Genetic and environmental influences on sensory gating of mid-latency auditory evoked responses: A twin study

Schizophrenia Research 89 (2007) 312 – 319 www.elsevier.com/locate/schres

Genetic and environmental influences on sensory gating of mid-latency auditory evoked responses: A twin study Andrey P. Anokhin a,⁎, Andrei B. Vedeniapin a , Andrew C. Heath a , Oleg Korzyukov b , Nashaat N. Boutros b a

b

Department of Psychiatry, Washington University School of Medicine, St.Louis, MO, USA Wayne State University School of Medicine, Department of Psychiatry and Behavioral Neurosciences, Detroit, MI, USA Received 22 March 2006; received in revised form 20 July 2006; accepted 14 August 2006 Available online 2 October 2006

Abstract A deficit in sensory gating measured by the suppression of P50 auditory event-related potential (ERP) has been implicated in the biological bases of schizophrenia and some other psychiatric disorders and proposed as a candidate endophenotype for genetic studies. More recently, it has been shown that gating deficits in schizophrenics extend to ERP components reflecting early attentive processing (the N1/P2 complex). However, evidence for heritability of sensory gating in the general population is very limited. Heritability of P50, N1, and P2 amplitudes and gating was estimated in 54 monozygotic and 55 dizygotic twin pairs using a dualclick auditory paradigm. Genetic model-fitting analysis showed high heritability of peak amplitudes of P50, N1, and P2 waves. Genetic influences on P50 gating (S2/S1) were modest, while heritability of N1 and P2 gating was high and significant. The alternative gating measure (S1–S2 difference) showed significant heritability for all three ERP components. Weak genetic influences on P50 gating ratio can be related to its poor test–retest reliability demonstrated in previous studies. These results suggest that gating measures derived from the N1/P2 wave complex may be useful endophenotypes for population-based genetic studies of the sensory gating function and its impairments in psychopathology. © 2006 Elsevier B.V. All rights reserved. Keywords: Auditory evoked potentials; P50; N100; Sensory gating; Twins; Heritability

1. Introduction An impaired ability to filter sensory information has been hypothesized as one of the core dysfunctions in schizophrenia and related spectrum of disorders and cognitive abnormalities (Freedman et al., 1983; Braff and Geyer, 1990). Filtering excessive information can protect limited processing resources against overloading, or “flooding”(Venables, 1964). It has been proposed ⁎ Corresponding author. Tel.: +1 314 286 2201; fax: +1 314 286 0092. E-mail address: [email protected] (A.P. Anokhin). 0920-9964/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2006.08.009

that one form of such inhibitory regulation is sensory gating, or modulation of the sensitivity to repetitive sensory information (Braff and Geyer, 1990; Adler et al., 1999; Freedman et al., 2000). The hypothesized sensory gating can be studied experimentally in humans using the so-called conditioning-testing paradigm, in which the subject is administered a pair of brief auditory stimuli (clicks) and brain evoked responses are recorded. Click stimuli elicit a cascade of neuroelectric response components reflecting sequential stages of information processing in the auditory pathways and the cerebral cortex. A distinct complex of ERP components

A.P. Anokhin et al. / Schizophrenia Research 89 (2007) 312–319

including P50, N1, and P2 peaks occurs between the early brainstem responses and late cognitive “endogenous” P300 response. Following the classification suggested by Buchsbaum (1977) and later by Roth et al. (1980), we refer to this wave complex as midlatency auditory evoked responses (MLAER). The P50 component reflects largely pre-attentive sensory processing, while N1 and P2 components reflect early attentional processing (relative to the late “endogenous” P300 component involving voluntary, controlled attention), although the distinction between preattentional and attention-dependent cognitive functions is not necessarily discrete and categorical (Braff and Light, 2004). When auditory click stimuli are presented in pairs with inter-click interval of less than 1 s, the P50 response to the second of the two identical clicks is suppressed, or gated out. According to the sensory gating hypothesis, the first stimulus elicits the initial excitatory response of the neuronal population giving rise to the P50 response and also activates inhibitory pathways attenuating the response to the second stimulus (Davis et al., 1966; Freedman et al., 1983; Adler et al., 1999). Studies have shown that P50 gating is diminished in schizophrenic patients and, importantly, in unaffected first degree relatives of schizophrenics compared to healthy controls, suggesting that P50 suppression can be a marker of familial and, possibly, genetic risk for the disorder (reviewed in Adler et al., 1999; Freedman et al., 2000; Bramon et al., 2004). Other studies reported a genetic linkage between P50 suppression index and the alpha7 neuronal nicotinic receptor subunit gene, as well as association between different mutations in this gene and P50 phenotype (reviewed in Freedman et al., 2003). These studies suggest that P50 suppression can serve as one of the candidate vulnerability markers, or endophenotypes for schizophrenia and schizophrenia spectrum disorders (Freedman et al., 2000). It was suggested that the gating deficit occurring at early stages of sensory processing is primary to more complex cognitive deficits occurring at later processing stages, such as problems with attention and perception, cognitive fragmentation, hypervigilance, etc. (Freedman et al., 1983; Braff and Geyer, 1990). However, findings in this field have not been entirely consistent: some studies failed to find any significant association between P50 gating and schizophrenia (Arnfred et al., 2003), or found this association only in a subgroup of patients (Jin et al., 1998). Other studies failed to find association between mutations in the nicotinic receptor gene and schizophrenia (Li et al., 2004) or genetic linkage with markers in this chromo-

