Common genetic influences on intelligence and auditory simple reaction time in a large Swedish sample

INTELL-01163; No of Pages 6 Intelligence xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Intelligence

Common genetic influences on intelligence and auditory simple reaction time in a large Swedish sample Guy Madison a,⁎, Miriam A. Mosing b,c, Karin J.H. Verweij b, Nancy L. Pedersen c, Fredrik Ullén b a b c

Department of Psychology, Umeå University, Umeå, Sweden Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden

a r t i c l e

i n f o

Article history: Received 7 April 2016 Received in revised form 18 September 2016 Accepted 6 October 2016 Available online xxxx Keywords: Intelligence Reaction time Simple reaction time Common genetic factors Heritability Timing Chronometric

a b s t r a c t Intelligence and cognitive ability have long been associated with chronometric performance measures, such as reaction time (RT), but few studies have investigated auditory RT in this context. The nature of this relationship is important for understanding the etiology and structure of intelligence. Here, we present a bivariate twin analysis of simple auditory RT and psychometric intelligence (measured by the Wiener Matrizen Test). The sample consisted of 1,816 complete twin pairs and 4623 singletons enrolled in the Swedish Twin Registry, who performed the tests online. The heritabilities were 0.54 and 0.21 for intelligence and RT, respectively, and the phenotypic correlation was −0.17, 47% of which was explained by common genetic variance. These results are comparable to those found for visual RT and for other cognitive tests, and add RT in the auditory modality to the small literature on common genetic influences across intelligence and other cognitive and chronometric variables. © 2016 Published by Elsevier Inc.

1. Introduction Cognitive ability has long been associated with various chronometric performance measures, such as reaction time (e.g., Galton, 1883; see also Jensen, 2002). This line of inquiry was initially concerned with the relation between psychometric intelligence (IQ) and simple reaction time (SRT) to visual or auditory stimuli, and was later extended to choice reaction time (CRT), reaction time variability, and measures of perceptual and cognitive speed. Reaction time is also associated with cognitive ageing (Deary & Der, 2005b; Fozard, Vercruyssen, Reynolds, Hancock, & Quilter, 1994) and motor timing (Holm, Ullén, & Madison, 2011). It may seem intuitively compelling that people who are able to respond faster are also brighter, but the nature of such a relationship remains obscure. Hick (1952) proposed a model based on information theory stating that reaction time be proportional to the number of bits of information that have to be processed in order to arrive at a decision, and that the time difference between a four-choice reaction task (2 bits of information) and a two-choice task (1 bit) should reflect the bit ratio. Individual differences in this entity would accordingly reflect differences in information processing speed. According to this theory, faster individuals could have a competitive edge by making a larger number of operations in the same amount of time, which may ⁎ Corresponding author at: Umeå University, Department of Psychology, Behavioral Sciences Building, Mediagränd 14, A-123, SE-901 87 Umeå, Sweden. E-mail address: [email protected] (G. Madison).

be increased by the fact that this time is often limited, as in a learning situation during a lecture. These ideas are discussed in depth by Jensen (1982, 2006) and Deary (2000), among others. Although Hick's quantitative predictions have not borne out empirically (Jensen, 1998; Rammsayer & Brandler, 2007), his model is of historical interest as an early attempt to explain relations between general cognitive ability and speed. Another type of explanation is that more efficient central nervous system communication is characterized by less variability in neural activity, which results in more accurate and stable cognitive representations over time, something which could manifest itself both as higher cognitive performance and as lesser variability or higher temporal resolution in perceptual and motor tasks (Madison, Forsman, Blom, Karabanov, & Ullén, 2009; Rammsayer & Brandler, 2007; Ullén, Forsman, Blom, Karabanov, & Madison, 2008; Ullén, Söderlund, Kääriä, & Madison, 2012). Regardless of the underlying mechanisms, a large empirical literature consistently indicates a chronometric-cognition relationship. For reviews, see Jensen (2006), Sheppard and Vernon (2008), and Rammsayer and Troche (2010). Two main conclusions can be drawn from this research. First, this relation is ubiquitous, and reflects an important phenomenon of great potential theoretical interest. It provides a window for observing the functional relationships between constructs that would seem to be theoretically unrelated, and hence an opportunity for experimentally exploring this relationship (Deary, 2001; Holm, Karampela, Ullén, & Madison, 2016; Holm, Ullén, & Madison, 2013;

http://dx.doi.org/10.1016/j.intell.2016.10.001 0160-2896/© 2016 Published by Elsevier Inc.

