Endophenotypes are quantitative, laboratory-based measures representing intermediate links in the pathways between genetic variation and the clinical expression of a disorder. Ideal endophenotypes exhibit deficits in patients, are stable over time and across shifts in psychopathology, and are suitable for repeat testing. Unfortunately, many leading candidate endophenotypes in schizophrenia have not been fully characterized simultaneously in large cohorts of patients and controls across these properties. The objectives of this study were to characterize the extent to which widely-used neurophysiological and neurocognitive endophenotypes are: 1) associated with schizophrenia, 2) stable over time, independent of state-related changes, and 3) free of potential practice/maturation or differential attrition effects in schizophrenia patients (SZ) and nonpsychiatric comparison subjects (NCS). Stability of clinical and functional measures was also assessed.
Participants (SZ n = 341; NCS n = 205) completed a battery of neurophysiological (MMN, P3a, P50 and N100 indices, PPI, startle habituation, antisaccade), neurocognitive (WRAT-3 Reading, LNS-forward, LNS-reorder, WCST-64, CVLT-II). In addition, patients were rated on clinical symptom severity as well as functional capacity and status measures (GAF, UPSA, SOF). 223 subjects (SZ n = 163; NCS n = 58) returned for retesting after 1 year.
Most neurophysiological and neurocognitive measures exhibited medium-to-large deficits in schizophrenia, moderate-to-substantial stability across the retest interval, and were independent of fluctuations in clinical status. Clinical symptoms and functional measures also exhibited substantial stability. A Longitudinal Endophenotype Ranking System (LERS) was created to rank neurophysiological and neurocognitive biomarkers according to their effect sizes across endophenotype criteria.
The majority of neurophysiological and neurocognitive measures exhibited deficits in patients, stability over a 1-year interval and did not demonstrate practice or time effects supporting their use as endophenotypes in neural substrate and genomic studies. These measures hold promise for informing the “gene-to-phene gap” in schizophrenia research.
Citation: Light GA, Swerdlow NR, Rissling AJ, Radant A, Sugar CA, Sprock J, et al. (2012) Characterization of Neurophysiologic and Neurocognitive Biomarkers for Use in Genomic and Clinical Outcome Studies of Schizophrenia. PLoS ONE 7(7): e39434. doi:10.1371/journal.pone.0039434
Editor: Hossein Fatemi, University of Minnesota, United States of America
Received: February 9, 2012; Accepted: May 23, 2012; Published: July 3, 2012
This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Funding: This work was supported by grants from National Institute of Mental Health (MH042228, MH079777, and MH065571) and supported by the Department of Veteran Affairs (VISN 22 Mental Illness Research, Education, and Clinical Center). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Dr. Light has received compensation from Astellas for one-time participation in a scientific advisory boad unrelated to this study. Dr. Geyer has received compensation from Abbott, Acadia, Addex, Cerca Insights, Merck, Omeros, and Takeda for work unrelated to this study, and holds an equity interest in San Diego Instruments (San Diego, CA). Dr. Swerdlow has received compensation from Neurocrine for work unrelated to this study. This does not alter the authors’ adherence to all PLoS ONE policies on sharing data and materials. The remaining authors report no financial relationships with commercial interests.
One prominent strategy for deconstructing complex, heritable neuropsychiatric disorders such as schizophrenia is to examine discrete, genetically determined “endophenotypes” that are part of the illness and detected in the laboratory rather than by “the naked eye” of the clinical interview . Endophenotypes may be useful for deconstructing the complexity of clinical, neural substrate, and genetic underpinnings of the disorder , . Several criteria for viable endophenotypes have been proposed , –. While there is some variability in the criteria, in general, endophenotypes are a subset of biomarkers that: 1) are associated with the illness, i.e., exhibit deficits in patients; 2) are stable over time; 3) are relatively independent of fluctuations in clinical symptoms; 4) show similar, though often lesser deficits in clinically unaffected family members; and 5) are heritable. This study addresses criteria 1–3 above. Thus, endophenotypes are quantitative, laboratory-based measures that represent intermediate links in the pathways between genetic variation and the clinical expression of the disorder that can uniquely inform the “gene-to-phene” knowledge gap.
Some widely used candidate neurophysiological endophenotypes in schizophrenia include prepulse inhibition of the acoustic startle reflex (PPI), P50 and N100 event-related potential amplitudes and gating, oculomotor antisaccade, mismatch negativity (MMN), and the P3a event-related potentials , , . Commonly used neurocognitive endophenotypes include measures of attention, working memory, verbal recall, perseverative thinking and rule learning in response to verbal feedback , .
Increasingly, these and other neurophysiological and neurocognitive measures are used as biomarkers in clinical trials for “proof of concept” studies designed to determine whether a drug has a detectable “neurobiological signal” or as outcome measures to determine if a drug improves cognition –. In this context, the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) expert panel has established criteria that are considered “essential” for measure selection in clinical trial studies designed to improve cognition in schizophrenia . These criteria are: 1) good test-retest reliability; 2) utility as a repeated measure (i.e., no practice effects); 3) a relationship to functional outcome; 4) a potential response to pharmacologic agents; and 5) practicality/tolerability. Thus, although endophenotypes and biomarkers share some common desirable characteristics , , further validation is required before treatment and clinical trial applications using these measures can be fully implemented.
Effect sizes (Cohen’s d) calculated from group main effects (Table 2) collapsed across time.