313

somal region (Neves-Pereira et al., 1998; Curtis et al., 1999). Some studies suggest that increased S2/S1 ratio (indicating poor gating) can in fact be related to smaller response to the first click (S1) in schizophrenics, rather than to deficient suppression of the response to the second click (e.g. Blumenfeld and Clementz, 2001; Clementz and Blumenfeld, 2001; Johannesen et al., 2005). Thus, the amplitude of P50 response to the first click alone (S1) can potentially serve as a marker of genetic vulnerability. In addition to P50, several studies have also reported similar gating abnormalities in other MLAER components including N1 and P2 related to early attentive stages of processing (Roth et al., 1980; Boutros et al., 1999; Clementz and Blumenfeld, 2001; Boutros et al., 2004a; Boutros et al., 2004b). Characterization of the heritability of gating indices derived from each of these components is important as these components likely reflect different physiological aspects of the sensory gating function (Boutros and Belger, 1999). Moreover, recent studies provided behavioral evidence suggesting that P50 gating and N1 gating correspond to different sensory processing phenomena. Using a self-report measure of sensory experiences, Kisley et al. (2004) have shown that poor P50 suppression was correlated with the dimension of Perceptual Modulation indicating filtering difficulties, while poor N1 suppression was correlated with the dimension of Over-Inclusion indicating increased awareness of background sounds (Kisley et al., 2004). Sensory gating deficits as measured by P50 suppression are not limited to schizophrenia and have been reported in a broader spectrum of conditions including schizotypal personality disorder (Cadenhead et al., 2000) and questionnaire-measured psychometric schizotypy (Croft et al., 2001), as well as in substance use disorders (Fein et al., 1996; Patrick and Struve, 2000; Marco et al., 2005). Thus, deficient gating of P50 and perhaps other components (N1, P2) may represent an endophenotype, or biological marker of genetic vulnerability, for schizophrenia and related spectrum of disorders and comorbidities, perhaps including addictions. One of the requirements to an endophenotype is its significant heritability. There have been two small twin studies of P50 suppression (Myles-Worsley et al., 1996; Young et al., 1996). To the best of our knowledge, there have been no genetic studies of other MLAER components (N100, P200) recorded in the dual click paradigm. Accordingly, the goal of the present study was to estimate heritability of the peak amplitudes and gating measures of MLAER including P50, N100, and P200 components in the general population using a community-based sample of twins.

314

A.P. Anokhin et al. / Schizophrenia Research 89 (2007) 312–319

2. Methods Due to space limitations, only a concise description of methods is presented here. For a more detailed description, see Supplementary methods online.

mastoid reference, and an averaged mastoid reference was digitally computed off-line. Electrical signals were amplified by Synamps bioamplifiers (Compumedics/ Neuroscan, El Paso, TX) with a bandpass of 0.05 to 70 Hz and stored on a computer hard disk for subsequent off-line analysis. The sampling rate was 1000 Hz.

