Abstract Background Delta band (1–4 Hz) neuronal responses support the precision and stability of auditory processing, and a deficit in delta band synchrony may be relevant to auditory domain symptoms in schizophrenia patients. Methods Delta band synchronization elicited by a 2.5 Hz auditory steady state response (ASSR) paradigm, along with those from theta (5 Hz), alpha (10 Hz), beta (20 Hz), gamma (40 Hz), and high gamma (80 Hz) frequency ASSR, were compared in 128 patients with schizophrenia, 108 healthy controls, and 55 first-degree relatives (FDR) of patients. Results Delta band synchronization was significantly impaired in patients compared with controls (F = 18.3, P < .001). There was a significant 2.5 Hz by 40 Hz ASSR interaction (P = .023), arising from a greater reduction of 2.5 Hz ASSR than of 40 Hz ASSR, in patients compared with controls. Greater deficit in delta ASSR was associated with auditory perceptual abnormality (P = .007) and reduced verbal working memory (P < .001). Gamma frequency ASSR impairment was also significant but more modest (F = 8.7, P = .004), and this deficit was also present in FDR (P = .022). Conclusions The ability to sustain delta band oscillation entrainment in the auditory pathway is significantly reduced in schizophrenia patients and appears to be clinically relevant. auditory, hallucination, working memory, schizophrenia, auditory steady state, delta, gamma, ASSR Introduction Many schizophrenia-related symptoms, such as auditory hallucination, speech disorganization, disorganized thoughts, and verbal working memory (VWM) deficits are in the auditory–verbal domain, suggesting that the schizophrenia disease process impacts the auditory processing pathway. Electrophysiological abnormalities in schizophrenia are consistently reported in patients during auditory paradigms such as auditory mismatch negativity,1,2 steady-state response,3,4 sensory gating,5,6 and word, language and speech processing.7,8 Auditory–verbal processing deficits in schizophrenia may thus be associated with fundamental electrophysiological deficits in the auditory processing network. Cortical oscillations are thought to play an important role in cognitive functioning, communication, and integration of information across different regions of the brain.9–11 In healthy subjects, low frequency oscillations (<10 Hz) regulate speech processing,12 where accurate perception of attended speech is associated with more precise delta band (1–4 Hz) neuronal responses13,14 than in other bands. These low frequency oscillations, especially in the delta band, appear to serve a stabilization and enhancement function while attending to, and during the processing of, auditory streams.15 Auditory selective attention is also associated with the entrainment of ongoing neuronal oscillations in the delta band, which modulates neuronal excitability in primary auditory cortex.16 We hypothesized that in schizophrenia a reduced ability to generate synchronous delta oscillations in response to auditory stimuli would disrupt auditory processing, leading to auditory–verbal domain problems. The auditory steady-state response (ASSR) can be used to test the integrity of cortical oscillatory activity.17–19 It is a robust activation paradigm to elicit frequency-specific auditory responses. It is generated using stimuli that are repeated (periodic) at a specified frequency, resulting in electroencephalographic neural entrainment at the presentation frequency. We used a 2.5 Hz (mean of 1–4 Hz) stimulus train to elicit ASSR in the delta band, and to test the joint hypotheses that (1) schizophrenia is associated with an inability to support delta synchronization, and (2) this impairment is associated with cognitive disturbances and other symptoms in the auditory–verbal domain. Our study is the first to investigate the delta (1–4 Hz) range ASSR in schizophrenia. Previous studies using attention-based paradigms have shown that delta entrainment is associated with clinical symptoms in schizophrenia.20 ASSR has appeal in translational research for studying intrinsic neurobiological abnormalities without the need to rely on explicit behavioral performance. The first study to investigate ASSR in the 20–40 Hz range in schizophrenia reported reduced ASSR at 40 Hz.3 Subsequent ASSR studies expanded the range down to 5–10 Hz or up to 80–160 Hz,21,22 and have generally confirmed ASSR deficits in schizophrenia4,21–28 primarily in the 40–80 Hz range. The strong interest in 40 Hz has also been justified by finding a reduced 40 Hz ASSR in the first-degree relatives (FDR) of schizophrenia patients.27,29 FDR typically have about a 10-fold increase in risk for schizophrenia compared with the general population, although risks do not necessitate a transition to psychosis as the rate of schizophrenia is about 10% in FDR.30,31 Given these findings, we investigated delta band ASSR, along with higher frequencies at 5, 10, 20, 40, and 80 Hz, with a focus on assessing the relationship between delta (2.5 Hz) vs gamma (40–80 Hz). This design also allowed an unbiased assessment across a very broad range of frequencies, to determine whether there may be frequency specificity in schizophrenia psychopathology. Methods Participants The study included 128 schizophrenia spectrum disorder (SSD) patients and 108 healthy controls (HC) (table 1). The Structured Clinical Interview for DSM was used to make Axis I diagnoses. All patients were recruited from outpatient clinics; media advertisements were used for HC. Subjects with medical and neurological illnesses, head injury, and substance dependence or abuse (except nicotine) were excluded. Six patients were not on antipsychotics, 19 on typical, 74 on one atypical, 18 on more than one