Key PointsQuestion
How robust are 40-Hz auditory steady-state response (ASSR) impairments in schizophrenia across different patient samples and paradigm designs?
Findings
In this meta-analysis of the 40-Hz ASSR in schizophrenia, including 20 studies and a total of 606 patients with schizophrenia and 590 healthy controls, 26 of 29 reported effects showed a reduction in ASSR measures in patients with schizophrenia compared with healthy controls.
Meaning
The 40-Hz ASSR spectral power and phase-locking deficits are robust in schizophrenia, suggesting that these measures could be useful probes for assessing circuit dysfunctions in the disorder.
Importance
The neurobiological mechanisms underlying circuit dysfunctions in schizophrenia remain poorly understood. The 40-Hz auditory steady-state response (ASSR) has been suggested as a potential biomarker for schizophrenia.
Objectives
To provide a meta-analytical insight into the presence of 40-Hz ASSR impairments in patients with schizophrenia and to examine the effects of the participant group, stimulus parameters, and analysis and recording techniques.
Data Sources
Searches were conducted in PubMed and reference lists of appropriate publications to identify relevant studies published from November 1999 to March 2016. Initial literature searches were performed with combinations of the following search terms: (1) auditory steady state response, (2) schizophrenia, (3) 40 Hz, (4) EEG, (5) MEG, and (6) steady state response.
Study Selection
Original articles reporting 40-Hz ASSR data on patients with schizophrenia (chronic or first episode) compared with healthy controls using electroencephalographic (EEG) and magnetoencephalographic (MEG) recordings.
Data Extraction and Synthesis
Hedges g effect sizes were calculated using sample sizes, P values, and/or Cohen d effect sizes from 20 studies. Effect size data were pooled using random-effects models. Publication bias was corrected for using funnel plots, the Egger regression test, and a trim and fill test. The contributions of study design parameters and participant characteristics were assessed using a mixed linear model approach and subsequent post hoc t tests. The present analysis was performed during the period from November 2015 to March 2016.
Main Outcomes and Measures
Random model Hedges g effect sizes for auditory steady-state amplitude and phase-locking measures from sensor/electrode and sources-space responses in EEG and MEG studies.
Results
Of the 20 studies analyzed (representing a total of 590 healthy controls and 606 patients with schizophrenia), 17 reported significant reductions in 40-Hz ASSR spectral power and/or phase locking in patients with schizophrenia compared with healthy controls (Hedges g effect: −0.58 [power] and −0.46 [phase]). Effect sizes from spectral power and phase-locking measures did not differ significantly (95% CI, −0.49 to 0.22; t = −0.80; P = .43). Stimulus characteristics and analysis methods were not associated with the findings of 40-Hz ASSR impairment in schizophrenia.
Conclusions and Relevance
The 40-Hz ASSR spectral power and phase-locking deficits are robust in schizophrenia, which suggests that these measures could be useful probes for assessing circuit dysfunctions in the disorder. Moreover, these findings should motivate large-scale studies of the longitudinal expression in patients with schizophrenia and at-risk populations, to further validate the 40-Hz ASSR as a potential biomarker.
Steady-state responses (SSRs) are evoked oscillatory responses that are entrained to the frequency and phase of temporally modulated stimuli. Steady-state responses can be detected noninvasively using electroencephalographic (EEG) and magnetoencephalographic (MEG) recordings in different sensory modalities. Auditory SSRs (ASSRs) show a peak frequency at around 40 Hz,1 whereas, in the visual domain, SSRs show clear peaks at around 10 Hz.2,3 Although the generation of the 40-Hz ASSR has been explained as the mere linear summation of transient event-related potentials (ERPs),4-6 dissimilarities between the ASSR and the synthetized ASSR, reconstructed from convoluted ERPs, suggest that the ASSR reflects the synchronization of endogenous oscillations.7,8 In this view, 40-Hz ASSRs reflect the propensity of neurons to oscillate at a particular gamma-band “resonant” frequency induced by external periodic stimulation. Supporting different neural mechanisms underlying ASSR and ERP generation, MEG studies locate the generators of the 40-Hz ASSR in medial areas of the primary auditory cortex, distinct from those underlying transient auditory components.9,10 In addition, EEG studies11 and multimodal imaging techniques more sensitive to distal generators have described subcortical and cerebellar contributions,12-14 along with a consistent right-hemispheric lateralization.15 Although the functional relevance is not fully understood, the 40-Hz ASSR has been suggested as a candidate mechanism underlying the fast temporal integration and resolution of auditory inputs.16-18
One important application of the 40-Hz ASSR has been in schizophrenia research because of the potential importance of gamma-band (30-200 Hz) oscillations in explaining cognitive deficits in the disorder.19,20 Gamma-band oscillations have been hypothesized to establish communication between distributed neuronal ensembles,21 the disturbances of which could underlie the pronounced cognitive and perceptual alterations in patients with schizophrenia.22 Moreover, this view is consistent with data indicating that rhythm-generating parvalbumin γ-aminobutyric acid (GABA) interneurons and N-methyl-d-aspartate (NMDA) receptors are dysfunctional in patients with schizophrenia,23,24 highlighting the potential of using noninvasively measured gamma-band oscillations to provide insights into circuit abnormalities in the disorder.