Please cite this article as: Madison, G., et al., Common genetic influences on intelligence and auditory simple reaction time in a large Swedish sample, Intelligence (2016), http://dx.doi.org/10.1016/j.intell.2016.10.001

2

G. Madison et al. / Intelligence xxx (2016) xxx–xxx

Holm et al., 2011; Karampela, Holm, & Madison, 2015; Ullén et al., 2012). Secondly, the chronometric-cognition relation varies substantially in strength across the plethora of different cognitive and chronometric tasks and methods (e.g., Jensen, 2006). It therefore remains important to replicate it in other samples and contexts, as well as with different modalities and methods, so as to map out the space of factors that influence it. Here, we attempt to assess the relationship between intelligence and auditory SRT, as well as the heritability of both constructs and their shared genetic influence. Simple and choice reaction time are the chronometric tasks that have been most widely used, and tend to have exhibited the most consistent results (Jensen, 2006). SRT is also more stable than CRT with respect to ageing and sex (Fozard et al., 1994), and also with respect to development, as CRT but not SRT decreases from late adolescence to the early twenties (Deary & Der, 2005b). Previous research has reported phenotypic correlations between RT and IQ throughout the range from −0.05 to −0.5. For visual SRT, correlations of −0.20 (Larson, 1989), − 0.31 (Deary, Der, & Ford, 2001), −0.26 (Jensen & Munro, 1979), and −0.53 (Thoma et al., 2006) have been reported. Very few studies have assessed phenotypic correlations between IQ and auditory RT, however. One reason may be that auditory choice reaction tasks with two, four, or even eight alternatives are more complex and difficult than their visual counterpart, because they typically require memorizing of sounds. As comparing simple and choice RT conditions is often an objective in RT studies, even simple auditory RT tasks may therefore be avoided. Yet, auditory RT and its relationships with other tasks should be of particular interest because humans are 45–60 ms faster to react to auditory than to visual stimuli, as reviewed by Niemi and Näätenen (1981). More recent studies confirm these figures, with 40 (Jaskowski, Jaroszyk, & HojanJezierska, 1990) and 47 ms difference (Shelton & Kumar, 2010). To our knowledge, only three studies consider both auditory RT and IQ. Agrawal and Kumar (1993) reported correlations of −0.40 for 50 males and −0.49 for 40 females, using Raven's SPM. Holm et al. (2011) found very small (non-significant) correlations between auditory SRT and Raven's SPM Plus among 112 university students (−0.050 and −0.065, for mean RT and RT variability) but substantially higher ones for CRT (−0.18) and CRT variability (−0.35), and Poon, Yu, and Chan (1986) found a −0.30 correlation between RT and Raven's SPM among 150 Chinese students. A common chronometric-cognitive factor, as indicated by the literature above, suggests that individual differences in both traits are mediated by quite general mechanisms, possibly ones that are under genetic control. Literature reviews and meta-analyses of the heritability of general intelligence (g) point to an estimate between 0.7 and 0.8 for adults in Western societies (e.g., Bouchard, 2004; Panizzon et al., 2014). A meta-analysis of the heritability of chronometric variables found 12 relevant studies (Beaujean, 2005), and an overall heritability estimate of 0.30 for simpler and 0.52 for more complex chronometric tasks. Six of these reported common genetic influence between some chronometric and some cognitive variable, with genetic correlations in the range −0.4 to −1.0 and proportion common explained variance in the range 0.6– 0.7. More recent studies not included in that meta-analysis report genetic correlations in the range −0.3 to −0.7 (Edmonds et al., 2008; Lee et al., 2012; Luciano et al., 2001; Luciano et al., 2004a; Luciano et al., 2005; Wainwright, Wright, Luciano, Geffen, & Martin, 2008) and proportion common explained variance in the range 0.2–1.0 (Luciano et al., 2001; Luciano et al., 2004a; Luciano et al., 2004b; Wainwright et al., 2008). In conclusion, auditory RT is understudied compared to visual RT, and the shared genetic influence for auditory RT and IQ is as yet undocumented. Here, we obtain heritability estimates of simple auditory reaction time (SART) and assess its phenotypic and genetic correlations with IQ in a genetically informative sample that is much larger than those used in previous studies. We predict a phenotypic correlation in the range −0.2 to −0.5, and that more than half of the shared variance between the traits to be explained by common genetic influences, based

