and education-matched control groups. In a third study, RTresiduals were calculated in a group of healthysubjects (n = 32) in order to split the entiregroup in a high impulsiveness subgroup (HI, n= 16) and a low impulsiveness subgroup (LI,n = 16) according to the method of Pailing etal. (2002). Participants performed a hybridflanker-Go/Nogo paradigm while a 64-channelEEG was recorded. Both, BPD patients andHI subjects showed reduced Ne/ERN ampli-tudes compared to their respective controlgroups, whereas OCD patients and LI sub-jects had enhanced Ne/ERN amplitudes. TheNe/ERN seems to reflect a continuum of com-pulsiveness (OCD patients and LI subjects)and impulsiveness (BPD patients and HI sub-jects) in a spectrum of obsessive-compulsivedisorders. It seems a reasonable target forfurther studies whether Ne/ERN amplitudesin psychiatric patients change with successfulpharmacological and/or psychotherapeutictreatment.

Effects of working memory load andmethylphenidate in healthy adults during avisual working memory task. P. Studer (1),S. Wangler (1), M. Diruf (1), O. Kratz (1), G.H. Moll (1) & H. Heinrich (1,2), (1) Child andAdolescent Psychiatry, University of Erlangen,Erlangen, Germany, (2) Heckscher-Klinikum,Munich, Germany.In the present study, we investigated howworking memory load affects the neuronalprocessing during the encoding, retention andretrieval phases of a visual working mem-ory task and how processing is modulatedby methylphenidate. Eleven adults partici-pated in a placebo-controlled, double-blind,

crossover study. They performed the taskin which they memorized the order of four,five or six pictures under methylphenidate(20 mg) and under placebo. Brain electricalactivity was recorded, eye movement artefactswere corrected using independent componentanalysis and repeated-measurement ANOVAswere calculated. The number of correct re-sponses decreased with incremental workingmemory load. In the encoding phase the P3amplitudes became smaller with increasingnumbers of pictures shown within one trial.Over the midline electrodes P3 amplitudesincreased with increasing memory load. Inthe retention period a slow negative waveoccurred with a fronto-central distribution. Theretrieval phase was characterized by a slownegative wave with a posterior topography.Medication neither influenced performancenor the different processing stages in oursmall sample. Increasing working memoryload was associated with P3 effects duringencoding. Methylphenidate does not seem tomodulate processing during a visual workingmemory task in healthy adults significantly.

sLORETA, eLORETA, and SWARM in thepresence of noise and multiple sources. M.Wagner, M. Fuchs & J. Kastner, CompumedicsNeuroscan, Hamburg, Germany.Standardized low resolution brain electro-magnetic tomography (sLORETA) retrievesthe sources of EEG and MEG data with lowlocalization error. Its outcome, however, isnot a distribution of currents modeling brainactivity but a statistical map thereof. As aconsequence, like EEG/MEG beamformersor fMRI images, it is not a solution to theunderlying inverse problem. Two new meth-ods promise to solve the inverse problem



by computing neuronal current flow instead ofstatistical maps, also with low localization errorfor point sources: sLORETA-weighted accu-rate minimum-norm (SWARM) and exact lowresolution brain electromagnetic tomography(eLORETA). In this contribution, we evaluatethe performance of these methods under thepresence of noise and with multiple, simulta-neously active sources. The same simulationsetup and data as in [1] and [2] are used.[1] M. Wagner, M. Fuchs, J. Kastner. 2004. Evalua-tion of sLORETA in the presence of noise and mul-tiple sources, Brain Topography 16:277-280.

[2] M. Wagner, M. Fuchs, J. Kastner. 2004. Eval-uation of sLORETA for MEG data in the presenceof noise and multiple sources, in: Biomag 2004,Eds. Halgren E, Ahlfors S, Hämäläinen M, CohenD. Boston, Biomag 2004 Ltd., p. 611.

Spatio-temporal organisation of quadraticphase- couplings in tracé alternant EEGpattern in full-term newborns. K. Schwab, P.Putsche, M. Eiselt & H. Witte, Institute of Med-ical Statistics, Computer Sciences and Docu-mentation, Friedrich Schiller University Jena,Jena, Germany.

The aim of this work was the investigation ofthe spatio-temporal organisation of quadraticphase-couplings (QPC) in the electroen-cephalographic ’trace alternant’ pattern duringquiet sleep. The investigation was carried outin a group of 6 clinically and neurologicallynormal, full-term newborns (8-channel-EEG:Fp1, Fp2, C3, C4, T3, T4, O1, O2). The quietsleep EEG was segmented by a trained physi-cian into 4 sec periods of burst and interburstpatterns. The Gabor expansion, a fast Fouriertransformation based method, was adaptedto examine the time-course of biamplitude,bicoherence and phase-bicoherence of both

burst and inter-burst patterns in the region ofinterest [1-1.5 Hz, 3.5-4.5 Hz]. The burst andthe interburst patterns are characterised bytemporally and topographically different QPCprofiles. All differences are dominant at Fp1followed by Fp2. There is a significant differ-ence (combined multiple and global tests) inthe QPC characteristics between both patternswithin the time period from 0.75 to 1.5 s afterthe pattern onset at Fp1. The maximal QPCin burst patterns can be observed during thistime period. In contrast to this finding, maximalQPC in interburst patterns are reached imme-diately after the onset and at 3 s. Additionally,a time-variant, parametric bispectral approachresulting in a so-called mean biamplitude(mBA) in the region of interest was adaptedto investigate the time-course of QPC withinthe continuous EEG recordings (total length ofrecordings 320 to 830 sec). The rhythmicityof the time course of the mBA is dominatedby the sequential alteration of burst and inter-burst patterns. The mBA course rises duringthe occurrence of a burst pattern. Within allnewborns, a periodicity of QPC of about 8-10s can be calculated. A comparison of the QPCevents and of the burst-interburst-pattern seg-mented by the trained physician shows thatthe periodicity of the QPC alteration persists inEEG epochs in which no burst patterns can bedetected (visually, according to the amplitudecriteria). In conclusion, it can be assumedthat the QPC rhythm of the TA is generatedby a pattern-spanning time-variant phase-locking process, and there are indications fora correspondence between the QPC rhythmand vegetative rhythms. This study showedthat advanced, time-variant analysis methodsquantifying QPC rhythms are able to add new



scientific information to the understanding ofnature, characteristics and significance of TAin the neonatal EEG.

Decomposition of multivariate time seriesby state space modelling. A. Galka (1,2,3),K. F. K. Wong (3), T. Ozaki (3), H. Muhle (4),U. Stephani (4) & M. Siniatchkin (4), (1) De-partment of Neurology, University of Kiel, Kiel,Germany (2) Institute of Experimental and Ap-plied Physics, University of Kiel, Kiel, Germany(3) Institute of Statistical Mathematics (ISM),Tokyo, Japan (4) Department of Neuropedi-atrics, University of Kiel, Kiel, Germany.In many situations, decomposition of multi-variate time series data into a set of sourcecomponents represents a useful and importantstep for preprocessing and analysis of time-resolved data in neuroscience. As a prior con-straint, the components may be assumed tobe mutually independent, which leads to In-dependent Component Analysis (ICA). In thiscontribution we propose to employ linear statespace modelling for the purpose of time se-ries decomposition, as an alternative to cur-rently available algorithms for ICA. State spacemodelling, a generalisation of classical ARMAmodelling, is well suited for exploiting the dy-namical information encoded in the temporalordering of time series data, while this infor-mation remains inaccessible to most ICA algo-rithms. Within the state space framework, theassumptions of mutual independence of com-ponents, of stationarity and of normality (gaus-sianity) can be relaxed. Using a Kalman Fil-ter, linear state space models can be employedeven for detecting strongly nonlinear dynam-ical processes. As a result, much more de-tailed decompositions become possible, andboth components with sharp power spectrum,

such as alpha components, sinusoidal artifactsor sleep spindles, and with broad power spec-trum, such as FMRI scanner artifacts or epilep-tic spiking components, can be separated. Re-sults from simulations and from practical EEGprocessing tasks are presented.

Randomization techniques specific for theanalysis of ERP data. T. Koenig & T. Dierks,Department of Psychiatric Neurophysiology,University Hospital of Psychiatry, University ofBern, Bern, Switzerland.ERP data is redundant in many ways. Chan-nels are intrinsically correlated across thescalp, similar scalp distributions are observedacross prolonged time intervals, and the mea-surements are typically repeated within andacross subjects. This redundancy is prob-lematic for statistics, because it obscures thereal degrees of freedom of the data, whichin return increases the possibility of falseconclusions due to repeated testing or overlyconservative corrections for multiple testing.These problems can be overcome by ran-domization statistics. Randomization statisticsallow the user to invent a suitable measure ofeffect size and then test the effect size in themeasured data against the null-hypothesis.The problem of redundancy of ERP data canthus be overcome by choosing specific effectsize measures that are global across the di-mension where the redundancy is expected.We present a series of techniques that applyglobal measures of effect size to ERP data.One series of tests investigates whether thereis evidence that at one moment in time, oneor several ERP topographies are systemati-cally related to the experimental design or tosome continuous predictor. Another seriesof tests investigates whether the distribution



of statistical effect sizes across the analysisperiod is likely to have occurred by chance,thus dealing with redundancy across time.Finally, a method is proposed that allows todetermine periods of consistent topographyacross repetitions (trials or subjects).


M. Doppelmayr et al. — Effects of SMR Feedback on the EEG Amplitude

Abstract

M. Doppelmayr, E. Weber, K. Hoedlmoser & W. Klimesch (Salzburg) — Effects of SMRFeedback on the EEG Amplitude Neurofeedback (NF) is a method to learn how to voluntarilyproduce and modulate specific brain rhythms, as assessed by the electroencephalogram (EEG).It is used both, in the treatment of a variety of disorders, such as epilepsy or attention deficit(hyperactivity) disorder (AD(H)D), as well as in healthy subjects to increase performance in aspecific type of task. However, especially for healthy subjects, there is a lack of empirical datafocusing on the effects of NF on the EEG. While there are some reports indicating that subjectsare able to learn to increase the amplitude within the sensorimotor rhythm frequency range(SMR; 12 - 15 Hz) by neurofeedback training (NFT), there are almost no studies describing theconcomitant effects on a resting EEG preceding and/or following NFT. In our study 20 healthysubjects participated in 25 NF sessions. The experimental group (SMR n=12) was instructed toincrease the SMR amplitude, while the control group (n=8) received a pseudo-NF (PNF). EEGwas recorded at C3 and C4 during the training as well as in resting conditions preceding andfollowing each NF session.With respect to amplitude changes, we found that, during training, SMR amplitude significantlyincreased in the experimental group, whereas it remained unchanged over time in the PNF-training. Our results indicate that a significant increase in SMR amplitude is present only duringthe training. However, there was a clear tendency that in a resting condition with eyes openSMR amplitude was also increased. This leads to the assumption that possibly after morethan 25 sessions, SMR NFT might exert effects on resting EEG parameters. As there wasno difference between experimental and control group concerning the motivation (measured byvisual analogue scales), it can be ruled out that the training-related increase is due to a placeboeffect.

Effects of SMR Feedback onthe EEG Amplitude

M. Doppelmayr, E. Weber, K. Hoedlmoser &W. Klimesch, Department of Physiological

Psychology, University of Salzburg,Hellbrunnerstr. 34, 5020 Salzburg, Austria

[email protected]

Introduction

The roots of neurofeedback (NF) reach backto the 1970s. In 1969 Joe Kamiya was thefirst to propose that the alpha rhythm of thehuman electroencephalogram (EEG) can beself-regulated. This was followed by a studyfrom Sterman, Howe, and Macdonald (1970),who reported that cats are able to learn to in-crease specific frequencies of their electroen-cephalogram (EEG). In a later experiment by



Sterman and Friar (1972), it was found thatthese cats showed a decreased sensitivity toepileptic seizures as induced by monomethyl-hydrazine. Although NF has not been the fo-cus of neuroscience for a long time, scientificinterest in different aspects of NF has been re-established in the last decade.

In general, it is assumed that NF is a typeof operant conditioning procedure used tomodulate EEG parameters such as amplitude,power and coherence (Sterman & Egner,2006). Subjects are connected to the EEGdevice and receive visual and/or acousticfeedback of the EEG parameter of interest.This feedback may be a circle, a bar, ormore sophisticated displays such as videogames. An increase or decrease in the rele-vant EEG parameter leads to changes in thisfeedback. Participants are supposed to altertheir individual EEG guided by the presentedfeedback (e.g. increase or decrease a bar),which, in turn, should lead to changes in brainrhythmicity by operant conditioning.

Meanwhile, there is a series of well pub-lished manuscripts that analyzed behaviouralchanges after a neurofeedback-intervention.Effectiveness of neurofeedback training (NFT)has been demonstrated in the treatment ofepilepsy (Kotchoubey et al., 1999, 2001;Strehl, Kotchoubey, Trevorrow, & Birbaumer,2005; Strehl, Trevorrow, et al., 2006; Sterman& Egner, 2006) and attention deficit (hyperac-tivity) disorder (AD(H)D) (Fuchs, Birbaumer,Lutzenberger, Gruzelier, & Kaiser, 2003;Monastra et al., 2006; Lubar, Swartwood,Swartwood, & O’Donnell, 1995; Strehl, Leins,et al., 2006). In addition, it is documented thatsigma NFT can increase spindle power dur-ing sleep (Berner, Schabus, Wienerroither, &

Klimesch, 2006) and that NFT exerts positiveeffects on sleep and memory (Hoedlmoseret al., 2008). Other studies reported themore or less successful treatment of mi-graine (Kropp, Siniatchkin, & Gerber, 2002),tinnitus (Dohrmann, Weisz, Schlee, Hart-mann, & Elbert, 2007), emotional dysfunctions(Hammond, 2005; Raymond, Sajid, Parkinson,& Gruzelier, 2005) and stroke rehabilitation(Doppelmayr, Nosko, Pecherstorfer, & Fink,2007). Furthermore, several NFT investiga-tions have been conducted demonstratingthe modification of cognitive (Hanslmayr,Sauseng, Doppelmayr, Schabus, & Klimesch,2005; Egner & Gruzelier, 2001; Gruzelier, Eg-ner, & Vernon, 2006) and creative (Gruzelier& Egner, 2005; Raymond et al., 2005) per-formance. Last but not least, modification ofbrain rhythms by means of NF is used for braincomputer interface technology (Birbaumer,2006; Birbaumer et al., 2006).

