You are on page 1of 11

Biological Psychology 75 (2007) 154–164

www.elsevier.com/locate/biopsycho

Common and specific genetic influences on EEG power
bands delta, theta, alpha, and beta
Brendan P Zietsch a,b,*, Jonathan L Hansen a,c, Narelle K Hansell a, Gina M Geffen d,
Nicholas G Martin a, Margaret J Wright a
a
Queensland Institute of Medical Research, Brisbane 4029, Australia
b
School of Psychology, University of Queensland, Brisbane, Australia
c
School of Medicine, University of Queensland, Brisbane, Australia
d
Cognitive Psychophysiology Laboratory, School of Psychology, University of Queensland, Brisbane, Australia
Received 11 September 2006; accepted 18 January 2007
Available online 25 January 2007

Abstract
It is difficult to study the genetic basis of psychological function/dysfunction due to its etiological complexity. Instead, we studied a biological
marker, EEG power, which is associated with various psychological phenotypes and is closer to gene function. Previous studies have consistently
demonstrated high heritability of EEG band power, but less is known about how common or specific genes influence each power band. For 519
adolescent twin pairs, spectral powers were calculated for delta, theta, alpha, and beta bands at bilateral occipital and frontal sites. All four bands
were entered into a multivariate genetic model, with occipital and frontal sites modelled separately. Variance was decomposed into additive (A) and
dominant (D) genetic factors, and common (C) and unique (E) environmental factors. Band heritabilities were higher at occipital (0.75–0.86) than
frontal sites (0.46–0.80). Both common and specific genetic factors influenced the bands, with common genetic and specific genetic factors having
more influence in the occipital and frontal regions, respectively. Non-additive genetic effects on beta power and a common environment effect on
delta, theta, and alpha powers were observed in the frontal region.
# 2007 Elsevier B.V. All rights reserved.

Keywords: EEG power; Genetic; Environmental; Band-specific; Common factor; Dominance

1. Introduction 1991; Salinsky et al., 1991)). As noted below, at the phenotypic
level it is related to psychiatric disorder (Knott et al., 2001;
The importance of genetic influences has been demonstrated Lazzaro et al., 1998; Pogarell et al., 2006; Porjesz et al., 1998),
for many psychological traits and disorders (Bouchard and as well as personality (Tran et al., 2006) and cognition (IQ, g)
Loehlin, 2001; Deary et al., 2006; Ehringer et al., 2006; (Giannitrapani, 1985; Schmid et al., 2002).
Leonardo and Hen, 2006; Preuss et al., 2004). To better One of the most common and straightforward ways of
understand these phenotypes, it is often advantageous to look at quantifying EEG is to divide the frequency spectrum into
a more basic, underlying, preferably biological trait (or discrete ranges (bands), and with a transformation of the data,
‘endophenotype’), as this will more directly reflect the determine the amplitude or ‘power’ of each range. EEG band
influence of the genome (Gottesman and Shields, 1972; power at a given site reflects the circuitry and function of the
Gottesman and Gould, 2003). The electroencephalogram underlying pyramidal cells (Schaul, 1998). It varies greatly
(EEG) provides a good endophenotype because it is largely between individuals (Vogel, 2000), is stable within an
genetically controlled (van Beijsterveldt and van Baal, 2002; individual in a given condition (e.g. at rest with eyes closed)
Vogel, 2000) and is stable over time (test-retest reliability is (Williams et al., 2005), and changes according to mental state
around 0.8 for both absolute and relative powers (Pollock et al., (Moretti et al., 2004), task demands (Klimesch, 1999), and age
(Li et al., 1996; McEvoy et al., 2001). Dividing the frequency
spectrum into delta, theta, alpha, and beta waves can be done
* Corresponding author. Tel.: +61 07 38453584; fax: +61 07 33620101. using fixed bands (e.g. the classical broad bands delta (up to
E-mail address: Brendan.Zietsch@qimr.edu.au (B.P. Zietsch). 4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–25 Hz)) or
0301-0511/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.biopsycho.2007.01.004

/ Biological Psychology 75 (2007) 154–164 155 individually aligned bands (ranges for each band are set relative influenced alpha at all brain regions. but overall. 2006). and the disadvantage of excessive data loss with the aligned method whether genes controlling delta. 1999. and this has been aimed to expand on the EEG literature by studying the linked more specifically to a decrease in attention (Mann et al. These abnormal patterns of resting EEG activity in a particular band ranged from . A multivariate analysis by Anokhin et al..89. specific regions of the brain. Other results of the study are Resting EEG power in specific bands has been associated derived from univariate analyses. B... the correlations were based on data from one Increased resting beta power has also been linked to depression central electrode.. For example. alpha 0. 2005.. Coan and Allen. ADHD has been linked to abnormally was low. Alpha and beta correlated negatively with neuroticism heterogeneity (the sample consisted of 213 twin pairs). it suggested both predisposition to becoming an alcoholic has been associated common and specific genetic factors contributed to the power with elevated resting beta power (Rangaswamy et al. resting alpha power for comparison to the current study: the heritabilities of all four was found to be positively associated with IQ (Doppelmayr bands were very high and similar to each other (delta 0. correlations were neither zero nor unity. and beta powers. whereas frontal lobes perform higher power. 2001). sex. and (2) investigate to what extent the same genes of bands. rather than a general change across the entire influence all four bands. and while this does not allow us to comprehen- basis of psychological function and dysfunction. and that there were no to an individual’s peak alpha frequency) (Klimesch. (2006) found that delta apparent trend of homogeneity of genetic influence across brain and theta powers were positively correlated with extraversion in regions.76. then its sively assess variation across brain areas. Zietsch et al. spheres. In doing this we also aimed to with obsessions and negatively with compulsions (Pogarell disentangle heterogeneity from homogeneity across hemi- et al.86). hemispheric differences in the genetic control of delta. by van Beijsterveldt et al... there were no mean differences or found (Gaser et al. (2005) calculated bi-variate genetic Various forms of psychopathology are associated with correlations between delta. Specifically. conversely. theta et al. and beta differ between outweighs any advantage of using individualised data. Smit et al. Lazzaro et al.55 to . From may be contrasted with the lower versus higher order occipito- the analysis it could be concluded that the same genes frontal differences. (2001) resting alpha power (Bruder et al. As the genetic (see Porjesz et al. and. determining the frontally increased resting delta but decreased alpha and beta extent to which they are influenced by a common factor versus powers. bands. . A more elaborate study. Tran et al. from homologous left and right sites. Many studies have determined the cognitive functions such as executive control. in males. modelling to the classical broad bands delta. For example. It is not known whether the resting EEG band powers. alpha. 2006). theta. 2004). whether the same genes influenced the different bands. A broadly attribute brain activity to posterior versus anterior general genetic factor influencing all frequencies uniformly regions. the present study (Lazzaro et al. and age. also suggested that common as well as specific genetic factors 2004). EEG is an important endophenotype for alcoholism genetic variance overlapped between bands. 1998). 2002). and their EEG band powers tend to correlate positively band-specific factors. and genetic influences on one band (usually alpha. (2005) have shown that for the purposes of genetic alpha. This suggests that if resting EEG power specific genetic effects are. and there was no personality have been found to be associated with specific effect of common environment. but beta power was not included and statistical power (Blackhart et al. for example. alpha. and beta of resting EEG. 2006). these brain regions are dramatically would provide few clues to the genetic basis of a different: the occipital lobes perform mainly basic functions in psychopathology which involved only an increase in beta visual perception. theta. although not in posterior sites. the main objectives were to The summarised evidence indicates that mental disorder (1) confirm that each EEG power band is mainly influenced by and individual differences in cognition are usually associated genetic factors and ascertain the mode of transmission of this with unusually high or low power in a specific band or group influence. and sexes is a reflection of the true nature males and negatively correlated with conscientiousness in of the genetic architecture or simply a lack of power to detect females. theta. Functionally. We applied multivariate structural equation each band. beta 0. multivariate architecture of the genetic and environmental 1992).P. indicating that 55% to 75% of the or bands. but they provide a good basis with various cognitive traits. particularly in the frontal lobes Using the largest twin sample to date.. high resting theta band activity. Smit et al. planning. power bands. However... A study by Smit et al. These functional differences may be reflected by Beijsterveldt and van Baal (2002) for a review). (2005) for a review).. although a significant relationship is not always 0. used multivariate modelling to determine to what extent the Thus. see van reasoning. the far more subtle hemispheric functional differences same genes affected EEG in different parts of the brain. or beta power.89. (1996). It left unanswered the questions of analysis there is little difference between the two. how important band- frequency spectrum. it allows us to more genetic influences should exhibit some band-specificity. Aspects of heritability differences between the sexes.75. so comparisons could not be made between (Knott et al. but these do differences in the EEG signal and its underlying genetic not reveal whether common or specific genes are influencing influences. theta. 1999). as has frontal hemispheric asymmetry in brain regions. Sufferers of obsessive-compulsive disorder exhibit influences on the four EEG power bands. We focussed on occipital and is to provide a good endophenotype for studying the genetic frontal sites. 1983. Recent findings have suggested that frontal resting EEG contributed to EEG power bands and event related potentials asymmetry may also predict future development of anxiety (ERPs).