To our knowledge, Kwon et al25 conducted the first 40-Hz ASSR study for patients with schizophrenia using EEG and reported reduced power and synchronization to 40-Hz stimulation. These initial findings have been replicated by several groups26 in EEG27-40 and MEG recordings.41-46 In addition, there is evidence that ASSR deficits may also occur at theta-36,44 and beta-band frequencies.42
Despite the possibility that the 40-Hz ASSR could constitute a potential biomarker for schizophrenia, the current findings have not been systematically evaluated in a meta-analysis. Therefore, the primary goal of this article is to establish the robustness of ASSR impairment across populations of patients with schizophrenia and to identify potential contributions of differences in recording and analysis parameters, as well as patient characteristics. Most ASSR studies use a similar task design, but there are differences in stimulus type, reported outcome measures, attentional load, and stimulus duration.26,39,47 There are also important differences in analysis and recording techniques,26 the impact of which is unclear. Finally, the majority of 40-Hz ASSR studies have focused on patients with chronic schizophrenia, and only a small number of groups have examined the first episode of psychosis and at-risk populations,32,40 early-onset schizophrenia,43 and first-degree relatives,28,34 raising the question whether differences exist in oscillatory responses across different stages of illness.
Searches were conducted in PubMed and reference lists of appropriate publications to identify relevant studies published from November 1999 to March 2016. Initial literature searches in PubMed were performed with combinations of the following search terms: (1) auditory steady state response, (2) schizophrenia, (3) 40 Hz, (4) EEG, (5) MEG, and (6) steady state response. The search yielded 42 potential articles that were considered for inclusion in the present meta-analysis if they met the following criteria: they were human studies, presented new data (ie, no reviews), used EEG or MEG to measure ASSRs, included sufficient statistical information (sample sizes and mean values and/or raw data and/or P values and/or effect sizes), included at least 1 sample of patients with schizophrenia and 1 sample of healthy controls, and reported measures of evoked spectral power and/or intertrial phase coherence (ITPC, also referred to as phase-locking factor), the latter representing stimulus evoked phase consistency across trials.48
Based on these criteria, 25 studies were excluded (reviews [n = 5], animal studies [n = 6], studies without patients with schizophrenia or studies that examined another sensory modality [n = 8], studies that did not report measures of spectral power and/or phase synchrony [n = 3], and studies that used a sample already included in the analysis [n = 3]). As a result, we identified 17 studies, 2 reports obtained by searches through the reference sections in the original 42 articles, and 1 study highlighted by a reviewer (for a total sample of 20 studies; Table 1). Three studies did not report relevant statistics for 1 or both outcome measures, but this information was provided to us by the authors.31,34,44
Thirteen articles from the main analysis and 1 excluded article29 investigated ASSRs to 20-Hz stimuli in addition to the 40-Hz stimulation. Effect size data were obtained from 11 of these studies and were analyzed in an exploratory manner. Where the direction of effect/no effect was not explicitly specified, the effect directionality was derived by examining the data figures. Furthermore, some of the studies reported induced power in addition to evoked measures.31,36-38,43,46 However, owing to the limited literature currently available, these data were not explored in the present analysis.
Effect size calculations were performed using the Comprehensive Meta-Analysis software (version 3.3.070). Hedges g effect sizes (mean1 − mean2/SDpooled) were calculated for each measure (spectral power and phase measures) in each study using sample sizes for each group and either a P value or the Cohen d value. The resulting effect sizes were plotted, and an overall random-effect size estimated. Some studies reported statistics from several conditions or time windows.39,40 In these cases, overall effects across conditions were included. In addition, secondary analyses were performed to investigate the effect of outliers more than 2 SD away from the mean effect size. Hedges g effect sizes were chosen for the comparison because the pooled standard deviation used to calculate Hedges g values is weighted by the number of participants in each group,49 making the measure less positively biased than the Cohen d or the Glass Δ and, therefore, suitable for measuring effect sizes in small samples.50
To control for reporting bias, effect size data were inspected visually using funnel plots. The theory underlying this approach is that small studies have greater standard errors.51 This was investigated using a funnel plot that is symmetrical if there is no bias. The degree of funnel-plot asymmetry was evaluated using the Egger regression test.52 To provide an estimation of the possible effect of unpublished studies, effect sizes were corrected using the “trim and fill” method.53 In 9 studies in which both phase and power measures were reported, we only included power effects in this analysis (resulting in 15 studies with power measures and 5 studies with phase measures). This approach was justified through our observation that there was no significant difference between power and phase effect sizes.
Further analyses of the effect of sample and study design were performed using R54 in Rstudio.55 Only 1 effect size was included from each article (15 studies with power measures and 5 studies with phase measures). P < .05 is considered to be statistically significant.
The impact of selected variables was evaluated by a mixed linear model approach using the R package nlme.56 Owing to the small overall sample size, the model was restricted to 4 independent fixed-effects variables and 1 random-effects variable. Because the majority of studies using MEG reported source-level data and the majority of studies using EEG reported sensor/electrode-level data, only one of these factors—analysis level (sensor/source)—was included in the model. The other selected variables were patient age, stimulus type (click trains vs amplitude-modulated tones), and stimulus duration. The R2 values of the model were calculated using the R package MuMIn,57 which allows for the calculation of R2 values adapted for mixed linear models.58
Because only 3 studies included patients who had a first episode of psychosis or early-onset patients, a comparison between patients who had a first episode of psychosis and patients with chronic schizophrenia was not possible. Instead, the sample was divided into 2 age groups based on the median age (39.8 years). Likewise, because only 5 different stimulus durations have been reported (475, 500, 1000, 1024, and 1500 milliseconds), durations were treated as categorical variables in the mixed linear model and were defined as either brief (≤500 milliseconds) or long (≥1000 milliseconds) in post hoc analyses. One study reported 2 conditions of stimulus duration,39 and these were both included in the post hoc t test evaluation of the effects of stimulus duration.