atypical, and 11 on a combination of atypical and typical antipsychotics. Patients on daily GABAergic hypnotics were excluded. Significant findings in SSD were re-examined in FDR of patients (n = 55) who have no psychosis. Seventy percentage of the FDR were from families of the patients in this study. All patient probands of the FDR were interviewed by SCID regardless of whether the patients were in the current study. Table 1. Demographic and Clinical Information HC (n = 108) SSD (n = 128) FDR (n = 55) HC Vs SSD HC Vs FDR F or χ2 P Value F or χ2 P Value Age mean (SD) 37.9 (13.8) 37.8 (13.1) 46.6 (13.6) 0.003 .96 15.1 <.001** % Male 65.7 67.2 31.6 0.06 .89 17.5 <.001** Verbal working memory 20.4 (4.4) 17.0 (5.3) 18.9 (5.1) 24.8 <.001** 3.4 .067 Auditory perception trait 3.8 (4.8) 19.4 (11.8) 4.8 (7.7) 140.3 <.001** 0.9 .35 Auditory perception state 1.1 (2.4) 9.8 (12.0) 3.8 (4.8) 46.4 <.001** 0.5 .46 BPRS n/a 40.4 (11.2) n/a n/a n/a n/a n/a HC (n = 108) SSD (n = 128) FDR (n = 55) HC Vs SSD HC Vs FDR F or χ2 P Value F or χ2 P Value Age mean (SD) 37.9 (13.8) 37.8 (13.1) 46.6 (13.6) 0.003 .96 15.1 <.001** % Male 65.7 67.2 31.6 0.06 .89 17.5 <.001** Verbal working memory 20.4 (4.4) 17.0 (5.3) 18.9 (5.1) 24.8 <.001** 3.4 .067 Auditory perception trait 3.8 (4.8) 19.4 (11.8) 4.8 (7.7) 140.3 <.001** 0.9 .35 Auditory perception state 1.1 (2.4) 9.8 (12.0) 3.8 (4.8) 46.4 <.001** 0.5 .46 BPRS n/a 40.4 (11.2) n/a n/a n/a n/a n/a Note: BPRS, Brief Psychiatric Rating Scale; HC, healthy controls; SSD, schizophrenia spectrum disorder patients; FDR, first-degree relatives of SSD patients. **Statistically significant. View Large Auditory Clinical and Cognitive Symptoms We developed the Auditory Perceptual Trait and State Scale (APTS) to measure perceptual abnormalities. The anomalies are rated for “trait,” defined as longitudinally experienced symptoms over one’s lifetime, and “state,” defined as symptoms recently experienced in the past week. The full scale is available at http://www.mdbrain.org/APTS.pdf. The APTS is self-rated. Its test–retest reliability was assessed in 41 participants about 4 months apart, which showed ICC = 0.81 for both the trait and state measures, suggesting good reliability. The Brief Psychiatric Rating Scale (BPRS) was used to rate overall symptoms. The APTS was administered to all participants; the BPRS only to patients. Finally, auditory– VWM was assessed using the digit sequencing task.27,29 ASSR Paradigm ASSRs were recorded in a sound-attenuated chamber while participants listened to click trains, delivered by headphones, at 2.5, 5, 10, 20, 40, and 80 Hz. Seventy-five stimulus trains (trials) each consisting of 15 clicks, with each click at 72 dB and of 1 ms duration were delivered at each stimulus frequency. The duration ranged from 6 s per train for 2.5 Hz, to 0.1875 s per train for 80 Hz. The inter-train interval was 0.7 s. Therefore, the durations for 2.5, 5, 10, 20, 40, and 80 Hz were 8.38, 4.69, 2.82, 1.89, 1.42, and 1.19 min, respectively, presented in 6 separate blocks separated by 2 min. The order of the blocks was randomized. This design allows steady-state neural entrainment for each frequency (figure 2 and supplementary figure S1). A hearing screening test excluded apparent hearing impairment. EEG was recorded using a 64 electrode Quick-Cap with sintered Ag/Ag chloride electrodes and a Neuroscan SynAmp2 (Compumedics, Charlotte, NC) at 1000 Hz with a 0.1–200 Hz bandpass filter. Impedance was kept below 5 kΩ. Offline, electrodes were average referenced, high-pass filtered at 0.8 Hz, and detrended. Ocular artifacts were removed using the time-shift-PCA algorithm, with ocular channels as references.32 The full-duration waveforms from each channel were epoched into 75 individual trials. Normalized ASSR Power While typical ASSR analysis uses individual channels (often CZ or FZ), we adapted signal processing techniques33,34 where individual EEG channels are spatially combined to maximize response reliability using the denoising source separation (DSS) algorithm.35–37 DSS is a blind source separation technique related to principal component analysis (PCA) and independent component analysis (ICA) but specifically designed for use with data from multi-trial evoked responses or narrowband signals. DSS works by enhancing stimulus-driven activity over stimulus-unrelated activity, with its components ordered according to their reliability35–37 [details in supplementary information]. Raw ASSR power was calculated as the magnitude squared of the Fourier transform at the stimulus frequency. The Fourier transform was calculated using concatenated trials rather than averaged trials to increase spectral resolution.38,39 Background power was calculated as average spectral power over 1 Hz width frequency bands (on either side of the stimulus frequency, after leaving a guard band of 0.5 Hz on either side). Normalized ASSR power was then calculated as the mean over DSS components of the ratio of raw ASSR power and respective background power. This normalization with respect to background power dramatically reduces subject-to-subject variability of frequency response profiles.38 This combined use of DSS and normalized ASSR power represents the 2 critical improvements over previous ASSR power extraction methods. The accompanying reduction of noise is particularly critical for low frequency ASSR, which is known to be more susceptible to background low frequency fluctuations.17,33 ASSR Phase Locking Value Phase locking value (PLV)40 has been extensively used in ASSR analysis. Increased variability of neural responses across trials reduces the PLV value toward 0, whereas increased reliability increases the value toward 1.41 First, intra-electrode PLV was calculated for