Studies investigating the treatment of disor-ders focus on participants that are cognitivelyor emotionally impaired, while research focus-ing on the enhancement of cognitive or cre-ative processes is based on healthy subjects.Depending on the type of treatment, differentfrequency bands are used for training: sen-sorimotor rhythm SMR (approximately 12 - 15Hz) (Sterman & Egner, 2006) or slow corticalpotentials (SCPs) for epilepsy (Kotchoubey etal., 2001) and AD(H)D or theta/beta ratio forAD(H)D (Leins et al., 2007), to give some ex-amples.

Although many investigations have focussedon the behavioural efficiency of NFT, there isless literature reporting the respective changesin the EEG. However, there are some stud-ies showing changes in the powerspectra af-



ter NFT in patients suffering from AD(H)Dor epilepsy (Fernández et al., 2007; Grin-Yatsenko, Kropotov, Ponomarev, Chutko, &Yakovenko, 2001; Kropotov, Ponomarev, &Grin-Yatsenko, 2001; Pineda et al., 2008). Incontrast to clinical populations, there is little, ifany literature concerning changes in the EEGafter NFT in healthy subjects. Therefore, it isan interesting question whether NFT changesEEG-characteristics only during the training orwhether the effects are longer lasting. Underthe assumption that the EEG reflects, at leastin part and indirectly, cognitive processes itis of outstanding importance to know whetherNFT modifies these processes only for short(Hanslmayr et al., 2005) time periods or in amore general and longer lasting way (Pinedaet al., 2008).

Vernon et al. (2003) demonstrated that sub-jects were able to increase their SMR (or atleast the SMR/theta and SMR/beta ratio) af-ter 8 sessions of NFT. A second group hadto learn to enhance theta activity and failed tochange the EEG in any aspect. Unfortunatelyneither SMR amplitude or power values (onlyratios) nor data relating to the accompanyingresting conditions were reported.

Further research concerning changes in EEGafter NFT was published by Egner, Zech, andGruzelier (2004) who used healthy subjects intwo experiments. In Experiment 1 the partici-pants received two different beta trainings fol-lowed by an alpha/theta training. The results,however, yielded no clear evidence that thosefrequency bands that had been trained demon-strated altered spectral properties.

In Experiment 2, subjects were assigned to ei-ther alpha/theta, low beta, or beta training. Forabsolute power analyses, none of the hypothe-

ses concerning the modification of the respec-tive frequency bands could be confirmed. Inaddition, relative power analyses only yieldedan effect on the alpha/theta training in the lowbeta band at prefrontal areas. Therfore, as theauthors conclude, the straightforward assump-tion that training in a specific frequency bandwould alter the EEG respectively could not beproven.

Cannon et al. (2007) used a new approach oflow resolution brain electromagnetic tomogra-phy (LORETA) - NF. Eight subjects had to in-crease 14 – 18 Hz beta power in the area ofthe anterior cingulate cortex (ACC) [for detailsof LORETA-NF see Congedo, Lubar, & Joffe,2004]. The results indicated that the subjectswere able to increase beta power in the ACCas well as adjacent areas, but no increase wasdetected in more distant locations.

As outlined above, little data exist relating tothe impact of NFT on amplitude and power inhealthy subjects; there is only one report con-cerning the effect of NFT on the resting EEGof healthy participants. Cho et al. (2008) ana-lyzed effects of alpha (8 - 12 Hz) NFT in 9 sub-jects during 16 sessions of NFT including pre-ceding resting conditions. They found a signifi-cant increase in alpha amplitude after 9 ses-sions, in both the training and in the restingcondition with eyes open.

Taken together, the effect of NFT on the EEGof healthy subjects has not been investigatedin depth. Therefore, from our point of view, itis necessary and highly important for the fu-ture application of NFT to address these ba-sic issues, especially if NFT is used in healthysubjects. Due to the fact that SMR is themost commonly used frequency of interest inNF studies, we decided to concentrate on this



rhythm. The main focus of this study, therefore,was to investigate a) whether healthy subjectsincrease their SMR-amplitude over the courseof 25 SMR-NFT sessions, and b) if thereby anychanges in SMR-amplitude are detectable dur-ing the resting periods preceding and followingeach NFT.

Methods

Subjects and Design

23 healthy subjects were carefully instructed,signed written informed consent and received250.- Euro for their participation. The experi-ment was conducted in line with the declara-tion of Helsinki. Subjects had to perform 25sessions of NFT (one each day for five weeks,excluding weekends). Each of these sessionsconsisted of 2 min rest with eyes open, 2 minrest with eyes closed, 8 3-min blocks of NFT,followed by 2 resting conditions such as at thebeginning.

Due to missing or bad data, 3 subjects had tobe excluded, so resulting in a sample of 20 par-ticipants (11 male), with a mean age of 29.3years (SD 10.5).

Subjects were randomly assigned to either anexperimental group (n =12, SMR) or to a con-trol group (n = 8, pseudoneurofeedback PNF).Participants in the experimental group weresupposed to increase SMR (12 – 15 Hz, re-ferred to as low beta by several authors) am-plitude during the NFT. The PNF group, on theother hand, was instructed to increase ampli-tude in different frequency bands in a pseu-dorandomized way (delta 2 – 3 Hz, theta 6– 7 Hz, beta 1 18 – 20, gamma 1 30 – 35Hz and gamma 2 40 – 45 Hz). In the PNFgroup, the respective frequency bands were

changed after each 3 min. block so no learningeffects within a specific frequency range wereexpected.Data were recorded by BioGraph Infinity-ProComp (Thought Technology R©) using asample rate of 256 Hz. Electrodes were lo-cated at C3 and C4 with reference on the leftearlobe. Ground electrode was placed on theright earlobe.

Neurofeedback training

Subjects were seated in front of a computermonitor displaying a green background, awhite counter on the top, the main feedbackin the centre, and bars representing the inhibitbands on the right hand side (see Figure 1).The main feedback in the centre consistedof an orange circle, indicating the rewardthreshold, including a white filled circle repre-senting the actual amplitude of the respectivefrequency band. Increases in the amplitudelead to an increase of the filled circle until thereward threshold was reached. If the rewardthreshold was exceeded for more than 250 msone point was added on the counter. After areward, no further reward was given for thenext 3 seconds. On the right hand side ofthe monitor three bars were depicted repre-senting the inhibit bands. Inhibit bands areused to prevent subjects from manipulatingEEG amplitude by eye blinks or modulatingamplitude estimates by voluntary muscle con-traction (e.g. m. masseter or temporalis).The frequency bands for these inhibit bandswere adjusted from 3 – 5 Hz (eye blinks), 22– 30 Hz and 45 – 60 Hz (muscle artefacts).If the amplitude in one of these bands wasincreased no positive feedback was provided.In general the ongoing EEG was band-pass



filtered to continuously extract the averageamplitude within the respective frequencybands. The average amplitude of C3 and C4was transformed into visual feedback (circleand bars).The reward threshold was set individually foreach of the training blocks. The idea was toprovide sufficient rewards within a 3 minuteperiod to keep the subjects motivated. Wetherefore decided to change the reward thresh-old if less than 20 or more than 30 rewardswere given within a 3 minute block. For thefirst training block each day the last thresholdvalue of the previous day was used. The re-ward threshold for the PNF was set similarly insuch a way that the subject received approxi-mately the same amount of positive feedback,independently of which frequency band theytrained.

Motivation

To evaluate whether motivational aspects aresimilar, we used a visual analogue rating scale,5 cm in length, indicating low motivation onthe left and high motivation on the right handside. Motivation was evaluated each day af-ter NFT by subjects marking that point on thescale which represented how motivated theywere. The distance was measured and the re-spective value [cm] was used for analysis.

Data Analyses

EEG data were exported for analysis to Scan4.3.3. (Neuroscan Inc.) software and visuallyinspected for artefacts. For each 3 min trainingblock as well as for the resting conditions pow-erspectra were calculated to extract the meanpower of Alpha (8 - 11 Hz), SMR (12 - 15 Hz),

Beta (16 - 25 Hz) and the Total frequency band(2 - 40 Hz). Root values of the powerspectra,representing amplitudes were calculated andthe respective values averaged across C3 andC4. Relative SMR amplitude was calculated bydividing SMR amplitude by the Total amplitude.

The original dataset comprised 8 NFT blocksplus 4 resting sessions for 25 days, i.e. 300datasets per person. To reduce the amountof data, we decided to compare the EEG val-ues from the beginning, t1 (average of day 1, 2and 3), with those from the middle of the train-ing, t2 (average of day 12, 13, and 14), andthose from the end of the training t3 (averageof day 23, 24, and 25). The values for threedays were averaged to reduce the substantialintraindividual variance.

To evaluate the effects of NFT on Alpha, SMR,Beta and relative SMR amplitude, 2-factorialANOVAs with the factors TIME (t1, t2, t3) andGROUP (SMR, PNF) were calculated. In addi-tion, an ANOVA comprising the factors TIME-ALL (day 1 to day 25) and GROUP (SMR,PNF) was conducted. To compare SMR ampli-tude changes in the resting condition with eyesopen before/after the training, a 3-factorialANOVA including the factors PRE/POST (be-fore the training, after the training), TIME (t1,t2, t3) and GROUP (SMR, PNF) was calcu-lated. The same process was performed forthe eyes closed condition.

To compare motivation between subjects andacross the training sessions, the motivationrating values provided after the first, secondand third training were averaged for t1; afterthe 12th, 13th, and 14th session for t2 and af-ter the 23rd, 24th and 25th for t3. Again, thesevalues were contrasted using a 2-way ANOVAwith the factors TIME (t1, t2, t3) and GROUP



Figure 1: The counter on the top increased every time the reward threshold, represented by an orange

circle in the middle of the green screen was exceeded for more than 250 ms. The amplitude of the relevant

EEG frequency was represented by the white circle within the threshold circle. This white circle increased

or decreased in response to amplitude changes of the selected frequency. On the right hand side, three

bars represented amplitude of the three inhibit frequency bands (3 - 5 Hz, 22 – 30 Hz, and 45 – 60 Hz).

(SMR, PNF).

Results

Kolmogorov-Smirnov tests were applied to testfor the normality of the distribution of the data,which was given in all cases. In addition,Greenhouse-Geisser corrected p-values arepresented if the sphericity assumption was vio-lated. The ANOVA for the SMR band address-ing the differences between the groups andthe changes during the course of the trainingsessions revealed a significant main effect forTIME F(2,36) = 8.776, p = 0.003 and a signif-icant interaction between TIME and GROUPF(2,36) = 3.703, p = 0.050. The main factorGROUP did not reach significance. These re-sults indicate that there is a general increaseof SMR amplitude during the training, steadilyincreasing from t1 to t3. As depicted in Fig-

ure 2, only the experimental group increasedSMR amplitude, whereas the control group re-mained at a relatively constant amplitude.Pairwise comparisons revealed a significant in-crease in SMR amplitude for the experimentalgroup from t1 to t2 (T=-3.787, df 11, p = 0.003)and from t1 to t3 (T=-3.452, df 11, p = 0.005)but no significant difference between t2 and t3.For the PNF group, no significant differencescould be observed.Using relative SMR amplitude as dependentvariable revealed no significant results, merelya marginal not significant interaction of TIMEand GROUP F(2,36) = 2.981, p = 0.076 withincreased relative SMR amplitude with TIMEfor the SMR group only.For the Alpha frequency range, the ANOVAresulted in a significant main effect for TIMEF(2,36) = 8.44, p = 0.003, indicating a steadyincrease in amplitude from t1, to t2 and t3, but



0,5

1

1,5

t1 t2 t3

SMR

-Am

plitu

de µ

V

Timecourse of SMR amplitude increase

SMRPNF

**

**

Figure 2: Significant interaction between TIME and GROUP, indicating that SMR amplitude selectively

increased in the SMR-training group, but not in the pseudo-neurofeedback group (PNF). The increase in

SMR amplitude in the SMR group is more pronounced for the time interval t1 to t2 as compared to t2 to

t3. The PNF group did not show any significant changes in the SMR amplitude. Error bars indicate the

standard error of means and asterisks indicate significant increases of SMR amplitude, with respect to t1

at 1% (**) level of significance.

no main effect for GROUP or interaction hasbeen found. Similarly, the ANOVA for the Betaband revealed only a significant main effect forTIME F(2,36) = 18.69, p <= 0.001, again indi-cating a steady amplitude increase from t1, tot2 and t3.

The ANOVA addressing specific changes inthe amplitude values for each day between thegroups revealed a significant main effect forTIME F(24,423) = 2.57, p = 0.017, but neithera significant effect for GROUP nor a significantinteraction. Figure 3 depicts all the amplitudevalues for both groups throughout the wholetraining process.

The 2 ANOVAs focussing on the differencesconcerning resting amplitude PRE/POST XTIME X GROUP yielded neither a significantmain effect for any of the factors nor a sig-nificant interaction. However, for eyes open a

strong tendency for TIME and GROUP F(2,36)= 3.219, p = 0.060 could be observed. The re-sult indicates, that in general, the SMR ampli-tude in the eyes open condition increases onlyfor the training group, but not for the PNF. Foreyes closed, no effects were observed.The ANOVA with the dependent measure MO-TIVATION indicated no differences betweengroups or sessions. The average value of bothgroups for motivation was 3.73 cm on the vi-sual analogue rating scale indicating a rela-tively high motivation throughout NFT.