The combined samples included a total of 606 patients with schizophrenia and 590 healthy controls (Table 1 and Table 2). In total, 15 studies included 40-Hz ASSR evoked power measures, 14 studies included 40-Hz ASSR phase measures (reported as ITPC or phase-locking factor), and 9 studies included both measures.
Effect size calculations revealed a range of Hedges g values from 0.69 to −1.50.33,41 In total, 3 effect sizes were greater than 0, suggesting an increase in patients with schizophrenia compared with healthy controls (g = 0.20 for spectral power in Hong et al28; g = 0.53 for spectral power and g = 0.69 for phase locking in Hamm et al33). One of these effects was statistically significant (ie, phase locking in Hamm et al33). The remaining 26 negative effect sizes reflect a reduction in ASSR measures in schizophrenia. The average Hedges g random-effect size was −0.58 for power measures and −0.46 for phase measures (Figure 1), indicating a moderately strong effect. Moreover, the initial analysis revealed that 3 effect sizes were more than 2 SD from the total mean effect size (ie, the total [SD] mean effect size was −0.55 [0.46]), and, therefore, they were treated as outliers (both power and phase in Hamm et al33 and power in Vierling-Claassen et al41). Without these outliers, the Hedges g random-effect sizes were −0.45 (13 studies with phase measures) and −0.59 (13 studies with power measures), suggesting a robust effect also in the absence of outliers.
The initial statistical comparison revealed that the effect sizes from spectral power and phase-locking measures did not differ significantly (95% CI, −0.49 to 0.22; t = −0.80; P = .43), justifying the inclusion of 1 single effect size from each study (15 studies with power measures and 5 studies with phase measures).
With regard to the differences between healthy controls and patients with schizophrenia, the mixed linear model revealed a significant effect of patient age (P = .03) and a strong trend toward a significant effect of stimulus duration (P = .05) on the 40-Hz ASSR measures. There was, however, no effect of stimulus type (P = .40) or analysis level (P = .79). The conditional R2 of this model was 0.94, indicating that the fixed and random variables combined explained more than 90% of the variance. The R2 of the fixed variables alone (marginal R2) was 0.54. When the same model was used to evaluate the data set without outliers,33,41 only the effect of age remained (patient age: P = .04; stimulus duration: P = .34; stimulus type: P = .27; analysis level: P = .84).
Post hoc t tests were used to investigate these results further. None of the effects remained statistically significant, but there was a slight trend toward stronger ASSR reductions in studies with younger participants and in studies with shorter stimulus durations. The group average Hedges g effect sizes for these group comparisons are as follows: −0.40 for patients older than 39.8 years of age and −0.67 for patients 39.8 years of age or younger (t = 1.41; P = .18; 95% CI, −0.14 to 0.69), −0.37 for long stimulus duration and −0.72 for brief stimulus duration (t = 1.50; P = .17; 95% CI, −0.18 to 0.90), −0.66 for click stimuli and −0.47 for amplitude-modulated tones (t = −0.83; P = .42; 95% CI, −0.70 to 0.32), and −0.53 for sensor level and −0.73 for source level (t = 0.75; P = .48; 95% CI, −0.42 to 0.82) (Figure 2). One outlier each was identified for the post hoc evaluations of the effects of analysis level33 and stimulus duration,28 and 2 for the age comparison,28,33 but removing these studies from the respective analyses did not alter the outcome. For secondary analysis without outliers, the group average Hedges g effect sizes for these group comparisons are as follows: −0.52 for patients older than 39.8 years of age and −0.77 for patients 39.8 years of age or younger (t = 1.85; P = .09; 95% CI, −0.04 to 0.56), −0.37 for long stimulus duration and −0.79 for brief stimulus duration (t = 1.82; P = .11; 95% CI, −0.11 to 0.96), and −0.66 for sensor level and −0.63 for source level (t = 0.16; P = .88; 95% CI, −0.40 to 0.35).
Plotting one 40-HZ ASSR effect size from each study in a funnel plot (Figure 3A) revealed some asymmetry, suggesting that the sample could be affected by reporting bias (Egger regression test: t = 2.20; P = .04; 95% CI, −2.73 to −0.07). Using the “trim and fill” method of Duval and Tweedie,53 we estimated 5 hypothetically missing studies, which suggests a slight bias in favor of studies reporting ASSR impairments in patients with schizophrenia. The addition of these studies adjusted the overall Hedges g random-effect size to −0.42 (Figure 3B).
The 20-Hz ASSR effect sizes ranged from 0.95 to −0.49.28,41 The overall Hedges g random effect size for these studies was 0.006 (P = .95). In this sample of data points, there was no significant reporting bias as measured by the Egger regression test (t = 0.59; P = .53; 95% CI, −3.26 to 1.80), yet the trim and fill test indicated 2 missing studies. The addition of these studies adjusted the Hedges g random-effect size to −0.11, thus still indicating a very small effect.