each stimulus frequency at each electrode (algorithm in SI). Based on topographic analysis showing that ASSR was strongest at fronto-central locations (figure 1), the PLV of 16 fronto-central electrodes (AF3, AFZ, AF4, F3, F1, FZ, F2, F4, FC3, FC1, FCZ, FC2, FC4, C1, CZ, C2) were averaged and used for the final PLV assessment. Fig. 1. View largeDownload slide Grand averages of topographies of normalized ASSR power for healthy controls, first-degree relatives, and schizophrenia spectrum disorder patients. Scaled based on lowest (blue) to highest power value (red) within each frequency. Refer to Results for statistical group differences. Fig. 1. View largeDownload slide Grand averages of topographies of normalized ASSR power for healthy controls, first-degree relatives, and schizophrenia spectrum disorder patients. Scaled based on lowest (blue) to highest power value (red) within each frequency. Refer to Results for statistical group differences. Statistics Data processing was performed without the knowledge of group and demographic information. Repeated measures ANOVA was performed to compare normalized ASSR power by stimulus frequency (6) and group (SSD vs HC). The Greenhouse-Geisser correction was applied. Significant effect was followed by post hoc comparison using Bonferroni correction (P < .008). If a significant difference was found for SSD vs HC, we then further tested whether the same frequency was significantly different between FDR vs HC (no further Bonferroni correction was applied). A similar analysis was followed for the PLV measure. Contributions of ASSR to clinical measures were examined using stepwise linear regression, where at each step the ASSR power at the 6 frequencies were the predictors and one clinical measure was the dependent variable. Multi-collinearity was examined using variance inflation factor (VIF).42 A regression model was considered significant if the overall model was significant at P <.05 and all predictors had VIF <5. All tests were 2-tailed. Results ASSR in SSD Patients Figure 1 shows that the spatial distribution of normalized ASSR power has a fronto-central accentuation; figure 2 shows grand average time courses of the ASSR responses from electrode FZ. Repeated measures ANOVA on normalized ASSR power extracted by DSS showed significant effects for stimulus frequency (F = 390.1, P < .001), group (SSD vs HC; F = 13.4, P < .001) and a frequency × group interaction (P = .039). Post hoc tests showed that the SSD group had significantly reduced power at 2.5 Hz (F = 18.3, P < .001), 5 Hz (F = 10.1, P = .002), 10 Hz (F = 9.9, P = .002), and 40 Hz (F = 8.7, P = .004), but not 20 Hz (P = .18) or 80 Hz (P = .03) after Bonferroni correction (figure 3A). When the analogous analysis was performed on ASSR responses without the use of DSS (eg, from the single electrode FZ) and without normalization, most findings of significance were lost: only 40 Hz ASSR showed nominally significant reduction in SSD compared with HC (P = .017), which was then lost after correcting for multiple comparisons (supplementary figure S2B). Fig. 2. View largeDownload slide Time-domain grand averages from electrode FZ. The vertical axis shows amplitude in μV. The vertical dotted lines indicate begin and end points of a stimulus train. Preferential entrainment in the delta (2.5 Hz) and gamma (40 Hz) bands can be seen in both the controls and schizophrenic patient group, and 2.5 Hz and 40 Hz stimuli are also associated with larger patient–control differences. Fig. 2. View largeDownload slide Time-domain grand averages from electrode FZ. The vertical axis shows amplitude in μV. The vertical dotted lines indicate begin and end points of a stimulus train. Preferential entrainment in the delta (2.5 Hz) and gamma (40 Hz) bands can be seen in both the controls and schizophrenic patient group, and 2.5 Hz and 40 Hz stimuli are also associated with larger patient–control differences. Fig. 3. View large Download slide Mean and SE of normalized power (in dB) and phase locking values (PLV). (A) Power at 2.5, 5, 10, and 40 Hz are significantly lower for patients than controls. (B) Replicable findings using PLV. (C and D) First-degree relatives showed replicable auditory steady state response reduction compared with controls only at 40 Hz. *Statistically significant. Effect sizes are tabulated in supplementary table 1. Fig. 3. View large Download slide Mean and SE of normalized power (in dB) and phase locking values (PLV). (A) Power at 2.5, 5, 10, and 40 Hz are significantly lower for patients than controls. (B) Replicable findings using PLV. (C and D) First-degree relatives showed replicable auditory steady state response reduction compared with controls only at 40 Hz. *Statistically significant. Effect sizes are tabulated in supplementary table 1. To formally test the frequency × group interaction between 2.5 and 40 Hz, ANOVA was repeated contrasting these frequencies. It showed significant group (P < .001) and interaction effects (P = .023), where the interaction was due to a greater reduction of 2.5 Hz ASSR than of 40 Hz ASSR, in patients compared with controls, as seen in figure 3A. Re-examining these findings using PLV, significant effects were seen for frequency (P < .001) and group (P < .001) without interaction (P = .09). Patients had reduced PLV at 2.5 Hz (F = 9.5, P = .002), 5 Hz (F = 8.2, P = .004), 10 Hz (F = 5.9, P = .016), 40 Hz (F = 5.3, P = .022) and 80 Hz (F = 4.0, P = .045) but not 20 Hz (P = .50). Findings from 2.5, 5, 10, and 40 Hz replicated power-based analyses and thus no further Bonferroni correction was applied (figure 3B). Therefore, reduced ASSR was found in 2.5, 5, 10, and 40 Hz in both power and phase based analysis. ASSR in FDR Age and sex were not