on previous studies of the relationship between IQ and visual RT and other auditory tasks. 2. Methods 2.1. Participants Data were sourced from a large cohort of 32,005 Swedish twins born between 1959 and 1985 (Lichtenstein et al., 2006), who were, during 2012 and 2013, invited to a web-based survey covering, among others, measures of IQ and auditory reaction time (see Ullén, Mosing, Holm, Eriksson, & Madison, 2014; Mosing, Pedersen, Madison, & Ullén, 2014 for further details about this survey). The total number of participants that started the survey was 11,543 and their age was 27–54 years (mean 40.7, SD 7.7). Zygosity was determined based on self-rated intra-pair resemblance, a method shown to be N 98% accurate when compared to zygosity status based on genotyping the twins registered at the Swedish Twin Registry (STR) (Lichtenstein et al., 2002). For further details on the STAGE cohort and zygosity determination in the Swedish Twin Registry see Lichtenstein et al. (2002, 2006). The present study received approval from the Regional Ethics Review Board in Stockholm (Dnr 2011/570-31/ 5, 2012/1107/32, 2012/2172-32). 2.2. Measures The web survey took between 55 and 90 min to complete and contained a wide range of instruments and questions. Many of these are described in previous publications (e.g., Mosing et al., 2014). Measures of IQ and simple auditory reaction time (SART) were used in the present study. The Wiener Matrizen Test (WMT) is a matrix-reasoning test similar to the Raven's Progressive Matrices (Formann & Piswanger, 1979). It was implemented in a Flash application that presented the pictures and recorded the participants' responses, and was discontinued after 25 min, according to the instructions for administration. Reaction time was measured by software that ran in a plugin to the participant's web browser. This code was run in Shockwave, if installed on the participant's computer, or else in Adobe Flash. If none of these multimedia engines were installed and the participant declined to have them installed, the SART test was not run. The SART application issued a sound through the computer's headphones or loudspeaker, and registered the response on the space key of the computer keyboard. Each sound was preceded by a 1.5 to 3.5 s foreperiod, randomly varied from a rectangular distribution and starting from the previous response. Some systematic differences were observed between operating systems and software versions which were corrected for, thus decreasing the sample variance and reducing possible influences of the computer that participants happened to use for completing the survey. These methodological details are described in Madison, Woodley of Menie, and Sänger, 2016 (see Supplemental data, Appendix 1.1 and 1.2). The SART trials had 10 and 25 runs, the median of which was subjected to analysis. The first trial was intended for training, but was used to estimate cross-trial reliability to 0.86. Within-trial split-half reliability was 0.95, based on the medians of the first and final 12 responses. 2.3. Statistical analyses All analyses were conducted using maximum likelihood procedures in the statistical package Mx (Neale, Boker, Xie, & Maes, 2006), correcting for relatedness. Variables were analysed as standardized raw data. The classical twin design was used to determine the extent to which covariation between SART and IQ is due to overlapping genetic and/or overlapping environmental variance. In essence, the variance in a trait and the covariance between traits are decomposed into additive

Please cite this article as: Madison, G., et al., Common genetic influences on intelligence and auditory simple reaction time in a large Swedish sample, Intelligence (2016), http://dx.doi.org/10.1016/j.intell.2016.10.001

G. Madison et al. / Intelligence xxx (2016) xxx–xxx

genetic (A), dominant genetic (D), shared environmental (C) and residual (E) influences, in which additive genetic variance includes the influence of the summed allelic effects and non-additive genetic effects include allelic interactions within and across genes (dominance and epistasis). Environmental influences shared within twin pairs that make them more similar to each other constitute shared environmental variance. Variance not shared within twin pairs, including, for example, unique environmental influences, idiosyncratic experiences, stochastic biological effects, and measurement error, constitute residual variance. Partitioning of variance into genetic and environmental components can be achieved because identical (monozygotic; MZ) twins share 100% of their genes while non-identical (dizygotic; DZ) twins share on average 50% of their segregating genes. The twin correlation should therefore be 1 for MZ pairs and 0.5 for DZ pairs if A were the only source of variance in a trait. If non-additive genetic influences were the only source of variance, a twin pair correlation of 1.0 for MZ pairs and, at most, 0.25 for DZ pairs would be expected (for an explanation, see Posthuma et al., 2003). Likewise would the twin correlation be 1 for both MZ and DZ pairs if C were the only source of variance in a trait, and it would be 0 for both MZ and DZ twin pairs if E were the only source of variance. Because A, C, and E influences predict different patterns of MZ and DZ twin pair correlations, structural equation modelling can be used to determine the best combination of influences on the data. It should be noted MZ correlations provide an indication of the extent to which individual differences in SART and the WMT can be explained by E influences (E = 1 – MZ correlation). E includes both measurement error and environmental factors not shared by twin pairs (e.g. idiosyncratic experiences, such as traumatic events, different peers, and stochastic biological effects, etc.). MZ correlations do therefore not necessarily reflect the reliability of the measures. Bivariate twin models are used to partition the covariance between traits into A, C, and E, or into A, D, and E, in the same way as for the variance in a single trait. We thus estimated the extent to which the observed correlation between SART and IQ is due to overlap in genetic influences, shared environmental influences, or residual factors (see, e.g., Neale & Cardon, 1992; Posthuma et al., 2003). It is not possible to estimate C and D simultaneously when including only twins reared together, as C and D are negatively confounded. The choice of an ACE or ADE model depends on the pattern of MZ and DZ correlations. It is common practice that when DZ twin correlations are at least half the MZ correlation, shared environmental influences are implied and an ACE model is fitted. If DZ twin correlations are less than half the MZ correlations, dominant genetic influences are implied and an ADE model is fitted. We fitted a bivariate Cholesky decomposition, from which we derived an estimate of the overlap in genetic influences between the traits. Preliminary analyses tested for effects of sex, age and zygosity on the means and variances on each trait (α = 0.01). 3. Results 3.1. Descriptive statistics The study sample contained 8257 individuals (mean age 40.1, SD 7.9), including 893 complete MZ (579 female, 314 male) and 923 complete DZ pairs (317 female, 195 male, 411 opposite sex) and 4623 single twins without a participating co-twin. Single twins were included as they contribute to the estimation of means and variances. There were significant mean differences between the sexes for SART and IQ (both p b 0.001). SART was 243.8 ms (SD = 47.0) for females and 239.2 ms (47.5) for males, while the mean number of correct items on the WMT were 13.6 (5.4) for males and 12.3 (5.1) for females. The standardized effect sizes for these differences were 0.10 for SART and 0.25 for IQ. There was also a significant age effect on IQ with lower scores with increased age, but not on SART (cf. Fozard et al., 1994). First,