identical) in Mx. Eye movement artefacts were removed from each analysis to reduce any bias on significance testing. indicated a striking concordance between the analogous genetic (A) path loadings in the O1 and O2 models. and epochs with checked whether swapping the order of model reduction steps made a sub- abnormal EEG patterns (>25% of amplitude at 0 mV) or extreme voltages stantive difference.4. between O1 and F3. For first with eyes closed and second with eyes open. to the point where the left and right hemisphere models appeared essentially equivalent. which uses a standardised Mahalanobis distance to compute a z-score for female pairs (MZF). 2006). P3. Which factor to retain was decided independently for each EEG band Wright et al. skin.2. As one twin did the psychometric session. then D was modelled. a history of head injury. 2001a). 250 mV for all other channels) were excluded. 0. outliers. Fp2. and a combination of measurement error and unique environ- mental effects not shared by co-twins (E). Participants Hanning window was applied to the first and last 5% of each epoch to prevent spectral leaking. and likewise in the F3 and F4 2. F4. EEG was recorded from 15 scalp loadings in frontal and occipital models. Prior to testing. Fp2. We tested the electrode (10–20) placement system and referenced to physically linked ears. Values outside the range of 3 to +3 indicated extreme families identical) female pairs (DZF). otherwise C was modelled. zygosity was assigned with an extremely low Multivariate genetic modelling partitioned the total variance into that due to probability of error (less than 104). to sit quietly and be relaxed. For the eyes closed condition frontal sites (F3. and for DZ twins. O2) a four-variable Cholesky model was specified in complementary session.1.0A) plus non-additive (1. 1. F4) a four-variable Cholesky model was specified in which participants were informed that the duration of the recording would be ACE components of variance contributed to delta. The equating all analogous left/right loadings (e.3.156 B. and includes six zygosity groups. and Fp2 were amplified could be dropped or reduced before reducing the A component further (if C with a factor 5K and all other channels with a factor 20K by Grass preamplifiers were dropped after fully reducing the A component it would be a less (model P511K). factors A2. and EOG. decimal places were used for analysis to enhance the possibility of finding any data were checked for normality and screened for univariate and multivariate genuine age effects. secondary schools as part of a study on the genetics of cognition (Wright et al. Cz. The DC component was removed from each epoch and a 2. so that only one can be estimated for a given variable in a given (ERPs) were recorded during a working memory task (Hansell et al. (2006). excluded so that results may be generalizable. 77 male first born (DZMF)). EEG data were converted to a frequency distribution Participants were adolescent twins recruited through South East Queensland ranging from 1 to 30 Hz (resolution 0. Prior to genetic modelling. we combined O1 and O2 into an octovariate Cholesky model (and electrode cap. A6. However. . 2001). for residual hemisphere-specific genetic variance (since all analogous left/ Software used for the recording determined that the maximum length of right loadings were equated as described above). each family. 145 opposite sex DZ (taking into account MZ and DZ similarities) in assumption testing analyses on pairs (68 female first born (DZFM). the which ACE components of variance were modelled for all four bands. F7. theta. Finally.5A + 0. written informed consent was obtained from all participants and their parents or guardians.01 to 30 Hz conservative test of C because variance is more constrained). and meter estimates are biased by outlying families (Tabachnick and Fidell. Band power was computed as the sum of the power frequency bins within each band. e. as maximum-likelihood para- MN.42 and 18. and then reduced the eyes open and 20 with eyes closed. the A1 loadings on delta left ground lead was located just anterior to the Fz electrode.16 (16. to test (6dB per octave) and a 50 Hz notch filter.g.0D) genetic factors plus Two parallel testing sessions were used in the Twin Cognition study common environment (1. psychophysiological session. we subsequently epoch using a dynamic regression algorithm (Molenaar. A4. and to focus on the monitor in front of them. Ethics approval for the study was obtained from the 2. at each of the frontal and occipital sites taken in a semi-darkened. 2001) and a psychophysiological session where event-related potentials confounded. F3. alpha 8–13 Hz. such that delta included frequencies ranging from 1. Twin pairs The frequency distribution was divided into the four standard broad bands were excluded from participation if parental report indicated that either twin had using fixed bands.3.P. The full sample consisted of 533 females and 505 males between the ages of Data were analysed using maximum likelihood (ML) estimation procedures 15. O1. C3. In the protocol: a psychometric assessment of processing speed and IQ (e. / Biological Psychology 75 (2007) 154–164 2. Ocular potentials and delta right) at once and comparing model fit. so on the outer canthus and the centre of the supraorbital ridge above the left eye. substance abuse or theta 4–8 Hz. 125 monozygotic (MZ. and ADE approximately 5 min. genetic factors – additive (A) and non-additive (D. dominance. theta. Statistical analysis Human Research Ethics Committee. Only Preliminary multivariate model fitting including all four EEG power bands data from the eyes closed condition are used in this study. electrically shielded.31). 1987).g. locations (Fp1.. Twenty 12-second blocks were recorded with tested whether factors A3. dependence. represented schematically in Fig. 2001. Next. To generate power spectra EEG data were model further by testing which individual hemispheric specific loadings had to divided into sixty 4-second epochs per condition using EEG analysis software be retained. epistasis) – and that due to non-genetic factors . Extreme families were AmpFLSTR1 Profiler1 PCR Amplification Kit and crosschecking with ABO.25D + 1. the co-twin undertook the than half the MZ correlation. F8. and were asked to relax and sit quietly with their eyes components of variance contributed to beta. their equality was not tested. to minimize any movement. non. and beta). Recordings were (delta. General procedure co-twins (C). of which the recording of resting EEG was a component.. 2001b) followed by the recording of resting EEG (Smit et al. For MZ twins the co-variance was defined as additive (1. Two 4 min recordings comprised the resting EEG. 114 MZ male pairs (MZM). alpha.g.5–4 Hz. C4. There was no such concordance between genetic path As described in Smit et al. separately. A5. Due to the narrow age range. but as a conservative test of C we tested whether it Impedance readings were all below 5 kV. or current use of medication with central nervous system effects. and A8 continuously recorded EEG was 12 s with a discontinuity of 2 s between were dropped. and it did not. Sex and age were entered into the closed.24  0. EOG. The tin electrodes were arranged according to the International similarly for F3 and F4). and Rh blood groups and/or phenotypic information (e. and alpha. 1999). Outlying families were detected and excluded by using the %P option (2006). hair. Recordings were filtered with a band pass filter of 0. and A7 could be dropped. EEG recording models. Zygosity was each power band at each site (univariate outliers) and for the multivariate determined by typing 9 independent polymorphic DNA markers using the models prior to model reduction (multivariate outliers).0C). For the frontal model where these could not be dropped we successive 12-second blocks.common environmental influence shared by 2. Luciano absence of data from separated twins or half siblings D and C are negatively et al. model.. Zietsch et al. equality of left and right hemisphere genetic loadings for each band by with the ear impedances matched at the beginning of the recording session. and beta 13–25 Hz. Queensland Institute of Medical Research. ages to two with the statistical package Mx (Neale et al.g. for example..25 Hz) using Fast Fourier Transformation (Niedermeyer. The ordering of model reduction steps was decided prior to EPTOR 1. This sample is almost identical to that reported in Smit et al. P4. and after a short break each twin completed the For occipital sites (O1. For the eyes open condition they were asked multivariate models as covariates. (Groot. neurological or psychiatric illness. Pz. O2) using an Therefore.0C. Methods (exceeding 1000 mV for Fp1.3.. Fz. Fp1. and sound-attenuated cubicle. Based on this. based on the ratio of MZ and DZ correlations – if the DZ correlation was less 2006).. 69 dizygotic (DZ. 1999). Common environment (C) (electro-oculogram or EOG) were recorded from single tin electrodes located loadings displayed no pattern suggesting left/right hemisphere equivalence. eye colour). 66 DZ male pairs (DZM).