The 40-Hz ASSR Impairments and Circuit Abnormalities in Schizophrenia
A key finding from the present meta-analysis is that there is consistent evidence for a robust reduction in both spectral power and phase-locking measures during the 40-Hz stimulation in patients with schizophrenia, highlighting both an impairment in the generation of high-frequency oscillations and the precise temporal coordination of rhythmic activity in response to entrainment of neural circuits. This pattern could potentially establish links with current data and models of circuit impairments in schizophrenia that have emphasized the contribution of deficits in parvalbumin GABAergic interneurons24 and/or dysfunctional NMDA receptors.23 The former has been shown to be mechanistically related to the emergence of gamma-band oscillations through optogenetic manipulations whereby a downregulation of parvalbumin cells leads to a reduction of rhythmic activity in the 40- to 70-Hz range.59 The hypofunctioning of NMDA receptors is associated with both an increase and a decrease of 40-Hz ASSRs depending on the level of channel occupancy of NMDA receptors.60 Together, these findings highlight the possibility that disturbances in excitation/inhibition-balance parameters are potential candidate mechanisms that might explain the 40-Hz ASSR impairments in patients with schizophrenia, which could provide important links to preclinical research.
The mixed linear model analysis showed an effect of stimulus duration on the pattern of 40-Hz ASSR effects, but this effect did not survive in the post hoc analysis. Moreover, any effects of stimulus type on power and ITPC measures were not found to contribute to 40-Hz ASSR impairments in schizophrenia. However, data from healthy controls suggest potential differences in some of these parameters. Notably, click-train stimuli have been associated with more robust responses than white-noise stimuli, and ITPC measures have shown improved test-retest reliability compared with spectral power measures.47 In light of these data, careful consideration of experimental parameters could optimize the ASSR measure across groups.
Analysis Level and Neuroimaging Method
An important question is whether effect sizes vary depending on the approach of analysis (source or sensor level), as well as the imaging technique (EEG or MEG) used. An initial review of the studies revealed that the analysis approach was closely linked to the neuroimaging technique used. All but one MEG study45 presented source-level data, and all but one EEG study38 presented sensor-level data. Although there was no significant effect size difference between source- and sensor-level studies, this difference has implications for the type of information available from MEG and EEG studies so far. Reconstruction of the source of the underlying generators of the 40-Hz responses could yield insights into the underlying brain regions and networks involved in the ASSR deficits in patients with schizophrenia. Moreover, evidence from Tan et al61 suggests that an MEG-informed reconstruction of the source significantly enhances signal-to-noise estimates for 40-Hz ASSR estimates during normal brain functioning. Finally, evidence from studies of healthy populations62 (comparing the sensitivity of EEG vs MEG for the measurement of gamma-band oscillations) have highlighted the improved detectability of high-frequency activity using MEG measurements.
Our analysis indicates a trend toward more pronounced differences in the 40-Hz ASSRs in younger patients with schizophrenia compared with older patients. This contrasts with existing evidence that has highlighted progressive reductions in EEG63 and magnetic resonance imaging parameters64 during the course of the disorder. However, the majority of current 40-Hz ASSR studies have been conducted in patients with chronic schizophrenia, and further research is required to examine the pattern and strength of ASSR deficits in patients with a first episode of psychosis and in at-risk populations. To date, one study has investigated ASSRs in participants at ultrahigh risk for psychosis, demonstrating the presence of reduced 40-Hz ASSRs41 and, thus, providing tentative support for the potential utility of the 40-Hz ASSR as a biomarker for early detection and diagnosis of schizophrenia.
Specificity of 40-Hz ASSR Deficits
Our explorative analysis of 20-Hz stimulation ASSR data showed no effect (Hedges g random effect of 0.006; P = .95), indicating that beta-range ASSRs are largely intact in patients with schizophrenia. Only one of the analyzed studies reported a statistically significant difference between patients with schizophrenia and healthy controls, with the patients with schizophrenia showing higher 20-Hz ASSR spectral power than healthy controls.41 In addition, Hamm et al44 examined ASSRs at a wide range of modulation frequencies (5, 20, 40, 80, and 160 Hz) and found that the 5-Hz ASSR was characterized by an improved sensitivity (compared with those in the higher-frequency ranges) to distinguish between healthy controls and patients with schizophrenia. Moreover, there is also evidence that both the amplitude and the phase locking at a 80-Hz ASSR stimulation is impaired in patients with schizophrenia,44,45 raising the question of the specificity of 40-Hz ASSR deficits.