matched between FDR and HC (table 1). However, neither age (P = .44) nor sex (P = .44) were significant in the repeated measures ANCOVA and so were removed. The results showed significant effects for stimulus frequency (F = 171.0, P < .001), group (F = 5.1, P = .025), and frequency × group interaction (P = .036) in FDR vs HC. Post hoc tests at the frequencies for which SSD and HC were significantly different (2.5, 5, 10, and 40 Hz) showed that FDR had lower ASSR power than HC at 40 Hz (F = 5.4, P = .022) but not at 2.5, 5, or 10 Hz (P = .28–.85) (figure 3C). Only the reduction in gamma band ASSR at 40 Hz was considered a replication of findings in patients. For PLV, age (P = .71) and sex (P = .88) were not significant. There was a significant stimulus frequency effect (F = 84.4, P < .001) and a frequency × group interaction (F = 3.5, P = .007). Post hoc tests showed that only the 40 Hz ASSR reduction (F = 5.5, P = .021) was replicated (figure 3D). In summary, findings were largely consistent between normalized power and PLV except with PLV generally having smaller effect sizes (figure 3; supplementary tables 1 and 2). In subsequent analyses, we opted to use only normalized power based ASSR. ASSR and VWM Working memory is impaired in SSD.43,44 While the SSD group had lower VWM compared with HC (P < .001, table 1), the FDR did not significantly differ from HC on VWM (P = .067). The regression model was significant in SSD (F = 15.8, ΔR2 = 11.8%, P < .001; all VIFs < 1.5) where only 2.5 Hz ASSR significantly contributed to VWM (t = 4.0, P < .001): patients with lower delta power showed worse VWM (r = .36, P < .001) (figure 4A). The correlation of 2.5 Hz ASSR with VWM was not significant in either the HC or FDR groups (supplementary figure S3). Fig. 4. View largeDownload slide Frequency-specific associations between verbal working memory and auditory steady state response (ASSR). (A) In schizophrenia spectrum disorder patients, higher 2.5 Hz ASSR was associated with better working memory. (B) In first-degree relatives, 40 Hz ASSR was associated with working memory. (C) The ASSR–working memory relationships (by their correlation coefficients: y axis) were strongly (negatively) associated with stimulus frequencies. Fig. 4. View largeDownload slide Frequency-specific associations between verbal working memory and auditory steady state response (ASSR). (A) In schizophrenia spectrum disorder patients, higher 2.5 Hz ASSR was associated with better working memory. (B) In first-degree relatives, 40 Hz ASSR was associated with working memory. (C) The ASSR–working memory relationships (by their correlation coefficients: y axis) were strongly (negatively) associated with stimulus frequencies. We calculated the correlation coefficients between VWM and ASSR at each frequency: 2.5 Hz: r = .34, P < .001; 5 Hz: r = .30, P = .001; 10 Hz: r = .22, P = .016; 20 Hz: r = .19, P = .040; 40 Hz: r = .20, P = .033; and 80 Hz: r = .15, P = .11. The relationship between ASSR and VWM, quantified through these correlation coefficients, was strongly linked to the stimulus frequency (r = −.95, P = .003) (figure 4C): the correlation between ASSR and VWM significantly decreases with increasing stimulus frequency. The model was also significant in FDR (F = 9.8, ΔR2 = 17.5%, P = .003; VIFs < 3.6) where only the 40 Hz ASSR significantly contributed to VWM (t = 3.13, P = .003) (figure 4B). The model was not significant in controls (model P > .05). ASSR and Auditory Perception Abnormality A regression model with APTS trait score as the dependent variable and normalized ASSR power as predictors in the SSD group was significant (F = 7.7, ΔR2=13.8%, P = .001; VIFs < 1.5). Only 2.5 Hz (t = −2.8, ΔR2 = 6.8%, P = .007) and 40 Hz (t = 3.6, ΔR2 = 6.9%, P = .001) normalized ASSR powers were significant predictors but in opposite directions: reduced 2.5 Hz and increased 40 Hz ASSR were associated with more longitudinally experienced auditory symptoms in SSD patients (figures 5B and 5C). The model was not significant for state auditory symptoms (P = .056) although the trends were the same. Fig. 5. View largeDownload slide Auditory perception as measured by Auditory Perceptual Trait and State Scale was significantly higher (*) in patients compared with controls as experience in lifetime (trait) or the past 7 days (state), but not in first-degree relatives compared with controls (A) Regression analyses reviewed that 2.5 Hz (B) and 40 Hz (C) auditory steady state response contributed to auditory perceptual trait score in patients but in opposite directions. Partial r refers to having partialled out the effects of 2.5 Hz for 40 Hz, or vice versa, in the regression analyses. Fig. 5. View largeDownload slide Auditory perception as measured by Auditory Perceptual Trait and State Scale was significantly higher (*) in patients compared with controls as experience in lifetime (trait) or the past 7 days (state), but not in first-degree relatives compared with controls (A) Regression analyses reviewed that 2.5 Hz (B) and 40 Hz (C) auditory steady state response contributed to auditory perceptual trait score in patients but in opposite directions. Partial r refers to having partialled out the effects of 2.5 Hz for 40 Hz, or vice versa, in the regression analyses. Medication and Other Clinical Measurements Chlorpromazine equivalent (CPZ) of antipsychotic dosages was not correlated with ASSR power at any frequency (all r < .13, all P > .20). BPRS total or psychosis score was not significantly correlated with ASSR power at any frequency (all r < .11, all P > .30). Discussion We found that delta and gamma ASSR were both reduced in patients, with delta showing a more pronounced reduction. Critically, reduced delta ASSR was associated