3

intra-individual SRT reliability was investigated by correlating the second SRT trial with the first, which was intended for training and not included in the following analyses. This cross-trial reliability coefficient was 0.86. For comparison, within-trial split-half reliability was 0.95, based on the medians of the first and final 12 responses in the second trial. The phenotypic correlation between SART and IQ was low but significant at r = −0.17 for each sex separately as well as across all participants (CI: −0.20; −0.15), which corresponds to −0.21 when correcting for the reliability of SART (0.86) and IQ (0.75) (Formann & Piswanger, 1979). Second, we investigated whether performance on the WMT might be related to a chronometric component, given that it is a timed test. The correlation between the WMT score and time to completion was −0.0047 (N = 4699; p = 0.749), excluding trials that were completed within 18.75 min (75% of the maximal time) and assumed to be voluntarily ended with no time constraint. 3.2. Twin analyses Twin correlations for the two variables are shown in Table 1. First, univariate general sex-limitation models were fitted for SART and IQ, allowing for qualitative and quantitative sex differences between males and females. As for SART, DZ twin correlations were less than half the MZ correlations suggesting potential dominant genetic effects, a univariate ADE model was fitted. For IQ the DZ correlations were slightly more than half the MZ correlations, suggesting potential shared environmental effects, so a univariate ACE model was fitted. Univariate modelling results are shown in Table 2 for males and females separately. As the univariate analysis showed that D was not significant for SART, a bivariate ACE Cholesky decomposition was fitted (non-additive genetic variation will mostly be captured by A). Model fitting results of the bivariate twin model are shown in Table 3. A full sex-limitation model was fitted first, allowing the estimates to quantitatively differ between males and females. This is standard procedure in twin modelling for minimizing unexplained variance, based on the observation that the sexes often differ. Equating the parameter estimates between males and females did not lead to a significant decrease in model fit, implying no quantitative sex-differences in the A, C, and E estimates. The full ACE model with the sexes equated is shown in Fig. 1. We tested the significance of the A and C variance components. C effects could be removed from the model without significant deterioration of model fit (nonsignificant pathways are presented as dashed lines in Fig. 1). The model suggests that both additive genetic and residual influences (including non-shared environmental influences) significantly contribute to the variation in SART and IQ and the covariation between the traits. According to Fig. 1, the heritability of SART was 0.21 (0.46 × 0.46) and that of IQ was 0.46 (0.71 × 0.71 + 0.18 × 0.18). Following the tracing rules of pathway analysis, it is possible to determine the A, C, and E parts of the phenotypic correlation: the genetic covariation is −0.08 (0.46 × − 0.18), the shared environmental covariation −0.03 (0.13 × −0.26), and the residual covariation −0.06 (0.88 × −0.07) which add up to the phenotypic correlation of −0.17. Overlapping genetic influences thus explain approximately 47% of the phenotypic correlation (− 0.08/−0.17). The genetic covariation based on the reduced model

Table 1 Twin pair correlations (95% confidence intervals). Zygosity

SART

IQ

MZ DZ MZ female MZ male DZ female DZ male DZ os

0.24 (0.18; 0.31) 0.11 (0.03; 0.18) 0.23 (0.15; 0.31) 0.27 (0.15; 0.38) 0.02 (−0.12; 0.16) 0.05 (−0.12; 0.21) 0.17 (0.06; 0.27)

0.58 (0.54; 0.62) 0.32 (0.27; 0.38) 0.58 (0.53; 0.63) 0.59 (0.52; 0.65) 0.35 (0.25; 0.44) 0.38 (0.25; 0.48) 0.27 (0.18; 0.35)

Note. MZ = monozygotic; DZ = dizygotic; DZ os = DZ opposite-sex; SART = simple auditory reaction time.