The 40-Hz ASSR abnormalities do not appear to be unique to patients with schizophrenia. First, there is emerging evidence that ASSRs may constitute an endophenotype for schizophrenia because the 40-Hz ASSR impairments are present in first-degree relatives.28,34 In addition, a small number of studies have investigated the ASSR in other patient groups, providing evidence for similar patterns of ASSR power- and phase-related impairments in patients with autism spectrum disorders65 or bipolar disorder,66-68 whereas ASSR deficits are not observed in patients with unipolar depression.68 The overlap of ASSR impairments in schizophrenia, autism spectrum disorders, and bipolar disorder is in line with the notion that these conditions may share phenotypic, structural, and genetic characteristics.69-71
Functional and Clinical Correlates
One important point concerning the 40-Hz ASSR findings in patients with schizophrenia is a possible link between cognitive deficits and clinical symptoms. Hamm et al45 found a negative correlation between high gamma (80-Hz) ASSR amplitude and negative symptoms. Spencer et al32 demonstrated a positive correlation between 40-Hz ASSR ITPC and positive symptoms in the left auditory cortex, whereas others have found a negative association between left-hemisphere high gamma (80-Hz) ASSR power and hallucinations.46
Given the role of gamma-band oscillations in facilitating cognition and perception,20,21 correlations may be expected between the 40-Hz ASSR impairments and cognitive deficits in patients with schizophrenia. Kirihara et al36 examined the relationship between ASSR measures and cognition in a large sample of patients with schizophrenia and found that only total theta (4-8 Hz) amplitude reductions were correlated with deficits in verbal memory in patients with schizophrenia, whereas no relationship was found with 40-Hz ASSRs. Accordingly, further studies need to systematically test the relationship between cognitive deficits and 40-Hz ASSR impairments in patients with schizophrenia.
Limitations and Future Directions
The current modest number of studies investigating the 40-Hz ASSR deficit in patients with schizophrenia requires further extension and replication. Specifically, the effect size of the ASSR deficit is currently lower than other electrophysiological indices of auditory dysfunctions in schizophrenia, such as the mismatch negativity50 (Hedges g = 0.95 for all patients with schizophrenia and Hedges g = 0.81 for patients with chronic schizophrenia) and P5072 (Cohen d = 1.28). Nonetheless, the current data warrant further investigations that should systematically examine the influence of recording and analysis parameters to improve effect sizes. Based on the current data, our meta-analysis can only provide preliminary findings with regard to the potential impact of recording techniques (EEG vs MEG) and analysis parameters (source vs sensor). In addition, further relationships with cognitive deficits, as well as clinical parameters, in patients with schizophrenia would significantly enhance the utility of the 40-Hz ASSR as an important index of auditory circuit functions in patients with schizophrenia. Finally, emerging evidence suggests that aberrant, intrinsic high-frequency activity could affect the generation of 40-Hz ASSRs,38 which could have important implications for the understanding of the underlying mechanisms of auditory circuit dysfunctions in patients with schizophrenia.
This systematic meta-analysis of the 40-Hz ASSR in patients with schizophrenia demonstrates that there is a consistent impairment of both amplitude and phase-locking measures in patients with schizophrenia across different stages of illness. Accordingly, the 40-Hz ASSR measure could constitute a potentially useful biomarker that could shed light on the neurobiology of circuit dysfunctions in patients with schizophrenia.
Accepted for Publication: August 24, 2016.
Corresponding Author: Peter J. Uhlhaas, PhD, Institute of Neuroscience and Psychology, University of Glasgow, Hillhead Str 58, Glasgow G12 8QB, Scotland (peter.uhlhaas@glasgow.ac.uk).
Published Online: October 12, 2016. doi:10.1001/jamapsychiatry.2016.2619
Author Contributions: Ms Thuné had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: Thuné, Recasens.
Statistical analysis: Thuné, Uhlhaas.
Administrative, technical, or material support: Uhlhaas.
Study supervision: Recasens, Uhlhaas.
Conflict of Interest Disclosures: Dr Uhlhaas has received research support from Lilly UK. No other disclosures are reported.
Funding/Support: This study was supported by the project MR/L011689/1 from the Medical Research Council (MRC). Ms Thuné is supported by a PhD studentship from the MRC doctoral training programme.
Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank Jordan Hamm, PhD (Postdoctoral Research Fellow, Neurotechnology Center, Department of Biological Sciences, Columbia University, New York, New York), Brian F. O’Donnell, PhD (Professor, Department of Psychological and Brain Sciences, Indiana University, Bloomington), Giri P. Krishnan, PhD (Assistant Project Scientist, University of California, San Diego), and Olga Rass, PhD (Postdoctoral Research Fellow, Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, Maryland), for providing additional data and/or effect sizes from their research. No compensation was received from a funding sponsor for these contributions.
2.Fawcett
IP, Barnes
GR, Hillebrand
A, Singh
KD. The temporal frequency tuning of human visual cortex investigated using synthetic aperture magnetometry.
Neuroimage. 2004;21(4):1542-1553.
PubMedGoogle ScholarCrossref 3.Herrmann
CS. Human EEG responses to 1-100 Hz flicker: resonance phenomena in visual cortex and their potential correlation to cognitive phenomena.
Exp Brain Res. 2001;137(3-4):346-353.
PubMedGoogle ScholarCrossref 4.Galambos
R, Makeig
S, Talmachoff
PJA. A 40-Hz auditory potential recorded from the human scalp.
Proc Natl Acad Sci U S A. 1981;78(4):2643-2647.
PubMedGoogle ScholarCrossref 5.Presacco
A, Bohórquez
J, Yavuz
E, Ozdamar
O. Auditory steady-state responses to 40-Hz click trains: relationship to middle latency, gamma band and beta band responses studied with deconvolution.
Clin Neurophysiol. 2010;121(9):1540-1550.
PubMedGoogle ScholarCrossref 6.Capilla
A, Pazo-Alvarez
P, Darriba
A, Campo
P, Gross
J. Steady-state visual evoked potentials can be explained by temporal superposition of transient event-related responses.