both with more severe longitudinally experienced auditory symptom “traits” and also poorer VWM. The observed reduction in gamma ASSR, on the other hand, which was also present in nonpsychotic FDR, was found to be more associated with the risk of SSD than with SSD itself. The finding of reduced 40 Hz ASSR in SSD replicates other studies3,23,24,45 while the finding of reduced delta ASSR is new. While delta band power may be significantly increased in SSD in resting EEG,46 delta band oscillations in SSD are also found to be significantly reduced in stimulus or behavior activated paradigms.22,47–53 Of particular relevance are findings of reduced delta power and fronto-temporal coherence during talking51 and auditory target detection52 in SSD. Unlike a task-related auditory paradigm depending on performance, the delta ASSR is a passive paradigm and its deficit may indicate difficulty in generating normal delta synchronization to auditory stimuli. Reduced ability to generate delta entrainment might serve as a tangible mechanism for the frequently observed auditory–verbal domain issues in SSD, as neural entrainment in the delta and theta band is critical for normal speech perception.12–14 Auditory hallucinations are experienced by most patients with SSD in their lifetime.54,55 We tested the hypothesis that an ASSR delta deficit contributes to their auditory symptoms, as low frequency temporal modulations (<4 Hz) are more critical in speech perception than faster (22–40 Hz) modulations.12–14 This hypothesis was supported. Patients have impaired ability to generate delta ASSR more so than any other frequency tested, and this deficit is associated with more severe longitudinally experienced auditory anomalies. We did not find this correlation with “state” symptoms in APTS or BPRS, perhaps due to fluctuations in state symptoms or to variability in treatment or treatment response. ASSR-based delta entrainment was significantly associated with VWM (figure 4A). The linear relationship between stimulus frequency and the ASSR–VWM correlation coefficients (r = −.95; figure 4C) further highlights a potentially prominent role of low frequency oscillations in the auditory cognitive system in patients with SSD. In evaluating the association between ASSR power and VWM, we observed that the most strongly correlated frequency band changed from delta with SSD to gamma with FDR (figure 4A vs figure 4B). In the auditory cortex, the amplitude of neural oscillations are controlled in a nested fashion, where delta (1–4 Hz) phase modulates theta (4–8 Hz) amplitude, and theta phase modulates gamma (30+ Hz) amplitude.56 This oscillatory hierarchy is thought to control baseline excitability.56 Under this assumption, we speculate that in individuals without a major deficit in delta generation, as in FDR, gamma band abnormality may yield a more apparent relationship with VWM (figure 4B). However, in individuals with major deficits in delta generation, as in the patients, delta deficits may exert a more fundamental role and thus stronger contribution to VWM (figures 4A and 4C). Figure 2 illustrates the “preferential” entrainment at 2.5 and 40 Hz using absolute power analysis even at a single electrode. The special 40 Hz entrainment in human brains is well known but the 2.5 Hz case is a new observation. Using normalization and DSS analysis, the reduction of 2.5 Hz ASSR for SSD was significant, but not when using simple spectral power; this may explain the lack of earlier observations. DSS and normalization reduce variability separately in delta ASSR (normalization does not improve gamma ASSR), allowing the delta reduction to even surpass the gamma reduction. Delta ASSR was not significantly different between FDR and controls, suggesting that this deficit does not indicate a genetic vulnerability for SSD. The finding of reduced 40 Hz ASSR in FDR replicated our previous finding,27 now in an independent, much larger cohort. Combined with another independent replication,29 the data support a 40 Hz ASSR deficit as a genetic biomarker for SSD. Greater 40 Hz ASSR (within overall reduction) in patients was associated with more auditory symptoms (figure 5C). This matches findings in the visual domain, where higher gamma during gestalt perception was associated with more visual hallucinations.57 Recent animal and human studies are converging to show that glutamatergic receptor antagonists increase gamma neural oscillations.58,59 A leading hypothesis in psychosis generation is excitatory glutamatergic receptor hypofunction, based on observations that glutamatergic receptor antagonism by phencyclidine and ketamine mimics aspects of schizophrenia symptomatology.60,61 Therefore, the link between higher gamma power and more visual and auditory symptoms could be through abnormal glutamatergic mechanisms. Gamma oscillations are generated by inhibitory GABAergic interneurons regulating excitatory glutamatergic pyramidal neurons.62,63 Abnormal GABAergic regulation of gamma is thought to underlie working memory deficits in SSD64–66 and is often hypothesized to be genetic in origin.67 Therefore, reduced gamma ASSR in FDR, and correlation with working memory in FDR (figure 4B), appear to support the hypothesis that gamma-working memory deficit confers risk for SSD. The neural mechanisms underlying delta oscillations are less well understood. Studies of sleep and waking state delta oscillations68 suggest that NMDA receptors play a role in maintaining these slow oscillations, and the NMDA receptor antagonist ketamine reduces slow wave 1–5 Hz oscillations.58 Whether reduced delta ASSR reflects an NMDA hypofunction origin of schizophrenia69 would require follow-up studies. Compared to conventional single-channel-based spectral