Effect sizes (Cohen’s d) of changes in neurocognitive and neurophysiological measures across the retest interval.
The long-term stability of many neurocognitive measures in schizophrenia outpatients appears to be well-established (e.g., , ). Relatively few studies, however, have examined the test-retest reliability of commonly used neurophysiological biomarkers in schizophrenia patients , , . Longitudinal studies of at least 6 months to one year are essential for disentangling state and trait influences and detecting enduring relationships among endophenotypes and clinical outcome measures (e.g., –). Thus, the aims of the present study were to characterize the extent to which a battery of 15 frequently used but not as yet fully validated candidate neurophysiological and neurocognitive measures fulfill many of the established criteria of endophenotypes and biomarkers of drug response including the extent to which measures are deficient in a large cohort of schizophrenia patients, stable over 1 year, and independent of symptom fluctuations in schizophrenia outpatients relative to nonpsychiatric comparison subjects (NCS). Secondary analyses also included assessments of potential practice effects (i.e., subjects’ performance improves due to increased familiarity with the test) and differential attrition (i.e., differences in baseline characteristics in patients who returned vs. failed to return for repeat testing).
We hypothesized that the heritable neurophysiological and neurocognitive measures (e.g., , ) would show deficits in schizophrenia patients, exhibit at least moderate (ICCs>0.60) test-retest stability with little evidence of practice/maturation effects or relationships to fluctuations in clinical symptoms as is commonly assumed in the genetics of schizophrenia literature. We also hypothesized that the schizophrenia patients that failed to return for repeat testing would differ in their baseline (T1) characteristics with “non-returners” being generally more symptomatic, showing worse neurocognitive performance and having poorer functional status vs. those patients who returned for re-testing after 1 year (T2).
This study was approved by the University of California, San Diego Human Research Protections Program Institutional Review Board. All participants were assessed and judged to be capable of providing informed consent and, after subjects were given a detailed description of study procedures, written consent was obtained per UCSD IRB-approved protocols (IRB# 071128 and 071831) prior to each testing session. Participants included 546 subjects: 341 schizophrenia patients and 205 nonpsychiatric comparison subjects (NCS). Schizophrenia patients were recruited from community residential facilities and via physician referral. Normal comparison subjects were recruited through newspaper and internet advertisements and fliers posted at the UCSD Medical Center. All subjects received a urine toxicology screen to rule out recent drug use. In addition, patients and NCS were assessed using the Structured Clinical Interview for DSM-IV . This interview was used to ensure that NCS did not meet criteria for an Axis I or Axis II Cluster A diagnosis Family history of psychiatric disorders was also assessed to ensure that NCS did not have a first-degree relative with a psychotic disorder . In addition, patients did not have a current Axis I diagnosis other than schizophrenia. All subjects were carefully screened to ensure they had never experienced a neurologic insult, such as significant head trauma and/or loss of consciousness as per our established methods , , . Audiometric testing was used to ensure that all participants could detect 40-dB tones at 1000 Hz for their data to be used in PPI, MMN, P3a, P50/N100 measures. There were neither statistically significant differences in hearing thresholds between the schizophrenia patient and NCS groups nor significant correlations between hearing thresholds and dependent measures. Data from subsets of these 546 participants who completed baseline testing were reported previously –, including preliminary genetic association findings of a subset of 219 subjects of European ancestry and 76 subjects of African ancestry recently published in this journal .
Intraclass correlation coefficients are shown for schizophrenia patients (blue; n = 163) and nonpsychiatric comparison subjects (red, n = 58). The mean retest interval was 364.57 (SD: 23.83) days.
Schizophrenia patients (n = 163) and NCS (n = 58) were retested using identical procedures and the same fixed order of tests after approximately 1 year (mean number of days (SD) SZ: 365.34 (26.44); NCS: 363.95 (18.13) days; t = −0.37, p = 0.71) in order to characterize the stability of the measures as well as relationships to changes in clinical symptoms. As shown in Figure 1, of the 178 patients who were not retested, reasons for not returning were as follows: 62% were unable to be re-contacted, 26% were not invited to return by the study staff and investigators (e.g., unable to tolerate testing procedures, removed electrodes and/or showed excessive artifact during testing), 9% declined to return, and 3 subjects were excluded on retest (positive toxicology screen, failed hearing test). Medications were not experimentally controlled in this study. Among the retested patients, 6 were not treated with an antipsychotic (AP), 14 were prescribed 1st generation AP, 113 patients 2nd generation AP, and 30 received a combination of 1st and 2nd Generation AP when they came in for their initial (T1) test session. Over the retest interval, 44 patients underwent a change in the AP, with the majority (n = 22) switching from one to another primary 2nd generation AP.
Clinical and Functional Assessment Measures
In the schizophrenia patients, clinical symptoms were assessed with the Scale for the Assessment of Negative Symptoms (SANS; ) and the Scale for the Assessment of Positive Symptoms (SAPS; ). Functional status was assessed using a modified version of the Global Assessment of Functioning Scale (GAF; ) and the Scale of Functioning (SOF; ). Functional capacity was assessed using the UCSD Performance Based Skills Assessment (UPSA; ). To evaluate the relative stability of the abbreviated version of the UPSA, UPSA-Brief scores were also derived from the full scores in accordance with established methods .