PLoS One. 2011;6(1):e14543.
PubMedGoogle ScholarCrossref 7.Başar
E, Rosen
B, Başar-Eroglu
C, Greitschus
F. The associations between 40 Hz-EEG and the middle latency response of the auditory evoked potential.
Int J Neurosci. 1987;33(1-2):103-117.
PubMedGoogle ScholarCrossref 8.Ross
B. Steady-state auditory evoked responses. In: Celesia
GG, ed. Handbook of Clinical Neurophysiology: Disorders of Peripheral and Central Auditory Processing. Vol 10. Amsterdam: the Netherlands: Elsevier; 2013:155-176.
9.Pantev
C, Roberts
LE, Elbert
T, Ross
B, Wienbruch
C. Tonotopic organization of the sources of human auditory steady-state responses.
Hear Res. 1996;101(1-2):62-74.
PubMedGoogle ScholarCrossref 10.Herdman
AT, Wollbrink
A, Chau
W, Ishii
R, Ross
B, Pantev
C. Determination of activation areas in the human auditory cortex by means of synthetic aperture magnetometry.
Neuroimage. 2003;20(2):995-1005.
PubMedGoogle ScholarCrossref 11.Herdman
AT, Lins
O, Van Roon
P, Stapells
DR, Scherg
M, Picton
TW. Intracerebral sources of human auditory steady-state responses.
Brain Topogr. 2002;15(2):69-86.
PubMedGoogle ScholarCrossref 12.Pastor
MA, Artieda
J, Arbizu
J, Marti-Climent
JM, Peñuelas
I, Masdeu
JC. Activation of human cerebral and cerebellar cortex by auditory stimulation at 40 Hz.
J Neurosci. 2002;22(23):10501-10506.
PubMedGoogle Scholar 13.Reyes
SA, Lockwood
AH, Salvi
RJ, Coad
ML, Wack
DS, Burkard
RF. Mapping the 40-Hz auditory steady-state response using current density reconstructions.
Hear Res. 2005;204(1-2):1-15.
PubMedGoogle ScholarCrossref 14.Steinmann
I, Gutschalk
A. Potential fMRI correlates of 40-Hz phase locking in primary auditory cortex, thalamus and midbrain.
Neuroimage. 2011;54(1):495-504.
PubMedGoogle ScholarCrossref 15.Ross
B, Herdman
AT, Pantev
C. Right hemispheric laterality of human 40 Hz auditory steady-state responses.
Cereb Cortex. 2005;15(12):2029-2039.
PubMedGoogle ScholarCrossref 16.Ross
B, Picton
TW, Pantev
C. Temporal integration in the human auditory cortex as represented by the development of the steady-state magnetic field.
Hear Res. 2002;165(1-2):68-84.
PubMedGoogle ScholarCrossref 17.Ross
B, Pantev
C. Auditory steady-state responses reveal amplitude modulation gap detection thresholds.
J Acoust Soc Am. 2004;115(5, pt 1):2193-2206.
PubMedGoogle ScholarCrossref 18.Baltus
A, Herrmann
CS. Auditory temporal resolution is linked to resonance frequency of the auditory cortex.
Int J Psychophysiol. 2015;98(1):1-7.
PubMedGoogle ScholarCrossref 22.Uhlhaas
PJ, Singer
W. Oscillations and neuronal dynamics in schizophrenia: the search for basic symptoms and translational opportunities.
Biol Psychiatry. 2015;77(12):1001-1009.
PubMedGoogle ScholarCrossref 23.Kantrowitz
JT, Javitt
DC. N-methyl-
d-aspartate (NMDA) receptor dysfunction or dysregulation: the final common pathway on the road to schizophrenia?
Brain Res Bull. 2010;83(3-4):108-121.
PubMedGoogle ScholarCrossref 24.Lewis
DA, Curley
AA, Glausier
JR, Volk
DW. Cortical parvalbumin interneurons and cognitive dysfunction in schizophrenia.
Trends Neurosci. 2012;35(1):57-67.
PubMedGoogle ScholarCrossref 25.Kwon
JS, O’Donnell
BF, Wallenstein
GV,
et al. Gamma frequency-range abnormalities to auditory stimulation in schizophrenia.
Arch Gen Psychiatry. 1999;56(11):1001-1005.
PubMedGoogle ScholarCrossref 26.O’Donnell
BF, Vohs
JL, Krishnan
GP, Rass
O, Hetrick
WP, Morzorati
SL. The auditory steady-state response (ASSR): a translational biomarker for schizophrenia.
Suppl Clin Neurophysiol. 2013;62:101-112.
PubMedGoogle Scholar 27.Brenner
CA, Sporns
O, Lysaker
PH, O’Donnell
BF. EEG synchronization to modulated auditory tones in schizophrenia, schizoaffective disorder, and schizotypal personality disorder.
Am J Psychiatry. 2003;160(12):2238-2240.
PubMedGoogle ScholarCrossref 28.Hong
LE, Summerfelt
A, McMahon
R,
et al. Evoked gamma band synchronization and the liability for schizophrenia.
Schizophr Res. 2004;70(2-3):293-302.
PubMedGoogle ScholarCrossref 29.Light
GA, Hsu
JL, Hsieh
MH,
et al. Gamma band oscillations reveal neural network cortical coherence dysfunction in schizophrenia patients.