C7 replaced by D1. for frontal N = 980. For clarity. and all were such that EEG power factor analyses of the genetic correlation matrices from the occipital and frontal decreased with age.86 (delta F3–F4) to .g. non-additive genetic var- iance (D) was modelled for beta (left and right) instead of common environment at both O1 ( p = .58–826.47 F3 27. Twenty-one Phenotypic correlations between left and right hemisphere additional pairs were identified as multivariate outliers for O1– (O1–O2.06–281.55 (16.68) 0. Preliminary analyses areas.07 F4 27. so a model reflecting this was tested. These variances were left free contain a right hemisphere specific loading.0007) and O2 ( p = . and 490 pairs for F3–F4. Distributions of the power bands at each brain region correlations/covariances were significantly higher than DZ twin were positively skewed. or between same sex and opposite sex DZ twin groups for delta.75) 7. The exceptions were for beta at the variance of EEG power bands delta. but for all other power bands the on transformed data unless otherwise specified. Additive and variances (using an a-level of .76 (33. For beta power transformation resulted in all values being defined and variables the DZ twin correlations at frontal (left and right) sites were less being less skewed and kurtotic. Note for the frontal model.99 Alpha O1 114.79) 3. but most effects were small. measures.21 (45.97 F4 7. alpha 3.60–90.68–139. where the variance of males and females differed hemispheres (squares). In addition. to vary between sexes for subsequent analyses.00–152. twin pairs did not differ significantly according to sex. Adding one in this sites (Dx21 ¼ 15:70 to 114:74. independent pathway models strong tendency for females to have higher beta power at the were tested. These models were compared to each other and the final reduced Twin correlations are presented in Table 2.0001).e.55 (3.e.92) 4. unique environmental (E) influences are not shown. 1.52 (40. Phenotypic correlations between occipital and frontal .02 (14.91 F3 45. and O1 and O2 for the occipital model).94 F4 19. and A8 being more variable than males. p < .24) 3. i. A6. and beta in the left and right occipital sites. Smit et al.79 O2 152. and thus log transformed (lg 10(x + 1)) correlations in all four bands at left and right frontal and occipital to improve the normality of the distributions.64) 3. theta.16 O2 14. the A2 loading on delta right.P. Other models were based on varimax rotations of occipital alpha power. All further analyses are based than half the MZ correlations. All age effects were significant except for factors loading on each band. alpha. with outliers excluded) Mean (S. and C8 replaced by D2). O1–O2).19 F3 20.94 (18.85–31.20) 1.87–77. Schematic representation of the octovariate base Cholesky model for the occipital and frontal models.95 (theta F3–F4 and theta 483 pairs for O1–O2. consisting of a general factor influencing all bands and four specific size and direction.65 F4 42.87) 5.01) according to birth order.67 O2 63.78 Beta O1 10.02 Theta O1 36.01 (12.00008). The wider confidence intervals for the same sex DZ groups than the Mean EEG power and standard deviations for each of the opposite sex DZ group was due to the smaller sample size (i.85 (136.40–1055. The composition of the final sample was ranging from . e. theta. Also. A2. Left and right hemisphere sites are included in a single model (F3 and Homogeneity of sampling was found with equality of means F4 for the frontal model. and overlapped.50 (11.58 (32.88 O2 47.48–84. The exception was a As well as by reducing the Cholesky model. with females (C). and 15 for F3–F4. 66 four frequency bands at occipital and frontal sites are presented and 69 pairs compared with 145 pairs. B.1.57) 2.20) 0. respectively).29–45.51–134. MZ twin in Table 1.40) 2.49–371. Results or beta at any site.) mV2 Range Delta O1 48.86–34.97) 8. Fig. Fifteen twin DZ correlation was half or greater than half of the MZ pairs were identified as univariate outliers at the assumptions correlation. (2005) suggested that common as well as specific occipital sites. A4. testing stage for O1–O2. 95% confidence intervals were wide.30 (15.23 (4.25 F3 6. and 14 for F3–F4. but they have fixed effects of sex and age significantly influenced some the same structure as A and C. / Biological Psychology 75 (2007) 154–164 157 Table 1 Descriptives for each EEG power band at each electrode site (from raw data.58) 1.13 (6. While there was a trend for DZ male correlations to be greater than DZ females.D. particularly in frontal 3. Covariances of Cholesky models using the Akaike’s Information Criterion (AIC) index.13 (101. Sex effects were significant for around a third of power bands. genetic (‘A’ circles) and common environmental (‘C’ circles) factors explain the zygosity group.57–288.95) 2.06–347. O2. F3–F4) within each frequency band were very high. and sex. models. Zietsch et al.70–133.15) 4. Age effects were stronger and more consistent in additive genetic factors contributed to each band. from which subsequent reduced models were derived.06 (8.27) 2.10 For occipital sites N = 966.