power analysis, this study employed the techniques of DSS, which integrates over channels, and normalization, which takes into account background power; both contribute separately to increase statistical power in the ASSR analysis. Normalization particularly improves ASSR analysis at lower frequencies, due to the 1/f nature (strong rise at low frequencies) of noisy background activity in electrophysiological recordings.70,71 DSS enhances amplitude contrast, due to its ability to optimally combine responses across electrodes and so extract ASSR responses with higher fidelity. DSS performs only spatial, not spectral filtering, and hence does not introduce artifacts associating with spectral filtering. An important limitation is that we did not test for the specificity of the findings. We tested auditory working memory, but not deficits in other cognitive domains. This limits interpretation regarding whether the correlations with clinical features were specific to auditory working memory or more general to other cognitive deficits. Reduced delta frequency ASSR in SSD might also arise from antipsychotic medications, though no correlations were found between delta frequency ASSR and current antipsychotic medication dosage. Finally, the failure to synchronize to delta frequency stimulation could also be related to abnormal baseline delta activity, although the DSS procedure was designed to account for this effect. In summary, the results from this study support that inadequate ability to sustain neural oscillatory responses in the lower frequency range may play a role in the auditory perceptual and cognitive deficit mechanisms in schizophrenia. The findings from this study support the use of delta range ASSR as part of the effort to build translational animal models to study the etiology of SSD. Supplementary Material Supplementary data are available at Schizophrenia Bulletin online. Funding This research was supported by the National Institutes of Health ([R01MH085646, U01MH108148, P50MH103222 Project 3, R01DA027680 to L.E.H.] and [R01DC014085 to J.Z.S.]) and by a University of Maryland Seed Grant (14-103 to L.E.H. and J.Z.S.). Acknowledgments The authors would like to thank Andrew Dimitrijevic for advice and Francisco Cervantes Constantino for help with the data analysis. Conflict of interest: L.E.H. reported receiving or planning to receive research funding or consulting fees from Mitsubishi, Your Energy Systems LLC, Sound Pharma, Neuralstem, Taisho Pharmaceutical, Heptares, and Pfizer. Other authors reported no biomedical financial interests or potential conflicts of interest. References 1. Javitt DC, Steinschneider M, Schroeder CE, Arezzo JC. Role of cortical N-methyl-D-aspartate receptors in auditory sensory memory and mismatch negativity generation: implications for schizophrenia. Proc Natl Acad Sci U S A . 1996; 93: 11962– 11967. Google Scholar CrossRef Search ADS PubMed 2. Light GA, Braff DL. Mismatch negativity deficits are associated with poor functioning in schizophrenia patients. Arch Gen Psychiatry . 2005; 62: 127– 136. Google Scholar CrossRef Search ADS PubMed 3. Kwon JS, O’Donnell BF, Wallenstein GVet al. Gamma frequency-range abnormalities to auditory stimulation in schizophrenia. Arch Gen Psychiatry . 1999; 56: 1001– 1005. Google Scholar CrossRef Search ADS PubMed 4. Hirano Y, Oribe N, Kanba S, Onitsuka T, Nestor PG, Spencer KM. Spontaneous gamma activity in schizophrenia. JAMA Psychiatry . 2015; 72: 813– 821. Google Scholar CrossRef Search ADS PubMed 5. Freedman R, Adler LE, Myles-Worsley Met al. Inhibitory gating of an evoked response to repeated auditory stimuli in schizophrenic and normal subjects. Human recordings, computer simulation, and an animal model. Arch Gen Psychiatry . 1996; 53: 1114– 1121. Google Scholar CrossRef Search ADS PubMed 6. Hong LE, Summerfelt A, Mitchell BDet al. Sensory gating endophenotype based on its neural oscillatory pattern and heritability estimate. Arch Gen Psychiatry . 2008; 65: 1008– 1016. Google Scholar CrossRef Search ADS PubMed 7. Ford JM, Krystal JH, Mathalon DH. Neural synchrony in schizophrenia: from networks to new treatments. Schizophr Bull . 2007; 33: 848– 852. Google Scholar CrossRef Search ADS PubMed 8. Kiang M, Kutas M, Light GA, Braff DL. An event-related brain potential study of direct and indirect semantic priming in schizophrenia. Am J Psychiatry . 2008; 165: 74– 81. Google Scholar CrossRef Search ADS PubMed 9. Başar E, Başar-Eroglu C, Karakaş S, Schürmann M. Gamma, alpha, delta, and theta oscillations govern cognitive processes. Int J Psychophysiol . 2001; 39: 241– 248. Google Scholar CrossRef Search ADS PubMed 10. Uhlhaas PJ, Pipa G, Lima Bet al. Neural synchrony in cortical networks: history, concept and current status. Front Integr Neurosci . 2009; 3: 17. Google Scholar CrossRef Search ADS PubMed 11. Ward LM. Synchronous neural oscillations and cognitive processes. Trends Cogn Sci . 2003; 7: 553– 559. Google Scholar CrossRef Search ADS PubMed 12. Ding N, Simon JZ. Emergence of neural encoding of auditory objects while listening to competing speakers. Proc Natl Acad Sci U S A . 2012; 109: 11854– 11859. Google Scholar CrossRef Search ADS PubMed 13. Ding N, Chatterjee M, Simon JZ. Robust cortical entrainment to the speech envelope relies on the spectro-temporal fine structure. Neuroimage . 2014; 88: 41– 46. Google Scholar CrossRef Search ADS PubMed 14. Zion Golumbic EM, Ding N, Bickel Set al. Mechanisms underlying selective neuronal tracking of attended speech at a “cocktail party”. Neuron . 2013; 77: 980– 991. Google Scholar CrossRef Search ADS PubMed 15. Ding N, Simon JZ. Cortical entrainment to continuous speech: functional roles and interpretations. Front Hum Neurosci . 