2.1. Participants 2.3. ERP analysis and peaks detection The original sample consisted of 226 female twin subjects (age 18–29) ascertained through birth records from the general population. After the exclusion of 27 subjects due to recording artifacts, drowsiness, or the lack of identifiable P50 response, the remaining full pairs (48 MZ and 40 DZ) were included in the final analysis of P50 data. The analysis of N1 and P2 data was based on a larger sample (54 MZ and 55 DZ pairs), since fewer subjects were excluded. N1 and P2 peaks are relatively large and can be measured reliably even when P50 is masked by artifacts and cannot be identified reliably. Therefore, subjects lacking distinctive P50 components but showing a measurable N1/P2 complex were included in the analysis of the latter. The study was approved by the Washington University Institutional Review Board, and after complete description of the study to the subjects a written informed consent was obtained. 2.2. Experimental procedure and EEG recording A paired clicks paradigm was used (Clementz et al., 1997) including 60 trials (pairs of clicks) with a 500 ms interval between the clicks and an average 8 s inter-trial interval (for details regarding the apparatus, procedure, and stimulus presentation characteristics, see Supplementary methods). The EEG was recorded using the left

Data were bandpass filtered using different filter settings for P50 and later components (N1 and P2), since the main frequency of the P50 component is about 40 Hz (Clementz et al., 1997), whereas the frequencies of the N1 and P2 components are substantially lower. For P50 peak detection, a bandpass of 10–60 Hz (24dB/octave roll-off, zero phase shift) was used. For N1 and P2 detection, data were filtered using 0.3–30 Hz bandpass (24-dB/octave roll-off, zero phase shift). Two alternative algorithms for P50 detection were used to ensure comparability of the present study with previous studies. First, the P50 was identified as the largest positive peak preceding the N100 potential (Boutros et al., 2004a). P50 amplitude was scored as the difference between P50 peak and the peak value of the preceding negative trough (N50). The second algorithm for P50 identification was based on a slightly different set of criteria described by Nagamoto et al. (1989), which has become more or less the standard in the literature. In most of the subjects comprising our sample, application of these two algorithms resulted in the detection of the same peaks, except 18 subjects. Nevertheless, heritability analyses were conducted separately for P50 suppression scores obtained using these two algorithms.

Fig. 1. P50 suppression effect: Grand-average for the entire sample (a); P50 components from two subjects — with a strong S2 suppression (b) and with a poor S2 suppression (c). Click onset is at 0 ms. Black thick line: S1 response waveform; grey line: S2 response.

A.P. Anokhin et al. / Schizophrenia Research 89 (2007) 312–319

The N100 component was identified as the largest negative deflection in the 80–150 ms post-stimulus window. The P200 component was identified as the largest positive deflection between 150 and 250 ms. The P200 amplitude was measured from its peak to the preceding N100 trough (Boutros et al., 2004a). For each MLAER component, the suppression of the response to the second stimulus (S1) was measured by

315

dividing the amplitude of S2 responses by the amplitude of S1 responses (S2/S1 ratio). Lower S2/S1 values indicate greater suppression and hence a stronger sensory gating and vice versa. For P50 suppression we used both raw S2/S1 index for each of the above criteria sets, as well as truncated S2/S1 values as an alternative gating index following the suggestion of Cadenhead et al. (2000) to prevent outliers from disproportionately affecting the

Table 1 Twin correlations and heritability estimates for mid-latency auditory ERP components and gating indices ERP variable P50, Method 1: S1 S2 S2/S1 S1–S2 S2/S1 (truncated to 1.0)

P50, Method 2: S1 S2 S2/S1 S1–S2 S2/S1 (truucated to 1.0)

rMZ

rDZ

a2 (95% CI)

c2 (95% CI)

e2 (95% CI)

χ2 (df = 4)

p

AIC

.53⁎⁎⁎ .43⁎⁎

.28⁎ .38

.54 ± .33

.28⁎

.22

2.52 ± 2.18 .52 ± .27

.45⁎⁎ .36⁎⁎

.13 .25

.58(.36–.74) .52(.28–.69) – .27 (.03–.47) – .41(.17–.60) .36 (.12–.55) –

– – .39 (.19–.55) – .25(.04–.44) – – .31(.11–.49)

.42(.26–.64) .48(.31–.72) .61 (.45–.81) .73 (.53–.97) .75(.56–.96) .59(.40–.83) .64 (.45–.88) .69(.51–.89)

3.90 9.46 10.50 5.89 5.36 7.01 6.06 5.95

.42 .05 .03 .21 .25 .14 .20 .20

− 4.10 1.46 2.50 − 2.11 − 2.64 − .99 − 1.94 − 2.05

5.08 ± 2.70 2.55 ± 1.74

.56⁎⁎⁎ .46⁎⁎

.30⁎ .42⁎⁎

.47 ± .35 2.76 ± 2.27 .45 ± .30

.30⁎⁎ .49⁎⁎ .35⁎

.19 .11 .23

.61(.40–.75) .56 (.33–.71) – .30(.05–.50) .46(.22–.64) .34(.11–.54) –

– – .42 (.27 –.56) – – – .31(.14–.46)