Please cite this article as: Madison, G., et al., Common genetic influences on intelligence and auditory simple reaction time in a large Swedish sample, Intelligence (2016), http://dx.doi.org/10.1016/j.intell.2016.10.001

4

G. Madison et al. / Intelligence xxx (2016) xxx–xxx

Table 2 Male and female ADE estimates for SART and ACE estimates for IQ based on univariate sexlimitation modelling (corrected for age and sex effects on the means). SART

A C D E

IQ

Females

Males

Females

Males

0.16 (0.00; 0.30) na 0.07 (0.00; 0.30) 0.77 (0.69; 0.85)

0.20 (0.00; 0.36) na 0.08 (0.00; 0.37) 0.73 (0.62; 0.84)

0.43 (0.23; 0.63) 0.16 (0.00; 0.33) na 0.41 (0.37; 0.47)

0.59 (0.24; 0.65) 0.00 (0.00; 0.32) Na 0.41 (0.35; 0.47)

Note. A = additive genetic, C = shared environmental, D = dominant genetic, E = nonshared environmental influences, SART = simple auditory reaction time.

(AE) is −0.12 (0.48 × −0.24), and explained 70% of the phenotypic correlation (−0.12/−0.17). Note that this is an overestimation as the covariance explained by C would move into A. 4. Discussion The aims of the present study were to determine the heritability of SART, as well as the phenotypic correlation between auditory RT and IQ and the extent to which it could be explained by overlapping genetic influences. The heritability for SART was 0.21, and that the phenotypic correlation (r = −0.17) was slightly below the lower end of that found for IQ and visual RT in previous studies. Approximately half of the correlation was due to genetic pleiotropy, which is also slightly lower than previous estimates for IQ and visual RT. We chose the full Cholesky decomposition and not a more parsimonious model as this model provides the unconstrained variance-covariance matrix and is most precise. As a consequence, the present estimate of the proportion of the phenotypic correlation due to A is somewhat lower than in studies employing AE models, which result in an overestimation of genetic influences. Our estimate of the proportion of the phenotypic correlation explained by genetic influences based on the AE model was 0.70, which is comparable to previous studies that dropped C influences. Aside from that, error variability related to the Internet-based data collection might also have caused lower than usual correlations. As mentioned, participants perform the tests under uncontrolled conditions on any computer they may have had access to. Differences in computer performance, peripheral hardware, and installed software may to some extent influence time-critical data, such as RT. Although considerable effort was made to minimize such variability as well as to quantify and control for it (Madison et al., 2016), these kind of data are inevitably somewhat inferior to laboratory data (cf. Reimers & Stewart, 2007; Wallace & Madison, 2012). Yet, we suspect that technical inaccuracy had a smaller influence than the uncontrolled environment of the participants, i.e. participants may have been disturbed or interrupted by their environment, notwithstanding our careful instructions to seek out a secluded space and a time with minimal distractions. As a result, a proportion of trials may not reflect the participant's true level of performance, and, inasmuch as this obtains to either RT or IQ, the true correlation will be attenuated for the whole sample. That the small correlations are due to measurement noise is consistent with IQ × RT correlations being typically higher in laboratory studies – as reviewed in the introduction approximately −0.30 to −0.45 for both auditory and visual RT. Accordingly, the genetic influence is also likely to be Table 3 Model fitting results for bivariate ACE Cholesky decomposition with simple auditory reaction time and IQ. AIC Full ACE model - sexes separately EQ male/female estimates Remove C from the model

−2LL

df

D− 2LL

D df

p-Value

9 3

0.93 0.67

11,046.17 41,028.17 14,991 11,031.81 41,031.81 15,000 3.641 11,027.36 41,033.36 15,003 1.551

Note. C = shared environmental influences.

Fig. 1. Full ACE Cholesky decomposition for simple auditory reaction time (SART) and IQ with male/female estimates equated. A depicts additive genetic influences, C are shared environmental influences, and E are residual influences. The boxes represent the observed variables and the circles the latent variables that influence the observed variables. A1 is a genetic factor that influences both traits; A2 only influences IQ. The same structure applies for the C and E factors. Dashed lines indicate non-significant pathways.

underestimated in the present study. Nevertheless, a substantial proportion of about half the covariance between the traits can be accounted for by common genetic influences, supporting the notion of a fundamental neural property underlying both chronometric and cognitive behavior. The other side of that coin is the considerable strength of obtaining such large numbers of participants made possible by Internet-based data collection, as laboratory testing typically incurs self-selection based on geographical factors. One relatively simple control one could do is to apply the same instruments to a smaller group of adults in a laboratory setting, to find out how much that affects the phenotypic correlations (cf. Holm et al., 2011). Another limitation of our study is that when only using twins raised together, it is not possible to estimate A, C, D, and E in one model, as this model is not identified. An extended twin-family design would have allowed us to distinguish between additive and dominant genetic effects and shared-environmental influences in one model, but the database did not include information about parents, siblings, or more distantly related kin. In conclusion, the present study is the first exploring auditory reaction time and its covariation with intelligence using a large genetically informative sample. It is of theoretical importance by showing that the genetic basis for the IQ and visual reaction time relation generalizes to auditory RT, and hence most likely to reaction time in general. This finding reinforces the connection between general intelligence and chronometric measures, as advocated by Jensen (2006, 2011), and further explored empirically by Rammsayer and colleagues. One particularly important implication is that RT provides a physiological biomarker of cognitive ability that is less contaminated by environmental factors than are intelligence tests, in which intelligence is indirectly measured through performance on a number of specific tasks. The Flynn effect is