2014; 8: 311. Google Scholar CrossRef Search ADS PubMed 16. Lakatos P, Musacchia G, O’Connel MN, Falchier AY, Javitt DC, Schroeder CE. The spectrotemporal filter mechanism of auditory selective attention. Neuron . 2013; 77: 750– 761. Google Scholar CrossRef Search ADS PubMed 17. Picton TW, John MS, Dimitrijevic A, Purcell D. Human auditory steady-state responses. Int J Audiol . 2003; 42: 177– 219. Google Scholar CrossRef Search ADS PubMed 18. 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. Google Scholar CrossRef Search ADS PubMed 19. Uhlhaas PJ, Singer W. Abnormal neural oscillations and synchrony in schizophrenia. Nat Rev Neurosci . 2010; 11: 100– 113. Google Scholar CrossRef Search ADS PubMed 20. Lakatos P, Schroeder CE, Leitman DI, Javitt DC. Predictive suppression of cortical excitability and its deficit in schizophrenia. J Neurosci . 2013; 33: 11692– 11702. Google Scholar CrossRef Search ADS PubMed 21. Tsuchimoto R, Kanba S, Hirano Set al. Reduced high and low frequency gamma synchronization in patients with chronic schizophrenia. Schizophr Res . 2011; 133: 99– 105. Google Scholar CrossRef Search ADS PubMed 22. Hamm JP, Gilmore CS, Picchetti NA, Sponheim SR, Clementz BA. Abnormalities of neuronal oscillations and temporal integration to low- and high-frequency auditory stimulation in schizophrenia. Biol Psychiatry . 2011; 69: 989– 996. Google Scholar CrossRef Search ADS PubMed 23. Light GA, Hsu JL, Hsieh MHet al. Gamma band oscillations reveal neural network cortical coherence dysfunction in schizophrenia patients. Biol Psychiatry . 2006; 60: 1231– 1240. Google Scholar CrossRef Search ADS PubMed 24. Spencer KM, Salisbury DF, Shenton ME, McCarley RW. Gamma-band auditory steady-state responses are impaired in first episode psychosis. Biol Psychiatry . 2008; 64: 369– 375. Google Scholar CrossRef Search ADS PubMed 25. Brenner CA, Kieffaber PD, Clementz BAet al. Event-related potential abnormalities in schizophrenia: a failure to “gate in” salient information? Schizophr Res . 2009; 113: 332– 338. Google Scholar CrossRef Search ADS PubMed 26. Clementz BA, Keil A, Kissler J. Aberrant brain dynamics in schizophrenia: delayed buildup and prolonged decay of the visual steady-state response. Brain Res Cogn Brain Res . 2004; 18: 121– 129. Google Scholar CrossRef Search ADS PubMed 27. Hong LE, Summerfelt A, McMahon Ret al. Evoked gamma band synchronization and the liability for schizophrenia. Schizophr Res . 2004; 70: 293– 302. Google Scholar CrossRef Search ADS PubMed 28. Kirihara K, Rissling AJ, Swerdlow NR, Braff DL, Light GA. Hierarchical organization of gamma and theta oscillatory dynamics in schizophrenia. Biol Psychiatry . 2012; 71: 873– 880. Google Scholar CrossRef Search ADS PubMed 29. Rass O, Forsyth JK, Krishnan GPet al. Auditory steady state response in the schizophrenia, first-degree relatives, and schizotypal personality disorder. Schizophr Res . 2012; 136: 143– 149. Google Scholar CrossRef Search ADS PubMed 30. Kendler KS, McGuire M, Gruenberg AM, O’Hare A, Spellman M, Walsh D. The Roscommon Family Study. I. Methods, diagnosis of probands, and risk of schizophrenia in relatives. Arch Gen Psychiatry . 1993; 50: 527– 540. Google Scholar CrossRef Search ADS PubMed 31. Erlenmeyer-Kimling L, Adamo UH, Rock Det al. The New York High-Risk Project. Prevalence and comorbidity of axis I disorders in offspring of schizophrenic parents at 25-year follow-up. Arch Gen Psychiatry . 1997; 54: 1096– 1102. Google Scholar CrossRef Search ADS PubMed 32. de Cheveigné A, Simon JZ. Denoising based on time-shift PCA. J Neurosci Methods . 2007; 165: 297– 305. Google Scholar CrossRef Search ADS PubMed 33. Wang Y, Ding N, Ahmar N, Xiang J, Poeppel D, Simon JZ. Sensitivity to temporal modulation rate and spectral bandwidth in the human auditory system: MEG evidence. J Neurophysiol . 2012; 107: 2033– 2041. Google Scholar CrossRef Search ADS PubMed 34. de Cheveigné A, Arzounian D. Scanning for oscillations. J Neural Eng . 2015; 12: 066020. Google Scholar CrossRef Search ADS PubMed 35. Särelä J, Valpola H. Denoising source separation. J Mach Learn Res . 2005; 6: 233– 272. 36. de Cheveigné A, Simon JZ. Denoising based on spatial filtering. J Neurosci Methods . 2008; 171: 331– 339. Google Scholar CrossRef Search ADS PubMed 37. de Cheveigné A, Parra LC. Joint decorrelation, a versatile tool for multichannel data analysis. Neuroimage . 2014; 98: 487– 505. Google Scholar CrossRef Search ADS PubMed 38. Elhilali M, Xiang J, Shamma SA, Simon JZ. Interaction between attention and bottom-up saliency mediates the representation of foreground and background in an auditory scene. PLoS Biol . 2009; 7: e1000129. Google Scholar CrossRef Search ADS PubMed 39. Xiang J, Simon J, Elhilali M. Competing streams at the cocktail party: exploring the mechanisms of attention and temporal integration. J Neurosci . 2010; 30: 12084– 12093. Google Scholar CrossRef Search ADS PubMed 40. Jervis BW, Nichols MJ, Johnson TE, Allen E, Hudson NR. A fundamental investigation of the composition of auditory evoked potentials. IEEE Trans Biomed Eng . 1983; 30: 43– 50. Google Scholar CrossRef Search ADS PubMed 41. Herrmann B, Henry MJ, Grigutsch M, Obleser J. Oscillatory phase dynamics in neural entrainment underpin illusory percepts of time. J Neurosci . 2013; 33: 15799– 15809. Google Scholar CrossRef Search ADS PubMed 42. Stevens JP. Applied Multivariate Statistics for the Social Sciences . 4th ed. Mahwah, NJ: Psychology Press; 2002. 43. Forbes NF, Carrick LA, McIntosh AM, Lawrie SM. Working memory in schizophrenia: a meta-analysis. Psychol Med . 2009; 39: 889– 905. Google Scholar CrossRef Search ADS PubMed 44. Barch DM, Berman MG, Engle Ret al. CNTRICS final task selection: working memory. Schizophr Bull . 2009; 35: 136– 152. Google Scholar CrossRef Search ADS PubMed 45. 