Biol Psychiatry. 2006;60(11):1231-1240.
PubMedGoogle ScholarCrossref 30.Spencer
KM, Salisbury
DF, Shenton
ME, McCarley
RW. γ-band auditory steady-state responses are impaired in first episode psychosis.
Biol Psychiatry. 2008;64(5):369-375.
PubMedGoogle ScholarCrossref 31.Krishnan
GP, Hetrick
WP, Brenner
CA, Shekhar
A, Steffen
AN, O’Donnell
BF. Steady state and induced auditory gamma deficits in schizophrenia.
Neuroimage. 2009;47(4):1711-1719.
PubMedGoogle ScholarCrossref 32.Spencer
KM, Niznikiewicz
MA, Nestor
PG, Shenton
ME, McCarley
RW. Left auditory cortex gamma synchronization and auditory hallucination symptoms in schizophrenia.
BMC Neurosci. 2009;10(1):85.
PubMedGoogle ScholarCrossref 33.Hamm
JP, Gilmore
CS, Clementz
BA. Augmented gamma band auditory steady-state responses: support for NMDA hypofunction in schizophrenia.
Schizophr Res. 2012;138(1):1-7.
PubMedGoogle ScholarCrossref 34.Rass
O, Forsyth
JK, Krishnan
GP,
et al. Auditory steady state response in the schizophrenia, first-degree relatives, and schizotypal personality disorder.
Schizophr Res. 2012;136(1-3):143-149.
PubMedGoogle ScholarCrossref 35.Kömek
K, Ementrout
GB, Walker
CP, Cho
RY. Dopamine and gamma band synchrony in schizophrenia—insights from computational and empirical studies.
Eur J Neurosci. 2012;36(2):2146-2155.
PubMedGoogle ScholarCrossref 36.Kirihara
K, Rissling
AJ, Swerdlow
NR, Braff
DL, Light
GA. Hierarchical organization of gamma and theta oscillatory dynamics in schizophrenia.
Biol Psychiatry. 2012;71(10):873-880.
PubMedGoogle ScholarCrossref 37.Roach
BJ, Ford
JM, Hoffman
RE, Mathalon
DH. Converging evidence for gamma synchrony deficits in schizophrenia.
Suppl Clin Neurophysiol. 2013;62:163-180.
PubMedGoogle Scholar 38.Hirano
Y, Oribe
N, Kanba
S, Onitsuka
T, Nestor
PG, Spencer
KM. Spontaneous gamma activity in schizophrenia.
JAMA Psychiatry. 2015;72(8):813-821.
PubMedGoogle ScholarCrossref 39.Hamm
JP, Bobilev
AM, Hayrynen
LK,
et al. Stimulus train duration but not attention moderates γ-band entrainment abnormalities in schizophrenia.
Schizophr Res. 2015;165(1):97-102.
PubMedGoogle ScholarCrossref 40.Tada
M, Nagai
T, Kirihara
K,
et al. Differential alterations of auditory gamma oscillatory responses between pre-onset high-risk individuals and first-episode schizophrenia.
Cereb Cortex. 2016;26(3):1027-1035.
PubMedGoogle ScholarCrossref 41.Vierling-Claassen
D, Siekmeier
P, Stufflebeam
S, Kopell
N. Modeling GABA alterations in schizophrenia: a link between impaired inhibition and altered gamma and beta range auditory entrainment.
J Neurophysiol. 2008;99(5):2656-2671.
PubMedGoogle ScholarCrossref 42.Wilson
TW, Hernandez
OO, Asherin
RM, Teale
PD, Reite
ML, Rojas
DC. Cortical gamma generators suggest abnormal auditory circuitry in early-onset psychosis.
Cereb Cortex. 2008;18(2):371-378.
PubMedGoogle ScholarCrossref 43.Teale
P, Collins
D, Maharajh
K, Rojas
DC, Kronberg
E, Reite
M. Cortical source estimates of gamma band amplitude and phase are different in schizophrenia.
Neuroimage. 2008;42(4):1481-1489.
PubMedGoogle ScholarCrossref 44.Hamm
JP, Gilmore
CS, Picchetti
NAM, Sponheim
SR, Clementz
BA. Abnormalities of neuronal oscillations and temporal integration to low- and high-frequency auditory stimulation in schizophrenia.
Biol Psychiatry. 2011;69(10):989-996.
PubMedGoogle ScholarCrossref 45.Tsuchimoto
R, Kanba
S, Hirano
S,
et al. Reduced high and low frequency gamma synchronization in patients with chronic schizophrenia.
Schizophr Res. 2011;133(1-3):99-105.
PubMedGoogle ScholarCrossref 46.Edgar
JC, Chen
Y-H, Lanza
M,
et al. Cortical thickness as a contributor to abnormal oscillations in schizophrenia?
Neuroimage Clin. 2013;4:122-129.
PubMedGoogle ScholarCrossref 47.McFadden
KL, Steinmetz
SE, Carroll
AM, Simon
ST, Wallace
A, Rojas
DC. Test-retest reliability of the 40 Hz EEG auditory steady-state response.
PLoS One. 2014;9(1):e85748.
PubMedGoogle ScholarCrossref 48.Roach
BJ, Mathalon
DH. Event-related EEG time-frequency analysis: an overview of measures and an analysis of early gamma band phase locking in schizophrenia.
