Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 1.
Published in final edited form as: Schizophr Res. 2019 Oct 12;219:34–46. doi: 10.1016/j.schres.2019.09.005

Electoretinographic evidence of retinal ganglion cell-dependent function in schizophrenia

Pantea Moghimi f,1,2, Nathalia Torres Jimenez a,c,d,1, Linda K McLoon a,c,d, Theoden I Netoff a,c,e, Michael S Lee d, Angus MacDonald III a,b,*, Robert F Miller a,c,d
PMCID: PMC7442157  NIHMSID: NIHMS1608737  PMID: 31615740

Abstract

Schizophrenia is a complex disorder that is diagnosed mainly with clinical observation and evaluation. Recent studies suggest that many people with schizophrenia have abnormalities in the function of the N-methyl-D-aspartate receptor (NMDAR). The retina is part of the central nervous system and expresses the NMDAR, raising the possibility of the early detection of NMDAR-related schizophrenia by detecting differences in retinal function. As a first-step, we used two non-invasive outpatient tests of retinal function, the photopic negative response (PhNR) of the light-adapted flash-electroretinogram (PhNR-fERG) and the pattern ERG (PERG), to test individuals with schizophrenia and controls to determine if there were measurable differences between the two populations. The PhNR-fERG showed that males with schizophrenia had a significant increase in the variability of the overall response, which was not seen in the females with schizophrenia. Additionally at the brightest flash strength, there were significant increases in the PhNR amplitude in people with schizophrenia that were maximal in controls. Our results show measurable dysfunction of retinal ganglion cells (RGCs) in schizophrenia using the PhNR-fERG, with a good deal of variability in the retinal responses of people with schizophrenia. The PhNR-fERG holds promise as a method to identify individuals more at risk for developing schizophrenia, and may help understand heterogeneity in etiology and response to treatment.

Keywords: Electroretinogram, Schizophrenia, NMDA receptor, Pattern ERG, PhNR, Retina, Retinal ganglion cells, Biomarker

1. Introduction

Schizophrenia is a complex disorder with variable onset of symptoms and progression. One strong association is the link to altered excitation and inhibition in cortical circuitry related to N-methyl-D-aspartate receptor (NMDAR) hypofunction (Balu and Coyle, 2015). While other potential underlying biochemical mechanisms have been hypothesized to cause schizophrenia, proteomic and genomic evidence suggests that a relatively large number of individuals diagnosed with schizophrenia have abnormalities in NMDAR-interacting molecules at the synapse (Föcking et al., 2015). Antagonists to the NMDAR, such as phencyclidine or ketamine, produce schizophrenia-like symptoms in healthy subjects and provide evidence for the receptor’s involvement in the disease (Javitt and Zukin, 1991; Driesen et al., 2013). These studies collectively provide significant support for the NMDAR hypofunction hypothesis, also known as the glutamate dysfunction hypothesis, in the onset and progression of schizophrenia (Olney and Farber, 1995; Coyle, 1996; Moghaddam and Javitt, 2012; Snyder and Gao, 2013).

A direct prediction of the NMDAR hypothesis is that the activity of neurons that express NMDAR is altered in people with schizophrenia. Since NMDARs are expressed by many neuronal types, including those in the retina (Shen et al., 2006), retinal electrophysiology may indicate abnormal physiological responses among individuals with schizophrenia. This study tested the hypothesis that NMDAR dysfunction could be directly and non-invasively measured by monitoring retinal neuronal responses to visual stimulation in people with schizophrenia.

One measure of retinal neuron activity in response to visual stimulation is electroretinography (ERG). The ERG is a non-invasive, low-cost test that records the electrical potentials evoked by visual stimuli and is widely used in clinical settings (McCulloch et al., 2015). ERG data are collected under different protocols with variable visual stimuli in order to capture activity of different retinal neurons. The two ERG tests we used in the present study were the full-field or flash ERG (fERG) and the pattern ERG (PERG). The fERG measures the retinal response to a uniform flash of light, while the PERG stimulates the retina using an alternating, contrast-reversing black and white pattern, usually a checkerboard. These two tests complement each other because they target different neuronal populations in the retina. This should allow visualization of significant functional changes in people with schizophrenia when compared to normal controls.

Various protocols of the fERG, such as dark- and light- adapted fERGs, have been used to compare retinal activity of people with schizophrenia and controls. In one study people with schizophrenia demonstrated altered retinal activity when compared to bipolar disorder (Balogh et al., 2008), but only in the “acute stage” of schizophrenia. Other studies showed decreased ERG a- or b-wave amplitudes in people with schizophrenia compared to healthy subjects (Lavoie et al., 2014; Silverstein and Rosen, 2015; Adams and Nasrallah, 2018; Demmin et al., 2018). These studies used fERG protocols that capture activity of the outer retinal layer, specifically the photoreceptors and bipolar cells. Only one study to date measured the activity of the inner retinal layers, the retinal ganglion cells (RGCs) and amacrine cells, in their assessment of people with schizophrenia compared to controls (Demmin et al., 2018). They used a light-adapted fERG that evokes a photopic negative response, the PhNR protocol of the fERG (PhNR-fERG). Evaluating the activity of the inner retinal layers in schizophrenia aids understanding of signal transmission to the brain.

It is well established that RGCs express functional NMDARs (Lukasiewicz and McReynolds, 1985; Aizenman et al., 1988; Massey and Miller, 1990) that contribute 12% to light-evoked responses from RGCs in non-human primates (Cohen and Miller, 1994). The NMDAR is also present on photoreceptors, amacrine cells, and bipolar cells, which play a modulatory role in neuronal excitation and inhibition (Dixon and Copenhagen, 1992; Harsanyi et al., 1996). These studies support the hypothesis that changes in neuronal function should be visible in both the fERG and the PERG. To examine RGC function specifically, we used fERG to measure the PhNR,PhNR-fERG, (Frishman et al., 2018) and the PERG (Bach et al., 2013). The PhNR-fERG has the added benefit of recording the electrical activity of RGCs in addition to photoreceptor and bipolar neurons. This is accomplished by modifying the traditional fERG from a white to a red flash on a blue background luminance (Viswanathan et al., 2000; Frishman et al., 2018). The resultant waveform consists of four components: 1) the a-wave, a corneal-negative deflection that captures activity of cone photoreceptors; 2) the b-wave, a subsequent corneal-positive deflection that captures activity of retinal bipolar cells; 3) the oscillatory potentials, high-frequency oscillations that are superimposed on the b-wave, capturing activity from the inner nuclear layer (Wachtmeister, 1998); and 4) the photopic negative response (PhNR), a corneal-negative deflection following the b-wave that captures activity of the RGCs (Machida, 2012). While earlier studies attributed the b-wave response to both bipolar and Müller cells, examination of mice whose Müller cells were functionally inactivated showed that the b-wave was unchanged, supporting the view that Müller cells are not involved in b-wave activity (Kofuji et al., 2000). The changes in amplitude and latency of each of these components may serve as biomarkers for disease state. Previous PhNR-fERG testing in people with schizophrenia revealed a significant difference in a-wave amplitude at brighter flash strengths, but borderline significance for PhNR amplitude (Demmin et al., 2018). An important question left unanswered was whether there were any functional changes to RGCs in these retinas. Despite no significant difference in the PhNR amplitude, the same study reported a significant correlation between the PhNR amplitude and patient symptom severity (Demmin et al., 2018). It may be that the flash strength used to evoke the PhNR in their study was insufficient for demonstrating RGC deficits in people with schizophrenia. Since PhNR amplitude positively correlates with flash strength (Frishman et al., 2018), recording at higher flash strengths might increase sensitivity to more subtle NMDAR dysfunction expressed in the retina. In the present study, we recorded PhNR-fERG test at three flash strengths in order to evaluate whether functional changes to RGCs are present in people with schizophrenia.

We also analyzed a component present in the fERG test rarely considered for evaluating differences in people with schizophrenia - the oscillatory potentials (OPs). The OPs ride on the ascending b-wave, but their cellular origin is more distributed within the inner retina rather than specifically due to photoreceptor or bipolar cells (Wachtmeister, 1998). The first study reporting differences in the OPs between people with schizophrenia and controls demonstrated greater amplitude variance in males with schizophrenia (Raese et al., 1982); however, that work was not replicated (Schechter et al., 1987). Since then, evaluation of the OPs in people with schizophrenia has not been reported, although new methods of extracting the OPs from the b-wave are now readily achievable (Gauvin et al., 2014) through the use of 75 Hz and 300 Hz bandpass filters (McCulloch et al., 2015). These studies evaluating OP in schizophrenia only measured amplitude in the time domain but did not perform a frequency-domain analysis. Given that OPs are collected during most fERG protocols - scotopic, photopic, and PhNR protocols - and provide information regarding RGC activity, we analyzed the OPs in our study. Time-frequency analysis in nonhuman primates showed that OPs can be deconstructed into two frequency bands, slow and fast, that distinguishes between activity of retinal neuron layers (Zhou et al., 2007). Pharmacological studies revealed that the low-frequency OPs, from 75 Hz to 100 Hz, originate from non-spiking activity of amacrine cells and retinal neurons with non-sodium spiking mechanisms; high-frequency OPs, from 100 Hz to 300 Hz, originate from spiking retinal ganglion cells and partially from amacrine cells. Analysis of the OPs holds promise for investigating inner retinal function in schizophrenia because they reflect different populations of retinal neurons, making the fERG an all-in-one test with information from all retinal layers.

In addition to the PhNR-fERG test, we evaluated PERG responses as a second measure of RGC function. To our knowledge, no studies have reported PERG differences in people with schizophrenia, although it is well established that RGCs contribute to generation of the PERG response (Holder, 2001). The PERG response is composed of three components that are named for the negative and positive neuronal responses and the expected times (ms) at which each peaks - N35, P50, and N95 (Holder, 2001). While the specific neuronal origins of each of these components are not fully understood (Porciatti, 2015), direct attribution to inner retinal neurons has been validated across species including humans (Ventura and Porciatti, 1984; Aldebasi et al., 2004), monkeys (Maffei et al., 1985; Viswanathan et al., 2000), and rodents (Miura et al., 2009). The PERG response is sensitive to retinal pathology, such as in glaucoma (Preiser et al., 2013) and diabetic retinopathy (Prager et al., 1990). As a well-established ophthalmological tool for assessing RGC function, PERG is arguably superior for assessing RGC dysfunction in schizophrenia.

One complicating factor is the role of sex in NMDAR activity. Rodent studies demonstrated differences in NMDAR activity in the brain between male and female rats (Woolley, 1998; Cyr et al., 2001; McEwen, 2002; McRoberts et al., 2007). Our recent study examining fERG in a control and mouse model of schizophrenia, a serine racemase knock-out which results in NMDAR hypofunction, reported significant differences between a- and b- wave amplitudes in male and female mice regardless of genotype or flash strength (Torres Jimenez et al., 2019). In the present study, we also included sex as a variable. A sex-dependent difference in NMDAR activity and fERG variables may relate to the numerous sex differences in schizophrenia, such as age of onset, symptom expression, and treatment response (Peterson, 1968; Loranger, 1984).

We compared retinal responses using PERG and PhNR-fERG protocols from a cohort of people with chronic schizophrenia and normal controls. For the PERG test, we analyzed N35, P50, and N95 amplitudes and latencies. For the PhNR-fERG test, we evaluated three different flash strengths and measured the a-wave, b-wave, and PhNR amplitude and implicit times. We also analyzed high-and low- frequency OPs to determine whether differences in the activity of inner retinal neurons were present in people with schizophrenia. All analyses included sex as a factor to investigate whether sex-related ERG abnormalities were present in people with schizophrenia.

2. Methods

2.1. Participants

We recruited 30 people with schizophrenia and 34 demographically-similar healthy controls, aged 18 to 65, within a large metropolitan area as reported in Table 1. Patient diagnosis was confirmed using the SCID-IV (First et al., 2002). The severity of psychotic symptoms was rated using the Brief Psychiatric Rating Scale (BPRS) (Ventura et al., 1993). The BPRS values for each of the people with schizophrenia was graphed based on age (Supplemental Fig. 1; Table 1). The BPRS showed that all subjects had a significant number of symptoms associated with schizophrenia, with a range of scores from 26 to 93. Most subjects were on an array of anti-psychotic medications; 6 people with schizophrenia (20%) were not using anti-psychotic medicine at the time of ERG recording. Among patients, the average duration of psychosis was greater than 10 years, and the average duration of medication use was greater than 5 years. Using multiple linear regression analyses, there were no significant differences in disease severity between males and females with schizophrenia nor was age a significant predictor for BPRS score. All participants gave informed consent. For 6 controls and 5 people with schizophrenia, minority data is unavailable. All procedures were done in accordance with a University of Minnesota IRB approved protocol and adhered to the principles of the Declaration of Helsinki.

Table 1.

Clinical and demographic characteristics of the participants. Data are mean (standard deviation). Two of the 30 subjects with schizophrenia did not have a BPRS score and were omitted from this analysis.

Subject Information
Scz (N = 30) HC(N = 35) Difference
Sex (% female) 46.70% 48.60% 3.90%
Age (M ± S.D.) 40.27 ± 11.53 39.97 ± 13.01 0.75%
Minority (%) 24% 23% 4.26%
BPRS score (M ± S.D.) 58.7 ± 15.99 0 -
Unmedicated 6 (20%) - -
Open in a new tab

2.2. Data collection

Participants came to three sessions, which included a clinical psychiatric evaluation and an ophthalmologic examination to evaluate for any eye diseases that could affect ERG outcomes. In the control group, 73.5% had best corrected visual acuity of 20/20 in each eye and in the group with schizophrenia 80% had best corrected visual acuity of 20/20. All subjects had clear vision and no history of eye disease that would affect the ERG data. ERG data were collected in the third session using the Diagnosys LLC (Boston, MA) system at a sampling frequency of 1.2 kHz. DTL fiber electrodes manufactured by Diagnosys LLC (Boston, MA) were utilized for all tests. Two types of ERG responses were collected: responses to a full-flash red light on a blue background (PhNR-fERG test) and responses to contrast-reversing checkerboard or bars (PERG test). The time of the PhNR-fERG and the PERG tests for each subject are shown in Supplemental Fig. 2. The majority were between 9:30 a.m. and 6:30 p.m., with only 2 subjects at 7:30 p.m. However, it should be noted that a number of recent studies looking at the ERG responses compared to time of day showed no changes in ERG amplitudes or implicit times due to time of day the recordings were taken (Heinemann-Vernaleken et al., 2000; Marcus et al., 2004) unless performed at very late evening times (e.g. 11 p.m.) (Rufiange et al., 2002; Lavoie et al., 2010). All of our recordings were within the time window shown not to be affected by time of day.

2.2.1. PhNR-fERG test

The pupils were dilated with Tropicamide (1%) (Sandoz, Princeton, NJ), and 50 red light flashes were presented against a blue background at a frequency of 4 Hz. Three different light intensity levels were used as stimuli. Table 2 summarizes both luminance and retinal illuminance of the visual stimuli. We used constant luminance level visual stimuli. Retinal illuminance was calculated as the constant luminance multiplied by the mean pupil area across subjects.

Table 2.

Parameters of the visual stimuli used for data collection. The average retinal luminance for each stimulus are calculated by multiplying luminance by average pupillary area across subjects and reported within brackets [ ].

PhNR-fERG stimulus conditions
Red flash strength Flash duration Blue background luminance Interstimulus duration
1 cd-s/m2 [60Td.s] 4 ms 10cd -s/m2 [600Td.s] 250 ms
5 cd-s/m2 [300Td.s] 4 ms 10cd -s/m2 [600Td.s] 250 ms
7 cd-s/m2 [420Td.s] 4 ms 10cd -s/m2 [600Td.s] 250 ms
Transient PERG stimulus conditions
Visual stimulus display Stimulus pattern Mean luminance Reversal rate
CRT Checkerboard 999 cd/m2 2.1 rps
LED Bar 999 cd/m2 8 rps
Open in a new tab

2.2.2. PERG test

We collected two datasets under the PERG protocol where pupil dilation was not used. For the first dataset, PERG responses were collected in response to a contrast reversing black and white checkerboard pattern at 0.8 degrees spatial frequency, with a luminance of 999 cd s/m2. Contrast was reversed at a rate of 2.1 reversals per second (1.05 Hz). Visual stimuli were presented on a CRT display at 1000 mm distance and at 100% contrast. The response to 150 contrast reversal for each stimulus was recorded for each subject.

Partway through the study, an LED display was implemented, which allowed for faster stimulation of the retina (Monsalve et al., 2017). Twenty-four controls (11 males and 13 females) and 25 people with schizophrenia (13 males and 12 females) returned for additional PERG recording with the LED display. The LED PERG data were collected using reverse contrasting black and white bars at a luminance of 999 cd s/m2 (Table 2). All stimuli had a spatial frequency of 0.5 degrees and were presented at 100% contrast. The subjects viewed the monitor at a distance of 57 cm. Contrast of each bar was reversed at a rate of 8 reversals per second (4 Hz). The response to 150 contrast reversal for each stimulus was recorded for each subject.

2.3. Data analysis

2.3.1. PhNR-fERG test

For each subject, single ERG traces in response to each light flash were averaged. Prior to averaging, noisy traces were identified and excluded. A trace was excluded if it deviated more than one standard deviation from the average in more than 50% of time points. Visual inspection of the traces showed that this criterion robustly identified noise.

The average waveform for each subject was used to measure amplitude and latency of the a-wave, b-wave, and PhNR components. Example waveforms are shown in Fig. 1. Prior to measuring any of the indices, each waveform was visually inspected. Subjects with noisy waveforms where at least one of the indices could not be identified with certainty were excluded (4 subjects with schizophrenia, 2 females; and 6 control subjects, 3 females). Each waveform was low-pass filtered using a digital 5th order Butterworth filter with cut-off frequency at 50 Hz to filter out and avoid contamination from oscillatory potentials (Weymouth and Vingrys, 2008). Amplitude of the a-wave component was measured as the trough of the first negative deflection after visual stimulus onset with respect to the baseline. Latency or implicit time of the a-wave component was measured as the occurrence time of the trough post stimulus. Amplitude of the b-wave component was calculated as the difference between the peak of the positive deflection and the a-wave trough. Implicit time of the b-wave component was calculated as occurrence time of the peak post-stimulus. We measured amplitude of the PhNR deflection using two indices previously used to report PhNR amplitude. The first index measured the negativity of the PhNR deflection with respect to baseline 72 ms after the onset of the visual stimulus (N72) (Kundra et al., 2016; Demmin et al., 2018). The PhNR reaches its maximum deflection on average around 72 ms post-stimulus (Kundra et al., 2016), and the amplitude of the waveform at this time point is commonly used as an index for PhNR amplitude. Previous studies, however, reported altered latencies in the photoreceptor responses of people with schizophrenia (Hebert et al., 2015). Consequently, we could not assume that the PhNR latency would occur at 72 ms in the subjects with schizophrenia. To compensate for this, we used an index that measures the trough of the PhNR component with respect to the baseline (Kundra et al., 2016; Demmin et al., 2018). Implicit time of PhNR was measured as time of the trough poststimulus. As a control, we repeated all statistical analyses with the a-wave and b-wave amplitudes measured using unfiltered waveforms and obtained qualitatively similar results.

Fig. 1.

Fig. 1.

Open in a new tab

Individual PhNR-fERG waveform for a single subject at a single flash (1 cd s/m2) intensity from a A) female control, (B) female with schizophrenia, (C) male control, and (D) male with schizophrenia with time from stimulus onset in milliseconds (ms) on the x-axis and amplitude (Amp.) in microvolts (μV) on the y-axis. Arrows indicate the a-wave, bwave, i-wave, and PhNR components of the PhNR-fERG.

To measure the strength of fast and slow OPs, total power in the 75–100 Hz and 100–300 Hz frequency bands were calculated (Zhou et al., 2007). OPs typically appear shortly after the stimulus onset and continue until 65 ms post-stimulus (Zhou et al., 2007). We calculated total power in each frequency band from the power spectral density of the waveform from 0 to 70 ms post-stimulus. We chose 70 ms instead of 65 ms to make sure the entire duration of OPs was included. Although strength of the OPs was quantified within a time window where other post-b-wave components lie, OPs were isolated from other components in the frequency-domain rather than the time-domain. I-wave and PhNR correspond to frequency bins around 11 Hz (Kundra et al., 2016); while OPs were post-75 Hz. Values from the two eyes were averaged for each subject.

2.3.2. PERG test

Single traces from each contrast reversal were averaged for each subject after exclusion of outliers. A trace was excluded if it deviated more than one standard deviation from the average in more than 60% of time points. A less stringent criterion to define outliers was used for the PERG test in comparison to the PhNR-fERG test because we had more traces per subject for the PERG, and therefore, the average was more robust. Visual inspection of the traces proved this criterion to be reliable. Average traces were subsequently filtered using a digital 5th order Butterworth bandpass filter with cut-off frequencies at 6 and 72 Hz to filter out low frequency drifts in the signal as well as high frequency noise. Amplitude and implicit times of each component were measured from the average trace from each subject. Subjects with noisy waveforms were identified through visual inspection of each waveform and excluded from further analysis. One female subject with schizophrenia was excluded from the data collected using CRT. Four control subjects (two female) were excluded from the data collected using LED. N35 amplitude was measured as the trough of the first negative deflection occurring 30–35 ms post-stimulus with respect to baseline. Occurrence time of the trough post-stimulus was measured as N35 implicit time. P50 amplitude was measured as the peak of the positive deflection following the N35 component with respect to baseline. P50 implicit time was measured as time of the peak post-stimulus. N95 was measured as trough of the negative deflection following the P50 deflection with respect to baseline. Occurrence time of the trough post-stimulus was used as the N95 implicit time. Amplitude and implicit times from the right and left eyes were averaged. When data from only one eye were available, values from the single eye were used.

2.4. Statistical analysis

For the statistical analyses on data collected from the PhNR-fERG response, we analyzed 11 dependent variables: a-wave, b-wave, and PhNR amplitude and implicit time; b/a ratio; low-frequency OPs (75e100 Hz), high-frequency OPs (100–300 Hz), and combined frequency OPs (75–300 Hz). For each of these dependent variables, we conducted a 3-way mixed ANOVA with two between-subject factors, disease state and sex, and one within-subject factor, flash strength. We performed Levene’s test for equal variances for each group of the between-subjects factor at each flash strength. To comply with the assumption of sphericity, Greenhouse-Geisser corrected statistics were reported. We tested simple two-way interactions (disease state·sex) at each level of flash strength and flash strength against one of the between-subject factors, (flash strength and sex) and (flash strength and disease state). We analyzed the main effects at each flash strength. Main effects from the two-way interactions were reported for all dependent variables attained at each flash strength. The variances for the a-wave, N72, and PhNR amplitudes were not equal, necessitating the use of Welch independent sample t-tests to evaluate differences in disease state at each level of flash strength.

For the statistical analyses on data collected from the PERG protocol, we analyzed 6 dependent variables: N37, P50, and N95 amplitude and implicit time. We performed a two-way between-subject ANOVA to examine the effects of disease state and sex on each dependent variable. Since interaction effects were not significant, analysis of main effect was conducted from pairwise-comparisons of unweighted marginal means. Analysis of the PERG recorded from CRT and LED visual display were conducted separately. Statistics were performed using SPSS Statistics 25 program (IBM, Armonk, NY).

3. Results

3.1. PhNR-fERG analyses

When the PhNR-fERG waveforms from all the tested subjects were graphed (Fig. 2), several generalizations could be made. First, variability in the strength of the neuronal responses were evident in the ERG waveforms at each flash strength. There were few overt changes in the ERG between females with schizophrenia and female controls (Fig. 2A,B). However, there was a huge variability in the waveforms in the male controls, and an even greater variability in the males with schizophrenia compared to male controls (Fig. 2C,D), as highlighted by the variance and standard deviation of each group of various dependent variables (Supplemental Table 1).

Fig. 2.

Fig. 2.

Open in a new tab

Filtered PhNR-fERGs at the brightest flash strength of 10 cd$s/m2 from (A) female controls, (B) females with schizophrenia, (C) male controls, and (D) males with schizophrenia. Each colour represents the PhNR-fERG from a single individual and black line represents the mean of each group. Time from stimulus onset in milliseconds (ms) is on the xaxis and amplitude in microvolts (mV) on the y-axis.

We analyzed the a- and b-wave amplitudes and b/a ratio (Supplemental Fig. 3, Table 3, Supplemental Table 2) and implicit times (Supplemental Table 3) for all subjects. There was no statistically significant difference in a-wave amplitude and implicit time between people with schizophrenia and controls at any flash strength (Table 3, Supplemental Table 3). For b-wave amplitude and b/a ratio, there were no significant 3-way nor 2-way interactions; however, there was a borderline b-wave main effect of disease state at the middle flash strength (Supplemental Table 2). While there were no statistically significant differences, b-waves trended toward reduced amplitude and delay at all flash strengths in people with schizophrenia in comparison to healthy controls. Since this effect was seen among males, females, and combined datasets, it warrants further investigation with a larger sample size. There were no significant 3-way or 2-way interactions in b-wave implicit time (Supplemental Table 3). While not significant in this study, a trend toward increased a-wave amplitude in male subjects with schizophrenia was observed. These data correlate with our study in the NMDAR hypofunction mouse, where significant differences were seen between male and female mice (Torres Jimenez et al., 2019). At the highest flash strength, the a-wave amplitude of NMDAR hypofunction mouse was significantly elevated in males, but not in females. These trends suggest that a clearer differentiation of these waveforms would be more likely at the higher flash strengths; our future study will use brighter flash strengths as well as increased sample size.

Table 3.

Statistical analysis of the PhNR-fERG, examining the amplitude of the a-wave, PhNR and PhNR negativity at 72 ms poststimulus (N72). Red indicates significance. Blue indicates approaching significance.

a-wave amplitude PhNR amplitude N72
Independent t-test
Disease state at each level of flash strength 1 cd.s/m2 t(53)=−0.456, p=0.650, d=.12 t(53)=−1.050, p=0.298, d=.28 t(53)=−0.565, p=0.575, d=.15
5 cd.s/m2 t(53)=−0.859, p=0.394, d=.23 t(53)=−1.615. p=0.112, d=.44 t(53)=−1.445, p=0.154, d=.39
7 cd.s/m2 t(33)=−1.650, p=0.108, d=.46 t(27)=−2.110, p=0.046, d=.59 t(28)=−2.110, p=0.044, d=.60
Open in a new tab

NMDARs are highly expressed by RGCs (Aizenman et al., 1988; Massey and Miller, 1990). Any functional change in people with schizophrenia would be evident in the PhNR and N72 amplitude, which are specifically associated with RGC function (Viswanathan et al., 2001; Li et al., 2005). At the brightest flash strength, the Welch’s t-test showed disease state differences in both the PhNR amplitude and N72 (Table 3). The mean PhNR amplitude was significantly different between people with schizophrenia and healthy controls t(27.265) = −2.092, p = 0.046 at the brightest flash strength (Fig. 3, Table 3). Mean PhNR amplitude was 40.6 mV higher (95%CI, −80.34 to −0.79) in people with schizophrenia than controls. Mean N72 amplitude was significantly different between people with schizophrenia and healthy controls t(28.148) = 2.110, p = 0.044 at the brightest flash strength (Fig. 3, Table 3). Mean N72 was 15.6 mV higher (95%CI, −6.4 to 37.5) in people with schizophrenia compared to controls (Fig. 3). The brightest flash strength was chosen because it was below the level of response saturation in humans (Binns et al., 2011), resulting in a PhNR amplitude that should not change at the brightest flash strength in healthy controls. The response saturation threshold appeared to be different in people with schizophrenia, as there was no change in the PhNR amplitude relative to flash strength in healthy controls but showed a continued increase in people with schizophrenia (Fig. 3, Table 3). As this flash strength was not tested by other investigators, it would be important to validate this stimulus condition and report variability (Supplemental Table 4). This suggests that use of brighter flash strengths should improve visualization of differences between people with schizophrenia and controls. Furthermore, to evaluate whether changes in the PhNR amplitude were due to changes in retinal ganglion cells or from more distal retina (i.e. photoreceptors, bipolar cells), a one-way Welch’s F test was conducted to analyze differences in PhNR/b-wave ratio across the 4 groups (Supplemental Table 4). The PhNR/b-wave ratio was statistically different across groups (F(3, 82.4) = 3.23, p = 0.027), with men with schizophrenia having a larger PhNR/b-wave ratio. However, none of the pairwise comparisons using the Games-Howell post-hoc test were significant. We show that in males with schizophrenia, the PhNR amplitude is enhanced but not due to a decreased response from bipolar cells, as would be reflected in the b-wave amplitude. There were no significant 3-way or 2-way interactions in PhNR implicit time (Supplemental Table 3).

Fig. 3.

Fig. 3.

Open in a new tab

Graphs represent the (A) PhNR trough amplitude, (B) PhNR implicit time, and (C) negativity of the PhNR deflection at 72 ms (N72) averaged for all subjects at each flash strength. Black represents the controls, dotted lines represent the people with schizophrenia, gray dotted lines represent all people with schizophrenia, blue represents the male people with schizophrenia, and red represents the females with schizophrenia. Ampl is amplitude. Impl. is implicit time. Cd s/m2 represents candelas per second per meters squared, a measure of constant luminance. Bars indicate standard error. HC is healthy controls; Scz is people with schizophrenia. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Oscillatory potentials (OP) were analyzed based on the frequency band split in three ways: between the frequencies of 75–100 Hz (low-frequency OPs), between the frequencies of 100–300 Hz (high-frequency OPs), and combination of the first two, between 75 and 300 Hz (Supplemental Fig. 4, Table 4, Fig. 4). While there were no significant differences in the 3-way or 2-way interactions in any of the OP bandwidths (Table 4), the 75–100 Hz bandwidth had a probability of 0.095, with effect of size <di>(ɳ2p = 0.39)</di>. While the two-way interactions were not significant, likely due to sample size and variability in the disease state, main effects analyses showed significant differences for sex regardless of disease state in high-frequency OPs and the combined OP data (Table 4). This suggests that OP analysis might identify males with schizophrenia, and future analyses should be cautious when pooling male and female ERG data.

Table 4.

Statistical analysis of the oscillatory potentials (OPs) derived from the PhNR-fERG for the low-frequency OPs, between 75 and 100 Hz, the high-frequency OPs between 100 and 300 Hz, and the total combined OPs between 75 and 300 Hz. Red indicates significance. Blue indicates approaching significance.

Low Frequency OPs (75-100 Hz) High Frequency OPs (100-300 Hz) Combined (75-300 Hz)
Flash Strength . Sex . Disease State F(2.0,119)=0.3, p=0.73, ηp2=0.005 F(1.7,102)=0.02, p=0.97, ηp2=0.000 F(1.8,110)=0.1, p=0.86, ηp2= 0.002
Simple two-way interactions
Flash Strength . Disease State F(2.0,119)=2.4, p=0.095 F(1.7,102)=1.4, p=0.249 F(1.8,110)=1.9, p=0.161
Flash Strength . Sex F(2.0,119)=0.1, p=0.874 F(1.7,102)=1.3, p=0.286 F(1.8,110)=0.9, p=0.390
Disease State. Sex at each level of flash strength 1 Cd.s/m2 F(1,60)=1.6, p=0.204 1 Cd.s/m2 F(1,60)=0.2, p=0.663 1 Cd.s/m2 F{1,60)=0.92, p=0.34
5 Cd.S/m2 F(1,60)=0.1, p=0.762 5 Cd.s/m2 F(1,60)=0.03, p=0.872 5 Cd.S/m2 F{1,60)=0.09, p=0.76
7 Cd.S/m2 F(1,60)=0.4, p=0.538 7 Cd.s/m2 F( 1,60)=0.1, p=0.767 7 Cd.s/m2 F{1,60)=0.2, p=0.66
Main Effects
Main Effects of the two-way interaction at each level of flash strength Disease State Sex Disease State Sex Disease State Sex
1 cd.s/m2 0.445 0.052 1 cd.s/m2 0.808 0.008 1 cd.s/m2 0.59 0.01
5 cd.s/m2 0.809 0.069 5 cd.s/m2 0.688 0.237 5 cd.s/m2 0.78 0.13
7 cd.s/m2 0.057 0.278 7 cd.s/m2 0.160 0.556 7 cd.s/m2 0.09 0.41
Open in a new tab

Fig. 4.

Fig. 4.

Open in a new tab

Graphs summarize the power analysis of the oscillatory potentials. (A–C) Mean total power of all oscillatory potentials between 75 and 300 Hz in microVolts (μV), (D–F) mean total power of oscillatory potentials between 75 and 100 Hz in mV, and (G-I) mean total power of oscillatory potentials between 100 and 300 Hz in mV. Cd·s/m2 represents candelas per second per meters squared, a measure of constant luminance. Bars represent standard error. HC is healthy controls; Scz is people with schizophrenia.

3.2. PERG analyses

PERG measurements were analyzed for both the CRT and the LED-generated data sets (Tables 5, 6; Figs. 5, 6) and showed a trend, but not significance, for differences between people with schizophrenia and controls. There were no significant differences between N35, P50, or N95 amplitude or implicit time with the evoked PERG using either the LED or CRT visual displays. The data trended toward significance for an interaction between disease state and sex for the N95 implicit time (p = 0.07) and a main effect of disease state for the N35 amplitude (p = 0.06) for the LED-derived data (Table 6). The CRT-derived data showed a borderline main effect of disease state (p = 0.054) for the N95 implicit time. Sex appears to be an important variable interacting with disease state on the N95 implicit time, but LED-display may be necessary to detect that interaction. Toward toward Due to the large and unexpected significant variance in the ERGs of people with schizophrenia (Fig. 2), a larger cohort of subjects would likely have generated statistical significance. These data, in combination with the PhNR results, suggest that differences in ganglion cell function are likely in people with schizophrenia compared to controls.

Table 5.

Statistical analysis for the PERG evoked using the CRT display, examining the N35 amplitude and implicit time, P50 amplitude and implicit time, and N95 amplitude and implicit time.

PERG evoked using CRT visual display
N35 amplitude P50 amplitude N95 amplitude
Disease State . Sex F(1,60) = 2.0, p = 0.164, ηp2 = 0.032 F(1,60) = 0.003, p = 0.955, ηp2 = 0.000 F(1,60) = 1.0, p = 0.331, ηp2 = 0.016
Main effects of disease F(1,60) = 2.1, p = 0.151 F(1,60) = 0.2, p= 0.625 F(1,60) = 0.1, p = 0.780
Main effects of sex F(1,60) = 0.4, p = 0.527 F(1,60) = 2.1, p = 0.157 F(1,60) = 1.2, p = 0.279
N35 implicit time P50 implicit time N95 implicit time
Disease State . Sex F(1,60) = 0.6, p = 0.430, ηp2 = 0.010 F(1,60) = 2.6, p = 0.113, ηp2 = 0.041 F(1,60) = 1.1, p = 0.302, ηp2 = 0.018
Main effects of disease F(1,60) = 0.2, p = 0.664 F(1,60) = 0.03, p = 0.875 F(1,60) = 3.9, p = 0.054
Main effects of sex F(1,60) = 0.004, p = 0.948 F(1,60) = 0.03, p = 0.874 F(1,60) = 0.1, p = 0.800
Open in a new tab

Table 6.

Statistical analysis for the PERG evoked using the LED display, examining the N35 amplitude and implicit time, P50 amplitude and implicit time, and N95 amplitude and implicit time. Blue indicates approaching significance.

PERG evoked using LED visual display
N35 amplitude P50 amplitude N95 amplitude
Disease State. Sex F( 1,45)=0.1, p=0.763, ηp2=0.002 F(1,45)=0.3, p=0.579, ηp2=0.002=0.007 F(1,45)=1.6, p=0.218, ηp2=0.002=0.034
Main Effects of Disease F{1,45)=3.7, p=0.062 F{1,45)=0.2, p=0.682 F(1,45)=2.7, p=0.108
Main Effects of Sex F(1,45)=2.0, p=0.167 F{1,45)=2.4, p=0.126 F(1,45)=0.1, p=0.795
N35 implicit time P50 implicit time 95 implicit time
Disease State. Sex F(1,45)=2.8, p=0.099, ηp2=0.002=0.059 F(1,45)=0.5, p=0.490, ηp2=0.002=0.011 F(1,45)=3.4, ηp2=0.002=0.071
Main Effects of Disease F(1,45)=0.2,p=0.673 F{1,45)=2.5, p=0.124 F(1,45)=0.2, p=0.638
Main Effects of Sex F(1,45)=0.0, p=0.991 F(1,45)=0.03, p=0.865 F(1,45)=0.2, p=0.638
Open in a new tab

Fig. 5.

Fig. 5.

Open in a new tab

Graphs of all PERG traces from the (A) healthy control female subjects examined with the CRT display (HC); (B) females with schizophrenia (Scz) examined with the CRT display; (C) control male subjects (HC) examined with the CRT display; (D) males with schizophrenia (Scz) examined with the CRT display; (E) control female subjects examined with the LED display (HC); (F) females with schizophrenia (Scz) examined with the LED display; (G) control male subjects (HC) examined with the LED display; (H) males with schizophrenia (Scz) examined with the LED display. Black lines depict the mean of the PERG response, thick red lines depict the mean of the PERG response for female subjects, and thick blue lines represent the mean of the PERG response for male subjects. X-axis shows time (ms) from pattern reversal. Amp is amplitude in microVolts (μV). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 6.

Fig. 6.

Open in a new tab

Analysis of the PERG using the LED and CRT displays. (A) N35 amplitude (Amp.) and (B) N35 implicit (Imp.) time; (C) P50 amplitude (Amp.) and (D) implicit (Imp.) time; and (E) N95 amplitude (Amp.) and (F) implicit (Imp.) time. Black open circles are control (HC) males subjects and blue open circles are males with schizophrenia (Scz). Black asterisks are control (HC) females and red asterisks are females with schizophrenia (Scz). Bars indicate standard error. Amplitude is in microVolts (μV) and implicit time is in milliseconds (ms). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

4. Discussion

The PhNR-fERG and PERG allow visualization of the function of different retinal neurons, with the PhNR-fERG showing a large amount of variability in men with schizophrenia specifically. The PhNR-fERG and PERG both assess activity of RGCs. Analysis of the PhNR-fERG showed a significant increase of PhNR and N72 amplitude at only the highest flash strength in people with schizophrenia, suggesting a shift in the saturation threshold of RGCs in people with schizophrenia compared to controls. Similar results were seen for the low-frequency OP analyses, where there was borderline significance of disease state, suggesting changes in amacrine cell function regardless of sex. The analysis of the data from the PERG-evoked responses were largely dependent on the visual display, with LED-derived data having a greater potential than CRT to reflect interactions between sex and disease state. This was evident in the N95 implicit time and the N35 amplitude relative to disease state. A larger sample is needed to show these observed trends had statistical significance.

The large variability in the PhNR-fERG responses in people with schizophrenia is likely a reflection of the complexity of the underlying biological causes of schizophrenia. Recent genetic and proteomic studies suggested that a large proportion of people with schizophrenia have abnormalities in postsynaptic signaling complexes (Kirov et al., 2012; Focking et al., 2015), and in particular differential expression of NMDA-interacting proteins, such as CYFIP2, SYNPO, SHANK3, ESYT and MAPK3. Receptor heterogeneity would be expected to result in differences in the manifestation of schizophrenia in different individuals. These data collectively suggest that developing a method for identifying subsets of people with significant changes to their PhNR-fERG may provide a means for a more personalized medicine approach to their treatment.

The PhNR-fERG provides insight into the function of various neuronal populations in the retina, including RGC function from the PhNR. While one study comparing PhNR amplitude and implicit time did not observe any differences in people with schizophrenia in comparison to controls (Demmin et al., 2018), our data showed a significant difference. One explanation for the former lack of significance of the PhNR is that the tested flash strength of 58 Td·s may have been too weak to distinguish between people with schizophrenia and controls. We hypothesize that evaluating a wider range of light intensities would shed light on dynamic modulation of NMDARs based on light intensity demands and would reveal NMDAR dysfunction. It is well-established that at increasing light intensities there is a decline in the b-wave amplitude, a phenomenon referred to as the Photopic Hill (Wali and Leguire, 1992, Kondo et al., 2000). We evaluated whether changes in the PhNR amplitude were due to changes from distal retina, as reflected by the a- and b-waves amplitude. Like the b-wave amplitude, the Naka-Rushton equation provides a good fit for intensity-response data in order to determine if there was a relationship between different light intensities and the response (Rufiange et al., 2003; Binns et al., 2011). Flash strengths greater than 2.5 cd·s/m2 would result in response saturation, and thus no increase in PhNR amplitude of controls would occur (Binns et al., 2011). Prior to our study, it was uncertain whether response saturation would occur in people with schizophrenia. We showed that increasing flash strength resulted in increased PhNR amplitude in the people with schizophrenia, with a statistically significant increase in the PhNR amplitude in people with schizophrenia compared to controls at the brightest flash strength of 7 cd$s/m2 (420 Td$s). Our N72 results were consistent with the PhNR results. We demonstrate that at brighter flash strengths, when the PhNR response should be saturated, there was an increase in the light-evoked potentials in the RGCs in people with schizophrenia. The same increase is not seen earlier in retinal processing (bipolar cells) because our a-wave and b-wave amplitude measurements from people with schizophrenia were not statistically significantly different from controls. However, Demmin et al. showed a statistically significant reduction in the a-wave amplitude in people with schizophrenia (Demmin et al., 2018), and we saw a trend of decreasing b-wave amplitude in people with schizophrenia. Greater sample sizes are necessary to understand distal retinal involvement in retinal ganglion cell changes as reflected by the PhNR. Our work suggests that people with schizophrenia have a different threshold of saturation. This is the first report of RGC layer dysregulation in schizophrenia. Even though most ganglion cells are responsive to NMDA (Shen et al., 2006), dopamine acts on almost all retinal ganglion cells (Zhang et al., 2008), so more work is needed to understand the how an imbalance in both NMDAR and dopamine can affect the ERG response. Furthermore, future studies would need to explore whether dysregulation of the retinal ganglion cells in schizophrenia is present in both the ON and OFF pathway, as retinal ganglion cells are characterized by their on-center vs. off-center surround. This can be investigated by measuring the PhNR at the off-set period during long flash stimulation (Horn et al., 2011).

The OPs, like the PhNR, shed light into inner retinal activity. We analyzed the frequencies of the OPs in three groups to allow distinction of two cell types in the inner retina, RGCs and amacrine cells (Zhou et al., 2007). While findings did not have statistical 3way nor 2-way interactions, it may in part due to our wide range of ages. OPs were reported to be sensitive to age (Dimopoulos et al., 2014), which would confound disease state-induced changes. However, OP analyses have potential to reflect early, age-related changes related to disease state that may not be reflected in other ERG variables, and our future studies will be directed at these analyses with a larger cohort of subjects. We saw a significant interaction between the disease state and sex for the high frequency and combined OP measurement. While the results only trended toward significant main effect of disease state at the highest flash strength only in the low-frequency OPs, this would suggest potential disruption of amacrine non-spiking activity (Zhou et al., 2007). Amacrine cells are the primary inhibitory input onto RGCs. As the NMDAR is involved in the transient response of on-off amacrine cells (Dixon and Copenhagen,1992), investigation of their involvement is important.

The two PERG data sets used different monitor types and visual stimuli. The latencies of all three PERG components were longer in the data collected using the CRT compared to the LED display. The CRT monitor generates images using an electron gun that sweeps across the screen, and at any given point in time, only parts of the contrast reversed image are presented to the eye. In contrast, the LED monitor synchronously reverses contrast over the entire screen, resulting in decreased latencies (Monsalve et al., 2017). Amplitude of the PERG components were also different across the two datasets, with PERG responses recorded with the CRT monitor having higher N95 amplitudes and lower P50 amplitudes than those with the LED monitor. PERG responses collected with an LED monitor were reported to have higher amplitudes than CRT monitors (Monsalve et al., 2017), which may explain the higher P50 amplitudes in the data collected using the LED monitor. N95 amplitude was, however, lower in the LED dataset, which was recorded at a higher reversal rate and N95 amplitude was shown to decrease as reversal rate increased (Ozdamar et al., 2014). PERG responses recorded with the CRT monitor showed a tendency toward significantly increased N95 implicit times in subjects with schizophrenia compared to controls. N95 implicit time of PERG responses recorded with the LED monitor had a borderline significant interaction between sex and disease state, where N95 implicit time was higher in males with schizophrenia than control males, with no differences in female subjects. PERG responses recorded using the LED monitor and not the CRT monitor showed a borderline significant attenuation of N35 amplitude in people with schizophrenia compared to controls. In the absence of a better understanding of the exact origin of the PERG components (Porciatti, 2015), interpreting these results is not straightforward. Previous studies showed that the amplitude, but not implicit time, of PERG components was reduced in diseases with damage to RGCs such as glaucoma (Preiser et al., 2013) and diabetes (Prager et al., 1990). In addition, pharmacological blocking of spiking activity of the RGCs in primates reduced the amplitude of the N95 component (Viswanathan et al., 2000; Luo and Frishman, 2011). However, diseases such as glaucoma and diabetes are associated with the death of RGCs rather than dysfunction specifically, as we hypothesize occurs in schizophrenia. Moreover, dysfunction of NMDARs would alter the activity of other neuronal populations in the retina whose activity modulates activity of RGCs. In glaucoma, selective death of RGCs occurs while activity of photoreceptors or bipolar cells remains intact (Viswanathan et al., 2000). Our ERG results might reflect overall changes in different parts of the retinal circuitry that resulted in altered activity of RGCs as captured by the PERG response. The difference we observed between the CRT and LED data could relate to the higher temporal precision of the LED monitor, which could reveal more subtle differences among the neuronal populations under study. The difference in reversal rate or pattern of the visual stimulus may alter the dynamics of NMDAR under one stimulation condition and not the other. Our results demonstrate the potential for the PERG to visualize functional differences underlying schizophrenia when larger sample sizes are used and show the complex effects elicited by choice of visual stimulus or display type on the results. It would be valuable in the future to subdivide subjects by duration of schizophrenia as well as treatment differences.

There are a number of limitations of this study. The data would be significantly strengthened with a larger study population and a wider range of flash strengths. When our data using the NMDA receptor hypofunction mouse model of schizophrenia showed significant differences based on sex, we disaggregated the ERG data in the present study; this, however, lowered the overall number of subjects based on sex. In addition, only a subset of the subjects had a PERG using the LED display; comparisons would have been more easily made had we been able to obtain this data from all the subjects originally tested using the CRT method of PERG acquisition. Finally, as with all studies using populations of individuals with schizophrenia, the medications varied between subjects. With a larger cohort of subjects, future studies should be able to analyze groups of subjects with schizophrenia on similar medications. We prepared scatterplots for the a-wave, b-wave, PhNR, and N72 amplitudes of the people with schizophrenia with the few nonmedicated subjects indicated in red (Supplemental Fig. 5), and for almost all data collected, these subjects were well within the measurements obtained from those on psychotropic medications. Some details, such as phase of menstrual cycle, could not be assessed as when the data were collected, it was unknown that sex was a critical variable. However, future studies would be strengthened by collecting menstrual cycle and birth control information. We cannot rule out a small variation in the ERG when morning recordings are compared to recordings made in the late afternoon or early evening (Rufiange et al., 2002; Marcus et al., 2004; Lavoie et al., 2010). The current study used 3 flash intensities that were in the lower and middle limb of the intensity-response data that fits the Naka-Rushton equation for photopic hill calculation, thus were below the levels when the phototopic hill would occur. Other studies used significantly greater ranges in light intensities, when this type of effect would certainly be present (Binns et al., 2011; Joshi et al., 2017). It would be beneficial to study the intensity-response function of both the b-wave amplitude and PhNR in people with schizophrenia by analyzing a wider range of flash-strengths. Future studies will more tightly control the time of ERG recordings to eliminate this potential variable. Similarly, with a larger cohort, it would be possible to control for differential disease severity (see Supplemental Fig. 1). Collectively, the current study identified key variables for defining how the ERG could be reliably used as a biomarker for schizophrenia.

The ERG as a biomarker for schizophrenia holds promise, albeit with differential impairments across sex. It is important to note that the retina has a number of neurons that express NMDAR, and future work will focus on which protocols will be best at differentiating between normal function and dysfunction as it relates to identification of individuals with different underlying causes of their schizophrenia.

Supplementary Material

Supplemental Table 1
Supplemental Table 2
Supplemental Table 3
Supplemental Table 4
Supplemental Figure 1
Supplemental Figure 2
Supplemental Figure 3
Supplemental Figure 4
Supplemental Figure 5

Acknowledgements

We want to thank Drs. Gail Summers, Cheryl Zabrowski, and Eric van Kuijk for examination of the subjects for visual system normalcy. We also want to thank Drew Miller and Terry Tanaka for administering the ERG tests on the subjects of this study.

Funding

This study was supported by NIH R21MH100622 (RFM & AM3), R23 EY025027 (RFM), NIH R01 EY15313 (LKM), NIH F31 MH106296 (NTJ), NIH T32 EY025187 (LKM), NIH P30 EY11375, the Brain Behavioral Research Foundation’s Sidney R. Baer, Jr. Prize (AM3), and the Minnesota Lions Vision Foundation. No funding source had input as to content, analysis or interpretation.

Footnotes

Declaration of competing interest

The authors declare no conflicts of interest.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.schres.2019.09.005.

3

This study is part of the PhD thesis of Nathalia Torres Jimenez.

References

  1. Adams SA, Nasrallah HA, 2018. Multiple retinal anomalies in schizophrenia. Schizophr. Res. 195, 3–12. [DOI] [PubMed] [Google Scholar]
  2. Aizenman E, Frosch MP, Lipton SA, 1988. Responses mediated by excitatory amino acid receptors in solitary retinal ganglion cells from rat. J. Physiol. 396, 75–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Aldebasi YH, Drasdo N, Morgan JD, North RV, 2004. S-cone, L þ M cone, and pattern electroretinograms in ocular hypertension and glaucoma. Vis. Res. 44 (24), 2749–2756. [DOI] [PubMed] [Google Scholar]
  4. Bach M, Mitchell G, Brigell MG, Hawlina M, Holder GE, Johnson MA, McCulloch DL, Meigen T, Viswanathan S, 2013. ISCEV standard for clinical pattern electroretinography (PERG): 2012 update. Doc. Ophthalmol. Adv. Ophthalmol. 126 (1), 1–7. [DOI] [PubMed] [Google Scholar]
  5. Balogh Z, Benedek G, Szabolcs K, 2008. Retinal dysfunctions in schizophrenia. Prog. NeuroPsychopharmacol. Biol. Psychiat. 32 (1), 297–300. [DOI] [PubMed] [Google Scholar]
  6. Balu DT, Coyle JT, 2015. The NMDA receptor ‘glycine modulatory site’ in schizophrenia: D-serine, glycine, and beyond. Curr. Opin. Pharmacol. 20, 109–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Binns AM, Mortlock KE, North RV, 2011. The relationship between stimulus intensity and response amplitude for the photopic negative response of the flash electroretinogram. Doc. Ophthalmol. 122 (1), 39–52. [DOI] [PubMed] [Google Scholar]
  8. Cohen ED, Miller RF, 1994. The role of NMDA and non-NMDA excitatory amino acid receptors in the functional organization of primate retinal ganglion cells. Vis. Neurosci. 11, 317–332. [DOI] [PubMed] [Google Scholar]
  9. Coyle JT, 1996. The glutamatergic dysfunction hypothesis for schizophrenia. Harvard Rev. Psychiat. 3 (5), 241–253. [DOI] [PubMed] [Google Scholar]
  10. Cyr M, Ghribi O, Thibault C, Morissette M, Landry M, Di Paolo T, 2001. Ovarian steroids and selective estrogen receptor modulators activity on rat brain NMDA and AMPA receptors. Brain Res. Rev. 37 (1–3), 153–161. [DOI] [PubMed] [Google Scholar]
  11. Demmin DL, Davis Q, Roche M, Silverstein SM, 2018. Electroretinographic anomalies in schizophrenia. J. Abnorm. Psychol. 127 (4), 417–428. [DOI] [PubMed] [Google Scholar]
  12. Dimopoulos IS, Freund PR, Redel T, Dornstauder B, Gilmour G, Sauvé Y, 2014. Changes in rod and cone-driven oscillatory potentials in the aging human retina. Invest. Ophthalmol. Vis. Sci. 55 (8), 5058–5073. [DOI] [PubMed] [Google Scholar]
  13. Dixon DB, Copenhagen DR, 1992. Two types of glutamate receptors differentially excite amacrine cells in the tiger salamander retina. J. Physiol. 449 (1), 589–606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Driesen NR, McCarthy G, Bhagwagar Z, Bloch M, Calhoun V, D’Souza DC, Gueorguieva R, He G, Ramachandran R, Suckow RF, Anticevic A, Morgan PT, Krystal JH, 2013. Relationship of resting brain hyperconnectivity and schizophrenia-like symptoms produced by the NMDA receptor antagonist ketamine in humans. Mol. Psychiatry 18 (11), 1199–1204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. First MB, Cutbbert B, Malison RD, Widiger T, 2002. Personality disorders and relational disorders: a research agenda for addressing crucial gaps in DSM In: Kupfer DJ, First MB, Regier DA (Eds.), A Research Agenda for DSM-V. American Psychiatric Association, Washington, DC, pp. 123–199. [Google Scholar]
  16. Föcking M, Lopez LM, English JA, Dicker P, Wolff A, Brindley E, Cagney G, Colter DR, 2015. Proteomic and genomic evidence implicates the postsynaptic density in schizophrenia. Mol. Psychiatry 20 (4), 424e–32. [DOI] [PubMed] [Google Scholar]
  17. Frishman L, Sustar M, Kremers J, McAnany JJ, Sarossy M, Tzekov R, Viswanathan S, 2018. ISCEV extended protocol for the photopic negative response (PhNR) of the full-field electroretinogram. Doc. Ophthalmol. 136 (3), 207–211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gauvin M, Lina JM, Lachapelle P, 2014. Advance in ERG analysis: from peak time and amplitude to frequency, power, and energy. Biomed. Res. Int. 2014, 246096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Harsanyi K, Wang Y, Mangel SC, 1996. Activation of NMDA receptors produces dopamine-mediated changes in fish retinal horizontal cell light responses. J. Neurophysiol. 75 (2), 629–647. [DOI] [PubMed] [Google Scholar]
  20. Hébert M, Merette C, Paccalet T, Emond C, Gagne AM, Sasseville A, Maziade M, 2015. Light evoked potentials measured by electroretinogram may tap into the neurodevelopmental roots of schizophrenia. Schizophr. Res. 162 (1e3), 294–295. [DOI] [PubMed] [Google Scholar]
  21. Heinemann-Vernaleken B, Palmowski A, Allgayer R, 2000. The effect of time of day and repeat reliability on the fast flicker multifocal ERG. Doc. Ophthalmol. 101 (3), 247–255. [DOI] [PubMed] [Google Scholar]
  22. Holder GE, 2001. Pattern electroretinography (PERG) and an integrated approach to visual pathway diagnosis. Prog. Retin. Eye Res. 20 (4), 531–561. [DOI] [PubMed] [Google Scholar]
  23. Horn FK, Gottschalk K, Mardin CY, Pangeni G, Jünemann AG, Kremers J, 2011. On and off responses of the photopic fullfield ERG in normal subjects and glaucoma patients. Doc. Ophthalmol. 122 (1), 53–62. [DOI] [PubMed] [Google Scholar]
  24. Javitt DC, Zukin SR, 1991. Recent advances in the phencyclidine model of schizophrenia. Am. J. Psychiatry 148 (10), 1301–1308. [DOI] [PubMed] [Google Scholar]
  25. Joshi NR, Ly E, Viswanathan S, 2017. Intensity response function of the photopic negative response (PhNR): effect of age and testeretest reliability. Doc. Ophthalmol. 135 (1), 1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kirov G, Pocklington AJ, Holmans P, Ivanov D, Ikeda M, Ruderfer D, Moran J, Chambert K, Toncheva D, Georgieva L, Grozeva D, Fjodorova M, Wollerton R, Rees E, Nikolov I, van de Lagemaat LN, Bayes A, Fernandez E, Olason PI, Böttcher Y, Komiyama NH, Collins MO, Choudhary J, Stefansson K, Stefansson H, Grant SG, Purcell S, Sklar P, O’Donovan MC, Owen MJ, 2012. De novo CNV analysis implicates specific abnormalities of postsynaptic signaling complexes in the pathogenesis of schizophrenia. Mol. Psychiatry 17 (2), 142–153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kofuji P, Ceelen P, Zahs KR, Surbeck LW, Lester HA, Newman EA, 2000. Genetic inactivation of an inwardly rectifying potassium channel (Kir4.1 subunit) in mice: phenotypic impact in retina. J. Neurosci. 20 (15), 5733–5740. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kondo M, Piao CH, Tanikawa A, Horiguchi M, Terasaki H, Miyake Y, 2000. Amplitude decrease of photopic ERG b-wave at higher stimulus intensities in humans. Jap. J. Ophthalmol. 44 (1), 20–28. [DOI] [PubMed] [Google Scholar]
  29. Kundra H, Park JC, McAnany JJ, 2016. Comparison of photopic negative response measurements in the time and time-frequency domains. Doc. Ophthalmol. 133 (2), 91–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lavoie J, Gagné A-M, Lavoie M-P, Sasseville A, Charron M-C, Hebert M, 2010. Circadian variation in the electroretinogram and the presence of central melatonin. Doc. Ophthalmol. 120, 265–272. [DOI] [PubMed] [Google Scholar]
  31. Lavoie J, Maziade M, Hébert M, 2014. The brain through the retina: the flash electroretinogram as a tool to investigate psychiatric disorders. Prog. Neuro Psychopharmacol. Biol. Psychiatry 48, 129–134. [DOI] [PubMed] [Google Scholar]
  32. Li B, Barnes GE, Holt WF, 2005. The decline of the photopic negative response (PhNR) in the rat after optic nerve transection. Doc. Ophthalmol. 111 (1), 23–31. [DOI] [PubMed] [Google Scholar]
  33. Loranger AW, 1984. Sex difference in age at onset of schizophrenia. Arch. Gen. Psychiatry 41 (2), 157–161. [DOI] [PubMed] [Google Scholar]
  34. Lukasiewicz PD, McReynolds JS, 1985. Synaptic transmission at N-methyl-Daspartate receptors in the proximal retina of the mudpuppy. J. Physiol. 367, 99–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Luo X, Frishman LJ, 2011. Retinal pathway origins of the pattern electroretinogram (PERG). Invest. Ophthalmol. Vis. Sci. 52 (12), 8571–8584. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Machida S, 2012. Clinical applications of the photopic negative response to optic nerve and retinal diseases. J. Ophthalmol. 2012 (397), 178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Maffei L, Fiorentini A, Bisti S, Hollander H, 1985. Pattern ERG in the monkey after section of the optic nerve. Exp. Brain Res. 59 (2), 423–425. [DOI] [PubMed] [Google Scholar]
  38. Marcus M, Cabael L, Marmor MF, 2004. Are circadian variations in the electroretinogram evident on routine testing? Doc. Ophthalmol. 108, 165–169. [DOI] [PubMed] [Google Scholar]
  39. Massey SC, Miller RF, 1990. N-methyl-D-aspartate receptors of ganglion cells in rabbit retina. J. Neurophysiol. 63 (1), 16–30. [DOI] [PubMed] [Google Scholar]
  40. McCulloch DL, Marmor MF, Brigell MG, Hamilton R, Holder GE, Tzekov R, Bach M, 2015. ISCEV standard for full-field clinical electroretinography (2015 update). Doc. Ophthalmol. 130 (1), 1–12. [DOI] [PubMed] [Google Scholar]
  41. McEwen B, 2002. Estrogen actions throughout the brain. Recent Prog. Horm. Res. 57, 357–384. [DOI] [PubMed] [Google Scholar]
  42. McRoberts JA, Li J, Ennes HS, Mayer EA, 2007. Sex-dependent differences in the activity and modulation of N-methyl-D-aspartic acid receptors in rat dorsal root ganglia neurons. Neuroscience 148 (4), 1015–1020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Miura G, Wang MH, Ivers KM, Frishman LJ, 2009. Retinal pathway origins of the pattern ERG of the mouse. Exp. Eye Res. 89 (1), 49e62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Moghaddam B, Javitt D, 2012. From revolution to evolution: the glutamate hypothesis of schizophrenia and its implication for treatment. Neuropsychopharmacology 37 (1), 4–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Monsalve P, Triolo G, Toft-Nielsen J, Bohorquez J, Henderson AD, Delgado R, Miskiel E, Ozdamar O, Feurer WJ, Porciatti V, 2017. Next generation PERG method: expanding the response dynamic range and capturing response adaptation. Transl. Vis. Sci. Technol. 6 (3), 5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Olney JW, Farber NB, 1995. Glutamate receptor dysfunction and schizophrenia. Arch. Gen. Psychiatry 52 (12), 998–1007. [DOI] [PubMed] [Google Scholar]
  47. Ozdamar O, Toft-Nielsen J, Bohorquez J, Porciatti V, 2014. Relationship between transient and steady-state electroretinograms: theoretical and experimental assessment. Invest. Ophthalmol. Vis. Sci. 55 (12), 8560–8570. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Peterson H, 1968. The normal B-potential in the single-flash clinical retinogram. A computer technique study of the influence of sex and age. Acta Ophthalmol. (Suppl. 99), 7–77. [PubMed] [Google Scholar]
  49. Porciatti V, 2015. Electrophysiological assessment of retinal ganglion cell function. Exp. Eye Res. 141, 164–170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Prager TC, Garcia CA, Mincher CA, Mishra J, Chu HH, 1990. The pattern electroretinogram in diabetes. Am J. Ophthalmol. 109 (3), 279–284. [DOI] [PubMed] [Google Scholar]
  51. Preiser D, Lagreze WA, Bach M, Poloschek CM, 2013. Photopic negative response versus pattern electroretinogram in early glaucoma. Invest. Ophthalmol. Vis. Sci. 54 (2), 1182–1191. [DOI] [PubMed] [Google Scholar]
  52. Raese JD, King RJ, Barnes D, Berger PA, Marmor MF, Hock P, 1982. Retinal oscillatory potentials in schizophrenia: implications for the assessment of dopamine transmission in man. Pharmacol. Bull. 18, 72–78. [Google Scholar]
  53. Rufiange M, Dumon M, Lachapelle P, 2002. Correlating retinal function with melatonin secretion in subjects with an early or late circadian phase. Invest. Ophthalmol. Vis. Sci. 43 (7), 2491–2499. [PubMed] [Google Scholar]
  54. Rufiange M, Dassa J, Dembinska O, Koenekoop RK, Little JM, Polomeno RC, Dumont M, Chemtob S, Lachapelle P, 2003. The photopic ERG luminanceresponse function (photopic hill): method of analysis and clinical application. Vis. Res. 43 (12), 1405–1412. [DOI] [PubMed] [Google Scholar]
  55. Schechter G, Hock P, Rodgers K, Pfefferbaum A, Marmor MF, Maurice R, 1987. Electroretinographic assessment in schizophrenia. Electroencephalogr. Clin. Neurosphysiol. Suppl. 40, 746–751. [PubMed] [Google Scholar]
  56. Shen Y, Liu XL, Yang XL, 2006. N-methyl-D-aspartate receptors in the retina. Mol. Neurobiol. 34 (3), 163–179. [DOI] [PubMed] [Google Scholar]
  57. Silverstein SM, Rosen R, 2015. Schizophrenia and the eye. Schizophr. Res. Cogn. 2 (2), 46–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Snyder MA, Gao WJ, 2013. NMDA hypofunction as a convergence point for progression and symptoms of schizophrenia. Front. Cell. Neurosci. 7, 31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Torres Jimenez N, Lines JW, Kofuji P, Wei H, Rankila A, Coyle JT, McLoon LK, Miller RF, 2019. Electroretinographic abnormalities and sex differences in an NMDAR hypofunction mouse model of schizophrenia: A and B wave analysis. Invest. Ophthalmol. Vis. Sci. (In re-review). [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Ventura L, Porciatti V, 1984. Pattern electroretinogram in glaucoma. Dev. Ophthalmol. 9, 133–139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Ventura J, Lukoff D, Nuechterlein KH, Liberman RP, Green M, Shaner A, 1993. Appendix 1: brief psychiatric rating scale (BPRS) expanded version (4.0) scales, anchor points and administration manual. Int. J. Meth. Psychiatr. Res. 3, 227–244. [Google Scholar]
  62. Viswanathan S, Frishman LJ, Robson JG, Walters JW, 2001. February The photopic negative response of the flash electroretinogram in primary open angle glaucoma. Invest Ophthalmol Vis Sci. 42 (2), 514–522. [PubMed] [Google Scholar]
  63. Viswanathan S, Frishman LJ, Robson JG, 2000. August The uniform field and pattern ERG in macaques with experimental glaucoma: removal of spiking activity. Invest Ophthalmol Vis Sci. 41 (9), 2797–2810. [PubMed] [Google Scholar]
  64. Wachtmeister L, 1998. Oscillatory potentials in the retina: what do they reveal. Prog. Ret. Eye Res. 17 (4), 485–521. [DOI] [PubMed] [Google Scholar]
  65. Wali N, Leguire LE, 1992. The photopic hill: a new phenomenon of the light adapted electroretinogram. Doc. Ophthalmol. 80 (4), 335–342. [DOI] [PubMed] [Google Scholar]
  66. Weymouth AE, Vingrys AJ, 2008. Rodent electroretinography: methods for extraction and interpretation of rod and cone responses. Prog. Ret. Eye Res. 27 (1), 1–44. [DOI] [PubMed] [Google Scholar]
  67. Woolley CS, 1998. Estrogen-mediated structural and functional synaptic plasticity in the female rat hippocampus. Horm. Behav. 34 (2), 140–148. [DOI] [PubMed] [Google Scholar]
  68. Zhang DQ, Wong KY, Sollars PJ, Berson DM, Pickard GE, McMahon DG, 2008. Intraretinal signaling by ganglion cell photoreceptors to dopaminergic amacrine neurons. Proc. Natl. Acad. Sci. 105 (37), 14181–14186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Zhou W, Rangaswamy N, Ktonas P, Frishman LJ, 2007. Oscillatory potentials of the slow-sequence multifocal ERG in primates extracted using the matching pursuit method. Vis. Res. 47 (15), 2021–2036. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Table 1
NIHMS1608737-supplement-Supplemental_Table_1.pdf (204.2KB, pdf)
Supplemental Table 2
NIHMS1608737-supplement-Supplemental_Table_2.pdf (71KB, pdf)
Supplemental Table 3
NIHMS1608737-supplement-Supplemental_Table_3.pdf (76.6KB, pdf)
Supplemental Table 4
NIHMS1608737-supplement-Supplemental_Table_4.pdf (251.4KB, pdf)
Supplemental Figure 1
NIHMS1608737-supplement-Supplemental_Figure_1.pdf (52.8KB, pdf)
Supplemental Figure 2
NIHMS1608737-supplement-Supplemental_Figure_2.pdf (61.5KB, pdf)
Supplemental Figure 3
NIHMS1608737-supplement-Supplemental_Figure_3.pdf (121.1KB, pdf)
Supplemental Figure 4
NIHMS1608737-supplement-Supplemental_Figure_4.pdf (197.5KB, pdf)
Supplemental Figure 5
NIHMS1608737-supplement-Supplemental_Figure_5.pdf (149.6KB, pdf)

RESOURCES