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: 2656– 2671. Google Scholar CrossRef Search ADS PubMed 46. Sponheim SR, Clementz BA, Iacono WG, Beiser M. Resting EEG in first-episode and chronic schizophrenia. Psychophysiology . 1994; 31: 37– 43. Google Scholar CrossRef Search ADS PubMed 47. 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: 2238– 2240. Google Scholar CrossRef Search ADS PubMed 48. Donkers FC, Englander ZA, Tiesinga PH, Cleary KM, Gu H, Belger A. Reduced delta power and synchrony and increased gamma power during the P3 time window in schizophrenia. Schizophr Res . 2013; 150: 266– 268. Google Scholar CrossRef Search ADS PubMed 49. Doege K, Bates AT, White TP, Das D, Boks MP, Liddle PF. Reduced event-related low frequency EEG activity in schizophrenia during an auditory oddball task. Psychophysiology . 2009; 46: 566– 577. Google Scholar CrossRef Search ADS PubMed 50. Basar-Eroglu C, Schmiedt-Fehr C, Mathes B, Zimmermann J, Brand A. Are oscillatory brain responses generally reduced in schizophrenia during long sustained attentional processing? Int J Psychophysiol . 2009; 71: 75– 83. Google Scholar CrossRef Search ADS PubMed 51. Ford JM, Mathalon DH, Whitfield S, Faustman WO, Roth WT. Reduced communication between frontal and temporal lobes during talking in schizophrenia. Biol Psychiatry . 2002; 51: 485– 492. Google Scholar CrossRef Search ADS PubMed 52. Ford JM, Roach BJ, Hoffman RS, Mathalon DH. The dependence of P300 amplitude on gamma synchrony breaks down in schizophrenia. Brain Res . 2008; 1235: 133– 142. Google Scholar CrossRef Search ADS PubMed 53. Bates AT, Kiehl KA, Laurens KR, Liddle PF. Low-frequency EEG oscillations associated with information processing in schizophrenia. Schizophr Res . 2009; 115: 222– 230. Google Scholar CrossRef Search ADS PubMed 54. Andreasen NC, Flaum M. Schizophrenia: the characteristic symptoms. Schizophr Bull . 1991; 17: 27– 49. Google Scholar CrossRef Search ADS PubMed 55. Sartorius N, Shapiro R, Jablensky A. The international pilot study of schizophrenia. Schizophr Bull . 1974;1: 21– 34. 56. Lakatos P, Shah AS, Knuth KH, Ulbert I, Karmos G, Schroeder CE. An oscillatory hierarchy controlling neuronal excitability and stimulus processing in the auditory cortex. J Neurophysiol . 2005; 94: 1904– 1911. Google Scholar CrossRef Search ADS PubMed 57. Spencer KM, Nestor PG, Perlmutter Ret al. Neural synchrony indexes disordered perception and cognition in schizophrenia. Proc Natl Acad Sci U S A . 2004; 101: 17288– 17293. Google Scholar CrossRef Search ADS PubMed 58. Hong LE, Summerfelt A, Buchanan RWet al. Gamma and delta neural oscillations and association with clinical symptoms under subanesthetic ketamine. Neuropsychopharmacology . 2010; 35: 632– 640. Google Scholar CrossRef Search ADS PubMed 59. Sullivan EM, Timi P, Hong LE, O’Donnell P. Reverse translation of clinical electrophysiological biomarkers in behaving rodents under acute and chronic NMDA receptor antagonism. Neuropsychopharmacology . 2015; 40: 719– 727. Google Scholar CrossRef Search ADS PubMed 60. 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: 108– 121. Google Scholar CrossRef Search ADS PubMed 61. Snyder MA, Gao WJ. NMDA hypofunction as a convergence point for progression and symptoms of schizophrenia. Front Cell Neurosci . 2013; 7: 31. Google Scholar CrossRef Search ADS PubMed 62. Bartos M, Elgueta C. Functional characteristics of parvalbumin- and cholecystokinin-expressing basket cells. J Physiol . 2012; 590: 669– 681. Google Scholar CrossRef Search ADS PubMed 63. Gonzalez-Burgos G, Lewis DA. NMDA receptor hypofunction, parvalbumin-positive neurons, and cortical gamma oscillations in schizophrenia. Schizophr Bull . 2012; 38: 950– 957. Google Scholar CrossRef Search ADS PubMed 64. Kim T, Thankachan S, McKenna JTet al. Cortically projecting basal forebrain parvalbumin neurons regulate cortical gamma band oscillations. Proc Natl Acad Sci U S A . 2015; 112: 3535– 3540. Google Scholar CrossRef Search ADS PubMed 65. Hines RM, Hines DJ, Houston CM, Mukherjee J, Haydon PG, Tretter V, Smart TG, Moss SJ. Disrupting the clustering of GABAA receptor alpha2 subunits in the frontal cortex leads to reduced gamma-power and cognitive deficits. Proc Natl Acad Sci U S A . 2013; 110: 16628– 16633. Google Scholar CrossRef Search ADS PubMed 66. Maldonado-Avilés JG, Curley AA, Hashimoto Tet al. Altered markers of tonic inhibition in the dorsolateral prefrontal cortex of subjects with schizophrenia. Am J Psychiatry . 2009; 166: 450– 459. Google Scholar CrossRef Search ADS PubMed 67. Straub RE, Lipska BK, Egan MFet al. Allelic variation in GAD1 (GAD67) is associated with schizophrenia and influences cortical function and gene expression. Mol Psychiatry . 2007; 12: 854– 869. Google Scholar CrossRef Search ADS PubMed 68. Neske GT. The slow oscillation in cortical and thalamic networks: mechanisms and functions. Front Neural Circuits . 2015; 9: 88. Google Scholar PubMed 69. Coyle JT. NMDA receptor and schizophrenia: a brief history. Schizophr Bull . 2012; 38: 920– 926. Google Scholar CrossRef Search ADS PubMed 70. Miller KJ, Sorensen LB, Ojemann JG, den Nijs M. Power-law scaling in the brain surface electric potential. PLoS Comput Biol . 2009; 5: e1000609. Google Scholar CrossRef Search ADS PubMed 71. Voytek B, Kramer MA, Case Jet al. Age-related changes in 1/f neural electrophysiological noise. J Neurosci . 2015; 35: 13257– 13265. Google Scholar CrossRef Search ADS PubMed © The Author(s) 2017. Published by Oxford University Press on behalf of the Maryland Psychiatric Research Center. All rights reserved. For permissions, please email: email@example.com
Schizophrenia Bulletin – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera