Genetic variation of individual alpha frequency (IAF) and alpha power in a large adolescent twin sample

Abstract

To further clarify the mode of genetic transmission on individual alpha frequency (IAF) and alpha power, the extent to which individual differences in these alpha indices are influenced by genetic factors were examined in a large sample of adolescent twins (237 MZ, 282 DZ pairs; aged 16). EEG was measured at rest (eyes closed) from the right occipital site, and a second EEG recording for 50 twin pairs obtained approximately 3 months after the initial collection, enabled an estimation of measurement error. Analyses confirmed a strong genetic influence on both IAF (h² = 0.81) and alpha power (h² = 0.82), and there was little support for non-additive genetic (dominance) variance. A small but significant negative correlation (− 0.18) was found between IAF and alpha power, but genetic influences on IAF and alpha power were largely independent. All non-genetic variance was due to unreliability, with no significant variance attributed to unique environmental factors. Relationships between the alpha and IQ indices were also explored but were generally either non-significant or very low. The findings confirm the high heritability for both IAF and alpha power, they further suggest that the mode of genetic transmission is due to additive genetic factors, that genetic influences on the underlying neural mechanisms of alpha frequency and power are largely specific, and that individual differences in alpha activity are influenced little by developmental plasticity and individual experiences.

Introduction

The EEG alpha rhythm, an oscillation with a frequency around 10 waves (cycles) per second, dominates the EEG power spectrum recorded from the brain during rest. Consequently it has become the anchor point for quantitative analysis of the EEG, and, due to the continuing interest in understanding brain processes in the resting state, which defines a baseline for brain activity, is one of the most widely studied physiological indices of brain function (Shaw, 2004). Indeed, this spontaneous brain activity, which is maximal with eyes closed and blocked with eyes open, has high intra-individual stability (Binnie et al., 2003, Fernandez et al., 1993), shows a considerable amount of variation among individuals (Klimesch, 1997, Posthuma et al., 2001), and is found to change with age (Li et al., 1996, McEvoy et al., 2001) and mental state (Moretti et al., 2004).

Individual variation in the alpha rhythm has been posited to reflect individual differences in working memory, attentional demands and/or arousal, and cognitive preparedness (Shaw, 2004), although further work is still required to fully clarify its functional significance. Notable is the body of evidence showing alpha frequency, the peak frequency within the spectral alpha band, to have a strong relationship with working memory performance (Klimesch, 1999), faster information processing (Klimesch et al., 1996), and that participants with superior memory performance have an alpha frequency approximately 1 Hz higher than age-matched controls (Klimesch, 1996, Klimesch, 1997). This is supported, for example, by the findings that alpha frequency increases from early childhood to adulthood and then decreases with age (e.g. Li et al., 1996, McEvoy et al., 2001) in a similar way to general cognitive performance, and of higher peak frequencies in children with higher reading performance compared with age-matched controls (Suldo et al., 2001). These cumulative findings suggest that the alpha rhythm should be related to general cognitive functioning as measured by intelligence tests. However, while alpha frequency has been found to be positively correlated with specific verbal and non-verbal abilities (Anokhin and Vogel, 1996), on balance it does not appear that alpha frequency reflects general intelligence (Anokhin and Vogel, 1996, Jausovec and Jausovec, 2000, Posthuma et al., 2001, Shaw, 2004). Likewise, while a relatively recent study showed a positive association between alpha power and general cognitive ability, the strength of the relationship being dependent on the type of IQ test (Doppelmayr et al., 2002), there are many reports indicating negative or inconsistent findings (e.g. Gaser et al., 1983).

An important generator or modulator of cortical alpha activity is the thalamus, with evidence of a close relationship between thalamic and EEG alpha activity mediated by cortico-thalamic loops (e.g. Lopes da Silva, 1999, Nunez et al., 2001). The occipital cortex is also thought to be important, with the hippocampus and the reticular formation posited to have a more general role. Indeed, several strands of work suggest that alpha oscillations are the result of widespread neuronal activity, and that alpha activity arises from multiple cortical generators (Basar, 1998, Basar, 1999). What is recorded at the scalp is thought to be a spatial average of a large number of components (i.e. the interaction of neural firing patterns generated by several circuits), with alpha activity being dependent on which components are the most highly synchronized over the largest area. More specifically, EEG alpha power is thought to reflect the number of neurons that discharge synchronously (Klimesch, 1999), and the higher the number of synchronously active neurons, the higher the amplitude of the alpha rhythm. There is also some indication of an inverse relationship between the amplitude of the alpha rhythm and alpha frequency, such that the higher the amplitude the slower the frequency of the alpha peak (Lopes da Silva et al., 1976, Pfurtscheller and Lopes da Silva, 1999, Singer, 1993), but more recent work suggests that each of these measures may capture different neural processes (Moretti et al., 2004).

Genetic studies, of which there have been a considerable number (reviewed by (Kuhlo, 1976, van Beijsterveldt and Boomsma, 1994, Vogel, 2000), the first study being in 1936 (Davis and Davis, 1936), all indicate extraordinary similarity of the alpha rhythm for MZ twin pairs and, where estimated, high heritability indicating that individual differences in alpha activity are to a large extent mediated by genetic influences. However, it is still not clear whether additive genetic or a combination of additive and non-additive genetic factors play a role in the genetic transmission, with a number of studies indicating a pattern of very low DZ co-twin correlations that are much less than half the corresponding MZ twin correlations (Christian et al., 1996, Lykken et al., 1974, Lykken et al., 1982, Posthuma et al., 2001, Stassen et al., 1999). While there are large differences across studies, especially with respect to age, EEG methodology, and genetic analysis, many of the studies are underpowered, especially for the detection of dominance as a large number of twin pairs are required. More recently, a meta analysis comprising five twin studies that measured individual alpha frequency (IAF), and eleven studies that measured alpha power (inclusion of studies based on overlap of EEG methodology and availability of heritability estimates), attempted to resolve this question of the importance of dominance (van Beijsterveldt and van Baal, 2002). For IAF a robust ‘meta’ heritability of 81% was indicated, with non-additive genetic factors shown to be important. However, for alpha power where it was not possible to equate estimates across studies, and an averaged heritability of 79% was calculated with no definitive test for dominance, non-additive genetic factors were indicated for adults, but for adolescents and children a purely additive genetic model was shown to be more likely.

The aim of the present study, therefore, was to utilise EEG data from a large sample of adolescent twin pairs, over half of whom were DZ twin pairs and all of the same age (237 MZ, 282 DZ pairs; aged 16), to further examine the extent to which individual differences in both alpha frequency and alpha power were influenced by genetic factors, and to provide additional information on the possible mode of genetic transmission. It also forms one of few studies to examine alpha frequency (i.e. IAF) and alpha power in the same sample, and the first to investigate whether any association between them is due to common genetic or environmental factors, or whether the substantial genetic influences on alpha frequency and alpha power are largely independent of each other. A final aim, since an assessment of psychometric IQ was available, was to explore associations between alpha indices and cognitive ability, and the extent to which any co-variations were genetically mediated.