Discussion

As outlined in the introduction, changes in theEEG spectrum in response to a specific NFTare assumed to lead to behavioural changes.In NFT, it is assumed that the relationship be-



1

1,1

1,2

1,3

1,4

1,5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

SMR

-Am

plitu

de µ

V

Timecourse of Training for 25 Days

SMR

PNF

Session

Figure 3: SMR amplitude for all 25 training sessions is depicted for both groups. Although there is no

significant TIME X GROUP interaction, this figure is shown for better presentation of the single day

results.

tween EEG pattern and cognitive processes iscausal rather than correlative. Using rTMS tomanipulate alpha power preceding a cognitivetask Klimesch, Sauseng, and Gerloff (2003)were able to demonstrate this causal relationexperimentally. The highly interesting findingwas that enhancement of alpha amplitude bythe means of rTMS at Fz or P6 lead to in-creased performance in the requested cuberotation task. These findings clearly demon-strate a functional relation between EEG am-plitude and behaviour (at least for the alphaband and cognitive performance).

Given this functional relationship, effectiveNFT should lead to an alteration in EEG-activity and thereby to behavioural modifi-cations. Similar to the results reported byHoedlmoser et al. (2008) our results clearlyindicate that healthy subjects are able to in-crease SMR amplitude during training in a

rather linear way. The increase of SMR ampli-tude was more pronounced from t1 to t2, ascompared to the time interval t2 to t3. Thisindicates that the strongest effects of trainingcan be achieved in approximately 12 to 15sessions, although further training continu-ously improves the performance. Cho et al.(2008) recently investigated the effects of NFTin the alpha band and reported, well in linewith our findings, a significant increase alreadyafter the 9th and 11th session.

Analyzing the relative SMR amplitude yieldedsimilar, but marginal non significant results, in-dicating a clear tendency in the same direction,namely an increase in relative SMR amplitudewith training, but only for the SMR-NFT group.

The increase of SMR and relative SMR am-plitude in our study is not due to any type ofplacebo effect or other variables, because thePNF group did not show this increase. In ad-



dition this result corresponds to reports fromAD(H)D studies documenting a modification ofthe trained EEG parameters or the ability tolearn to modulate SCPs (Kotchoubey et al.,1999; Strehl et al., 2005). Similarly Cannonet al. (2007) found a learning effect in the betaband when subjects had to increase the powerin the ACC.

The results for the Alpha and Beta band in-dicated no group related effects. In both fre-quency bands, amplitude increased with thenumber of the trainings. For the Alpha band,this may be interpreted in a way that the sub-jects were more relaxed, and therefore exhib-ited higher amplitudes.

However, the question is whether thesechanges in brain rhythms are stable or disap-pear within a very short time after the actualtraining. The analysis of the resting EEG indi-cated that there is no clear, highly significantchange in SMR amplitude over time. How-ever for the resting condition with eyes open,a strong tendency emerged demonstratingan amplitude increase for the SMR grouponly. This can be interpreted in two ways.Either there is a long lasting effect of the SMRtraining resulting in generally increased SMRamplitude or this increase in SMR amplitudeis situation-specific and occurs only in prepa-ration for (during) and after the training. Thefact that this effect was only observed for eyesopen, but not for eyes closed, leads us to theassumption that it might be situation-specific.The fact that these findings only reached the6% level of significance, causes the questionwhether using more than 25 sessions mightshow more clear cut results.

The reported tendency of increased restingSMR amplitudes in the SMR group is well

in line with reports from clinical samples,where NFT related power changes have beendemonstrated to remain for a longer time(Kropotov et al., 2007; Vernon, Frick, & Gruze-lier, 2004). Similar to our results, Cho et al.(2008) reported a significant increase in alphaamplitude during a resting condition with eyesopen (recorded before each training session).

The main goal of our study was to investigatewhether SMR-NFT in healthy subjects is ca-pable of increasing SMR amplitude during thetraining as well as in resting situations preced-ing or following the training. We used a con-trol group design to account for placebo effectsand compared subjects instructed to increaseSMR amplitude with participants who had toincrease different frequency bands (PNF). Theresults clearly indicated that, during the actualtraining, SMR amplitude steadily increasedonly in the SMR group, but not in the con-trol group. Therefore, we can conclude thathealthy subjects are very capable of learningto increase SMR amplitude in the course of 25training sessions. For changes in the concomi-tant resting EEG, the results were less clear.While there was no increase in SMR ampli-tude for the resting condition with eyes closed,there was a strong tendency for group-specificincreases in the eyes open condition.

Acknowledgements

This study was supported by OENB grant Ju-biläumsfonds 12276. The authors would like tothank Ch. Brem, K. Brandecker, A. Braunes-berger, R. Hauser, and J. Laferton for theirsupport in recording the data.



References

Berner, I., Schabus, M., Wienerroither, T.,& Klimesch, W. (2006). The sig-nificance of sigma neurofeedback train-ing on sleep spindles and aspects ofdeclarative memory. Appl PsychophysiolBiofeedback , 31, 97–114.

Birbaumer, N. (2006). Breaking the silence:brain-computer interfaces (BCI) for com-munication and motor control. Psy-chophysiology , 43, 517–532.

Birbaumer, N., Weber, C., Neuper, C.,Buch, E., Haapen, K., & Cohen, L.(2006). Physiological regulation of think-ing: brain-computer interface (BCI) re-search. Prog Brain Res, 159, 369–391.

Cannon, R., Lubar, J., Congedo, M., Thornton,K., Towler, K., & Hutchens, T. (2007).The effects of neurofeedback training inthe cognitive division of the anterior cin-gulate gyrus. Int J Neurosci , 117 , 337–357.

Cho, M., Jang, H., Jeong, S., Jang, I., Choi, B.,& Lee, M. (2008). Alpha neurofeedbackimproves the maintaining ability of alphaactivity. Neuroreport , 19, 315—317.

Congedo, M., Lubar, J., & Joffe, D. (2004).Low-resolution electromagnetic tomog-raphy neurofeedback. IEEE Transactionson Neural Systems and RehabilitationEngineering, 12, 387—397.

Dohrmann, K., Weisz, N., Schlee, W., Hart-mann, T., & Elbert, T. (2007). Neuro-feedback for treating tinnitus. Prog BrainRes, 166, 473–485.

Doppelmayr, M., Nosko, H., Pecherstorfer, T.,& Fink, A. (2007). An attempt to increasecognitive performance after stroke with

neurofeedback. Biofeedback , 35, 126–130.

Egner, T., & Gruzelier, J. (2001). Learnedself-regulation of EEG frequency compo-nents affects attention and event-relatedbrain potentials in humans. Neuroreport ,12, 4155–4159.

Egner, T., Zech, T. F., & Gruzelier, J. H. (2004).The effects of neurofeedback training onthe spectral topography of the electroen-cephalogram. Clin Neurophysiol , 115,2452–2460.

Fernández, T., Harmony, T., Fernández-Bouzas, A., Díaz-Comas, L., Prado-Alcalá, R., Valdés-Sosa, P., et al. (2007).Changes in EEG current sources in-duced by neurofeedback in learning dis-abled children: an exploratory study.Applied Psychophysiology and Biofeed-back , 32, 169–183.

Fuchs, T., Birbaumer, N., Lutzenberger, W.,Gruzelier, J. H., & Kaiser, J. (2003).Neurofeedback treatment for attention-deficit/hyperactivity disorder in children:a comparison with methylphenidate.Appl Psychophysiol Biofeedback , 28, 1–12.

Grin-Yatsenko, V., Kropotov, J., Ponomarev, V.,Chutko, L., & Yakovenko, E. (2001). Ef-fect of biofeedback training on sensori-motor and beta-sub-1 EEG rhythms onattention parameters. Human Physiol-ogy , 27 , 259–268.

Gruzelier, J., & Egner, T. (2005). Critical val-idation studies of neurofeedback. ChildAdolesc Psychiatr Clin N Am, 14, 83–104.

Gruzelier, J., Egner, T., & Vernon, D. (2006).Validating the efficacy of neurofeedback



for optimising performance. Prog BrainRes, 159, 421–431.

Hammond, D. (2005). Neurofeedback treat-ment of depression and anxiety. J. AdultDev., 12, 131–137.

Hanslmayr, S., Sauseng, P., Doppelmayr, M.,Schabus, M., & Klimesch, W. (2005). In-creasing individual upper alpha power byneurofeedback improves cognitive per-formance in human subjects. Appl Psy-chophysiol Biofeedback , 30, 1–10.

Hoedlmoser, K., Pecherstorfer, T., Gruber, G.,Anderer, P., Doppelmayr, M., Klimesch,W., et al. (2008). Instrumental condition-ing of human sensorimotor rhythm (12-15 Hz) and its impact on sleep as well asdeclarative learning. Sleep, 31, 1401–1408.

Klimesch, W., Sauseng, P., & Gerloff, C.(2003). Enhancing cognitive perfor-mance with repetitive transcranial mag-netic stimulation at human individual al-pha frequency. Eur J Neurosci , 17 ,1129–1133.

Kotchoubey, B., Strehl, U., Holzapfel, S.,Blankenhorn, V., Fröscher, W., & Bir-baumer, N. (1999). Negative potentialshifts and the prediction of the outcomeof neurofeedback therapy in epilepsy.Clin Neurophysiol , 110, 683–686.

Kotchoubey, B., Strehl, U., Uhlmann, C.,Holzapfel, S., König, M., Fröscher, W.,et al. (2001). Modification of slow cor-tical potentials in patients with refractoryepilepsy: a controlled outcome study.Epilepsia, 42, 406–416.

Kropotov, J., Grin-Yatsenko, V., Ponomarev,V., Chutko, L., Yakovenko, E., & Nik-ishena, I. (2007). Changes in EEG spec-

trograms, event-related potentials andevent-related desynchronization inducedby relative beta training in ADHD chil-dren. Journal of. Neurotherapy , 11, 3–11.

Kropotov, J., Ponomarev, V., & Grin-Yatsenko,V. (2001). EEG-biocontrol method intreating the attention deficit hyperactivitydisorder in children. Human Physiology ,27 , 496–504.

Kropp, P., Siniatchkin, M., & Gerber, W.-D. (2002). On the pathophysiologyof migraine–links for ”empirically basedtreatment” with neurofeedback. ApplPsychophysiol Biofeedback , 27 , 203–213.

Leins, U., Goth, G., Hinterberger, T., Klinger,C., Rumpf, N., & Strehl, U. (2007). Neu-rofeedback for children with ADHD: acomparison of SCP and Theta/Beta pro-tocols. Appl Psychophysiol Biofeedback ,32, 73–88.

Lubar, J. F., Swartwood, M. O., Swartwood,J. N., & O’Donnell, P. H. (1995). Eval-uation of the effectiveness of EEG neu-rofeedback training for ADHD in a clini-cal setting as measured by changes inT.O.V.A. scores, behavioral ratings, andWISC-R performance. Biofeedback SelfRegul , 20, 83–99.

Monastra, V., Lynn, S., Linden, M., Lubar, J.,Gruzelier, J., & La Vaque, T. (2006).Electroencephalographic biofeed-back in the treatment of attention-deficit/hyperactivity disorder. Journal ofNeurotherapy , 9, 5–34.

Pineda, J., Brang, D., Hecht, E., Edwards, L.,Carey, S., Bacon, M., et al. (2008). Pos-itive behavioral and electrophysiological



changes following neurofeedback train-ing in children with autism. Researchin Autism Spectrum Disorders, 2, 557–581.

Raymond, J., Sajid, I., Parkinson, L. A., &Gruzelier, J. H. (2005). Biofeedback anddance performance: a preliminary inves-tigation. Appl Psychophysiol Biofeed-back , 30, 64–73.

Sterman, M. B., & Egner, T. (2006). Foun-dation and practice of neurofeedback forthe treatment of epilepsy. Appl Psy-chophysiol Biofeedback , 31, 21–35.

Sterman, M. B., & Friar, L. (1972). Sup-pression of seizures in an epileptic fol-lowing sensorimotor EEG feedback train-ing. Electroencephalogr Clin Neurophys-iol , 33, 89–95.

Sterman, M. B., Howe, R. C., & Macdonald,L. R. (1970). Facilitation of spindle-burst sleep by conditioning of electroen-cephalographic activity while awake. Sci-ence, 167 , 1146–1148.

Strehl, U., Kotchoubey, B., Trevorrow, T., &Birbaumer, N. (2005). Predictors ofseizure reduction after self-regulation ofslow cortical potentials as a treatment ofdrug-resistant epilepsy. Epilepsy Behav ,6, 156–166.

Strehl, U., Leins, U., Goth, G., Klinger,C., Hinterberger, T., & Birbaumer, N.(2006). Self-regulation of slow corticalpotentials: a new treatment for childrenwith attention-deficit/hyperactivity disor-der. Pediatrics, 118, e1530–e1540.

Strehl, U., Trevorrow, T., Veit, R., Hinterberger,T., Kotchoubey, B., Erb, M., et al. (2006).Deactivation of brain areas during self-regulation of slow cortical potentials in

seizure patients. Appl PsychophysiolBiofeedback , 31, 85–94.

Vernon, D., Egner, T., Cooper, N., Compton, T.,Neilands, C., Sheri, A., et al. (2003). Theeffect of training distinct neurofeedbackprotocols on aspects of cognitive perfor-mance. Int J Psychophysiol , 47 , 75–85.

Vernon, D., Frick, F., & Gruzelier, J. (2004).Neurofeedback as treatment for ADHD:A methodological review with implica-tions for future research. Journal of Neu-rotherapy , 8, 53–82.