Schizophr Bull. 2008;34(5):907-926.
PubMedGoogle ScholarCrossref 49.Erickson
MA, Ruffle
A, Gold
JM. A meta-analysis of mismatch negativity in schizophrenia: from clinical risk to disease specificity and progression.
Biol Psychiatry. 2016;79(12):980-987.
PubMedGoogle ScholarCrossref 50.Enzmann
D. Notes on effect size measures for the difference of means from two independent groups: the case of Cohen’s d and Hedges’ g. January 12, 2015. Tech Rep. doi:
10.13140/2.1.1578.2725 51.Sterne
JAC, Sutton
AJ, Ioannidis
JPA,
et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials.
BMJ. 2011;343:d4002.
PubMedGoogle ScholarCrossref 52.Egger
M, Davey Smith
G, Schneider
M, Minder
C. Bias in meta-analysis detected by a simple, graphical test.
BMJ. 1997;315(7109):629-634.
PubMedGoogle ScholarCrossref 53.Duval
S, Tweedie
R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis.
Biometrics. 2000;56(2):455-463.
PubMedGoogle ScholarCrossref 55.Welcome to RStudio—open source and enterprise-ready professional software for R. RStudio website.
https://www.rstudio.com. Accessed September 8, 2016.
56.Pinheiro
J, Bates
D, DebRoy
S, Sarkar
D; R Core Team. Nlme: linear and nonlinear mixed effect models. R package version 3.1-122.
http://CRAN.r-project.org/web/packages/nlme. Published August 19, 2015. Accessed September 8, 2016.
59.Sohal
VS, Zhang
F, Yizhar
O, Deisseroth
K. Parvalbumin neurons and gamma rhythms enhance cortical circuit performance.
Nature. 2009;459(7247):698-702.
PubMedGoogle ScholarCrossref 60.Sivarao
DV, Chen
P, Senapati
A,
et al. 40 Hz auditory steady-state response is a pharmacodynamic biomarker for cortical NMDA receptors.
Neuropsychopharmacology. 2016;41(9):2232-2240.
PubMedGoogle ScholarCrossref 61.Tan
HR, Gross
J, Uhlhaas
PJ. MEG-measured auditory steady-state oscillations show high test-retest reliability: a sensor and source-space analysis.
Neuroimage. 2015;122:417-426.
PubMedGoogle ScholarCrossref 62.Muthukumaraswamy
SD, Singh
KD. Visual gamma oscillations: the effects of stimulus type, visual field coverage and stimulus motion on MEG and EEG recordings.
Neuroimage. 2013;69:223-230.
PubMedGoogle ScholarCrossref 63.Salisbury
DF, Shenton
ME, Griggs
CB, Bonner-Jackson
A, McCarley
RW. Mismatch negativity in chronic schizophrenia and first-episode schizophrenia.
Arch Gen Psychiatry. 2002;59(8):686-694.
PubMedGoogle ScholarCrossref 64.Kasai
K, Shenton
ME, Salisbury
DF,
et al. Progressive decrease of left Heschl gyrus and planum temporale gray matter volume in first-episode schizophrenia: a longitudinal magnetic resonance imaging study.
Arch Gen Psychiatry. 2003;60(8):766-775.
PubMedGoogle ScholarCrossref 65.Wilson
TW, Rojas
DC, Reite
ML, Teale
PD, Rogers
SJ. Children and adolescents with autism exhibit reduced MEG steady-state gamma responses.
Biol Psychiatry. 2007;62(3):192-197.
PubMedGoogle ScholarCrossref 66.Rass
O, Krishnan
G, Brenner
CA,
et al. Auditory steady state response in bipolar disorder: relation to clinical state, cognitive performance, medication status, and substance disorders.
Bipolar Disord. 2010;12(8):793-803.
PubMedGoogle ScholarCrossref 67.Oda
Y, Onitsuka
T, Tsuchimoto
R,
et al. Gamma band neural synchronization deficits for auditory steady state responses in bipolar disorder patients.
PLoS One. 2012;7(7):e39955.
PubMedGoogle ScholarCrossref 68.Isomura
S, Onitsuka
T, Tsuchimoto
R,
et al. Differentiation between major depressive disorder and bipolar disorder by auditory steady-state responses.
J Affect Disord. 2016;190:800-806.
PubMedGoogle ScholarCrossref 69.Hill
SK, Reilly
JL, Keefe
RSE,
et al. Neuropsychological impairments in schizophrenia and psychotic bipolar disorder: findings from the Bipolar-Schizophrenia Network on Intermediate Phenotypes (B-SNIP) study.
Am J Psychiatry. 2013;170(11):1275-1284.
PubMedGoogle ScholarCrossref 70.Radeloff
D, Ciaramidaro
A, Siniatchkin
M,
et al. Structural alterations of the social brain: a comparison between schizophrenia and autism.
PLoS One. 2014;9(9):e106539.
PubMedGoogle ScholarCrossref 72.de Wilde
OM, Bour
LJ, Dingemans
PM, Koelman
JH, Linszen
DH. A meta-analysis of P50 studies in patients with schizophrenia and relatives: differences in methodology between research groups.
Schizophr Res. 2007;97(1-3):137-151.
PubMedGoogle ScholarCrossref