66 0.85–.88 (theta–delta at O1).40–.73–.25–.70) 0.85) 0.68) 0.64) 0.73 frontal models.72) 0.42 (.87) 0.36–.36–.43) N = maximum number of twin pairs for band and electrode. indicating a significant influence of common (alpha O1–F3.62) 0.81–.70–.75) 0. d. F4) sites MZF (N = 125) MZM (N = 114) DZF (N = 69) DZM (N = 66) DZOS (N = 145) MZ (N = 239) DZ (N = 280) Delta O1 0.56 AIC = 0.82 (. Additive effect on beta.23–.57) O2 0. In the occipital region.55 (. and one sizable factor loading on the left and right compared with frontal sites. .6.2.50) O2 0.31–.88 0.50 0. AIC = 151.44 (.76 d.50 and specific’ model (2LL = 19619.40–.68) 0. non-additive genetic effects unique environmental influences.45) 0.56) 0.59 (. especially in the case of delta– hemisphere of delta.80) 0. Sequential dropping of C factors (from lowest to phenotypic correlations between bands (Table 3) were highest) indicated one very small factor loading on the left- generally around 0.71–. and beta at left and right occipital (O1.56 (.32 (.50) Theta O1 0.56) 0.27–.703. making them suitable for multivariate hemisphere of all four bands in the occipital model (models 4a analysis.29–.72) 0.73 0.20–.77 (. AIC = 188.f.75) 0.86 (.78–. magnitude.53 (.30–.45–.56) 0.85) 0.71) 0.33 (.81) 0.59) 0.23 (. Theta 0.47) 0.83) 0.35 (.81 (.72–.90.54) F3 0.29–.77–.20–.72–.30–.04–.74 (.71–. theta.45 (.56) 0.61 (.49 (.80–. d.39 (.31–.40–.43) 0. Correlations were generally higher at occipital to 4g).44 (.76–.12.92) 0.80) 0.76–.80 (.70–.48) 0.28 (.50) 0.f.23–.75–.88) 0.f.40–.84) 0. = 7754.78–.66 0.01–.56) 0.49 (.12–. In addition. Zietsch et al.82) 0. (2LL = 19574.50) F4 0.68 (.83) 0.70 (.69 (.63–. Delta Theta Alpha Beta Delta Theta Alpha Beta AIC = 81. ranging from 0.43) 0.38–.f. O2) and frontal (F3.59 d.79–. Corresponding genetic structure.83 (.158 B. and theta O2–F4).47 (.84) 0.61) 0.82) 0. Correlations for right hemisphere are fit poorly compared with the final reduced Cholesky all within . the frontal models. and.40.54 (delta O1–F3) to 0.86 (.38–.84) 0.85 0.72–.43) F4 0.65) 0.87) 0.32–. alpha.91) 0.69–.57 (.57) 0.42 (. alpha.44–.41 (.22 (. equated without significant loss of fit in both occipital and The alternative independent pathway models that were frontal models (model 2). AIC = 10.78 (.60 (.32 (.67–. Alpha 0.60 (.f. and a non-additive regions are presented in Tables 4 and 5.74–. the environment.28 (. d.65) 0.57) Alpha O1 0.72 (.53) 0.54) 0.73a (.31–. are left out of the figures and presented in Table 6 loadings on the left and right hemisphere of each band could be instead.27 (.58) 0.87 (.83–.41–.07–. theta.85) 0.83 (.09–.43–.31–.54) 0.56 (.66) 0.78 (.57) F3 0.30 (.44 (.78–.70) 0.83–.23–.41 (.02–.83) 0.48) 0. bands Independent assessment of each hemisphere-specific loading in the frontal model (models 7a to 7e) revealed that two were Results of genetic modelling for the occipital and frontal significant: an additive effect on alpha.27) and the Maximum likelihood phenotypic (upper triangle) and genetic (lower triangle) two factor varimax-based model (2LL = 24522.62 0.29–.89) 0.77) 0.79 (.67 0.56 (.62–. but left to vary for subsequent analyses.86) 0.46 (.37 (.68) 0.76 (.58 0.71) 0.89) 0.87) 0.90) 0.66) 0.53) 0.86 (.28–. and beta at occipital and frontal sitesa d.86 (.24–.49 (.45–.50) 0.41 (.83 (.57) 0.82) 0. being higher at occipital compared with frontal. Beta 0. respectively. theta.49 0.70–.19–.83) 0.03 of values given.78 (.80 (.90) 0.797).51 (.58) 0.85) which were overall the highest in (models 4a to 4d). alpha O2–F4.42 (.44 (. The factors containing these ranged from .61–.39) 0.87) 0.09–.25 (.71) 0.48) 0.56 (. AIC = 173. and alpha in the frontal model theta correlations (0.33.78a (. indicating that hemisphere- 3. The final best-fitting occipital and frontal genetic influences played a large role in both occipital and models are shown in Figs. d. For clarity. in the frontal model the ‘general Delta 0.90) factor varimax-based models all a Values given are for left hemisphere.74–.49 (.73 0.42) 0.44 (. = 7861. sites were lower.78) 0. a Variance equated between zygosities to estimate twin correlation. = 7854.79 (.78 (.57–.65) 0.83–.82) 0.68–.42 (.10–42) 0.89) 0.89 (.86 (.86 (.78–.33 (. Similarly.59 (.41a (.83) 0.45–.35 (.89) 0.42 (.48) 0.85–.32–.27–.39a (.21) and three (2LL = 19801.61 (.10–. = 7753.75 0.58 0.21–.44 (.66) both fit poorly compared with the Occipital Frontal final reduced Cholesky (2LL = 24249.45) 0. and small. AIC = 37.32–.31 (. correlations for delta. Genetic correlations (Table 3) showed a similar Residual hemisphere-specific genetic loadings were very pattern.73–. = 7862.86 (.83) 0.73–.43) 0.59 0.13–48) 0.75 (.87) 0.52) Beta O1 0.34–.45–.40) F4 0.72–.16.15–.52 (.37 (.41–.78) 0.69) 0.23.30–. = 7858.81) 0.69) 0.18–.39–. Variance components modelling of EEG power in four specific effects were only important to the frontal model.31 (. Dropping the entire C component tested did not provide good fits to the data compared to the final (model 3) led to a significant loss of fit for both occipital and reduced Cholesky models.11–.40–. the largest being 0.22–.73–.60) and both the two (2LL = 19777.48 (.29–.80–.49 (.86) 0.86) 0.28 (. 2 and 3.P.51 (. additionally.30 (.42 (.83 (.57 0. / Biological Psychology 75 (2007) 154–164 Table 2 Twin correlations (95% confidence intervals in parentheses) for each zygosity group for delta.51) O2 0.29–.64 0.82) 0.78 (. respectively.48 (alpha–delta at F4) to .61 (.30–.f.46 (. loadings could be dropped in the occipital model (model 5) but not in the frontal model (model 5).76 0. left in a full Cholesky had a large influence on frontal beta.81) F4 0.82–.50–.77–.83–.47–.89 (.82–.82 (.77 (.25–.78 (.85 (.81–.68) 0.47 (.21.89) 0.88) 0.77 (.83 (.63 0.81 (.57) O2 0.48) F3 0.48) F3 0.80 (.60 (.31–.72–.29–.83 (.67 0.67).f. the model with one general genetic factor and four band specific genetic Table 3 factors (2LL = 24317.79 (.66–.64 0. = 7751.55 (.