A. J. Fallgatter et al. — Impact of COMT Genotype on Prefrontal Brain Functioning

Abstract

A. J. Fallgatter, Ch. G. Baehne, A.-Ch. Ehlis, M. M. Plichta, M. M. Richter, M. J. Her-rmann, A. Conzelmann, P. Pauli, Ch. Jacob, & K.-P. Lesch (Würzburg) — Impact of COMTGenotype on Prefrontal Brain Functioning in Adult ADHDThe enzyme catechol-O-methyltransferase (COMT) is involved in the degradation of dopamineand has been shown to modulate prefrontal brain function. The COMT Met variant displays de-creased COMT enzyme activity corresponding to higher concentrations of synaptic dopaminein prefrontal regions of the brain as compared to the Val variant. 200 adult ADHD patients and115 healthy volunteers were assessed with a 21 channel EEG while performing a response inhi-bition task. Topographical event-related potential (ERP) measures (centroids) in the P300 timerange as well as the NoGo-anteriorisation (NGA), an electrophysiological marker of prefrontalbrain function, were determined and checked for an association with COMT genotypes. Thestudy aimed at uncovering the impact of COMT genotype on prefrontal brain function in ADHDpatients and healthy controls. We obtained very different results in a preliminary analysis of 50ADHD patients vs. 50 healthy controls and in the final analysis of the whole sample: While forthe preliminary sample, ADHD patients with assumed low prefrontal dopamine levels displayeda significantly lower NGA as compared to those with a presumably strong prefrontal dopaminer-gic tone (“gene-dose effect”; Val/Val < Met/Met, p <0.01; Val/Met < Met/Met, p <0.05), whichwas in accordance with our a-priori hypotheses, we could not replicate a significant impact ofCOMT genotype on the NGA in the final ADHD sample. In conclusion, the findings obtained fromthe preliminary study must be considered to reflect a type-I error. This, however, is not likely toresult from the smaller sample size per se, since cell-size was sufficient to meet the criteria ofthe applied statistical tests. Results from studies on the impact of COMT genotype (and otherpolymorphisms) on neural and behavioral measures need independent replication before strongconclusion can be drawn.

Impact of COMT Genotype onPrefrontal Brain Functioning

in Adult ADHD

A. J. Fallgatter∗1, Ch. G. Baehne∗1, A.-Ch.Ehlis1, M. M. Plichta1, M. M. Richter1, M. J.Herrmann1, A. Conzelmann2, P. Pauli2, Ch.

Jacob1 & K.-P. Lesch1

∗ Both authors contributed equally to the work

1Department of Psychiatry andPsychotherapy, 2Department of Biological and

Clinical Psychology and Psychotherapy,University of Würzburg, Füchsleinstraße 15,

97080 Würzburg, [email protected]



Introduction

A catecholaminergic dysregulation hypothesisis proposed in at least some neurobiologicalsubtypes of ADHD (Castellanos & Tannock,2002). But how are the etiology and/or thesymptomatology of ADHD affected by varia-tions in catecholaminergic genes? The linkbetween genes and symptoms is not neces-sarily a direct one, thus endophenotypes maybe an adequate concept to uncover the influ-ence of genes in a disorder, as they are cor-related with the disease and may be more di-rectly linked to the genetic disposition of an in-dividual than the phenotypic behavior (Almasy& Blangero, 2001). One of the proposedendophenotypes for ADHD is a deficit in re-sponse control (Castellanos & Tannock, 2002;Sonuga-Barke, 2002).

Go/NoGo tasks are suited to measure theendophenotype of response control as theyrequire the preparation and execution of re-sponses to predefined target-stimuli (Go) andthe inhibition of the anticipated motor responsein the non-target condition (NoGo). The NoGoanteriorisation (NGA) is an electrophysiologi-cal measure of response control based on Go-and NoGo event-related potentials (ERPs)elicited during a Continuous PerformanceTest (CPT) (Bokura, Yamaguchi, & Kobayashi,2001; Fallgatter, Brandeis, & Strik, 1997; Fall-gatter et al., 2000; Fallgatter & Strik, 1999).This topographical ERP-parameter (NGA)has been shown to be stable (Fallgatter etal., 1997; Fallgatter, Mueller, & Strik, 1999),highly reliable (short- and long-term test-retestreliability; Fallgatter, Aranda, Bartsch, & Her-rmann, 2002; Fallgatter et al., 2001) and to beindependent of age and gender (Fallgatter etal., 1999). Topographical ERP measures such

as the NGA are suitable to detect effects ofgenetic variants affecting neurotransmitters,as has already been shown for the serotonintransporter (Fallgatter, Ehlis, et al., 2004)and dysbindin (Fallgatter et al., 2006) genevariants in healthy subjects. Moreover, NGApatterns seem to be altered in ADHD patients.The NGA was found to be reduced in adultpatients with personality disorders and ADHDduring childhood, whereas patients with per-sonality disorders only did not differ from thehealthy control group (Fallgatter et al., 2005).

The functioning of the prefrontal cortex (PFC)is strongly dependent on dopamine (Seamans& Yang, 2004). The COMT enzyme is crit-ically involved in degrading dopamine (andother catecholamines) in frontal brain areas(Grossman, Emanuel, & Budarf, 1992). Acommon functional single nucleotide polymor-phism (Val158Met) in COMT (located at 22q11)affects synaptic breakdown of dopamine in thePFC (Chen et al., 2004; Goldberg & Wein-berger, 2004). This COMT variation is causedby a single base pair change (G→ A) at aminoacid position 158 (or 108 respectively). Va-line (Val) is substituted by methionine (Met)(Lotta et al., 1995). This substitution exerts asignificant effect on the enzymatic activity ofCOMT. The Met variant results in a more ther-molabile enzyme which catabolizes dopamineless rapidly than the Val variant (Lachman etal., 1996; Lotta et al., 1995). The variants ex-ert a co-dominant effect (Chen et al., 2004).COMT impacts on dopamine degradation pri-marily in prefrontal brain areas as there is apaucity of dopamine transporters (DAT) in thisregion (Chen et al., 2004; Gogos et al., 1998).

High activity COMT variants seem to be as-sociated with personality characteristics and



diseases that are associated with ADHD orat least exhibit similar features like a higherdegree of extraversion (Reuter & Hennig,2005) or substance abuse (Vandenbergh,Rodriguez, Miller, Uhl, & Lachman, 1997).Patients with 22q11 microdeletion syndrome(which includes deletion of COMT) frequentlyshow ADHD symptoms (Gothelf et al., 2003).While some studies support the hypothesisthat COMT variants are associated with ADHD(e.g., Eisenberg et al., 1999; Qian et al.,2003), other investigations do not (e.g., Barret al., 1999; Bellgrove et al., 2005; Bobb etal., 2005; Hawi, Millar, Daly, Fitzgerald, &Gill, 2000; Kirley et al., 2002; Manor et al.,2000; Payton et al., 2001; Turic et al., 2005).Concerning cognition and/or behavior, moststudies come to the conclusion that Met allelesare more favorable to prefrontal functioningand prefrontally mediated performance (e.g.,Bearden et al., 2004; Bilder et al., 2002; Blasiet al., 2005; Frias et al., 2005; Diamond,Briand, Fossella, & Gehlbach, 2004; Egan etal., 2001; Gallinat et al., 2003; Goldberg etal., 2003; Malhotra et al., 2002), while othersreport more favorable results for Val allele car-riers (Baker, Baldeweg, Sivagnanasundaram,Scambler, & Skuse, 2005; Bellgrove et al.,2005), or cannot find an impact of COMT onexecutive function measures (Ho, Wassink,O’Leary, Sheffield, & Andreasen, 2005; Millset al., 2004; Taerk et al., 2004). Results seemto be task dependent (Bilder, Volavka, Lach-man, & Grace, 2004) and also appear to beinfluenced by the environmental context likeageing (Gothelf, Furfaro, Penniman, Glover,& Reiss, 2005) and are perhaps dependenton diagnosis. We hypothesized that Met al-lele and Val allele carriers significantly differ

regarding their prefrontal cortex function (asindicated by the NGA) due to the modulat-ing effect of dopamine. Due to the assumedinvolvement of the dopaminergic system inthe pathophysiology of ADHD (Swanson etal., 2007), we furthermore expected the exactimpact of COMT genotype to differ betweendiagnostic groups.

Methods

Participants

The investigation was approved by the EthicsCommittee of the University of Wuerzburgand was in accordance with the Declara-tion of Helsinki. Over a data acquisitionperiod of about 4 years, a total of 200 adultADHD (DSM-IV) in- and outpatients of theDepartment of Psychiatry of the University ofWuerzburg as well as 115 healthy control par-ticipants were included in the study after writ-ten informed consent was obtained. Patientsand healthy controls were interviewed withthe Structured Clinical Interview for DSM-IV(SCID-I) and the Structured Clinical Interviewfor DSM-IV personality disorders (SCID-II) byan experienced psychiatrist for diagnosis anddifferential-diagnosis of psychiatric disorders.All adult ADHD patients had a history of ADHDsymptomatology during childhood accordingto self-report and Wender-Utah Rating Scale(short version WURS-k validated for a Germanpopulation, Retz-Junginger et al., 2002).

Exclusion criteria were age below 18 andabove 60 years (actual age ranges of the finalsample: ADHD group 18–56 years, controlgroup 18–56 years), current medication withmethylphenidate or any other ADHD remedy,as well as serious somatic disorders. In the



ADHD group, only one patient was treatedwith antidepressants (mirtazapine and mapro-tiline, i.e. a combination of tetracyclic andserotonergic/noradrenergic medication); threeother patients received low doses of a ben-zodiazepine as PRN medication. Excludingthese patients from the data analyses did notsignificantly affect any of the below reportedfindings. No further medication affecting thecentral nervous system was administered.Patients and controls were only included in thefinal data analysis if their EEG data yielded asufficient number of artifact-free EEG epochs(> 20) and if the error rates for the Go-NoGotask were acceptable (omission and commis-sion errors < 20).

Due to the rather long data acquisition period(> 4 years), an interim analysis was performedafter 2 years, when 50 ADHD patients and 50controls could be included. Following the endof the data acquisition period, the analyseswere repeated for the complete sample of 200ADHD patients and 115 control participants.The results of both analyses are reported be-low, and differing findings will be discussed.

Within both ADHD samples (n=50 and n=200,respectively), the genotype distribution was inHardy-Weinberg equilibrium (χ2

1=0.35, p=0.55and χ2

1=1.61, p =0.20, respectively; with 47 %and 49 % Val alleles, respectively) and geno-type groups did not differ significantly for thedistribution of gender, handedness, or ADHDsubtype composition according to DSM-IV(see Table 1a and Table 1b; all χ2 values <

2.60, p >0.35). Similarly, genotype distributionwas in Hardy-Weinberg equilibrium for bothhealthy control samples (n=50 and n=115,respectively: χ2

1=0.08, p=0.78 and χ21=0.68,

p=0.41, respectively; with 50 % and 52 %

Val alleles, respectively), and COMT groups,again, did not differ significantly in gender orhandedness distribution (all χ2values < 2.20;p >0.30).

Across diagnostic groups, the genetic sub-groups were comparable for age and genderdistribution, both for the preliminary (gender:χ2

2 =0.17, p=0.92; handedness: Fisher’s ex-act test p=0.68) and for the final data analy-sis (gender: χ2

2 =0.62; p=0.73; handedness:χ2

2 =0.39; p=0.82). Overall, healthy controlsand ADHD patients also did not differ signifi-cantly regarding these variables (gender: χ2

1

=3.25, p=0.07 and χ21=0.90, p=0.34; handed-

ness: Fisher’s exact test p=0.70 and χ21=0.23,

p=0.63; for the preliminary and final data anal-ysis, respectively). In an ANOVA analysis ofthe variable “age”, main effects of COMT geno-type, diagnosis, or the interaction of geno-type and diagnosis did not indicate any sig-nificant between-group differences, neither forthe preliminary (all F -values < 1.25, p-values> 0.3) nor for the final study sample (F <0.70,p >0.4). Demographics and genotype compo-sition of both samples are listed in Table 1 (aand b).

The Structured Clinical Interview of DSM-IV(SCID-I) (Wittchen, Wunderlich, Gruschwitz,& Zaudig, 1997) and the Structured ClinicalInterview of DSM-IV personality disorders(SCID-II) (Fydrich, Renneberg, Schmitz, &Wittchen, 1997) was administered by anexperienced psychiatrist for all participants.Controls were only accepted without any axisI or II morbidities. As assessed by meansof SCID I interview, 25 ADHD patients of thepreliminary and 91 patients of the final samplewere diagnosed with a current axis I disorder(see Table 1 a,b for genotype distribution). Co-



Tab

le1:

Dem

ogra

phic

Cha

ract

eris

tics

(a)

Fin

alsa

mple

(age:

mea

n(s

d),

sd=

standard

dev

iati

on;y=

yea

rs;n=

num

ber

;%

=per

centa

ge

)

AD

HD

contr

ols

all

(n=

200)

Val/

Val

(n=

53)

Val/

Met

(n=

91)

Met

/M

et(n

=56)

all

(n=

115)

Val/

Val

(n=

33)

Val/

Met

(n=

53)

Met

/M

et(n

=29)

age,

mea

n(s

d),

y34.6

(9.8

)34

(9.3

)34.6

(9.8

)35.3

(10.3

)35.5

(10.3

)35.1

(10.4

)35.3

(9.5

)36.5

(11.7

)

Gen

der

,%

male

53

45

55

55

47

55

40

52

left

handed

,n

16

28

611

45

2

axis

Idis

ord

er,%

46

45

43

50

00

00

axis

IIdis

ord

er,%

73

68

69

84

00

00

Att

enti

on

pro

ble

ms,

n51

13

22

16

00

00

Hyper

act

ive

pro

ble

ms,

n22

513

40

00

0

Com

bin

edpro

ble

ms,

n127

35

56

36

00

00

(b)

Pre

lim

inary

sam

ple

(age:

mea

n(s

d),

sd=

standard

dev

iati

on;y=

yea

rs;n=

num

ber

;%

=per

centa

ge)

AD

HD

contr

ols

all

(n=

50)

Val/

Val

(n=

10)

Val/

Met

(n=

27)

Met

/M

et(n

=13)

all

(n=

50)

Val/

Val

(n=

13)

Val/

Met

(n=

24)

Met

/M

et(n

=13)

age,

mea

n(s

d),

y33.8

(9.8

)37.8

(11.0

)32.1

(9.4

)34.2

(9.6

)34.3

(10.5

)35.9

(11.1

)33.6

(9.7

)34.0

(12.0

)

Gen

der

,%

male

56

60

59

46

38

39

29

54

left

handed

,n

30

12

41

21

axis

Idis

ord

er,%

50

30

56

54

00

00

axis

IIdis

ord

er,%

74

60

74

85

00

00

Att

enti

on

pro

ble

ms,

n17

210

50

00

0

Hyper

act

ive

pro

ble

ms,

n4

13

00

00

0

Com

bin

edpro

ble

ms,

n29

714

80

00

0



morbid disorders were mood disorders (n=8),anxiety disorders (n=8), and substance abuseor addiction (n=9) in the preliminary sample,and mood disorders (n=31), anxiety disorders(n=27), substance abuse or addiction (n=27),OCD (n=1), post-traumatic stress disorder(n=3) and eating disorders (n=2) in the finalADHD sample. In both samples the mostcommon personality disorder (axis II) washistrionic personality disorder (n=17 and n=60patients, respectively), followed by narcissisticpersonality disorder (n=12 vs. n=54). In thepreliminary sample, a total of 37 patients werediagnosed with at least one comorbid person-ality disorder, whereas 146 patients of the finalADHD sample fell into this category.