.39(.25–.60) .44 (.29–.66) .58 (.44–.73) .70(.50–.95) .54(.36–.78) .66 (.46–.89) .69 (.54–.86)

3.64 12.23 17.68 .811 3.84 3.30 4.37

.46 .02 .001 .94 .43 .51 .36

−4.36 4.23 9.68 − 7.19 − 4.16 − 4.70 − 3.63

Mean ± SD (μV)

AE: CE: AE: CE: AE: CE:

AE: CE:

AE: CE:

5.02 ± 2.74 2.51 ± 1.74

N1, absolute amplitude: S1 S2 S2/S1 S1–S2

− 11.86 ± 7.39 − 4.68 ± 3.50 .65 ± .30 − 7.18 ± 6.29

.78⁎⁎⁎ .67⁎⁎⁎ .44⁎⁎⁎ .63⁎⁎⁎

.50⁎⁎⁎ .40⁎⁎ .20 .42⁎⁎⁎

.76(.64–.84) .65(.50–.76) .45(.22–.63) .64(.48–.76)

– – – –

.24 .35 .55 .36

(.16–.36) (.24–.50) (.37–.78) (.25–.52)

5.50 2.51 2.18 2.98

.24 .54 .70 .56

− 2.50 − 5.49 − 5.82 − 5.02

N1, peak-to-peak: S1 S2 S2/S1 S1–S2

− 16.58 ± 8.92 − 6.63 ± 4.13 .59 ± .23 − 9.94 ± 7.22

.76⁎⁎⁎ .62⁎⁎⁎ .58⁎⁎⁎ .75⁎⁎⁎

.35⁎⁎ .26⁎ .32⁎⁎ .32⁎⁎

.73(.61–.82) .61(.42–.74) .57(.38–.70) .70(.57–.79)

– – – –

.27 .39 .43 .30

(.18–.40) (.26–.58) (.30–.62) (.20–.44)

1.61 7.38 2.42 6.25

.81 .12 .66 .18

− 6.39 − .62 − 5.58 − 1.75

.76⁎⁎⁎ .62⁎⁎⁎ .57⁎⁎⁎ .75⁎⁎⁎

.30⁎ .48⁎⁎⁎ .32⁎⁎ .21

.71(.57–.80) .64(.48–.75) .54(.35–.67) .68(.54–.78)

– – – –

.29 .36 .46 .32

(.20–.43) (.25–.52) (.33–.65) (.22–.46)

5.17 7.41 2.61 10.06

.27 .12 .63 .04

− 2.83 − .59 − 5.39 2.06

P2, peak-to-peak: S1 S2 S2/S1 S1–S2

27.19 ± 12.88 8.88 ± 4.16 .37 ± .19 18.31 ± 11.39

Notes: 1) rMZ and rDZ are intrapair correlations for MZ and DZ twins, respectively. Variance component estimates are provided for the best-fitting “AE” model. If “CE” model also showed a good fit, variance component estimates are provided for both models. Variance components are as follows: a2 is the proportion of total phenotypic variance explained by genetic factors (heritability), c2 is the proportion of variance attributed to the shared environmental influences, and e2 is the proportion of variance due to non-shared (individually specific) environmental factors; 95% confidence intervals of the maximum likelihood estimates of the variance components are shown in brackets. Chi-square (df = 4) shows the goodness of fit, with lower χ2 values and, respectively, higher p-values indicating a better fit; AIC is Akaike's Information Criterion, with lower values indicating better fit. Significance levels (1-tailed): ⁎p b 0.05; ⁎⁎p b 0.01; ⁎⁎⁎p b 0.001. 2) Methods 1 and 2 refer to the two alternative algorithms used for the identification of P50: Boutros et al. (2004a,b) and Nagamoto et al. (1989), correspondingly (see the text for more detail); 3) truncated S2/S1 values (following Cadenhead et al., 2000) were used as an alternative gating index; 4) Absolute peak amplitude of N1 was computed relative to pre-stimulus baseline, and peak-to-peak N1 was computed relative to the preceding P50 peak; 5) All results pertain to the vertex electrode location (Cz); 6) The number of pairs for P50 analysis: 48 MZ, 40 DZ; for N1 and P2: 54 MZ, 55 DZ.