61 17 0. the additive (A) and non- additive (D) genetic loadings on beta could not be assessed independently. In contrast.06 7853 13.28 7852 0.14 6 0. Loadings from this common variance.91 7743 2. oblique arrows in Figs.01 7813 2 Equate left and right hemisphere genetic loadings 1 19556. i. theta 0.64 for theta. being 0.51 16 0. frontal: 0. alpha.001 5 Drop A2.60 7841 0.827 7e Drop D2 loading on beta right** 7a 19588.14 7845 4. and A8 each contain a hemisphere-specific loading (right hemisphere loading without the corresponding left hemisphere loading for that band).059 7a Drop A8 loading on beta right** 6 19574.41 7856 41. the ‘alpha’ factor (A5. A6. D2LL Dd.001 4f Drop right hemisphere C1 loadings 4d 24233.993 3 Drop C (omnibus drop) 2 24272. the ‘theta’ factor (A3 in in both occipital and frontal regions common genes influence occipital.e. b C loadings of left and right hemisphere could not be equated (D2LL = 23. A4. specifics (i.27 11 <0. delta left.50 7695 2 Equate additive genetic (A) loadings on left and right hemisphere of each band 1 24221. A2 in frontal) accounted for 19% and 39% of the band delta.32 7852 0.e.f. theta. (43%).f. a factor analysis on the additive genetic Based on the final multivariate models. delta left.064 3 Drop C (omnibus drop)b 2 19591.50 additive and 0.334 4e Drop C1 4d 24272. A8 4c 19603.46 for delta.54 36 0. 2 and 3) on theta.46 1 0.000 4b Drop C2 4a 19557. accounting for relatively ‘beta’ factor (A7.66 7726 0. A4.f. alpha. heritability estimates correlation matrix confirmed that one factor accounted for more for the occipital region were delta 0. A7 4c 19574.54 7747 51.000 8 retain genetic right-hemisphere loadings on alpha (additive) and beta (non-additive) 4c 19574.24 21 0. 2 and 3).85. alpha. and beta.61 7836 0. B. a Bands entered into Cholesky in order of frequency with left hemisphere preceding right.197 9 0.045 4a Drop C8. C4 (all loadings zero) 2 19556.985 4c Drop C3 4b 24223. heritabilities tended to be frontal (46%) model. A4. of the genetic variance in the occipital model (67%) than in the and beta 0. A5.72 6 <0. and beta bands.202 *A2.05 1 0. theta.00 1 1. / Biological Psychology 75 (2007) 154–164 159 Table 4 Goodness-of-fit statistics for multivariate models of occipital EEG bands delta. For the frontal region. alpha right. p-value a 1 Cholesky decomposition: ACE (delta. theta right. A6.920 7d Drop A2 loading on delta right** 6 19574.61 7830 26.00 6 1. one factor).963 4c Drop C3 4b 19562.000 7c Drop A4 loading on theta right** 6 19574.70 7753 15. common environment was much more influential in the frontal Table 5 Goodness-of-fit statistics for multivariate models of frontal EEG bands delta. In the occipital bands (occipital: 0. indicating that and frontal models. The best-fitting model is in bold vs 2LL d.86. theta right. D-2LL Dd. p < .99 5 0. A6) 41% and 10%. alpha right.30 non-additive heritabilities. 0. delta right. and beta. beta left.01 1 0.000 4b Drop C4 4a 24221.00 10 1.e.53 4 0. C5 (all loadings zero) 2 24221.73 7852 17. In the frontal model. a more of the genetic variance. C7. delta right. and the genetic factor are substantially higher.54 7747 40.00 7721 0. A notable feature of both best-fitting models is a common and beta. A4. C5.000 7b Drop A6 loading on alpha right** 6 19591. theta left. = 3. so the non-additve loading was assessed with the additive loading fixed at zero.41–0.35 4 0.e.P.63 4 <0. and 0.80 10 0. even though this excluded the band.87. A4) 39% and 42%.27 7851 12.68).75. For 7e.0001) . i. theta left.027 4a Drop C6. and A8 each contain a hemisphere-specific loading (right hemisphere loading without the corresponding left hemisphere loading for that band).78 1 0. somewhat lower.54 7747 38.339 4d Drop C1b 4c 19591. theta. consisting of lower the percentage of total genetic influence accounted for by 0.56 7739 7.001 6 Drop A3. alpha.91 6 0. alpha). vertical arrows in Figs.00 7711 5.f.33 7854 12. alpha 0. alpha left.59. beta right.672 4g Drop left hemisphere C1 loadings 4f 24272. A6.99 7 0.e. alpha left. a Bands entered into Cholesky in order of frequency with left hemisphere preceding right.27 7852 0.57 7732 1.52–0. p-value a 1 ACE Cholesky decomposition 24215. ** Each hemisphere-specific loading was assessed for significance independently (7a–7d).78. respectively.85 7851 29. beta right.98 8 <0.928 4d Drop C2 4c 24231.85 7851 35.106 * A2. A6. with some small (2–13% of variance) cross-path additive genetic factor (A1) that loads strongly on all four EEG loadings (i. theta. Furthermore.66 5 0. broad heritability was 0. in the occipital model (64% of the specific non-additive factor accounted for a further 30% of the influence from all genetic factors) than in the frontal model variance in beta.001 5 Drop A2. ADE (beta) 19530. Common environmental influences accounted for only 1 to The remaining additive genetic factors were essentially 2% of the total variance in the occipital model. A8* 4f 24249. beta left.73 specific non-additive factors in the latter (including these would for alpha. Dd. Zietsch et al. to reduce any bias of significance testing.80. The best-fitting model is in bold vs 2LL d.f. i. For beta. C6.

Discussion Frontal E1 E2 E3 E4 E5 E6 E7 E8 These results support previous research showing that the Delta left 0.32 by common genetic factors influencing all bands in both Beta right 0. Structural equation model with additive genetic (‘A’ circles) and common environmental (‘C’ circles) factors explaining the variance of EEG power bands delta.06 0.21 hemispheres.18 0.03 0.09 0. / Biological Psychology 75 (2007) 154–164 Fig. brain regions were also shown to be influential: in particular.48 the variance at occipital sites.14 0. Beta left 0.09 0. Path coefficients are standardised such that the squared path coefficient indicates the percentage of variance accounted for.08 0. Delta right 0. Table 6 Parameter estimates for the influence of unique environment (E) on left and model. right and left hemisphere EEG power bands are ing the variance of EEG power bands delta. genetic factors specific to bands and Loadings greater than j0.13 0.g.32 variance in all power bands and brain regions studied. Zietsch et al. be explained Beta left 0.15 0. 3. 2.20 0.00 0.04 0.27 0. 2000). and 4–7% in alpha.37 hemisphere.10 0. but there were no Theta left 0.08 0.17 0.10 0. and left hemispheres of each band were equated. alpha.46 1996.09 0. As additive genetic factor loadings on right not shown in this figure but in Table 6. Also for clarity.11 0.20 0.11 0.17 0.22 0.07 0. and 95% con- Cholesky model.23 0. but as described in the methods were analysed in an eight- variable Cholesky model. . theta. and common environmental (‘C’ circles) factors explain- each band were equated. and beta (squares) in collapsed into one square (e. Alpha right 0.11 0.23 0.05).15 delta having a relatively higher E component in both occipital Alpha left 0.33 0.39 0.25 0.36 Genetic factors accounted for a substantial amount of the Theta right 0.10 0..12 0. 2002.21 Alpha left 0.04j are significant ( p < . the delta square represents delta left and delta right). the unique environmental (E) influences are not shown in this figure but in Table 6. at least in part.04 0. but as described in the methods were analysed in an eight-variable coefficient indicates the percentage of variance accounted for.g.01 0.21 0.21 0. aside from Theta right 0.19 0. As additive genetic factor loadings on right and left hemispheres of genetic (‘D’ circles).22 0. Vogel.25 0. theta. van Beijsterveldt and van Baal.21 0.21 0.22 and frontal. Theta left 0. and beta EEG bands bands. the delta square represents delta left and delta the frontal region. Also for clarity.08 0.64 human EEG is a highly heritable trait (van Beijsterveldt et al.35 Beta right 0.09 0. and 95% confidence intervals are in Fig.25 Alpha right 0. accounting for 13–18% of the variance in delta and theta right delta.160 B.01 0.00 0. alpha.25 striking patterns to the variation within this range. theta.14 Modelling suggested that this may.26 0. Structural equation model with additive genetic (‘A’ circles). and beta (squares) in the occipital region. and 11 to 40% at frontal.11 0.27 0.01 0.31 0.16 0.25 4.13 0.04 0. Path coefficients are standardised such that the squared path right). depending on band and Delta right 0. Occipital E1 E2 E3 E4 E5 E6 E7 E8 Unique environment and error accounted for 11 to 23% of Delta left 0. However. right and left hemisphere EEG power bands are collapsed into one square (e.09 0.17 0.28 0. the unique environmental (E) influences are fidence intervals are in parentheses. alpha.P. dominant parentheses.