Electrophysiological investigation

For the electrophysiological investigation theparticipants sat in an electrically shielded,dimly lit room and performed a CPT, whilethe ongoing EEG was recorded. 400 let-ters were presented in a pseudo-randomizedorder one at a time for 200 ms with an inter-stimulus interval of 1650 ms. Subjects wereinstructed to press a response button withtheir right hand whenever the letter O (primer)was followed by the letter X (Go-condition).Even though the response hand (left, right)was not counterbalanced across participants,left-handed subjects did not differ significantlyfrom right-handed ones for any of the behav-ioral measures (see below; Mann-WhitneyU ≥219.0, Z <1.5, p >0.15). Speed andaccuracy were equally emphasized. In thestimulus set of 400 letters (12 different letters:A, B, C, D, E, F, G, H, J, L, O, X) there were114 primer stimuli (O), 57 Go trials (O followedby X) and 57 NoGo trials (O followed by any

other letter than X). 172 distractor letters (otherletters, or X without a preceding O) completedthe stimulus set. Whenever necessary forunderstanding, the participants first performeda short training session; this was the case inless than 10% of both patients and controls.This CPT version took about 13 minutes. Re-sponses were registered with ERTS software(Experimental Run Time System, BeriSoftCooperation, Frankfurt/Main, Germany).Following the protocol established in our pre-vious studies (e.g., Fallgatter et al., 1997,2005; Fallgatter, Herrmann, et al., 2004), theEEG was recorded from 21 scalp electrodesplaced according to the extended international10-20 system, with three additional electrodesto monitor eye movements. The recordingreference was placed between Fz and Cz.Electrode impedances were below 5 kΩ.Recordings were performed with a 32-channelDC BrainAmp amplifier (Brain Products, Mu-nich, Germany) and the software Brain VisionRecorder (version 1.01 b; Brain Products,Munich, Germany). The A/D rate was 1000 Hzand the hardware filter was set to a bandpassof 0.1–70 Hz.

Data analysis

Behavioral data were the number of correctlyexecuted Go trials, the number of buttonpresses after NoGo trials (commission NoGo)and the number of button presses in othertrials (after the primer O or any distractorcondition, commission). Furthermore, meanreaction times of correct responses as well asthe individual standard deviation of reactiontimes as a measure of response variabilitywere calculated.Data analysis was performed offline with the



program Vision Analyzer (Brain Products, Mu-nich, Germany). A low-pass filter of 30 Hz wasused. After transforming the data to the av-erage reference, a correction for ocular arti-facts (Gratton, Coles, & Donchin, 1983) wasexecuted. The data were then segmented ac-cording to the Go and NoGo conditions of theCPT (-100 to 700 ms after stimulus presenta-tion). Only trials with a correct behavioral re-sponse were considered. Segments were re-jected when amplitudes exceeded ±50µV inany of the EEG channels or if the absolute dif-ference of two values in the segment exceeded200 µV. The maximal allowed voltage step be-tween two consecutive sampling points was50µV. The remaining segments were finally av-eraged to one Go and one NoGo event-relatedpotential per subject if each ERP contained atleast 20 epochs.

The anterior-posterior location of the positivecentroid (the amplitude-weighted center ofgravity of the positive brain-electrical field)(Lehmann, 1987) was calculated for the GoERPs according to the individual P300 peaklatency at Pz and for the NoGo ERPs accord-ing to the individual P300 peak latency at Czin a time frame of 277–434 ms (based on thestudy of Fallgatter et al., 1997). Visual inspec-tion of the grand average curves indicatedthat this particular P300 time frame was alsoadequate for the current data set. The cen-troid locations were quantified by a coordinatesystem defined by a two-dimensional delin-eation of the electrode array, with the digits 1to 5 indicating the electrode positions in theanterior-posterior and the left-right direction,respectively (see Figure 1). As a measure ofNoGo-anteriorisation the difference ygo – ynogo

was taken, since only the anterior-posterior

direction is of interest (1 = prefrontal elec-trode position Fpz, 5 = occipital electrodeposition Oz etc.; locations somewhere in be-tween two electrode positions were expressedby respective decimal numbers). The NGAtherefore represents the geometrical distancebetween the Go and NoGo centroid on ananterior-posterior axis (see Figure 1).

Genotyping

Genomic DNA was extracted from ethylene di-amine tetraacetic acid (EDTA) blood using theQIAamp Blood Kit (Qiagen, Hilden, Germany).Standard polymerase chain reaction (PCR)procedures modified from a previously pub-lished protocol (Egan et al., 2001) were usedfor genotyping COMT Val158Met. Primerswere 5’-GGG GCC TAC TGT GGC TAC TC(forward) and 5’-TTT TTC CAG GTC TGA CAACG (reverse). PCR reactions were performedin a reaction volume of 25 µl, including approxi-mately 50 ng of genomic DNA, 10 pmol of eachprimer, 2.5 mM of each dNTP, 0.75 mM MgCl2,and 1 U of Taq DNA Polymerase. Anneal-ing temperature was 58˚C (35 cycles). PCRproducts were digested with NlaIII (at least3h at 37˚C; fragment sizes wildtype G1947,114 bp; 1947A variant, 96bp and 13 bp) andsubsequently visualized on a 4% agarose gel.G1947 corresponds to the high-activity Val158allele, while 1947A codes for the low-activityMet variant. COMT allele frequencies for theADHD group were 47 % Val alleles vs. 53% Met alleles (Val/Val n=10, Val/Met n=27,Met/Met n=13) in the preliminary sample and49 % Val alleles vs. 51 % Met alleles (Val/Valn=53, Val/Met n=91, Met/Met n=56) in the fi-nal study sample. The allele frequencies forthe healthy control group were 50 % Val vs.



50 % Met alleles (Val/Val n=13, Val/Met n=24,Met/Met n=13) in the preliminary sample and52 % Val vs. 48 % Met alleles (Val/Val n=33,Val/Met n=53, Met/Met n=29) in the final studysample.

Statistical analysis

The behavioral variables considered were notnormally distributed (all Kolmogorov-SmirnovZ >1.2, all p-values < 0.2). Therefore, Mann-Whitney-U and Kruskal-Wallis tests were usedto test differences between the diagnosticgroups and between COMT genotypes in thewhole sample and in the different diagnostic orgenotypic groups. To correct for multiple test-ing (35 tests, p <0.05), Bonferroni correctionwas performed (p <0.0014). Only significantresults will be reported.For the NGA, an analysis of covariance (AN-COVA) was performed with COMT genotype(Val/Val vs. Val/Met vs. Met/Met) and diag-nosis (ADHD patients vs. healthy controls)as independent variables. We included ageand gender as covariates because severalstudies revealed a sexually dimorphic (Chenet al., 2004; Domschke et al., 2004; Go-gos et al., 1998) or age dependent impact ofCOMT genotypes (Gothelf et al., 2005). Asno age and gender effects on NGA could beshown in this ANCOVA, an analysis of vari-ance (ANOVA) on NGA was performed withthe same independent variables. To eluci-date significant interaction effects, separateANOVAs were run for the different diagnosticgroups with the independent variable COMT

genotype. T-tests were used for post hoc test-ing as well as for calculating the impact of diag-nosis in the different COMT genotype groups.Polynomial regression was used to analyze the

possible relationships between the number ofmet-alleles and NGA for the diagnostic groupsseparately.For the Go and NoGo centroid localizations,a 2×3 ANCOVA for repeated measurementswas conducted with the within-subject factorcondition (Go, NoGo) and the between-subjectfactors diagnosis and COMT-genotype. Forthe same reasons mentioned above we addedthe covariates age and gender. The ANCOVAshowed no significant impact of the factor gen-der and only a main effect for the factor agewith more anterior centroids in elderly subjectsirrespective of condition, which is in line withprevious findings (Fallgatter et al., 1999; Pf-efferbaum, Ford, Wenegrat, Roth, & Kopell,1984). We, therefore, conducted and reportbelow the results of a 2×3 ANOVA for repeatedmeasurements. As age showed no differencesconcerning genotypes, diagnosis or their inter-action (see above), the ANOVA for repeatedmeasurements without covariates seemed tobe justified.Post hoc analyses were conducted byANOVAs separated by condition, genotypeor diagnostic group. Significant results werespecified by means of two-tailed post-hoct-tests for matched or independent sam-ples. For repeated measurements ANOVAs,Greenhouse-Geisser correction was usedwhenever necessary. For t-tests, deviationsfrom variance homogeneity were tested byLevene test and corrections were performedwhen required.



Results

Preliminary analysis (50 ADHD patients vs.50 healthy controls)

Performance

Overall comparison between groups showedthat ADHD patients made significantly moreomission errors (U=750.0, p <0.001) andtended to make more commission errors (but-ton presses after an O or a distractor letter;U=842.0, p <0.0018) than healthy controls.No further difference was statistically signif-icant after Bonferroni correction. For detailssee Table 2.

ERP data

NoGo-Anteriorisation (NGA)

ADHD patients did not differ significantly fromhealthy controls in NGA (F1,94=1.03, p=0.31;Table 3b). However, in six patients and onlyone healthy control no NGA could be found.This difference was close to a statistical trend(Fisher’s exact test p=0.11).

Across all participants independent of diagno-sis, genotype had a significant effect on theNGA (F2,94=3.55, p <0.05): Val/Val carriershad a significantly smaller NGA than Val/Met(t72 = −2.12, p <0.05) and a trend to asmaller NGA as compared to Met/Met carri-ers (t47 = −1.95, p=0.06), with no signifi-cant difference between the latter two groups(t75 = −0.51, p=0.61). For uncovering the sig-nificant interaction effect of genotype and di-agnosis (F2,94=4.32, p <0.05), the impact ofgenotype was tested separately for both di-agnostic groups. In the ADHD group, COMT

genotype had a statistically significant impacton the NGA (F2,47=5.25, p <0.01), with sig-nificantly higher mean values in Met/Met car-

riers as compared to patients with a Val/Met(t38 = −2.06, p <0.05) or Val/Val geno-type (t21 = −2.95, p <0.01). A trend for asmaller NGA in Val/Val than Val/Met was found(t35 = −1.90, p=0.066). In the control group,no effect of genotype on NGA was detected(F2,47=1.68, p=0.20). Based on the polyno-mial regression, the relationship between thenumber of Met-alleles and the NGA for theADHD group can be best described as linear(linear model: F1,48=10.89; R2=.185; b=0.43;p <0.01; quadratic model: F -change < 0.01;b = −0.01; p=0.99; change in significance:p=0.99) and for the healthy control group asquadratic (linear model: F1,48=0.02; R2=.002;b = −0.02, p=0.87; quadratic model: F -change = 3.30; R2=0.07; b = −0.26, p <0.10;change in significance: p <0.10).

When subdividing the study sample into differ-ent genotypes to uncover differences betweendiagnostic groups, significant results could befound for the Val/Met group with a smaller NGAin the ADHD as compared to the control group(t49 = −2.19, p <0.05). In the Val/Val groupa trend was pointing to a smaller NGA in theADHD cohort (t21 = −1.82, p=0.08). No sig-nificant difference between diagnostic groupswas found for Met/Met allele carriers (t24=1.53,p=0.14) (for details see Table 3b and Figure 2).

Centroids In additional exploratory analyses,NoGo centroids were found to be more ante-rior than Go centroids across the whole studysample (main effect “condition”: F1,94=202.26,p <0.001; Table 3b and Figure 3). The maineffect of diagnosis revealed that overall ADHDpatients had more anterior centroid locationscompared to healthy controls (F1,94=8.86,p <0.01). The main effect of genotype wassignificant as well (F2,94=8.17, p <0.001) with



Figure 1: Grand average curves of the Go (Pz; black line) and Nogo (Cz; red line) condition for the

preliminary sample (n=100). Maps illustrate the distribution of the positive brain electrical field at the

respective peak of the P300. Schematic illustration of the quantification of the NGA as the geometrical

distance between Go and Nogo centroid on the anterior-posterior axis.

Val/Val allele carriers showing more anteriorcentroid locations compared to both Val/Met(t72 = −2.56, p <0.05) and Met/Met allelecarriers (t47 = −3.51, p <0.001).