316

A.P. Anokhin et al. / Schizophrenia Research 89 (2007) 312–319

results: all values over 1.0 (facilitation) were set to 1.0 (no suppression). Thus, a total of four S2/S1 indices were used: according to Boutros et al. (2004a) – Method 1, raw and truncated scores, and according to Nagamoto et al. (1989) – Method 2, raw and truncated scores. 2.4. Genetic analysis of twin data For each of the ERP measures, we computed Pearson correlations between the members of twin pairs (intrapair correlations) separately for MZ and DZ twins. To estimate the relative contribution of genetic and environmental sources to the total phenotypic variance of ERP measures (heritability), we performed a biometrical genetic analysis using a model-fitting approach, which has become a standard in twin genetic research (Neale and Cardon, 1992; Rijsdijk and Sham, 2002) and allows to estimate three sources of the phenotypic variance: additive genetic influences (A), non-additive genetic influences (D) or environmental influences shared by family members (C), and individually unique (unshared) environmental influences (E). For more details, see Supplementary methods online. 3. Results 3.1. General considerations Examples of P50 waves (S1 and S1 responses) illustrating strong and poor suppression of S2 are presented on Fig. 1 (note that the amplitudes of N1 and P2 components are attenuated on this waveform due to filter settings that are optimal for P50 detection but not

for N1 and P2). Data are presented for the vertex (Cz) electrode location, consistent with previous P50 gating literature. Mean values and standard deviations of peak amplitudes and S2/S1 suppression ratios (Table 1) are in good agreement with previous studies. Twin correlations and parameter estimates from the best fitting models for all studied ERP variables are presented in Table 1. MZ twin correlations were significant for all studied variables, although the values varied in a broad range (.28 to .78). DZ correlations were generally lower than MZ correlations (.13 to .50) with a few exceptions, where MZ and DZ correlations were approximately equal. Generally, higher MZ than DZ resemblance suggests genetic influences, while equal MZ and DZ resemblance suggests shared (e.g. familial) environmental influences. 3.2. P50 sensory gating Model fitting results showed substantial heritability of the amplitude of P50 wave elicited by the conditioning click (S1), suggesting that 58% of the total inter-individual variance can be attributed to genetic factors. The results further suggest significant familial transmission effect on P50 elicited by the test click (S2), however the model fitting analysis was unable to discriminate between genetic and shared environmental effects (both AE and CE models showed approximately equal fit to the data). Genetic influences on P50 suppression (S2/S1) were modest but significant, however, shared environmental transmission model also fit well. The alternative gating measure (S1–S2 difference) showed higher MZ twin correlation and

Fig. 2. N1 suppression effect: Grand-average for the entire sample (a); N1 components from subjects — with a strong S2 suppression (b) and with a poor S2 suppression (c). Click onset is at 0 ms. Black thick line: S1 response waveform; grey line: S2 response.

A.P. Anokhin et al. / Schizophrenia Research 89 (2007) 312–319

significant heritability, but this measure appears to be highly correlated with S1 amplitude ( r = 0.79, p b 0.001), suggesting that individual variability in this measure is largely driven by S1. 3.3. Early-attentive (N100, P200) gating Waveforms of the N1/P2 complex are presented on Fig. 2. Amplitudes of both N1 and P2 components elicited by the conditioning click (S1) showed very high heritability, with 71–76% of the total variance explained by genetic influences. As for the test stimulus (S2), heritability of N1 amplitude was also high, however, genetic influences on P2 amplitude could not be distinguished from shared environmental influences (CE model also showed a good fit). N1 and P2 gating indices (S2/S1) showed high and significant heritability (54–57%). The S1–S2 measure also showed high heritability, but, as in the case of P50, this measure correlated highly with S1 amplitudes (r = 0.86 and 0.96 for N1 and P2, respectively). It should be noted that confidence intervals for P50 and N1/P2 heritability estimates overlap, and demonstrating significant differences in heritability would require a substantially larger sample. 4. Discussion The present study extends previous studies on heritability of P50 suppression by utilizing a substantially larger twin sample and by applying the modelfitting approach to the estimation of genetic and environmental influences. This study also demonstrated for the first time the heritability of gating effects at early attentive stages of auditory processing (N1 and P2 waves) that have been associated with schizophrenia in recent studies (Boutros et al., 2004a,b). Our results indicate substantial heritability of the amplitude of P50 response to the first click (S1), but only a modest heritability of S2/S1 P50 gating (around 30%). The latter finding disagrees with the two previous twin studies that reported substantial heritability of P50 suppression. This discrepancy could result from the difference between the twin samples, including the sample size, age, gender, and inclusion/exclusion criteria. Firstly, previous studies used substantially smaller twin samples the total number of pairs was 27 (15 MZ and 12 DZ) in the study of Young et al. (1996) and 39 (26 MZ and 13 DZ) in the study of Myles-Worsley et al. (1996). Our study was based on a considerably larger sample of twins (88 pairs, including 48 MZ and 40 DZ pairs included in the final analyses), and intrapair twin