. monozygotic twin correlations were more than double EEG in the left and right hemispheres. 2006). which are themselves influenced by independent genetic (DeFelipe and Farinas.. additive genetic (A). is as a potential endophenotype for these conditions.and site-specific genetic frontal sites. 1998). 2005). It also the presence of a highly significant effect of common implies that there may be neural substrates or generators whose environment. From an evolutionary perspective.. These genetic components may be important underlying cortex (Schaul. no other study has found a significant GABAergic.e. This can be (i.P. infants (Orekhova et al. 1982.. as error. and in separate left and right frontal models. which is called epistasis anterior regions than in the posterior regions. The proportion of genetic variance 2005). 1996). while the beta range involves the action (particularly as a substantial portion of E may be explained of GABAA (Whittington et al. decreased with age.. Such an (Lykken. 1998). 1992. multivariate non-additive factor on beta was highly significant (30% of the genetic analysis (and a factor analysis) indicated that one variance).. D is likely to include dominance relationships lying EEG has more band-specificity. In our To a large degree the same genetic factors were influencing findings. and in contrast to previous research. sub- determine the distance from the electrode (on the scalp) to the stantial non-additive genetic effects suggest that the phenotype current generators (underlying pyramidal cells). effects (Lykken. In both occipital and frontal regions. in between genes at different loci. which is negligible in size in the occipital model influence on EEG is specific to certain frequency ranges or but substantial in the frontal model. across the frequency range. 2002). Genetic linkage (gene-finding) studies will only be common factor could account for a large proportion of the able to accurately determine the sample size required for a genetic variance in all bands. 2003). and a further occipito-frontal difference. The results of past (D). Different neural networks (e. 1999). and that frontal EEG The finding of band-specificity in the genetic effect on EEG asymmetry has been shown to have low but significant fits with the research indicating band-specific effects in heritability (Anokhin et al. It may be that the small psychiatric disorders such as alcoholism (Porjesz et al. much more Further genetic complexity present in frontal but absent in structurally complex. As such. That is. accounting for less than 2% of the variance in each band. involves gene effects that research which demonstrates that the neural structure of the combine configurally rather than additively. 1990). the necessary. lateralised genetic effects on alpha comprise genes that also 2005). 2003). Valverde. disorder (Pogarell et al. Nieuwenhuys. 2005). In frontal regions. a more complex genetic when viewed in the context of research suggesting a link influence on frontal compared with occipital EEG would be between frontal EEG alpha asymmetry and predisposition to expected. Lykken et al. 1982). affecting EEG in consideration was subject to intense selection pressure at power over the entire frequency spectrum. thereby enhancing EEG power A striking result. or is more complex. but some more recent research has human EEG.. we refer to occipital and non-additive genetic effect on beta was large and significant. and obsessive/compulsive influence depressive disorders. / Biological Psychology 75 (2007) 154–164 161 non-additive genetic influence was found only for beta power at A striking example of a band.g. EEG is partly a reflection of the pyramidal cell circuitry in the but significant. This general factor may reflect certain level of statistical power if non-additive genetic effects basic structural features such as skull thickness. 2006). depressive disorders (Bruder et al. non-additive genetic which was not present at the occipital sites. failed to detect any significant effect (Smit et al. Another polygenic mechanism that can create interpretation would make sense in light of neuroanatomical a D effect. it some stage (Merila and Sheldon.e. and unique environment (E) research have hinted at genetic dominance in the beta band (van factors were all found to have important influences on the Beijsterveldt et al. B. polygenic. only specifying left or right hemisphere when the more powerful left–right combined final frontal model. dashed arrow from A6) and beta many times more complex and have far greater functional (i.g. D) effect. As EEG power is a complex trait and probably interpreted as indicating that the genetic architecture under. rather than an effect of unique environment).. and have up to 23 times more spines than occipital regions was found in the form of hemisphere-specific those in occipital areas. cholinergic. as such. In this case that could reflect neural properties or processes that are influenced phenotype may be the balance in neural excitation–inhibition by the same set of genes and have a broad effect on power homeostasis that beta power seems to reflect (Porjesz et al. ADHD (Lazzaro et al. This means that neuronal circuits are genetic loadings on alpha (i. Since small. Zietsch et al. common environment (C). an cerebral cortex is very heterogeneous (Elston. but there was not the the dizygotic twin correlation for beta in both left and right same concordance for genetic influence in the occipital and frontal areas. 2000). emergenic trait relies on the configuration of its component Pyramidal cells are the most ubiquitous neurons in the cortex traits. These loadings were very capacity in frontal than occipital areas (Elston.. the slow waves to be accepted that beyond early childhood the environment delta and theta have been associated with cholinergic systems plays little or no role in determining the human EEG (Steriade et al.. Aside from a study on bands. . Alternatively. frontal sites.. dashed arrow from DRH) bands. a common influence on frontal EEG was the large and highly significant environment factor was found to influence the power bands at non-additive genetic effect on the beta band in the frontal lobe. and those in frontal areas are larger.. thalamo-cortical. dopaminergic) may be responsible for electrical influence of common environment on EEG. a frontal lobes of the brain. 2003) which found a C effect that cortico-cortical) or neurotransmitter systems (e. that this one general factor could account for differed markedly There are several possible explanations for the non-additive between occipital (64%) and frontal (43%) regions. It therefore seems rhythms of different frequencies. which would are accounted for. For example. 1992). called emergenesis.e. However. 1994.