Please cite this article as: Madison, G., et al., Common genetic influences on intelligence and auditory simple reaction time in a large Swedish sample, Intelligence (2016), http://dx.doi.org/10.1016/j.intell.2016.10.001

G. Madison et al. / Intelligence xxx (2016) xxx–xxx

a topical example of how this distinction may be useful. Flynn, (2009) has reported secular increases in psychometric intelligence test performance on the order of three points per decade. Several lines of evidence suggest that this reflects environmental factors that act on the phenotypic level, such as the content and length of schooling and the proliferation of information technology (e.g., Armstrong & Woodley, 2014; Baker et al., 2015; e.g., Meisenberg & Woodley, 2013; Woodley, te Nijenhuis, Must, & Must, 2014). Because RT is unlikely to be affected by those kinds of factors, it may be useful for understanding the nature of these trends, for example if it remains constant or trends in the same or opposite direction across time (e.g., Madison, Woodley of Menie, & Sänger, 2016; Woodley, te Nijenhuis, & Murphy, 2014). Because of its independence of such environmental factors, RT can likewise be useful for examining differences and trends in intelligence across age (e.g., Deary & Der, 2005a; Woodley, Madison, & Charlton, 2014), cultural, economic, and historical circumstances (e.g., Lynn & Vanhanen, 2012), and other conditions likely confounded with such environmental factors. Future research may also address the genetic etiology of the relationships between IQ and subcomponents of the RT task, such as variability and movement time, in order to better understand what exactly is common about them in behavioral terms as well as in the neural substrate. It would also be useful to consider the common genetic influence on IQ and RT variability, as their phenotypic correlation is often at least as high as for RT itself (Holm et al., 2011; Jensen, 1992; Jensen, 2006; Rammsayer & Troche, 2010). Acknowledgements The present work was supported by the Bank of Sweden Tercentenary Foundation (M11-0451:1), the Swedish Scientific Council (5212010-3195) and the Sven and Dagmar Salén Foundation. We would like to thank the twins for their participation. References Agrawal, R., & Kumar, A. (1993). The relationship between intelligence and reaction time as a function of task and person variables. Personality and Individual Differences, 14, 287–288. Armstrong, E. L., & Woodley, M. A. (2014). The rule-dependence model explains the commonalities between the Flynn effect and IQ gains via retesting. Learning and Individual Differences, 29, 41–49. Baker, D., Eslinger, P., Benavides, M., Peters, E., Dieckmann, N., & Leon, J. (2015). The cognitive impact of the education revolution: A possible cause of the Flynn Effect on population IQ. Intelligence, 49, 144–158. Beaujean, A. A. (2005). Heritability of cognitive abilities as measured by mental chronometric tasks: A meta-analysis. Intelligence, 33, 187–201. Bouchard, T. J. (2004). Genetic influence on human psychological traits. A survey. Current Directions in Psychological Science, 13, 148–151. Deary, I. J. (2000). Looking down on human intelligence: From psychometrics to the brain. vol. 34, Oxford, UK: Oxford University Press. Deary, I. J. (2001). Human intelligence differences: Towards a combined experimentaldifferential approach. Trends in Cognitive Sciences, 5, 164–170. Deary, I. J., & Der, G. (2005a). Reaction time explains IQ's association with death. Psychological Science, 16, 64–69. Deary, I. J., & Der, G. (2005b). Reaction time, age, and cognitive ability: Longitudinal findings from age 16 to 63 years in representative population samples. Aging, Neuropsychology, and Cognition, 12, 187–215. Deary, I. J., Der, G., & Ford, G. (2001). Reaction times and intelligence differences. A population-based cohort study. Intelligence, 29, 389–399. Edmonds, C. J., Isaacs, E. B., Visscher, P. M., Rogers, M., Lanigan, J., Singhal, A., et al. (2008). Inspection time and cognitive abilities in twins aged 7 to 17 years: Age-related changes, heritability and genetic covariance. Intelligence, 36, 210–225. Flynn, J. R. (2009). What is intelligence?: Beyond the Flynn effect. Cambridge, UK: Cambridge University Press. Formann, A. K., & Piswanger, K. (1979). Wiener Matrizen-Test. Ein Rasch-skalierter sprachfreier Intelligenztest. Weinheim: Beltz. Fozard, J. L., Vercruyssen, M., Reynolds, S. L., Hancock, P. A., & Quilter, R. E. (1994). Age differences and changes in reaction time: The Baltimore Longitudinal Study of Aging. Journal of Gerontology, 49, 179–189. Galton, F. (1883). Inquiries into human faculty and its development. London: Macmillan. Hick, W. E. (1952). On the rate of gain of information. Quarterly Journal of Experimental Psychology, 4, 11–26. Holm, L., Ullén, F., & Madison, G. (2011). Intelligence and temporal accuracy of behavior: Unique and shared associations between intelligence, reaction time and motor timing. Experimental Brain Research, 214, 175–183.

5

Holm, L., Ullén, F., & Madison, G. (2013). Motor and executive control in repetitive timing of brief intervals. Journal of Experimental Psychology: Human Perception and Performance, 39, 365–380. Holm, L., Karampela, O., Ullén, F., & Madison, G. (2016). Executive control and working memory is involved in sub-second repetitive motor timing. Biological Psychology. Jaskowski, P., Jaroszyk, F., & Hojan-Jezierska, D. (1990). Temporal-order judgments and reaction time for stimuli of different modalities. Psychological Research, 52, 35–38. Jensen, A. R. (1982). Reaction time and psychometric g. In H. J. Eysenck (Ed.), A model for intelligence (pp. 93–132). New York: Springer. Jensen, A. R. (1992). The importance of intraindividual variation in reaction time. Personality and Individual Differences, 13, 869–881. Jensen, A. R. (1998). The suppressed relationship between IQ and the reaction time slope parameter of the Hick function. Intelligence, 26, 43–52. Jensen, A. R. (2002). Galton's legacy to research on intelligence. Journal of Biosocial Science, 34, 145–172. Jensen, A. R. (2006). Clocking the mind: Mental chronometry and individual differences. Oxford, UK: Elsevier. Jensen, A. R. (2011). The theory of intelligence and its measurement. Intelligence, 39, 171–177. Jensen, A. R., & Munro, E. (1979). Reaction time, movement time, and intelligence. Intelligence, 3, 121–126. Karampela, O., Holm, L., & Madison, G. (2015). Shared timing variability in eye and finger movements increases with interval duration: Support for a distributed timing system below and above one second. Quarterly Journal of Experimental Psychology, 68, 1965–1980. Larson, G. E. (1989). A brief note on coincidence timing. Intelligence, 13, 361–367. Lee, T., Mosing, M. A., Henry, J. D., Trollor, J. N., Lammel, A., Ames, D., et al. (2012). Genetic influences on five measures of processing speed and their covariation with general cognitive ability in the elderly: The older Australian twins study. Behavior Genetics, 42, 96–106. Lichtenstein, P., de Faire, U., Flodérus, B., Svartengren, M., Svedberg, P., & Pedersen, N. L. (2002). The Swedish Twin Registry: A unique resource for clinical, epidemiological and genetic studies. Journal of Internal Medicine, 252, 184–205. Lichtenstein, P., Sullivan, P. F., Cnattingius, S., Gatz, M., Johansson, S., Carlström, E., et al. (2006). The Swedish Twin Registry in the third millenium: An update. Twin Research and Human Genetics, 9, 875–882. Luciano, M., Smith, G. A., Wright, M. J., Geffen, G. M., Geffen, L. B., & Martin, N. G. (2001). On the heritability of inspection time and its covariance with IQ: A twin study. Intelligence, 29, 443–457. Luciano, M., Wright, M. J., Geffen, G. M., Geffen, L. B., Smith, G. A., & Martin, N. G. (2004a). A genetic investigation of the covariation among inspection time, choice reaction time, and IQ subtest scores. Behavior Genetics, 34, 41–50. Luciano, M., Wright, M. J., Geffen, G. M., Geffen, L. B., Smith, G. A., & Martin, N. G. (2004b). Multivariate genetic analysis of cognitive abilities in an adolescent twin sample. Australian Journal of Psychology, 56, 79–88. Luciano, M., Posthuma, D., Wright, M. J., de Geus, E., Smith, G. A., Geffen, G. M., et al. (2005). Perceptual speed does not cause intelligence, and intelligence does not cause perceptual speed. Biological Psychology, 70, 1–8. Lynn, R., & Vanhanen, T. (2012). National IQs: A review of their educational, cognitive, economic, political, demographic, sociological, epidemiological, geographic and climatic correlates. Intelligence, 40, 226–234. Madison, G., Forsman, L., Blom, Ö., Karabanov, A., & Ullén, F. (2009). Correlations between general intelligence and components of serial timing variability. Intelligence, 37, 68–75. Madison, G., Woodley of Menie, M. A., & Sänger, J. (2016). Secular slowing of auditory simple reaction time in Sweden (1959–1985). Frontiers in Human Neuroscience. Meisenberg, G., & Woodley, M. A. (2013). Are cognitive differences between countries diminishing? Evidence from TIMSS and PISA. Intelligence, 41, 808–816. Mosing, M. A., Pedersen, N. L., Madison, G., & Ullén, F. (2014). Genetic pleiotropy explains associations between musical auditory discrimination and intelligence. PLoS ONE, 9, e113874. Neale, M. C., & Cardon, L. R. (1992). Methodology for genetic studies of twins and families. Dordrecht, The Netherlands: Kluwer Academic Publishers B.V. Neale, M. C., Boker, S. M., Xie, G., & Maes, H. H. (2006). Mx: Statistical modeling (7 ed.). Richmond, VA: Department of Psychiatry. Niemi, P., & Näätenen, R. (1981). Foreperiod and simple reaction time. Psychological Bulletin, 89, 133–162. Panizzon, M. S., Vuoksimaa, E., Spoon, K. M., Jacobson, K. C., Lyons, M. J., Franz, C. E., et al. (2014). Genetic and environmental influences on general cognitive ability: Is g a valid latent construct? Intelligence, 43, 65–76. Poon, P. W. F., Yu, W. Y., & Chan, J. W. C. (1986). Correlation between auditory reaction time and intelligence. Perceptual and Motor Skills, 63, 375–378. Posthuma, D., Beem, A. L., de Geus, E. J., van Baal, G. C. M., von Hjelmborg, J. B., Iachine, I., et al. (2003). Theory and practice in quantitative genetics. Twin Research, 6, 361–376. Rammsayer, T. H., & Brandler, S. (2007). Performance on temporal information processing as an index of general intelligence. Intelligence, 35, 123–139. Rammsayer, T. H., & Troche, S. (2010). Effects of age and the relationship between response time measures and psychometric intelligence in younger adults. Personality and Individual Differences, 48, 49–53. Reimers, S., & Stewart, N. (2007). Adobe Flash as a medium for online experimentation: A test of reaction time measurement capabilities. Behavior Research Methods, 39, 365–370. Shelton, J., & Kumar, G. P. (2010). Comparison between auditory and visual simple reaction times. Neuroscience and Medicine, 1, 30–32. Sheppard, L., & Vernon, P. A. (2008). Intelligence and speed of information-processing: A review of 50 years of research. Personality and Individual Differences, 44, 535–551.