317

correlations were less likely to be affected by a few influential observations. Secondly, some characteristics of the samples are different among the three studies. In the study of Young et al. (1996) the subjects were volunteers participating in a genetic study of alcohol response, and the age range was very broad (21– 51 years). The Myles-Worsley et al. (1996) study used a population-based study of twins ascertained through birth registry, however, in contrast to our study, it included both adults and children and the age range was also very broad (10–39 years). Finally, the two previous studies have included both genders, while our study included females only. A previous study has shown less efficient P50 and N1 gating in women compared to men due to increased S2 response (Hetrick et al., 1996). However, the effect of gender on heritability of P50 gating remains unclear and warrants further studies. Apart from these sample differences, there is another and perhaps more plausible explanation for the present finding of low heritability of P50 gating. Previous studies have shown a poor temporal stability of P50 suppression ratio across recording sessions, with test– retest reliability coefficients ranging from 0 to 0.27 (Boutros et al., 1991; Cardenas et al., 1993; Smith et al., 1994), although one study noted that test–retest reliability can be improved by using dipole modeling of the P50 source (Cardenas et al., 1993). Even within one session, test–retest correlation across trial blocks was zero, both in normal control and schizophrenic samples (Clementz et al., 1997). Taken together, these studies provide little support for the assumption that P50 suppression ratio is a stable and reproducible trait-like characteristic, at least when it is assessed using conventional ERP peak measurements. Importantly, in the studies cited above, S1 wave amplitudes showed a good test–retest reliability (r = 0.6–0.7 in normal controls), which is well consistent with twin correlations and heritability estimates obtained in the present study. Results obtained for ERP components associated with early attentive stage of auditory processing (N1 and P2) indicate substantial heritability of both peak amplitudes and S2/S1 suppression indices. The S2/S1 gating of the P2 wave (measured as N1–P2 peak-topeak difference) showed the highest heritability. The difference measures (S1–S2) showed higher twin correlations and, accordingly, greater familial or genetic effects for all three ERP components. However, the limitation of the difference measure is that it may largely reflect the variability of S1 response. There is some evidence to suggest that ERP components evaluated here may be underlied by dissociable neural substrates (Grunwald et al., 2003).

318

A.P. Anokhin et al. / Schizophrenia Research 89 (2007) 312–319

Moreover, a recent study has shown that P50 and N1 suppression correlate with distinct dimensions of a selfreport measure of sensory experiences (Perceptual Modulation and Over-Inclusion, respectively) (Kisley et al., 2004). Therefore, there is a possibility that these ERP components may represent relatively independent factors of risk for psychopathology. Elucidation of specific neuronal processes underlying gating mechanism at different stages of sensory processing, as well as specific genes involved in the determination of individual differences should be a matter of future studies. In summary, the present study showed substantial heritability of the peak amplitudes of P50, N1, and P2 waves. However, heritability of P50 gating ratio index (S2/S1) was low and it was difficult to distinguish between genetic and shared environmental transmission. Genetic influences on the gating measures derived from the N1/P2 wave complex were high and significant, suggesting that these measures may be useful endophenotypes for population-based genetic studies of the gating function and its impairments in psychopathology. Acknowledgements This work was supported by the grants DA00421 and DA018899 from the National Institute on Drug Abuse. This work was also partially supported by grant MH58784 from the National Institute of Mental Health. The authors thank Dr. Fred Struve for his contribution to the rating of the ERP components, Dr. Simon Golosheykin for his assistance with ERP analyses, and Erin Myers for her role in the data collection. Appendix A. Supplementary methods Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j. schres.2006.08.009. References Adler, L.E., Freedman, R., Ross, R.G., A.Waldo, M.C., 1999. Elementary phenotypes in the neurobiological and genetic study of schizophrenia. Biol. Psychiatry 46 (1), 8–18. Arnfred, S.M., Chen, A.C., Glenthoj, B.Y., Hemmingsen, R.P., 2003. Normal p50 gating in unmedicated schizophrenia outpatients. Am. J. Psychiatry 160 (12), 2236–2238. Blumenfeld, L.D., Clementz, B.A., 2001. Response to the first stimulus determines reduced auditory evoked response suppression in schizophrenia: single trials analysis using MEG. Clin. Neurophysiol. 112 (9), 1650–1659. Boutros, N.N., Belger, A., 1999. Midlatency evoked potentials attenuation and augmentation reflect different aspects of sensory gating. Biol. Psychiatry 45 (7), 917–922.