Neurophysiological and Neurocognitive Measures
The following neurophysiological and neurocognitive measures were assessed using our established parameters in a fixed battery: PPI , , startle habituation , P50 and N100 amplitudes and suppression measures , , MMN and P3a amplitudes , , , , , oculomotor antisaccade , simple auditory attention (Letter-Number Span Forward; LNS-Forward; , working memory (Letter-Number Span Re-order; LNS-Reorder; , ), immediate verbal recall (California Verbal Learning Test-2, Standard Form; CVLT-II; ), perseverative thinking and rule learning in response to verbal feedback (Wisconsin Card Sorting Test-64; WCST-64 ). The test order was as follows: PPI and startle habituation, oculomotor antisaccade, LNS-Forward, LNS-Reorder, CVLT-II, WCST-64 followed by a lunch break. After the break, particpants underwent EEG testing for P50 and N100 measures followed by MMN and P3a. For EEG testing, a 40 channel NeuroScan NuAmps system was used with sintered Ag/AgCl electrodes arranged in an electrode cap (EasyCap) with a forehead ground and nose reference. For N100, MMN, and P3a measures, amplitudes were measured relative to a 100 msec prestimulus baseline.
Mismatch negativity and P3a.
Stimulation, recording, and analysis techniques for calculating MMN and P3a amplitudes followed our previously established methods , , , , , . Subjects were presented with binaural stimulation (1 kHz computer-generated square wave stimuli, 85 dB[A] SPL, 1 msec rise/fall) with a fixed stimulus onset-to-onset asynchrony of 500 msec. Standard (P = 0.90; 50 msec duration) and deviant (P = 0.10; 100 msec duration) stimuli were presented in a pseudorandom order while participants watched a silent cartoon video. Signals were digitized at a rate of 1 kHz with system acquisition filter settings at 0.5–100 Hz. Testing was terminated after a minimum of 225 artifact-free responses to deviant stimuli was collected using the same automated procedures as described above. MMN and P3a waveforms were generated by subtracting ERPs in response to standard tones from the ERPs generated in response to the deviant tones. The MMN and P3a amplitudes were calculated from electrode Fz as the mean voltage from 135–205 and 250–300 msec ranges, respectively, consistent with established methods , , , .
Subjects were seated in a reclining chair in an upright position. The session began with a 5-min acclimation period with 70 dB[A] white noise that continued as the background throughout the session. All startle pulse stimuli were 40-msec 115 dB[A] bursts of white noise. Prepulse stimuli consisted of 20-msec noise bursts 15 dB above the 70 dB[A] background, presented 30, 60 or 120 msec prior to the onset of the startling stimulus. Our previous studies have demonstrated PPI in the 60 msec condition to be optimal for detecting deficits in schizophrenia patients , , , . Five pulse-alone trials were presented at the beginning and end of the session to assess habituation. In each of Blocks 2–3, there were 8 pulse-alone and 8 of each of the three prepulse trial-types presented in a pseudorandom order with a 9–23 sec (15 sec average) intertrial interval.
P50 and N100 amplitudes and suppression.
Subjects were tested a reclining chair to minimize myogenic artifacts. EEG data collection procedures, electrode locations, and data processing steps were performed following our established methods , , , , . P50 and N100 processing was performed offline at electrode Cz , , . Auditory click pairs (1 msec duration, 93 dB, 500 msec inter-click interval, 10 sec inter-pair interval) were presented to subjects. Testing was terminated after a minimum of 120 click pairs free of gross muscle or eye blink artifacts and were obtained using an automated threshold filter of +/−100 uVolts. Final data processing and EEG analyses were conducted offline and blind to group membership. Blink and baseline corrections were performed, followed by a second and fully automated artifact rejection screening for residual artifact exceeding +/−70 uVolts. Waveform averages were then generated and filtered using 24 dB/octave rolloff, (FIR) filters at 10–50 Hz for P50 and with a 30 Hz lowpass for N100 peak detections. P50 and N100 amplitudes in response to the first (S1) and second (S2) stimuli were selected automatically by a computer algorithm and subsequently reviewed offline by a research technician blind to subject diagnosis. For P50, the largest peak relative to a preceding negative trough in the 40–80 msec range was selected for S1 and S2 responses. S2 responses were further constrained to be within 10 msec of S1 responses . N100 peaks relative to baseline were selected in the 65–135 msec range . Individual P50 and N100 amplitudes in response to each of the clicks as well as the S1–S2 amplitude differences and percent suppression (1-[S2/S1]*100) were assessed.
Antisaccade (oculomotor inhibition).
Oculomotor recordings were obtained in a quiet, darkened room with subjects seated in front of a flat screen monitor with head stabilized using a chin rest or bite bar. Horizontal eye movements were measured using infra-red oculography (Eye Trak Model 310 eye movement monitor, Applied Science Laboratories, Waltham, MA). Stimuli consisted of square subtending about 0.35 degrees of visual angle. A stimulus was presented at central fixation, per our established methods , . Following a random 2.4 to 3.6 sec interval, the fixation stimulus was turned off. Two hundred ms prior to fixation extinction, peripheral cues were illuminated at 10 or 15 degrees of visual angle to the right or left of center. After fixation extinction, this stimulus remained present for another 800 ms. Finally a 500 ms duration stimulus was presented indicating the location of a correctly performed antisaccade. Subjects were instructed to move their eyes as quickly and accurately as possible to the cue’s mirror image (same amplitude, opposite direction). Prior to commencing oculomotor testing, antisaccade cues were presented and subjects were required to point to, in addition to looking at, the proper location of gaze, insuring that all subjects understood task instructions. Antisaccade cues were yellow, all other stimuli were blue. Antisaccade data is analyzed with computerized pattern recognition software. After artifactual responses are removed the primary dependent measure is proportion of correct responses divided by the total number of (artifact-free) responses.