Please cite this article as: Madison, G., et al., Common genetic influences on intelligence and auditory simple reaction time in a large Swedish sample, Intelligence (2016), http://dx.doi.org/10.1016/j.intell.2016.10.001

6

G. Madison et al. / Intelligence xxx (2016) xxx–xxx

Thoma, R. J., Yeo, R. A., Gangestad, S., Halgren, E., Davis, J., Paulson, K. M., et al. (2006). Developmental instability and the neural dynamics of the speed–intelligence relationship. NeuroImage, 32, 1456–1464. Ullén, F., Forsman, L., Blom, Ö., Karabanov, A., & Madison, G. (2008). Intelligence and variability in a simple timing task share neural substrates in the prefrontal white matter. The Journal of Neuroscience, 28, 4239–4243. Ullén, F., Söderlund, T., Kääriä, L., & Madison, G. (2012). Bottom-up mechanisms are involved in the relation between accuracy in timing tasks and intelligence — further evidence using manipulations of state motivation. Intelligence, 40, 100–106. Ullén, F., Mosing, M. A., Holm, L., Eriksson, H., & Madison, G. (2014). Psychometric properties and heritability of a new online test for musicality, the Swedish Musical Discrimination Test. Personality and Individual Differences, 63, 87–93. Wainwright, M. A., Wright, M. J., Luciano, M., Geffen, G. M., & Martin, N. G. (2008). Genetic covariation among facets of openness to experience and general cognitive ability. Twin Research and Human Genetics, 11, 275–286.

Wallace, A., & Madison, G. (2012). The timing accuracy of general purpose computers for experimentation and measurements in psychology and the life sciences. The Open Psychology Journal, 5, 53. Woodley, M. A., te Nijenhuis, J., Must, O., & Must, A. (2014a). Controlling for increased guessing enhances the independence of the Flynn effect from g: The return of the Brand effect. Intelligence, 43, 27–34. Woodley, M. A., te Nijenhuis, J., & Murphy, R. (2014b). Is there a dysgenic secular trend towards slowing simple reaction time? Responding to a quartet of critical commentaries. Intelligence, 46, 131–147. Woodley, M. A., Madison, G., & Charlton, B. G. (2014c). Possible dysgenic trends in simple visual reaction time performance in the Scottish Twenty-07 cohort: A reanalysis of Deary & Der (2005). Mankind Quarterly, 55, 110–124.

Please cite this article as: Madison, G., et al., Common genetic influences on intelligence and auditory simple reaction time in a large Swedish sample, Intelligence (2016), http://dx.doi.org/10.1016/j.intell.2016.10.001