Boutros, N.N., Overall, J., Zouridakis, G., 1991. Test–retest reliability of the P50 mid-latency auditory evoked response. Psychiatry Res. 39 (2), 181–192. Boutros, N.N., Belger, A., Campbell, D., D'Souza, C., Krystal, J., 1999. Comparison of four components of sensory gating in schizophrenia and normal subjects: a preliminary report. Psychiatry Res. 88 (2), 119–130. Boutros, N.N., Korzyuko, O., Oliwa, G., Feingold, A., Campbell, D., McClain-Furmanski, D., Struve, F., Jansen, B.H., 2004a. Morphological and latency abnormalities of the mid-latency auditory evoked responses in schizophrenia: a preliminary report. Schizophr. Res. 70 (2–3), 303–313. Boutros, N.N., Korzyukov, O., Jansen, B., Feingold, A., Bell, M., 2004b. Sensory gating deficits during the mid-latency phase of information processing in medicated schizophrenia patients. Psychiatry Res. 126 (3), 203–215. Braff, D.L., Geyer, M.A., 1990. Sensorimotor gating and schizophrenia. Human and animal model studies. Arch. Gen. Psychiatry 47 (2), 181–188. Braff, D.L., Light, G.A., 2004. Preattentional and attentional cognitive deficits as targets for treating schizophrenia. Psychopharmacology (Berl) 174 (1), 75–85. Bramon, E., Rabe-Hesketh, S., Sham, P., Murray, R.M., Frangou, S., 2004. Meta-analysis of the P300 and P50 waveforms in schizophrenia. Schizophr. Res. 70 (2–3), 315–329. Buchsbaum, M.S., 1977. The middle evoked response components and schizophrenia. Schizophr. Bull. 3 (1), 93–104. Cadenhead, K.S., Light, G.A., Geyer, M.A., Braff, D.L., 2000. Sensory gating deficits assessed by the P50 event-related potential in subjects with schizotypal personality disorder. Am. J. Psychiatry 157 (1), 55–59. Cardenas, V.A., Gerson, J., Fein, G., 1993. The reliability of P50 suppression as measured by the conditioning/testing ratio is vastly improved by dipole modeling. Biol. Psychiatry 33 (5), 335–344. Clementz, B.A., Blumenfeld, L.D., 2001. Multichannel electroencephalographic assessment of auditory evoked response suppression in schizophrenia. Exp. Brain Res. 139 (4), 377–390. Clementz, B.A., Geyer, M.A., Braff, D.L., 1997. P50 suppression among schizophrenia and normal comparison subjects: a methodological analysis. Biol. Psychiatry 41 (10), 1035–1044. Croft, R.J., Lee, A., Bertolot, J., Gruzelier, J.H., 2001. Associations of P50 suppression and desensitization with perceptual and cognitive features of “unreality” in schizotypy. Biol. Psychiatry 50 (6), 441–446. Curtis, L., Blouin, J.L., Radhakrishna, U., Gehrig, C., Lasseter, V.K., Wolyniec, P., Nestadt, G., Dombroski, B., Kazazian, H.H., Pulver, A.E., Housman, D., Bertrand, D., Antonarakis, S.E., 1999. No evidence for linkage between schizophrenia and markers at chromosome 15q13-14. Am. J. Med. Genet. 88 (2), 109–112. Davis, H., Mast, T., Yoshie, N., Zerlin, S., 1966. The slow response of the human cortex to auditory stimuli: recovery process. Electroencephalogr. Clin. Neurophysiol. 21 (2), 105–113. Fein, G., Biggins, C., MacKay, S., 1996. Cocaine abusers have reduced auditory P50 amplitude and suppression compared to both normal controls and alcoholics. Biol. Psychiatry 39 (11), 955–965. Freedman, R., Adler, L.E., Waldo, M.C., Pachtman, E., Franks, R.D., 1983. Neurophysiological evidence for a defect in inhibitory pathways in schizophrenia: comparison of medicated and drug-free patients. Biol. Psychiatry 18 (5), 537–551. Freedman, R., Adams, C.E., Adler, L.E., Bickford, P.C., Gault, J., Harris, J.G., Nagamoto, H.T., Olincy, A., Ross, R.G., Stevens, K. E., Waldo, M., Leonard, S., 2000. Inhibitory neurophysiological

A.P. Anokhin et al. / Schizophrenia Research 89 (2007) 312–319 deficit as a phenotype for genetic investigation of schizophrenia. Am. J. Med. Genet. 97 (1), 58–64. Freedman, R., Olincy, A., Ross, R.G., Waldo, M.C., Stevens, K.E., Adler, L.E., Leonard, S., 2003. The genetics of sensory gating deficits in schizophrenia. Curr. Psychiatry Rep. 5 (2), 155–161. Grunwald, T., Boutros, N.N., Pezer, N., von Oertzen, J., Fernandez, G., Schaller, C., Elger, C.E., 2003. Neuronal substrates of sensory gating within the human brain. Biol. Psychiatry 53 (6), 511–519. Hetrick, W.P., Sandman, C.A., Bunney, W.E., Jin, Y., Potkin, S.G., White, M.H., 1996. Gender differences in gating of the auditory evoked potential in normal subjects. Biol. Psychiatry 39 (1), 51–58. Jin, Y., Bunney, W.E., Sandman, C.A., Patterson, J.V., Fleming, K., Moenter, J.R., Kalali, A.H., Hetrick, W.P., Potkin, S.G., 1998. Is P50 suppression a measure of sensory gating in schizophrenia? Biol. Psychiatry 43 (12), 873–878. Johannesen, J.K., Kieffaber, P.D., O'Donnell, B.F., Shekhar, A., Evans, J.D., Hetrick, W.P., 2005. Contributions of subtype and spectral frequency analyses to the study of P50 ERP amplitude and suppression in schizophrenia. Schizophr. Res. 78 (2–3), 269–284. Kisley, M.A., Noecker, T.L., Guinther, P.M., 2004. Comparison of sensory gating to mismatch negativity and self-reported perceptual phenomena in healthy adults. Psychophysiology 41 (4), 604–612. Li, C.H., Liao, H.M., Chen, C.H., 2004. Identification of molecular variants at the promoter region of the human alpha 7 neuronal nicotinic acetylcholine receptor subunit gene but lack of association with schizophrenia. Neurosci. Lett. 372 (1–2), 1–5. Marco, J., Fuentemilla, L., Grau, C., 2005. Auditory sensory gating deficit in abstinent chronic alcoholics. Neurosci. Lett. 375 (3), 174–177. Myles-Worsley, M., Coon, H., Byerley, W., Waldo, M., Young, D., Freedman, R., 1996. Developmental and genetic influences on the P50 sensory gating phenotype. Biol. Psychiatry 39 (4), 289–295.

319

Nagamoto, H.T., Adler, L.E., Waldo, M.C., Freedman, R., 1989. Sensory gating in schizophrenics and normal controls: effects of changing stimulation interval. Biol. Psychiatry 25 (5), 549–561. Neale, M.C., Cardon, L.R., 1992. Methodology for Genetic Studies of Twins and Families. Kluwer Academic Publishers, Dordrecht. Neves-Pereira, M., Bassett, A.S., Honer, W.G., Lang, D., King, N.A., Kennedy, J.L., 1998. No evidence for linkage of the CHRNA7 gene region in Canadian schizophrenia families. Am. J. Med. Genet. 81 (5), 361–363. Patrick, G., Struve, F.A., 2000. Reduction of auditory P50 gating response in marihuana users: further supporting data. Clin. Electroencephalogr. 31 (2), 88–93. Rijsdijk, F.V., Sham, P.C., 2002. Analytic approaches to twin data using structural equation models. Brief. Bioinform. 3 (2), 119–133. Roth, W.T., Horvath, T.B., Pfefferbaum, A., Kopell, B.S., 1980. Eventrelated potentials in schizophrenics. Electroencephalogr. Clin. Neurophysiol. 48 (2), 127–139. Smith, D.A., Boutros, N.N., Schwarzkopf, S.B., 1994. Reliability of P50 auditory event-related potential indices of sensory gating. Psychophysiology 31 (5), 495–502. Venables, P., 1964. Input dysfunction in schizophrenia. In: Maher, B.A. (Ed.), Progress in Experimental Personality Research. Academic Press, New York, pp. 1–47. Young, D.A., Waldo, M., Rutledge 3rd, J.H., Freedman, R., 1996. Heritability of inhibitory gating of the P50 auditory-evoked potential in monozygotic and dizygotic twins. Neuropsychobiology 33 (3), 113–117.