PASW Statistics version 18 was used for all statistical tests, with significance defined as p<0.01. This α-level reflects that there are 5 classes of endophenotypes under investigation (neurocognitive, antisaccade, startle/PPI, P50/N100, and MMN/P3a) and represents a reasonable balance of possible Type I and Type II errors for the large sample of subjects. Differential attrition was assessed by one-way Analyses of Variance (ANOVA) to determine if there were significant differences on baseline demographic, clinical, cognitive, and functional characteristics between those subjects who returned vs. failed to return for follow-up testing at one year. Subsequent analyses focused on the participants that returned for follow-up assessments. Effect-sizes (Cohen’s d) are presented for all group comparisons. Repeated measures ANOVA were performed with session and group, respectively, as within- and between-subjects factors. This analytic strategy enables the evaluation of whether: 1) schizophrenia patients differ from NCS on the key measures (i.e., a group main effect; effect size calculations shown in Figure 2); 2) the measures exhibit overall stability of means (i.e., a test session main effect; effect size calculations shown in Figure 3); and 3) the patients and NCS differ in the amount of change across the 1 year interval (group by test session interaction).
Stability was assessed via Type C intraclass correlation coefficients (ICC) using a consistency definition from a two-way mixed-effects model to provide an overall index of stability of T1 and second test (T2) measures in the patient and control groups separately. The ICC is a conservative estimate of test–retest reliability, because it is sensitive to group mean changes over time in addition to intra-subject variability. The following descriptors of reliability coefficients were used in accordance with established guidelines : “Low” 0–0.1; “Slight” 0.11–0.40; “Fair” 0.41–0.60; “Moderate” 0.61–0.80; “Substantial” 0.81–1.0.
To determine whether changes in candidate endophenotypes were independent of fluctuations in clinical symptoms, T1–T2 change (Δ) values for endophenotypes and positive and negative symptoms over the retest interval were calculated. Pearson correlations were assessed between Δ values in endophenotype and SAPS and SANS.
Evaluation of Differential Attrition
Since differential attrition can confound the interpretation of results, demographic, clinical, cognitive, and functional characteristics of schizophrenia patients at intake who returned (n = 163) vs. failed to return (n = 178) for follow-up testing were examined (α = 0.01; 80% power to detect d = 0.35 effect) prior to conducting primary analyses. Although this study aimed to retest only 100 SZ and 50 NCS after 1 year, efforts were made to retest as many SZ patients as possible. We anticipated that there may be some important pre-existing differences in the baseline characteristics of retested vs. non-retested patients that could account for the attrition rate and potentially compromise the validity of the retest results. Across 33 variables assessed (see Table 1), no significant differences were observed between these two groups of patients in demographic characteristics, global symptom ratings, functional scales, cognitive, or neurophysiological variables. Although there were no differences on global positive or negative symptom summary ratings in the schizophrenia patients, inspection of the 9 individual SAPS and SANS symptom ratings revealed that retested patients had slightly less severe ratings of delusions (means (SD) retested = 3.05, SD = 1.78; non-retested = 3.63 (1.51); F = 10.68, df = 1,338, p<0.001; d = 0.35). The groups did not differ in their level of independence in their community living situation, the distribution of living environments (i.e., board and care facilities, assisted independent living programs, independent living), SOF scores, GAF ratings, UPSA total scores, neurocognitive, or neurophysiological performances. Exploratory analyses of baseline demographic, cognitive, and neurophysiological variables in the NCS group revealed that the retested NCS were slightly older than the non-retested subjects (means (SD) retested = 45.97, SD = 10.71; non-retested = 40.13 (11.53); F = 11.4, df = 1,204, p<0.001; d = 0.50), but no other significant differences were observed across 24 statistical comparisons in this group. Thus, attrition does not appear to be substantially associated with key dependent variables for either the schizophrenia patients or the NCS.
Group Differences and Evaluation of Potential Practice/Interval Effects
Significant deficits were present in the schizophrenia patients across all of the neurocognitive tests, antisaccade, PPI, N100 amplitudes, MMN, and P3a (Figure 2). Consistent with previous findings, no significant group differences were present on startle magnitude or habituation, and PPI in the 30 and 120 msec conditions . In contrast to our expectations, significant deficits were not detected for any P50 measures, or N100 S1–S2 amplitude difference or suppression variables. Marginally significant test session effects were present on the CVLT-II test of immediate verbal recall with both groups showing improvements in their T2 assessments (Table 2, Figure 3). This improvement roughly equated to recalling an additional 1–4 words (out of 80 possible) over the retest interval. While no significant group by test session interactions were present across measures, only the CVLT-II approached significance (F1,220 = 5.21, p = 0.023).
Evaluation of One-Year Test-Retest Stability
The majority of neurocognitive and neurophysiological measures demonstrated moderate to substantial one-year stability in both NCS and schizophrenia patients (Figure 4) with average ICCs across all measures exceeding 0.75. Indeed, 37 of 42 reliability assessments in the present study exceeded 0.60. Consistent with previous studies, P50 and N100 ratio measures did not show high stability. Similarly, percent startle habituation was not stable in the schizophrenia patients. The clinical and functional characteristics (Figure 5) of the schizophrenia patients also demonstrated substantial stability over the one year retest interval (mean ICCs = 0.83).
Assessment of State Independence of Endophenotypes
As shown in Table 2, the schizophrenia patients exhibited significant clinical symptoms and functional impairments. Positive and negative clinical symptoms and functional assessment ratings were consistent across the 1 year retest interval. T1–T2 Δ scores in clinical symptoms (SAPS Δ Mean: 0.43, SD: 3.58; SANS Δ Mean = −0.45, SD = 3.53) and individual endophenotypes were calculated in order to assess the extent to which variation in clinical symptoms were associated with changes in endophenotypes across the retest interval. Changes in P3a were modestly associated with changes in positive (r = 0.21, p = 0.036) and negative (r = 0.26, p = 0.007) symptoms. Changes in negative symptoms were associated with changes in P50 S1 amplitude (r = 0.19, p = 0.045) and P50 S1–S2 difference (r = 0.21, p = 0.03). None of the remaining neurocognitive or neurophysiological measures were significantly associated with changes in positive or negative symptoms (all r<0.12, all p>0.10).
The majority of neurophysiological and neurocognitive measures examined in this study fulfill the criteria as valid endophenotypes for genomic studies and as robust biomarkers in clinical studies. Specifically, these candidate endophenotypes are: 1) associated with illness as they exhibit significant deficits in schizophrenia patients); 2) stable over 1 year in both patients and controls; 3) relatively insensitive to modest fluctuations in clinical symptoms; and 4) suitable for use as repeated measures since they do not show practice effects. Figure 6 summarizes the cumulative effect size ratings across of these 4 criteria, via a Longitudinal Endophenotype Ranking System (LERS; cf. ). Related studies have confirmed that many of these measures show significant deficits in clinically unaffected relatives of schizophrenia patients, are heritable and informative for the identification of genetic variation in schizophrenia , , , . In addition, each measure has independently been the focus of intense scientific inquiry that has, in some cases, produced a rich literature describing informative animal models with predictive and construct validity  and detailed underlying neural and molecular substrates , .
Deficits in Patients
Mismatch negativity, P3a, PPI, oculomotor antisaccade, and neurocognitive measures demonstrated significant deficits in schizophrenia patients and stability over the retest interval. A few of the neurophysiological measures are known to be sensitive to antipsychotic medication effects, which may have contributed to the results. Specifically, deficits in PPI and P50 suppression measures are known to be opposed by second generation antipsychotic medications , , –. Since this was a naturalistic study designed to validate the use of these measures in even larger-scale genomic studies where medications are virtually never experimentally controlled, subjects were not stratified on the basis of their medication type, gender, or smoking status–all factors known to influence PPI, P50, and other measures , –. It is therefore possible that these or other confounding factors may have contributed to the failure to detect significant and reliable P50 and N100 amplitudes and gating deficits in the schizophrenia patients. The modest reliability of ratio-based measures of P50 and N100 gating is consistent with previous studies of normal subjects over relatively brief retest intervals –. The results of this study support examining constituent S1 and/or S2 component amplitudes or non-ratio S1–S2 ERP amplitude difference measures. The Consortium on the Genetics of Schizophrenia (COGS) study which utilized a separate family-based sample of subjects also found that P50 and N100 individual response amplitudes and S1–S2 difference scores performed substantially better than gating ratios in heritability ,  and genomic analyses .
Stability Over 1-year
The stability of many of the neurophysiologic and neurocognitive measures in schizophrenia patients was somewhat higher-than-expected given the naturalistic study design that allowed for changes in medications and clinical status over the 1 year follow-up period. For example, 27% of patients underwent an addition, subtraction, or substitution of their primary antipsychotic medication (without considering the many dosage or adjunctive medication changes) over the test-retest interval. This robust stability raises the question of whether participants in this study were higher-functioning, asymptomatic, and therefore not representative of a “real world” community sample. An answer to this key question resides in the patient data: the retested vs. non-retested patients did not differ substantially in baseline characteristics. Many of the re-tested patients were neither high functioning nor asymptomatic (see Table 1). For example, 41% of the retested patients received maximal clinical severity ratings on hallucination items from the SAPS, 70% received maximum anhedonia and avolition ratings and, in terms of real-world function, 52% of patients required assistance with basic financial management.
With respect to the stability of clinical assessment measures, functional status and capacity are becoming more widespread in schizophrenia research. Following the MATRICS initiative –, these types of assessments have been proposed as co-primary endpoints in clinical trial studies of potential cognitive-enhancing interventions. This study significantly advances the growing literature on the psychometric properties of functional outcome measures, including their long-term stability. For example, the stability of the Scale of Functioning (SOF) has not been reported previously. The substantial stability of the SOF (ICC = 0.88) supports its use in studies designed to characterize the global psychosocial and community functioning (e.g., , . This scale provides important information about community functioning milestones such as occupational status, social relationships, and level of independence in living situations that can be targeted by novel pharmacologic and nonpharmacologic interventions –. Likewise, the finding of high 1-year stability of the UPSA and the abbreviated UPSA-Brief is consistent with and replicates recent reports over both short  and longer  retest intervals.