To further analyze the significant interaction ofcondition × genotype × diagnosis (F2,94=4.32,p <0.05), separate post-hoc ANOVAs wereconducted for ADHD patients and healthycontrols. For the ADHD sample, analysesrevealed a significant main effect of COMT

genotype (F2,47=6.68, p <0.01) as well asa significant interaction genotype × condi-tion (F2,47=5.25, p <0.01): Overall, Val/Valallele carriers had significantly more anteriorcentroid locations compared to both Val/Met(t35 = −2.81, p <0.01) and Met/Met allelecarriers (t21 = −4.65, p <0.001), but – con-sidering both conditions separately – this

genotype effect was found only for the Gocentroid (F2,47=9.64, p <0.001; NoGo cen-troid: F2.47=1.30, p=0.28). The ADHD Val/Valgroup had significantly more anterior Go cen-troids than the Val/Met (t35 = −2.97, p <0.01)or Met/Met (t12 = −4.66, p <0.001) group,and patients carrying Val/Met had significantlymore anterior Go centroids than Met/Met allelecarriers (t38 = −2.32, p <0.05). Unlike in thepatient group, neither the main effect for geno-type (F2,47=2.09, p=0.14) nor the interactionbetween condition and genotype (F2,47=1.68,p=0.20) led to significant results in healthycontrols.

Splitting the sample into different genotypegroups and comparing the diagnostic groupsfor the Go and NoGo condition separately, cen-troid differences were revealed that appear to



Figure 2: Effect of COMT genotype on NoGo-Anteriorisation (NGA) separated by diagnostic groups

(preliminary sample). NGA is given in electrode positions (y-axis) and error bars represent standard

error of the mean. Asterisks represent p-values (*=p <0.05; **=p <0.01). Indicated are the differences

between genotypes in the ADHD group and the significant difference between healthy controls and ADHD

patients in the Val/Met genotype group.

underlie the NGA effects reported above: Inthe Val/Val group, ADHD patients had moreanterior Go centroid locations than healthycontrols (t21 = −2.59, p <0.05); in the Val/Metgroup, this effect was reduced to a statisticaltrend (t49 = −1.87, p=0.07). In Met/Met al-lele carriers the Go centroid did not differ sig-nificantly between diagnostic groups (t24=0.11,p=0.92), however, a marginally significant dif-ference was found for the NoGo centroid, withmore anterior centroid locations in ADHD pa-tients as compared to healthy controls (t24 =−1.74, p=0.096).

Final analysis (200 ADHD patients vs. 115healthy controls)

Performance

Analysis of the final study sample confirmedthe finding of a significantly higher number of

commission errors (button presses after an Oor a distractor letter; U=8554.5, p <0.001)and omission errors (U=8686.0, p <0.001) inADHD patients as compared to healthy con-trols. Moreover, the individual reaction timeswere more variable in ADHD patients than inhealthy controls as indicated by significantlylarger standard deviations for correct Go re-sponses (U=8062.0, p <0.001). No further dif-ference was statistically significant after Bon-ferroni correction. For details see Table 2a.

ERP data

NoGo-Anteriorisation (NGA)

In sharp contrast to the above-mentioned pre-liminary findings, for the final study samplethe 2×3 ANOVA conducted to analyze theNGA showed no significant main effects ofdiagnosis (F1,309=1.73, p=0.19) or genotype(F2,309=1.11, p=0.33) and no significant inter-action between the two factors (F2,309=0.19,



Table

2:Perform

ancedata

(a)

Fin

alsa

mple

(all

varia

bles:

mea

n(sd

),sd

=sta

ndard

dev

iatio

n,m

s=m

illiseconds,

n=

num

ber)

AD

HD

contro

ls

all

(n=

200)

Val/

Val

(n=

53)

Val/

Met

(n=

91)

Met/

Met

(n=

56)

all

(n=

115)

Val/

Val

(n=

33)

Val/

Met

(n=

53)

Met/

Met

(n=

29)

com

missio

nerro

rs(O

/distra

ctor;

n,sd

)1.1

2(1

.70)

1.1

7(1

.87)

1.0

9(1

.40)

1.1

3(1

.98)

0.4

7(0

.82)

0.6

1(1

.03)

0.4

7(0

.75)

0.3

1(0

.66)

com

missio

nerro

rs(N

oG

o;n,sd

)0.2

0(0

.68)

0.0

9(0

.30)

0.3

0(0

.80)

0.1

4(0

.72)

0.1

3(0

.39)

0.1

5(0

.36)

0.0

9(0

.35)

0.1

7(0

.47)

om

ission

errors

(n,sd

)2.0

2(3

.08)

1.6

4(2

.73)

2.2

9(3

.09)

1.9

3(3

.38)

0.9

0(1

.59)

1.3

3(1

.59)

0.5

3(1

.51)

1.0

7(1

.62)

reactio

ntim

e(m

s)494.0

7(1

20.7

3)

495.1

2(1

13.6

5)

498.0

2(1

20.2

3)

486.6

6(1

29.5

8)

459.2

7(1

11.6

6)

475.5

3(1

18.6

6)

448.6

7(9

7.7

6)

460.1

4(1

28.0

4)

Sd

ofrea

ction

times

114.8

8(5

2.1

5)

114.9

6(5

3.0

2)

116.0

6(5

2.9

6)

112.8

9(5

0.8

5)

89.9

6(4

4.3

6)

97.7

9(5

3.1

5)

83.5

9(3

6.6

0)

92.6

8(4

6.2

6)

(b)

Prelim

inary

sam

ple

(all

varia

bles:

mea

n(sd

),sd

=sta

ndard

dev

iatio

n,m

s=m

illiseconds,

n=

num

ber)

AD

HD

contro

ls

all

(n=

50)

Val/

Val

(n=

10)

Val/

Met

(n=

27)

Met/

Met

(n=

13)

all

(n=

50)

Val/

Val

(n=

13)

Val/

Met

(n=

24)

Met/

Met

(n=

13)

com

missio

nerro

rs(O

/distra

ctor;

n,sd

)1.0

8(1

.43)

1.7

0(1

.49)

1.0

0(1

.52)

0.7

7(1

.09)

0.3

6(0

.63)

0.2

3(0

.44)

0.3

8(0

.58)

0.4

6(0

.88)

com

missio

nerro

rs(N

oG

o;n,sd

)0.1

4(0

.41)

0.0

0(0

.00)

0.2

6(0

.53)

0.0

0(0

.00)

0.0

6(0

.24)

0.0

0(0

.00)

0.0

4(0

.20)

0.1

5(0

.38)

om

ission

errors

(n,sd

)2.5

6(3

.29)

2.4

0(4

.03)

2.3

3(3

.06)

3.1

5(3

.34)

0.8

8(1

.73)

0.6

9(1

.03)

0.8

3(1

.97)

1.1

5(1

.91)

reactio

ntim

e(m

s)488.0

2(1

22.9

9)

466.1

2(1

27.9

5)

485.3

0(9

8.1

7)

510.5

4(1

66.2

7)

476.8

7(1

16.2

6)

470.8

6(1

06.6

1)

460.7

5(9

6.9

3)

512.6

3(1

54.5

2)

Sd

ofrea

ction

times

108.9

2(5

0.9

0)

99.8

8(4

7.8

2)

102.1

4(4

3.0

8)

129.9

6(6

4.7

5)

89.6

6(4

1.8

6)

86.4

1(5

0.1

3)

82.4

3(3

1.5

0)

106.2

5(4

8.2

2)



Figure 3: Effect of COMT genotype on Go and NoGo centroids separated by diagnostic groups (pre-

liminary sample). Centroids are given in electrode positions (y-axis) in anterior posterior direction and

error bars represent standard error of the mean. Asterisks represent p-values (*=p <0.05; **=p <0.01;

***=p <0.001). Indicated are the differences in Go centroids between genotypes in the ADHD group

and the significant difference between healthy controls and ADHD patients in the Val/Val genotype group.

p=0.83; see Table 3a). Adding the factors ageand gender as covariates did not noticeablychange these results (all F -values < 1.70, p> 0.20).CentroidsThe exploratory analysis of Go and NoGo cen-troids confirmed the finding of more anteriorcentroid locations in the NoGo as comparedto the Go condition across study groups (maineffect “condition”: F1,309=496.97, p <0.001;see Table 3a). Moreover, the main effectof diagnosis again reached statistical signifi-cance indicating an overall more anterior cen-troid localization in ADHD patients comparedto healthy controls (F1,309=9.18, p <0.01).However, none of the remaining findings re-ported above could be replicated in the finalstudy sample: Neither the main effect of geno-type (F2,309=2.03, p=0.13) nor the interactionbetween genotype and condition (F2,309=1.11,

p=0.33) or condition, genotype and diagno-sis (F2,309=0.19, p=0.83) reached statisticalsignificance. In accordance with the above-mentioned findings, no significant interactioneffect of genotype and diagnosis could befound (F2,309=0.72, p=0.49).

Comparability

To test whether the obvious differences in re-sults between the preliminary and the finaldata analysis could be attributed to formaldifferences between the two study samples,we compared both samples for differencesin demographic and clinical variables. Wefurthermore compared the preliminary sam-ple with the subsample of participants addi-tionally included in the final study sample (inboth cases separately for patients and con-trols). We did not find any significant differ-



ences in age (t <1.15, p >0.25), IQ (t <1.15,p >0.25), gender (χ2 <1.20, p >0.25; ex-cept for the comparison of preliminary con-trol subjects and controls additionally includedin the final study sample: χ2=2.85, p=0.09)or handedness distribution (Fisher’s exact testp >0.35), or the distribution of COMT geno-types (χ2 <2.25, p >0.3) for any of thesecomparisons. Moreover, the different patientsamples did not differ significantly regardingtheir mean BDI (t <0.85, p >0.4) or WURSscore (t <0.35; p >0.7), mean number of per-sonality disorders (t <0.20, p >0.8) or thenumber of patients diagnosed with comorbidaxis I (χ2 <0.55, p >0.45) or axis II disorders(χ2 <0.05, p=0.85). They were also well com-parable regarding their ADHD subtype compo-sitions (χ2 <2.80, p >0.25).

Discussion

The aim of this study was to investigate the im-pact of COMT genotype on brain functioning ina response control task in adult ADHD patientsas compared to healthy controls. Analysis ofthe behavioral data revealed significant differ-ences in performance between ADHD patientsand control participants with patients perform-ing worse. Differences were found in commis-sion errors when participants directly pressedafter an O or a distractor letter (= failed inhibi-tion) and in omission errors, perhaps pointingto a more impulsive or inattentive answeringstyle in ADHD. Furthermore, in the final ADHDsample, reaction times were found to be morevariable than in healthy control subjects, possi-bly indicating fluctuations and lapses in atten-tion (Leth-Steensen, Elbaz, & Douglas, 2000;Molen, Somsen, & Jennings, 1996). COMT

genotype had no significant influence on any

of the behavioral measures.

On a neurophysiological level, we obtainedvery different results for the preliminary analy-sis of 50 ADHD patients and 50 healthy con-trols on the one hand, and the final sampleof 200 patients and 115 controls on the otherhand. Even though general electrophysiolog-ical findings could be demonstrated for bothsamples (more anterior centroids in NoGo thanin Go trials, see Fallgatter et al., 1997; moreanterior centroids in ADHD patients particu-larly for the Go condition, (see Baehne et al.,2008, Fallgatter et al., 2005; Fallgatter & Her-rmann, 2001), a significant impact of COMT

genotype on ERP markers of response con-trol was found in the preliminary analysis, butcould not be replicated in the final study sam-ple.

Within the preliminary sample of 50 ADHD pa-tients, a linear, gene dose dependent COMT

effect on the NGA was uncovered: Patientswith the most active COMT enzyme and there-fore the presumably lowest dopaminergic tonein prefrontal areas (Val/Val) were characterizedby the smallest NGA, those with the least ac-tive COMT enzyme and therefore the presum-ably highest dopaminergic tone (Met/Met) hadthe highest NGA, the heterozygotes lying inbetween (see Figures 2 and 4; Figure 4 mod-ified from Fallgatter & Lesch, 2007; Goldman-Rakic, Muly, & Williams, 2000). Patients car-rying two copies of the Met allele were alsocharacterized by an NGA that was not signif-icantly different from the values obtained inthe healthy control sample, whereas the re-maining two ADHD groups (Val/Met, Val/Val)showed at least a statistical trend for reducedNGA values. These findings are immediatelyplausible, as the very same correlation be-



Table

3:C

entroidsand

NG

A

(a)

Fin

alsa

mple

(all

varia

bles:

mea

n(sd

);cen

troid

sand

NG

Ain

electrode

positio

ns)

AD

HD

contro

ls

all

(n=

200)

Val/

Val

(n=

53)

Val/

Met

(n=

91)

Met/

Met

(n=

56)

all

(n=

115)

Val/

Val

(n=

33)

Val/

Met

(n=

53)

Met/

Met

(n=

29)

Cen

troid

sG

o3.6

6(0

.48)

3.5

4(0

.64)

3.7

0(0

.41)

3.7

1(0

.41)

3.8

0(0

.35)

3.7

6(0

.39)

3.8

2(0

.30)

3.8

3(0

.37)

Cen

troid

sN

oG

o3.0

7(0

.37)

3.0

1(0

.36)

3.0

8(0

.36)

3.1

0(0

.39)

3.1

3(0

.30)

3.1

3(0

.28)

3.0

9(0

.25)

3.2

0(0

.38)

NG

A0.5

9(0

.50)

0.5

3(0

.62)

0.6

2(0

.46)

0.6

1(0

.45)

0.6

7(0

.38)

0.6

2(0

.38)

0.7

3(0

.35)

0.6

3(0

.43)

(b)

Prelim

inary

sam

ple

(all

varia

bles:

mea

n(sd

);cen

troid

sand

NG

Ain

electrode

positio

ns)

AD

HD

contro

ls

all

(n=

50)

Val/

Val

(n=

10)

Val/

Met

(n=

27)

Met/

Met

(n=

13)

all

(n=

50)

Val/

Val

(n=

13)

Val/

Met

(n=

24)

Met/

Met

(n=

13)

Cen

troid

sG

o3.5

9(0

.48)

3.1

1(0

.48)

3.6

3(0

.46)

3.8

7(0

.21)

3.7

9(0

.39)

3.6

3(0

.48)

3.8

3(0

.31)

3.8

6(0

.43)

Cen

troid

sN

oG

o2.9

9(0

.35)

2.8

4(0

.24)

3.0

4(0

.37)

2.9

9(0

.37)

3.0

9(0

.32)

3.0

2(0

.28)

3.0

3(0

.23)

3.2

6(0

.43)

NG

A0.6

0(0

.48)

0.2

8(0

.50)

0.5

8(0

.41)

0.8

8(0

.47)

0.7

0(0

.38)

0.6

2(0

.39)

0.8

0(0

.30)

0.5

9(0

.49)



tween COMT and NGA values has also beenreported in a group of patients with disor-ders from the schizophrenia spectrum (Ehlis,Reif, Herrmann, Lesch, & Fallgatter, 2007). Inthis study, patients with a Met/Met genotypehad a significantly higher mean NGA than pa-tients with a Val/Met or Val/Val genotype. Thisfinding was accompanied by a trend for pro-longed reaction times in the interference condi-tion of the Stroop task in Val/Met as comparedto Met/Met carriers, which supports the inter-pretation of an impact of COMT genotype onprefrontal brain function in this group of pa-tients. Since both ADHD and schizophreniaspectrum disorders have been associated withabnormalities in dopaminergic neurotransmis-sion within the prefrontal cortex, a similar in-fluence of COMT in both diagnostic groupsseems convincing. In the preliminary healthycontrol sample, no significant impact of COMT

genotype could be found and the relation be-tween NGA and COMT genotypes was numer-ically best described by an inverted u-shape(Mattay et al., 2003; Meyer-Lindenberg et al.,2005) with both homozygote genotypes dis-playing smaller values of the NGA than theheterozygote genotype (see Figures 2 and 4),even if these differences did not reach statisti-cal significance.