Minnix. 545–554. 806–814. 2006... 46–50. Genetic correlation between the P300 event-related brain potential and the EEG power spectrum.I. finding of sex effects on EEG is Bruder... Heath.A. Heine. Bouchard. have lower EEG band power than younger participants.M. alpha.. but other hood and adolescence: a study of twins and their siblings. Y. this study shows that genetic influences are parameters. Selikowitz..J.. S. T. as well as Ann Eldridge and Marlene Grace for collecting environments in visual input.J. 2006. A79801419.A. but they Weissman. Rhee. predict future development of anxiety and depression? A preliminary study. S. so that both D and C cannot be estimated for any manuscript. 1999). detect C than previous studies. Hille. decreased through development.. R. Age and sex effects in the EEG: development of the normal child. V.. However. R.. D may play a hidden Frontier Science Program (grant number: RG0154. with the exception of occipital alpha.A. Allen.C.B.M. mixed. R. A common environment at frontal but not occipital sites.C. Cerebral Cortex 13. 1983. Nomura.. 1992).E. G. P. J. Genetic influences take the form of more lasting and important role.. Can EEG asymmetry patterns global impact).. and may explain why we have Importantly.J. This is Ehringer.C. 2001. beta power). and further work suggests that developmental changes in 1124–1138. observed at the occipital sites (a testing effect would have a Blackhart. more complex circuitry in enriched environments than in impoverished ones (see Renner and Rosenzweig (1987)). European Journal of Human Genetics 14.T. McCarthy. direction of the effect was such that older participants tended to EEG alpha power and intelligence.162 B.. Genes. Genetics of intelligence. M. Kline. Progress range.. 563–607. Young. Stadler. 2001. C. six of the 16 power bands (at left and right occipital and Clarke. 243–273.. which has higher power Coan. bands. I. In hemispheric activity in offspring at risk for depressive disorders. This work was supported by grants from the single variable.. role where C is observed (frontal delta. T.. Behavior used for all twin pairs. Acknowledgements Although the circuitry in the visual (occipital) cortex is also susceptible to environmental influence (Wiesel and Hubel.. Myers. Behavior human brain (Good et al. Biological Psychology 67. Cortex. 2003. Thanks also go to Matt Downey for processing the unsurprising. Correlating EEG are complex (Klimesch. and alpha powers Anokhin. M. T. M... 7–49. E. In all cases the Doppelmayr.. Craig. A. 2006. E. Bacher. J. and The Human the measured D effect). 289–295..P.M. in the occipital sites. Grillon.. P. there is not likely to be a differential between home study. G. J. M. to Sri Shekar. Hewitt. 690–700. Genetic and environmental contributions to common psychopathologies of child- consistent with a study by Martinovic et al. but we found no evidence for sex Deary.E.. Journal of research has found that while slow waves delta and theta Abnormal Child Psychology 34.K.. Von Lucadou-Muller.H. van Baal. / Biological Psychology 75 (2007) 154–164 our results suggest that environmental influences may play a previous work suggests. Poelhuber. testing effects manifesting as a C effect. 2001). Spinath.. A corollary of this is that there is no References evidence for band-specificity of the C effect in this study.. (the exception being right frontal delta..N. J.. evolution. 1992. R. J. Biological Psychology methodological consideration is the possibility of experimental 71.. Farinas. EEG beta power exhibits non-additive inheritance found it where other studies have not...P. and It is important to note that the influence of common Kate Morley for their helpful insights. 1–17.. G. and. 2001. Loehlin. F. DP0212016). 2005. Electroencephalography and Clinical Neurophysiology 55. The authors are grateful to the twins who took part in the 1963).R. Clinical Neurophy- frontal). vice versa. DeFelipe. and to the anonymous environment is confounded with the influence of non-additive reviewers for their excellent comments and suggestions on the genetic effects. Corley. . C. The pyramidal neuron of the cerebral cortex: Although our adolescent sample had a very narrow age morphological and chemical characteristics of the synaptic inputs. Bates.. C may also play a smaller role (hence depressing A79906588. 2004. Boomsma.. 2006.C. J. more complex and more heterogeneous across the cortex than 493–504. (1998). and beta powers Elston. the former more influential right combined multivariate analysis yielded greater power to at occipital sites and the latter more influential at frontal sites.. Gaser. 2001. Warner. and one might expect EEG Genetics 31. Zietsch et al. C. but this is unlikely Anokhin.. I.C. cognition and the cell: new insights into actually increased (Clarke et al..P. E.. A. I.P. Leite. J. and personality. particularly ment having a significant and substantial impact on frontal EEG developmental animal research. Biological the present study. J. given that the same testing conditions and experimenter were Grant. There are sex differences in the structure and function of the Biological Psychology 72.. Electroencephalographic measures of regional are not consistently found (see Vogel (2000) for a review). EEG and IQ: a new look at an old problem using computerized EEG In summary. D. as the one C factor influencing frontal delta. theta. 328–335. W... Michelle Luciano. 1973). This means that where D is observed (frontal Australian Research Council (grant numbers: A79600334. Barrya. so a negligible occipital C effect is the data. a significant mean effect on sex was found at psychiatry 57. Carl Anderson.J. D. Genetic and environmental may also have an unseen influence on beta power. Females were found to have a larger variance in beta mediator of emotion. G. de Geus. Klimesch. The large sample size and left– common and band-specific effects. EEG data.M.. 2002. W. with common environ- neural plasticity in response to environmental cues. theta and alpha) (Neale and Cardon. and that a substantial C effect was not Genetics 31. A.J. with these generally indicating lower power for males siology 112. the pyramidal neuron and prefrontal function. a few studies report a sex effect of some kind. Frontal EEG asymmetry as a moderator and in males). differences in the genetic effects.. showing that brains develop power. and the environment plays a effect on frontal EEG is consistent with research demonstrating greater role than previously thought. J. A. Verleger... 289–302.. Matousek and Petersen. we found significant age effects at all sites for all EEG in Neurobiology 39.. Another influences on frontal EEG asymmetry: a twin study.1998-B).. van Beijsterveldt. data to reflect this. Intelligence 30.C.