It is possible that the assessment of patients across more dynamic phases of illness, such as during the conversion from prodromal states to first episode schizophrenia – or during clinical exacerbations in already diagnosed schizophrenia patients might yield stability coefficients that were lower than what was observed in this more chronic outpatient cohort. Indeed, one could argue that the high stability observed in chronic patients is attributable to the fact that many of the patients have reached the nadir of their illness and are no longer experiencing dramatic fluctuations in the underlying neural networks that contribute to clinical symptoms and their related neurophysiological and neurocognitive function. It therefore remains an open question as to which if any of these measures will also serve as robust vulnerability markers in high risk populations , .
Independent of Fluctuations in Clinical Symptoms
Neurophysiological and neurocognitive measures were relatively insensitive, accounting for <5% of the variance, to modest fluctuations in in clinical symptoms. The one exception was with the P3a event-related potential where changes in both positive and negative symptoms were correlated with changes in amplitude over the test-retest interval, consistent with previous findings , . The absence of associations with symptom changes does not, however, invalidate the use of these measures as biomarkers of clinical, cognitive, or functional response to therapeutic interventions. Medications were not systematically controlled in this study, and thus changes in clinical symptoms were not necessarily ones that would optimally be “biomarker-sensitive” or indicative of a therapeutic drug effect. Many of the characteristics of the present measures such as deficits in patients, stability, utility as a repeated measure, are also criteria for biomarkers of response to experimentally controlled pharmacologic and non-pharmacologic interventions . In fact, the Cognitive Neuroscience Treatment Research to Improve Cognition in Schizophrenia (CNTRICS) expert consensus panel has determined that only 2 neurophysiological measures–MMN and PPI– are “already mature” and among the most promising biomarkers for use in multi-site clinical trials .
In conclusion, the cumulative pattern of results suggests that these widely used neurophysiological and neurocognitive biomarkers are robust, reliable, state-independent, and therefore valid endophenotypes for ongoing genomic research studies where endophenotypes are used to fill the “gene-to-phene” knowledge gap. Genomic analyses from a subset of these participants have been recently reported in this journal . Family-based genetic associations using these measures have also been reported for a separate sample from the multi-site COGS study . Future studies are needed to assess the utility of these measures for predicting conversion to psychosis, as biomarkers of response to pharmacologic and nonpharmacologic ,  treatments, for tracking disease progression across the course of illness , and for the delineation of schizophrenia-related abnormalities across genomic and neural networks , .
Conceived and designed the experiments: GAL NRS MAG DLB. Performed the experiments: GAL JS AJR MP. Analyzed the data: GAL AJR JS AR CAS. Wrote the paper: GAL NRS AJR AR CAS JS MP MAG DLB.
- 1. Gottesman, II, Gould TD (2003) The endophenotype concept in psychiatry: etymology and strategic intentions. Am J Psychiatry 160: 636–645.
- 2. Greenwood TA, Lazzeroni LC, Murray SS, Cadenhead KS, Calkins ME, et al. (2011) Analysis of 94 Candidate Genes and 12 Endophenotypes for Schizophrenia From the Consortium on the Genetics of Schizophrenia. Am J Psychiatry.
- 3. Greenwood TA, Light GA, Swerdlow NR, Radant AD, Braff DL (2012) Association analysis of 94 candidate genes and schizophrenia-related endophenotypes. PLoS One 7: e29630.
- 4. Braff DL, Light GA (2005) The use of neurophysiological endophenotypes to understand the genetic basis of schizophrenia. Dialogues Clin Neurosci 7: 125–135.
- 5. Bearden CE, Freimer NB (2006) Endophenotypes for psychiatric disorders: ready for primetime? Trends Genet 22: 306–313.
- 6. Turetsky BI, Calkins ME, Light GA, Olincy A, Radant AD, et al. (2007) Neurophysiological endophenotypes of schizophrenia: the viability of selected candidate measures. Schizophr Bull 33: 69–94.
- 7. Greenwood TA, Braff DL, Light GA, Cadenhead KS, Calkins ME, et al. (2007) Initial heritability analyses of endophenotypic measures for schizophrenia: the consortium on the genetics of schizophrenia. Arch Gen Psychiatry 64: 1242–1250.
- 8. Gur RE, Calkins ME, Gur RC, Horan WP, Nuechterlein KH, et al. (2007) The Consortium on the Genetics of Schizophrenia: neurocognitive endophenotypes. Schizophr Bull 33: 49–68.
- 9. Glahn DC, Almasy L, Blangero J, Burk GM, Estrada J, et al. (2007) Adjudicating neurocognitive endophenotypes for schizophrenia. Am J Med Genet B Neuropsychiatr Genet 144: 242–249.
- 10. Keefe RS, Bilder RM, Davis SM, Harvey PD, Palmer BW, et al. (2007) Neurocognitive effects of antipsychotic medications in patients with chronic schizophrenia in the CATIE Trial. Arch Gen Psychiatry 64: 633–647.
- 11. Olincy A, Harris JG, Johnson LL, Pender V, Kongs S, et al. (2006) Proof-of-concept trial of an alpha7 nicotinic agonist in schizophrenia. Arch Gen Psychiatry 63: 630–638.
- 12. Light GA, Malaspina D, Geyer MA, Luber BM, Coleman EA, et al. (1999) Amphetamine disrupts P50 suppression in normal subjects. Biol Psychiatry 46: 990–996.
- 13. Umbricht D, Koller R, Vollenweider FX, Schmid L (2002) Mismatch negativity predicts psychotic experiences induced by NMDA receptor antagonist in healthy volunteers. Biol Psychiatry 51: 400–406.
- 14. Lavoie S, Murray MM, Deppen P, Knyazeva MG, Berk M, et al. (2008) Glutathione precursor, N-acetyl-cysteine, improves mismatch negativity in schizophrenia patients. Neuropsychopharmacology 33: 2187–2199.
- 15. Swerdlow NR, van Bergeijk DP, Bergsma F, Weber E, Talledo J (2009) The effects of memantine on prepulse inhibition. Neuropsychopharmacology 34: 1854–1864.
- 16. Green MF, Nuechterlein KH, Gold JM, Barch DM, Cohen J, et al. (2004) Approaching a consensus cognitive battery for clinical trials in schizophrenia: the NIMH-MATRICS conference to select cognitive domains and test criteria. Biol Psychiatry 56: 301–307.
- 17. Cho RY, Ford JM, Krystal JH, Laruelle M, Cuthbert B, et al. (2005) Functional neuroimaging and electrophysiology biomarkers for clinical trials for cognition in schizophrenia. Schizophr Bull 31: 865–869.
- 18. Braff DL, Light GA (2004) Preattentional and attentional cognitive deficits as targets for treating schizophrenia. Psychopharmacology (Berl) 174: 75–85.
- 19. Heaton RK, Gladsjo JA, Palmer BW, Kuck J, Marcotte TD, et al. (2001) Stability and course of neuropsychological deficits in schizophrenia. Arch Gen Psychiatry 58: 24–32.
- 20. Kurtz MM (2005) Neurocognitive impairment across the lifespan in schizophrenia: an update. Schizophr Res 74: 15–26.
- 21. Calkins ME, Iacono WG, Curtis CE (2003) Smooth pursuit and antisaccade performance evidence trait stability in schizophrenia patients and their relatives. Int J Psychophysiol 49: 139–146.
- 22. Hall MH, Schulze K, Rijsdijk F, Picchioni M, Ettinger U, et al. (2006) Heritability and reliability of P300, P50 and duration mismatch negativity. Behav Genet 36: 845–857.
- 23. Mathalon DH, Ford JM, Pfefferbaum A (2000) Trait and state aspects of P300 amplitude reduction in schizophrenia: a retrospective longitudinal study. Biol Psychiatry 47: 434–449.
- 24. Light GA, Braff DL (2005) Stability of mismatch negativity deficits and their relationship to functional impairments in chronic schizophrenia. Am J Psychiatry 162: 1741–1743.
- 25. Clementz BA, Sponheim SR, Iacono WG, Beiser M (1994) Resting EEG in first-episode schizophrenia patients, bipolar psychosis patients, and their first-degree relatives. Psychophysiology 31: 486–494.
- 26. First MB, Spitzer RL, Gibbon M, Williams JB (1995) Structured Clinical Interview for DSM-IV Axis I Disorders – Patient Edition (SCID-I/P, Version 2.0); Department NYBR, editor. New York: New York State Psychiatric Institute.
- 27. Maxwell E (1992) Manual for the FIGS. Bethesda, MD: National Institutes of Mental Health.
- 28. Light GA, Braff DL (2005) Mismatch negativity deficits are associated with poor functioning in schizophrenia patients. Arch Gen Psychiatry 62: 127–136.
- 29. Light GA, Hsu JL, Hsieh MH, Meyer-Gomes K, Sprock J, et al. (2006) Gamma band oscillations reveal neural network cortical coherence dysfunction in schizophrenia patients. Biol Psychiatry 60: 1231–1240.
- 30. Braff DL, Light GA, Swerdlow NR (2007) Prepulse inhibition and P50 suppression are both deficient but not correlated in schizophrenia patients. Biol Psychiatry 61: 1204–1207.
- 31. Kiang M, Braff DL, Sprock J, Light GA (2009) The relationship between preattentive sensory processing deficits and age in schizophrenia patients. Clin Neurophysiol 120: 1949–1957.
- 32. Light GA, Swerdlow NR, Braff DL (2007) Preattentive sensory processing as indexed by the MMN and P3a brain responses is associated with cognitive and psychosocial functioning in healthy adults. J Cogn Neurosci 19: 1624–1632.
- 33. Andreasen NC (1984) Scale for the Assessment of Negative Symptoms (SANS). Iowa City, Iowa: University of Iowa.
- 34. Andreasen NC (1984) Scale for the Assessment of Positive Symptoms (SAPS). Iowa City, Iowa: University of Iowa.
- 35. Hall RC (1995) Global assessment of functioning. A modified scale. Psychosomatics 36: 267–275.
- 36. Rapaport MH, Bazetta J, McAdams LA, Patterson J, Jeste DV (1996) Validation of the Scale of Functioning in older outpatients with schizophrenia. American Journal of Geriatric Psychiatry 4: 218–228.
- 37. Patterson TL, Goldman S, McKibbin CL, Hughs T, Jeste DV (2001) UCSD Performance-Based Skills Assessment: development of a new measure of everyday functioning for severely mentally ill adults. Schizophr Bull 27: 235–245.