An exploratory analysis of Go and NoGo cen-troids indicated that it was the Go ERP thatwas particularly affected by the COMT effectwithin the preliminary group of ADHD patients,where Val allele carriers showed a significantlymore anterior position of the Go centroid ascompared to carriers of the Met allele. Sincethe Go condition is usually associated with rel-atively little prefrontal activation, this may bedue to an ineffective signal processing during

correctly performed Go trials (frontal noise) inthe ADHD Val/Val group, possibly due to de-creased dopamine levels. In this respect, thefrontal P300 may be considered a measure offrontal noise including broadband and unse-lective oscillatory features, even though somecorrelated signal-like activity might contributeto the frontal P300-component (Baudena, Hal-gren, Heit, & Clarke, 1995; Ehlis et al., 2007;Winterer et al., 2004). Based on the as-sumption that the frontal P300 can be con-sidered as an estimator of “neuronal noise”and that dopamine is in part responsible forthe reduction of prefrontal noise, Gallinat etal. (2003) compared the fronto-central P300amplitude in an auditory oddball task for thethree COMT genotypes in schizophrenic pa-tients and healthy controls. In line with thefrontal noise hypothesis and our own data re-ported above, the group homozygous for theMet allele was characterized by significantlylower frontal P300 amplitudes as compared toVal/Val allele carriers.

In summary, the impact of COMT genotype onERP measures of response control found inour preliminary sample of ADHD patients isin line with previous findings and the study’sa-priori hypotheses. Furthermore, the differ-ential impact of COMT in ADHD patients ver-sus healthy controls fits recent models on theassumed relation between dopaminergic toneand prefrontal brain functioning (see Figure4). Our interim analysis of a reasonably largesample of 50 ADHD patients and 50 controlparticipants, therefore, yielded consistent andplausible results.

However, these results could not be confirmedwhen analyzing the final study sample of 200ADHD patients and 115 healthy control sub-



Figure 4: Hypothetical correlation between prefrontal brain function (as indicated by observed values of

the NGA in the preliminary sample; Y-axis) and the hypothesized dopamine concentration (as indicated

by COMT genotype; X-axis) in Val/Val, Val/Met, and Met/Met carriers of ADHD patients and healthy

controls. Assuming that healthy controls are all located in a normal range of dopamine concentration,

it is hypothesized that ADHD Val/Val allele carriers are outside of the optimal range at the lower end of

dopamine concentration. ADHD Met/Met allele carriers are presumed to have dopamine concentrations

in the normal range.



jects. In this final analysis, COMT geno-type was found to have no significant impacton neurophysiological markers of cognitive re-sponse control, neither in ADHD patients nor inhealthy controls. These very different findingscould not be attributed to formal differencesbetween the two samples that were very wellcomparable regarding numerous demographicand clinical variables (see Results section).

Therefore, it must be concluded that the find-ings obtained from the preliminary study reflecta type-I error. This, however, is not likely to re-sult from the smaller sample size per se, sincecell-size was sufficient for the applied statisti-cal approaches and simulations demonstratedthat our statistical analyses were not associ-ated with a type-I-error inflation (data on re-quest). Results from studies on the impactof COMT genotype (and other polymorphisms)on neural and behavioral measures need inde-pendent replications before strong conclusioncan be drawn.

Limitations of our study concern the relativelybroad age range that was considered for thepresent study sample (inclusion criterion: 18-60 years; actual age range of the final sample:18-56 years) and the high incidence of comor-bid disorder. Since the NGA – our primary tar-get parameter – has been shown to be largelyindependent of the subjects’ age (Fallgatter etal., 1999), we do not think that our results weresignificantly affected by this potential source ofvariability. Regarding the issue of comorbidity,the frequent occurrence of comorbid disordersis a very common problem in ADHD studiesreflecting the actual clinical situation (Kordon& Kahl, 2004; Sobanski, 2006). Nevertheless,it cannot be excluded that our findings were af-fected by the high incidence of comorbid axis I

and axis II disorder. Another critical point con-cerns the fact that, due to a very strict artifactand error criterion, a total of 30 originally ex-amined ADHD patients could not be includedin the data analyses. This might have causeda selection bias since patients with particu-larly severe symptoms of hyperactivity (proneto cause EEG motion artifacts) and / or inatten-tion (prone to make errors) might have beensystematically excluded from the present pa-tient sample, which might have led to a corre-sponding distortion of the results.

In contrast to our previous study (Fallgatter etal., 2005), we did not find a clear indicationof a reduced NGA in ADHD patients. It maybe speculated that the clinical symptomatologywas more severe in the previous forensic pop-ulation with severe personality disorders andadditional indications for an ADHD symptoma-tology during childhood (Fallgatter et al., 2005)than in the non-forensic ADHD sample investi-gated in the current study.

In summary, we did not find any consis-tent indication of a significant impact of theVal158Met polymorphism of COMT on theNGA as a neurophysiological marker of frontallobe function. A positive finding of a prelimi-nary analysis has to be considered to be theresult of a type-I error, which could, however,not be attributed to an excessively small sam-ple size or unequal cell sizes for the threegenotype groups. In conclusion, results fromimaging genetics studies need independentreplication before strong conclusions can bedrawn, even with sample sizes that are con-sidered to be reasonably large within imaginggenetics literature.



Acknowledgements

We thank all participants. We would like tothank I. Gröbner, M. Harder, M. Heine, C.Jambor, T. Renner, J. Romanos, I. Schaf-frath, and N. Steigerwald for excellent tech-nical and methodical assistance. This studywas supported by the Deutsche Forschungs-gemeinschaft (Grant KFO 125/1-1 to A.J.F.and K.P.L.)

References

Almasy, L., & Blangero, J. (2001). Endopheno-types as quantitative risk factors for psy-chiatric disease: rationale and study de-sign. Am J Med Genet , 105, 42–44.

Baehne, C. G., Ehlis, A.-C., Plichta, M. M.,Conzelmann, A., Pauli, P., Jacob, C., etal. (2008, Apr). Tph2 gene variantsmodulate response control processes inadult adhd patients and healthy individ-uals. Mol Psychiatry . Available fromhttp://dx.doi.org/10.1038/mp.2008.39

Baker, K., Baldeweg, T., Sivagnanasundaram,S., Scambler, P., & Skuse, D. (2005).COMT Val108/158 Met modifies mis-match negativity and cognitive function in22q11 deletion syndrome. Biol Psychia-try , 58, 23–31.

Barr, C. L., Wigg, K., Malone, M., Schachar,R., Tannock, R., Roberts, W., et al.(1999). Linkage study of catechol-o-methyltransferase and attention-deficithyperactivity disorder. Am J Med Genet ,88, 710–713.

Baudena, P., Halgren, E., Heit, G., & Clarke,J. M. (1995). Intracerebral potentials torare target and distractor auditory and vi-

sual stimuli. III. Frontal cortex. Electroen-cephalogr Clin Neurophysiol , 94, 251–264.

Bearden, C. E., Jawad, A. F., Lynch, D. R.,Sokol, S., Kanes, S. J., McDonald-McGinn, D. M., et al. (2004). Effects of afunctional COMT polymorphism on pre-frontal cognitive function in patients with22q11.2 deletion syndrome. Am J Psy-chiatry , 161, 1700–1702.

Bellgrove, M. A., Domschke, K., Hawi, Z., Kir-ley, A., Mullins, C., Robertson, I. H., etal. (2005). The methionine allele ofthe COMT polymorphism impairs pre-frontal cognition in children and adoles-cents with ADHD. Exp Brain Res, 163,352–360.

Bilder, R. M., Volavka, J., Czobor, P., Mal-hotra, A. K., Kennedy, J. L., Ni, X., etal. (2002). Neurocognitive correlates ofthe COMT Val(158)Met polymorphism inchronic schizophrenia. Biol Psychiatry ,52, 701–707.

Bilder, R. M., Volavka, J., Lachman, H. M.,& Grace, A. A. (2004). The catechol-O-methyltransferase polymorphism: re-lations to the tonic-phasic dopamine hy-pothesis and neuropsychiatric pheno-types. Neuropsychopharmacology , 29,1943–1961.

Blasi, G., Mattay, V. S., Bertolino, A., Elvevåg,B., Callicott, J. H., Das, S., et al. (2005).Effect of catechol-O-methyltransferaseval158met genotype on attentional con-trol. J Neurosci , 25, 5038–5045.

Bobb, A. J., Addington, A. M., Sidransky, E.,Gornick, M. C., Lerch, J. P., Greenstein,D. K., et al. (2005). Support for associa-tion between ADHD and two candidate



genes: NET1 and DRD1. Am J MedGenet B Neuropsychiatr Genet , 134B,67–72.

Bokura, H., Yamaguchi, S., & Kobayashi, S.(2001). Electrophysiological correlatesfor response inhibition in a Go/NoGotask. Clin Neurophysiol , 112, 2224–2232.

Castellanos, F. X., & Tannock, R.(2002). Neuroscience of attention-deficit/hyperactivity disorder: the searchfor endophenotypes. Nat Rev Neurosci ,3, 617–628.

Chen, J., Lipska, B. K., Halim, N., Ma,Q. D., Matsumoto, M., Melhem, S.,et al. (2004). Functional analy-sis of genetic variation in catechol-O-methyltransferase (COMT): effects onmRNA, protein, and enzyme activity inpostmortem human brain. Am J HumGenet , 75, 807–821.

Diamond, A., Briand, L., Fossella, J., &Gehlbach, L. (2004). Genetic and neu-rochemical modulation of prefrontal cog-nitive functions in children. Am J Psychi-atry , 161, 125–132.

Domschke, K., Freitag, C. M., Kuhlenbäumer,G., Schirmacher, A., Sand, P., Ny-huis, P., et al. (2004). Association ofthe functional V158M catechol-O-methyl-transferase polymorphism with panic dis-order in women. Int J Neuropsychophar-macol , 7 , 183–188.

Egan, M. F., Goldberg, T. E., Kolachana, B. S.,Callicott, J. H., Mazzanti, C. M., Straub,R. E., et al. (2001). Effect of COMTVal108/158 Met genotype on frontal lobefunction and risk for schizophrenia. ProcNatl Acad Sci U S A, 98, 6917–6922.

Ehlis, A.-C., Reif, A., Herrmann, M. J., Lesch,K.-P., & Fallgatter, A. J. (2007). Impactof catechol-o-methyltransferase on pre-frontal brain functioning in schizophreniaspectrum disorders. Neuropsychophar-macology , 32, 162–170.

Eisenberg, J., Mei-Tal, G., Steinberg, A., Tar-takovsky, E., Zohar, A., Gritsenko, I., etal. (1999). Haplotype relative risk studyof catechol-O-methyltransferase (COMT)and attention deficit hyperactivity disor-der (ADHD): association of the high-enzyme activity Val allele with ADHDimpulsive-hyperactive phenotype. Am JMed Genet , 88, 497–502.

Fallgatter, A. J., Aranda, D. R., Bartsch, A. J.,& Herrmann, M. J. (2002). Long-termreliability of electrophysiologic responsecontrol parameters. J Clin Neurophysiol ,19, 61–66.

Fallgatter, A. J., Bartsch, A. J., Strik,W. K., Mueller, T. J., Eisenack, S. S.,Neuhauser, B., et al. (2001). Test-retestreliability of electrophysiological param-eters related to cognitive motor control.Clin Neurophysiol , 112, 198–204.

Fallgatter, A. J., Brandeis, D., & Strik,W. K. (1997). A robust assessmentof the NoGo-anteriorisation of P300 mi-crostates in a cued Continuous Perfor-mance Test. Brain Topogr , 9, 295–302.

Fallgatter, A. J., Ehlis, A.-C., Rösler, M., Strik,W. K., Blocher, D., & Herrmann, M. J.(2005). Diminished prefrontal brain func-tion in adults with psychopathology inchildhood related to attention deficit hy-peractivity disorder. Psychiatry Res,138, 157–169.

Fallgatter, A. J., Ehlis, A.-C., Seifert, J.,



Strik, W. K., Scheuerpflug, P., Zillessen,K. E., et al. (2004). Altered responsecontrol and anterior cingulate functionin attention-deficit/hyperactivity disorderboys. Clin Neurophysiol , 115, 973–981.