S. The American Psychologist 47... Emergenesis: 79. S. McEvoy. W... Brain Research: Steriade. M..F. variate Behavioral Research 22. large adolescent twin sample. 2001. Porjesz... / Biological Psychology 75 (2007) 154–164 163 Giannitrapani.I. Current Genomics 5. Genetic architecture of fitness and nonfitness of cerebral rhythmic activities. V...B.O. Geffen.J. Neale. Bauer.A.. N. Porjesz. H.J. Ashburner. 993–1018.. Vantage Point.. Smith.W.K.P. L. Scherb...O.. New York. J. Mahoney.R. 87–92. K. F.. S. Neuroimage 14.. S. Whitmont. A.. Cerebral asymmetry and the effects of sex and handed. G. W. B... Cognitive Brain Research 11. Basic mechanisms Merila.M. Cooper. comparative analysis of the neocortex. T.. 1991. Chorlian.. Folkerts. Miniussi. M.. C.. R. C.G. Petersen. R. Wong. Babiloni. Clinical EEG (Electroencephalography) 29..V. 363–376. C. A.. A. of background EEG across the power spectrum. B. Quantified EEG activity in adolescent attention A. Begleiter.. B. Almasy.T. Simultaneous EEG and EDA phenotypic marker for a predisposition to alcoholism: preliminary results measures in adolescent attention deficit hyperactivity disorder. 103–109. W.. Heredity 83. Lubar. Personality traits and its association with resting regional brain activity.. Clinical Neurophysiology 113. K. Reich.W. R. Academic Press. P.B. I. L. Posthuma..... R.. Schizophrenia and Genetics: A Twin Study Nieuwenhuys. C. A. M. Dosen. 603–614. Kuperman. D. Dal Forno. Automation Smit.. Klimesch. S.M. Begleiter... 2006. Niedermeyer. E. Lazzaro. 2004. Cassetta.. International from the COGA project. O’Connor. Padmanabhapillai. Rosenzweig.J.J..A. 758–780. Behavior Genetics 31. Schaul. 2001. 37–42. 1985.. K. 2006. graphy. C. 685–700. Baltimore.C.. I. K.. 562–573. van Beijsterveldt.. Neurology 8.. E. Llinas. Preuss. E. American Journal of Human Genetics 58. 2001. Research with twins: the concept of emergenesis. R... Watzke. J.. Ristanovic. Boker S. N.S.L. Kargere Press.. F. T.. 1996.. Amplitude of visual P3 event-related potential as a Barry. VA 23298. Friston. Hauke. Genetics of affective and anxiety disorders.M. Geffen.. Psychophysiology 42. Rice... Pediatric EEG parameters and intelligence test variables in school-age children. M.. U. 1973.R.. L.. Rohrbaugh. E. Litke. P. S. Schultz. Barnow. 123–134... D. E. Report of IFCN committee on basic mechanisms. Clarke. Rangaswamy.I. 2002. graphy. P.. H. U.J. M.C.M. 231–248.G. Revue Neurologique 34. structural organisation and synaptology.... Tellegen.. owiak.. Gloor. Gould. Tirsch. McGue. 76–81.... Van Eerdewegh. Evans. Frontal EEG characters in aging and the Rangaswamy.K.E.. J. S. G. Plahn. Ferreric. B. 2005.. I. 329–353. I.. R. genetic traits that may not run in familes.. J. EEG power spectra of Smit. E. Zaudig. L. 235–243.G. baugh... P. New York. K. Salinari. D.B.N.M. asymmetry and coherence in male depression. adult human brains. 1647–1656. R. Rodriguez.). G..F.. Oken. 382–392. Journal of Psychophysiology 62. P. jects.. Salinsky. logical signals of working memory in normal aging. Gordon. normal preadolescent twins.G. Morehead. Lanuzza. Geffen. 1999. 601–612. New York. H.. correlativity with some cognitive abilities... Effects on Brain and Behavior. 2001. Mesulam. Bauer.M. Smith. Reich. R. VCU Box 900126..C. M.. Dynamic factor analysis in the frequency domain: Tabachnick.H.. A.. A. Herit- 636–645. N. Methodology for genetic studies of twins and Heritability of human brain functioning as assessed by electroencephalo- families. etymology and strategic intentions.. Lopes da Silva. L. Dosen. Porjesz. study of alcoholism. Psychophysiology 40.. Clarke. C. 481–508. Pellouchoud.. 2003. M. L. Journal of Psychophysiology 51..A... Heritability of Clinical Electroencephalography.. Kamarajan. Gottesman. S. fourth ed... Nobili. S. Brain Research Reviews 29. Structure of the cerebral cortex. G.H.. Hen. K.. M. Lim. 117–137. Molenaar.H. 1992....C. Williams & Wilkins. A.. 1998. Martin.J. D. Cardon. Craig. T.J.. Wright.. E. New York.. Geffen.M. D. 2005. ability and ‘‘environmentability’’ of electroencephalogram in infants: the Groot.J. G. 1992.. 2001. Clinical Neurophysiology 116. W. L...J. 361–373..M. Edenberg..D.A. V. Geus. Reliability of topographic memory performance: a review and analy.S. Y. Johnsrude.M. 2003. J. Jovanovic. Lopes de Silva. Malykh. 1998.. J. L. 123–140.. Moretti.. 1999. 22. 307–337. ments... Genetic influence on ERP slow wave measures of working correlations in patients with obsessive-compulsive disorder. Miller. Blangero. B. R. N.0. 1998. T.. Li. W. F. S. 2004. Meares. E... H. C.S. 1994. 2001.. Lyness. Boomsma. 1991. seventh ed. Mx: Statistical Modeling.. J. Stroganova. frequency. V.S. de Geus. Neurophysio. The fundamental neural mechanisms of electroencephalo- 1565–1577.. Electroencephalography and Clinical Neu- traits: empirical patterns and development of ideas. Rossini.1. H. EEG power. G. 2004. Psy. Mulert. Correlation between spectral activity disorder: a controlled study with clinical implications. L.C. Needham Heights. Vrije Universiteit van Amsterdam. Mavrogiorgou. Boord. Mann. J.. F. Foroud. Alcoholism: Clinical and Experimental Research Journal of Psychophysiology 34.. Enriched and Impoverished Environ- a twin study. rophysiology 76. Maes H. (Eds. 1972.. Using Multivariate Statistics. 2006. Goate. 101–107. mann.. In: Kellaway. Wang. M. Plahn. Henson. Clinical Applications. Rohr- Lazzaro. D. Allyn & Bacon. 1999. An overview of its evolutionary development. International Journal of Psychophysiology 60.. C. Juckel.. Geffen. On the heritability of inspection time and its covariance with IQ: Renner...G. G. Muenchen.. 30–36.T. M.A.. G. O. O’Connor.S. Gender differences of quantitative EEG Genetic variation of individual alpha frequency (IAF) and alpha power in a maturation. F. M. I. M. Neale M. causal modeling of multivariate psychophysiological time series. Neurophysiologie Clinique 28. M. Intelligence 29. Petersen. EEG alpha and theta oscillations reflect cognitive and Pollock. Schmid.. MA. E. M. Li. Hegerl.C. S.N. B. Jones.. Electroencephalography and Clinical Neurophysiology 106. S.. Schneider.. R. The utility of neurophysiological markers in the deficit hyperactivity disorder... T.. Wright. Frequency analysis of the EEG in normal 61. Gevins. pp.D. Chorlian. E.1... 1987. R. P. Kuperman. 1987.. International memory. Posikera.. G.. Multi.R. Stimus. Electroencephalography: Basic ness on brain structure: a voxel-based morphometric analysis of 465 normal Principles.. T. Gordon... D.. N. 2006. H. Binetti. M. D. Gordon. 299–308.. 215–224.A.B. 1999. B.C. American Journal of Psychiatry 160.. W... Resting EEG in offspring of male alcoholics: beta frequencies. Sheldon. Electroencephalography and Clinical Neurophysiology Lykken. Martinovic. 2002. Anatomy and Embryology Gottesman. 2006.... B. Pogarell.. 75–102... Individual analysis of EEG frequency and band Valverde. Hansell. Hansell. quantitative EEG amplitude in healthy late-middle-aged and elderly sub- 169–195.. Kluwer.B. Connell.. Boomsma.K.D. Ferri. 1317–1323. and Related Fields.. 239–251.. Psychiatry...A..A.... Xie G. Jiao.. Clinical Neurophysiology 115.. children and adolescents. Department of Good.E. L. P. ..... P.. M. Acta Psychologica Sinica 28..J. M. I. A.J.. twin study. Z. Boston.. Li. N. Lykken. Knott. D. Springer-Verlag. International Luciano. Molenaar. I. Kennedy. M. Psychiatry Research 106. Amsterdam. D. P. 443–457. Meares. Y. Martin. Shields. Bouchard. 691–697. 727–741. EPTOR 0... Smith. 20–26.. Fidell.. Electroencephalography and Clinical Neurophysiology 79... Current perspectives in genetics and genomics of alcohol Annual Review of Psychology 57. D...V. Intrinsic organization and power in mild Alzheimer’s disease. C. International Journal of Psychophysiology Matousek. 1992. Zimmer- Leonardo. I. C. dependence... S. D. Zimmerman. Zietsch et al. Wright. Jones.B.J. Whitmont. Test/retest reliability in EEG chophysiology 19. Raven Press.J. Moller. frequency analysis. Begleiter. A. G. 1999.. F.I.A. 1982. Orekhova. Frack. C. Sorbell. The endophenotype concept in psychiatry: 190.. The Electrophysiology of Intellectual Functions. A. 1996. Quantitative analysis of EEG in boys with attention-deficit-hyper. 1990. Sun. Martin. Richmond... 1998. M..J.. J. J. Symptom-specific EEG power N.E. Tran..

2000.J. Journal of Neurophysiology 26..H.. Wright. Smith..H. N.. Geffen. Ermentrout. Miyake.. Gordon. Martin. D... De Geus. B. 1017.. Wiesel. Hasegawa..N. 111–138.E. International Journal of Psychophysiology 38. N.. M. Traub. 315–336.. E. 2001a. Geffen. Smith. 1605–1630... Zietsch et al. N. 2002.... D. International Journal of Vogel. 1963. Behavior Genetics 31. 1003– N. Twin and family studies of the Williams. C. M.. Boomsma. A. Luciano. T. Springer Verlag. 48–56. Hubel.D. L.. Whittington. . Rowe. Twin Research 4. Genetic influence on the variance in P3 amplitude and latency. Genetics and the Electroencephalogram. Neuroscience 115. J.R.M.. Ando. Hiraishi. F. Single-cell responses in striate cortex of Wright. C.. Van Baal.164 B. L.. G.. E. Kopell.B. kittens deprived of vision in one eye.M. G. / Biological Psychology 75 (2007) 154–164 van Beijsterveldt. M. Y.. Clark.... N. S. E.K.. 2000.A. R.C.G. E. T.. van Baal... Simms.. neurophysiological test battery: ‘‘neuromarker’’. G. Geffen. Martin.. R. M. D. 555–565. Inhibition-based rhythms: experimental and mathematical observations on L. Ono. G.P.. human electroencephalogram: a review and a meta-analysis.. Biological 2005. Paul.A. Buhl.. 2001b. Genetics of network dynamics.. D. C. The test-retest reliability of a standardized neurocognitive and Psychology 61. cognition: outline of a collaborative twin study. G.. Kanba.. Hansell. Geffen. Berlin. K. Posthuma.H. Hansell..