Fallgatter, A. J., Esienack, S. S., Neuhauser,B., Aranda, D., Scheuerpflug, P., & Her-rmann, M. J. (2000). Stability of lateevent-related potentials: topographicaldescriptors of motor control comparedwith the P300 amplitude. Brain Topogr ,12, 255–261.

Fallgatter, A. J., & Herrmann, M. J. (2001).Electrophysiological assessment of im-pulsive behavior in healthy subjects.Neuropsychologia, 39, 328–333.

Fallgatter, A. J., Herrmann, M. J., Hohoff,C., Ehlis, A.-C., Jarczok, T. A., Freitag,C. M., et al. (2006). DTNBP1 (dysbindin)gene variants modulate prefrontal brainfunction in healthy individuals. Neuropsy-chopharmacology , 31, 2002–2010.

Fallgatter, A. J., Herrmann, M. J., Roemm-ler, J., Ehlis, A.-C., Wagener, A., Hei-drich, A., et al. (2004). Allelic variationof serotonin transporter function modu-lates the brain electrical response for er-ror processing. Neuropsychopharmacol-ogy , 29, 1506–1511.

Fallgatter, A. J., & Lesch, K. P. (2007). 22q11.2deletion syndrome as a natural modelfor COMT haploinsufficiency-relateddopaminergic dysfunction in ADHD. Int JNeuropsychopharmacol , 10, 295–299.

Fallgatter, A. J., Mueller, T. J., & Strik, W. K.(1999). Age-related changes in the brainelectrical correlates of response control.Clin Neurophysiol , 110, 833–838.

Fallgatter, A. J., & Strik, W. K. (1999). The

NoGo-anteriorization as a neurophysio-logical standard-index for cognitive re-sponse control. Int J Psychophysiol , 32,233–238.

Frias, C. M. de, Annerbrink, K., Westberg,L., Eriksson, E., Adolfsson, R., & Nils-son, L.-G. (2005). Catechol O-methyltransferase Val158Met polymor-phism is associated with cognitive perfor-mance in nondemented adults. J CognNeurosci , 17 , 1018–1025.

Fydrich, T., Renneberg, B., Schmitz, B.,& Wittchen, H.-U. (1997). SKIDII. Interviewheft. Strukturiertes Klin-isches Interview fur DSM-IV. AchseII (Persönlichkeitsstörungen). Einedeutschsprachige, erweiterte Bear-beitung der amerikanischen Originalver-sion des SCID-II von: First, M. B.,Spitzer, R. L., Gibbon, M., Williams,J. B. W., Benjamin, L. (Version 3/96).Göttingen: Hogrefe.

Gallinat, J., Bajbouj, M., Sander, T.,Schlattmann, P., Xu, K., Ferro, E. F., etal. (2003). Association of the G1947ACOMT (Val(108/158)Met) gene poly-morphism with prefrontal P300 duringinformation processing. Biol Psychiatry ,54, 40–48.

Gogos, J. A., Morgan, M., Luine, V., San-tha, M., Ogawa, S., Pfaff, D., et al.(1998). Catechol-O-methyltransferase-deficient mice exhibit sexually dimorphicchanges in catecholamine levels and be-havior. Proc Natl Acad Sci U S A, 95,9991–9996.

Goldberg, T. E., Egan, M. F., Gscheidle, T.,Coppola, R., Weickert, T., Kolachana,B. S., et al. (2003). Executive sub-



processes in working memory: rela-tionship to catechol-O-methyltransferaseVal158Met genotype and schizophrenia.Arch Gen Psychiatry , 60, 889–896.

Goldberg, T. E., & Weinberger, D. R. (2004).Genes and the parsing of cognitive pro-cesses. Trends Cogn Sci , 8, 325–335.

Goldman-Rakic, P. S., Muly, E. C., & Williams,G. V. (2000). D(1) receptors in prefrontalcells and circuits. Brain Res Brain ResRev , 31, 295–301.

Gothelf, D., Furfaro, J. A., Penniman, L. C.,Glover, G. H., & Reiss, A. L. (2005).The contribution of novel brain imagingtechniques to understanding the neuro-biology of mental retardation and devel-opmental disabilities. Ment Retard DevDisabil Res Rev , 11, 331–339.

Gothelf, D., Gruber, R., Presburger, G., Dotan,I., Brand-Gothelf, A., Burg, M., et al.(2003). Methylphenidate treatment forattention-deficit/hyperactivity disorder inchildren and adolescents with velocar-diofacial syndrome: an open-label study.J Clin Psychiatry , 64, 1163–1169.

Gratton, G., Coles, M. G., & Donchin, E.(1983). A new method for off-line re-moval of ocular artifact. Electroen-cephalogr Clin Neurophysiol , 55, 468–484.

Grossman, M. H., Emanuel, B. S., & Budarf,M. L. (1992). Chromosomal mapping ofthe human catechol-O-methyltransferasegene to 22q11.1—-q11.2. Genomics,12, 822–825.

Hawi, Z., Millar, N., Daly, G., Fitzgerald, M., &Gill, M. (2000). No association betweencatechol-O-methyltransferase (COMT)gene polymorphism and attention deficit

hyperactivity disorder (ADHD) in an Irishsample. Am J Med Genet , 96, 282–284.

Ho, B.-C., Wassink, T. H., O’Leary, D. S.,Sheffield, V. C., & Andreasen, N. C.(2005). Catechol-O-methyl transferaseVal158Met gene polymorphism inschizophrenia: working memory, frontallobe MRI morphology and frontal cere-bral blood flow. Mol Psychiatry , 10, 229,287–229, 298.

Kirley, A., Hawi, Z., Daly, G., McCarron, M.,Mullins, C., Millar, N., et al. (2002).Dopaminergic system genes in ADHD:toward a biological hypothesis. Neu-ropsychopharmacology , 27 , 607–619.

Kordon, A., & Kahl, K. G. (2004). Attention-deficit/hyperactivity disorder (ADHS) inadulthood. Psychother Psychosom MedPsychol , 54, 124–136.

Lachman, H. M., Papolos, D. F., Saito,T., Yu, Y. M., Szumlanski, C. L., &Weinshilboum, R. M. (1996). Hu-man catechol-O-methyltransferase phar-macogenetics: description of a func-tional polymorphism and its potential ap-plication to neuropsychiatric disorders.Pharmacogenetics, 6, 243–250.

Leth-Steensen, C., Elbaz, Z. K., & Douglas,V. I. (2000). Mean response times,variability, and skew in the responding ofadhd children: a response time distribu-tional approach. Acta Psychol (Amst),104, 167–190.

Lotta, T., Vidgren, J., Tilgmann, C., Ulma-nen, I., Melén, K., Julkunen, I., etal. (1995). Kinetics of human solu-ble and membrane-bound catechol O-methyltransferase: a revised mechanismand description of the thermolabile vari-



ant of the enzyme. Biochemistry , 34,4202–4210.

Malhotra, A. K., Kestler, L. J., Mazzanti, C.,Bates, J. A., Goldberg, T., & Goldman, D.(2002). A functional polymorphism in theCOMT gene and performance on a testof prefrontal cognition. Am J Psychiatry ,159, 652–654.

Manor, I., Kotler, M., Sever, Y., Eisenberg, J.,Cohen, H., Ebstein, R. P., et al. (2000).Failure to replicate an association be-tween the catechol-O-methyltransferasepolymorphism and attention deficit hy-peractivity disorder in a second, indepen-dently recruited Israeli cohort. Am J MedGenet , 96, 858–860.

Mattay, V. S., Goldberg, T. E., Fera, F., Hariri,A. R., Tessitore, A., Egan, M. F., et al.(2003). Catechol O-methyltransferaseval158-met genotype and individual vari-ation in the brain response to am-phetamine. Proc Natl Acad Sci U S A,100, 6186–6191.

Meyer-Lindenberg, A., Kohn, P. D., Kolachana,B., Kippenhan, S., McInerney-Leo, A.,Nussbaum, R., et al. (2005). Mid-brain dopamine and prefrontal function inhumans: interaction and modulation byCOMT genotype. Nat Neurosci , 8, 594–596.

Mills, S., Langley, K., Bree, M. V. den,Street, E., Turic, D., Owen, M. J., etal. (2004). No evidence of associationbetween Catechol-O-Methyltransferase(COMT) Val158Met genotype and per-formance on neuropsychological tasksin children with ADHD: a case-controlstudy. BMC Psychiatry , 4, 15.

Molen, M. W. van der, Somsen, R. J., & Jen-

nings, J. R. (1996). Does the heart knowwhat the ears hear? a heart rate analy-sis of auditory selective attention. Psy-chophysiology , 33, 547–554.

Payton, A., Holmes, J., Barrett, J. H., Hever,T., Fitzpatrick, H., Trumper, A. L., etal. (2001). Examining for associationbetween candidate gene polymorphismsin the dopamine pathway and attention-deficit hyperactivity disorder: a family-based study. Am J Med Genet , 105,464–470.

Pfefferbaum, A., Ford, J. M., Wenegrat, B. G.,Roth, W. T., & Kopell, B. S. (1984).Clinical application of the P3 componentof event-related potentials. I. Normal ag-ing. Electroencephalogr Clin Neurophys-iol , 59, 85–103.

Qian, Q., Wang, Y., Zhou, R., Li, J., Wang, B.,Glatt, S., et al. (2003). Family-basedand case-control association studies ofcatechol-O-methyltransferase in atten-tion deficit hyperactivity disorder suggestgenetic sexual dimorphism. Am J MedGenet B Neuropsychiatr Genet , 118B,103–109.

Retz-Junginger, P., Retz, W., Blocher, D., Wei-jers, H. G., Trott, G. E., Wender, P. H.,et al. (2002). [Wender Utah rating scale.The short-version for the assessment ofthe attention-deficit hyperactivity disor-der in adults]. Nervenarzt , 73, 830–838.

Reuter, M., & Hennig, J. (2005). As-sociation of the functional catechol-O-methyltransferase VAL158MET polymor-phism with the personality trait of ex-traversion. Neuroreport , 16, 1135–1138.

Seamans, J. K., & Yang, C. R. (2004). The



principal features and mechanisms ofdopamine modulation in the prefrontalcortex. Prog Neurobiol , 74, 1–58.

Sobanski, E. (2006). Psychiatric comorbidity inadults with attention-deficit/hyperactivitydisorder (ADHD). Eur Arch PsychiatryClin Neurosci , 256 Suppl 1, i26–i31.

Sonuga-Barke, E. J. S. (2002). Psychologicalheterogeneity in AD/HD–a dual pathwaymodel of behaviour and cognition. BehavBrain Res, 130, 29–36.

Swanson, J. M., Kinsbourne, M., Nigg, J.,Lanphear, B., Stefanatos, G. A., Volkow,N., et al. (2007). Etiologic subtypesof attention-deficit/hyperactivity disorder:brain imaging, molecular genetic and en-vironmental factors and the dopaminehypothesis. Neuropsychol Rev , 17 , 39–59.

Taerk, E., Grizenko, N., Amor, L. B., Lageix,P., Mbekou, V., Deguzman, R., et al.(2004). Catechol-O-methyltransferase(COMT) Val108/158 Met polymorphismdoes not modulate executive function inchildren with ADHD. BMC Med Genet ,5, 30.

Turic, D., Williams, H., Langley, K., Owen, M.,Thapar, A., & O’Donovan, M. C. (2005).A family based study of catechol-O-methyltransferase (COMT) and attentiondeficit hyperactivity disorder (ADHD). AmJ Med Genet B Neuropsychiatr Genet ,133B, 64–67.

Vandenbergh, D. J., Rodriguez, L. A.,Miller, I. T., Uhl, G. R., & Lachman,H. M. (1997). High-activity catechol-O-methyltransferase allele is more preva-lent in polysubstance abusers. Am J MedGenet , 74, 439–442.

Winterer, G., Coppola, R., Goldberg, T. E.,Egan, M. F., Jones, D. W., Sanchez,C. E., et al. (2004). Prefrontal broadbandnoise, working memory, and genetic riskfor schizophrenia. Am J Psychiatry , 161,490–500.

Wittchen, H.-U., Wunderlich, U., Gruschwitz,S., & Zaudig, M. (1997). SKID I.Strukturiertes Klinisches Interview fürDSM-IV. Achse I: Psychische Störun-gen. Interviewheft und Beurteilungsheft.Eine deutschsprachige, erweiterte Bear-beitung der amerikanischen Originalver-sion des SCID I. Göttingen: Hogrefe.


Ankündigungen – Announcements

Ankündigungen – Announcements

• ISBET Meeting

The annual meeting of the International Society for Brain Electromagnetic Topography (IS-BET) will take place in Kyoto, Japan from September 29 to October 2, 2009.

Information and Registration at: http://bfe.kuee.kyoto-u.ac.jp/ISBET2009/index.html

• Deutsches EEG/EP Mapping Meeting (DMM)

Das 18. Deutsche EEG/EP Mapping Meeting findet vom 16. bis 18. Oktober 2009 inSchloss Rauischholzhausen statt.

Angenommen werden freie Beiträge zu Themen der hirnelektrischen Aktivität des Men-schen in Form von Vorträgen oder Postern.

Schwerpunkte

– Symposium ”Elektrophysiologie kognitiver Veränderungen im Alter ”, organisiert vonN. Wild-Wall, Dortmund (Sprecher: B. Kopp, Braunschweig, R. Mager, Basel, M.Pietschmann, Berlin, J. Zöllig, Zürich, u.a.).

– ”ADHD Neurofeedback Research Network Meeting”, organized by D. Brandeis,Zürich/Mannheim and H. Heinrich, Erlangen/München.

Information und Anmeldung unter: http://www.med.uni-giessen.de/physio/


kognitive neurophysiologie des menschen human cognitive ... · illustrations: all ﬁgures should...

Documents