- 38. Mausbach BT, Harvey PD, Pulver AE, Depp CA, Wolyniec PS, et al. (2010) Relationship of the Brief UCSD Performance-based Skills Assessment (UPSA-B) to multiple indicators of functioning in people with schizophrenia and bipolar disorder. Bipolar Disord 12: 45–55.
- 39. Meincke U, Light GA, Geyer MA, Braff DL, Gouzoulis-Mayfrank E (2004) Sensitization and habituation of the acoustic startle reflex in patients with schizophrenia. Psychiatry Res 126: 51–61.
- 40. Turetsky BI, Greenwood TA, Olincy A, Radant AD, Braff DL, et al. (2008) Abnormal auditory N100 amplitude: a heritable endophenotype in first-degree relatives of schizophrenia probands. Biol Psychiatry 64: 1051–1059.
- 41. Kiang M, Light GA, Prugh J, Coulson S, Braff DL, et al. (2007) Cognitive, neurophysiological, and functional correlates of proverb interpretation abnormalities in schizophrenia. J Int Neuropsychol Soc 13: 653–663.
- 42. Jahshan C, Cadenhead KS, Rissling AJ, Kirihara K, Braff DL, et al. (2011) Automatic sensory information processing abnormalities across the illness course of schizophrenia. Psychol Med. pp. 1–13.
- 43. Radant AD, Dobie DJ, Calkins ME, Olincy A, Braff DL, et al. (2007) Successful multi-site measurement of antisaccade performance deficits in schizophrenia. Schizophr Res 89: 320–329.
- 44. Horan WP, Braff DL, Nuechterlein KH, Sugar CA, Cadenhead KS, et al. (2008) Verbal working memory impairments in individuals with schizophrenia and their first-degree relatives: findings from the Consortium on the Genetics of Schizophrenia. Schizophr Res 103: 218–228.
- 45. Perry W, Heaton RK, Potterat E, Roebuck T, Minassian A, et al. (2001) Working memory in schizophrenia: transient “online” storage versus executive functioning. Schizophr Bull 27: 157–176.
- 46. Stone WS, Tsuang MT, Faraone SV, Glatt SJ, Braff DL, et al. (in submission). Impaired learning and memory on the California Verbal Learning Test (CVLT-II) in subjects with schizophrenia and their first-degree relatives: Findings from the Consortium on the Genetics of Schizophrenia. (COGS).
- 47. Braff DL, Geyer MA, Light GA, Sprock J, Perry W, et al. (2001) Impact of prepulse characteristics on the detection of sensorimotor gating deficits in schizophrenia. Schizophr Res 49: 171–178.
- 48. Swerdlow NR, Sprock J, Light GA, Cadenhead K, Calkins ME, et al. (2007) Multi-site studies of acoustic startle and prepulse inhibition in humans: initial experience and methodological considerations based on studies by the Consortium on the Genetics of Schizophrenia. Schizophr Res 92: 237–251.
- 49. Light GA, Braff DL (2001) Measuring P50 suppression and prepulse inhibition in a single recording session. Am J Psychiatry 158: 2066–2068.
- 50. Light GA, Geyer MA, Clementz BA, Cadenhead KS, Braff DL (2000) Normal P50 suppression in schizophrenia patients treated with atypical antipsychotic medications. Am J Psychiatry 157: 767–771.
- 51. Olincy A, Braff DL, Adler LE, Cadenhead KS, Calkins ME, et al. (2010) Inhibition of the P50 cerebral evoked response to repeated auditory stimuli: Results from the Consortium on Genetics of Schizophrenia. Schizophr Res.
- 52. Shrout PE (1998) Measurement reliability and agreement in psychiatry. Stat Methods Med Res 7: 301–317.
- 53. Glahn DC, Curran JE, Winkler AM, Carless MA, Kent JW, et al. (2012) High dimensional endophenotype ranking in the search for major depression risk genes. Biol Psychiatry 71: 6–14.
- 54. Hall MH, Schulze K, Rijsdijk F, Kalidindi S, McDonald C, et al. (2009) Are auditory P300 and duration MMN heritable and putative endophenotypes of psychotic bipolar disorder? A Maudsley Bipolar Twin and Family Study. Psychol Med 39: 1277–1287.
- 55. Swerdlow NR, Braff DL, Taaid N, Geyer MA (1994) Assessing the validity of an animal model of deficient sensorimotor gating in schizophrenic patients. Arch Gen Psychiatry 51: 139–154.
- 56. Woods SP, Delis DC, Scott JC, Kramer JH, Holdnack JA (2006) The California Verbal Learning Test–second edition: test-retest reliability, practice effects, and reliable change indices for the standard and alternate forms. Arch Clin Neuropsychol 21: 413–420.
- 57. Wynn JK, Green MF, Sprock J, Light GA, Widmark C, et al. (2007) Effects of olanzapine, risperidone and haloperidol on prepulse inhibition in schizophrenia patients: a double-blind, randomized controlled trial. Schizophr Res 95: 134–142.
- 58. Adler LE, Olincy A, Cawthra EM, McRae KA, Harris JG, et al. (2004) Varied effects of atypical neuroleptics on P50 auditory gating in schizophrenia patients. Am J Psychiatry 161: 1822–1828.
- 59. Nagamoto HT, Adler LE, Hea RA, Griffith JM, McRae KA, et al. (1996) Gating of auditory P50 in schizophrenics: unique effects of clozapine. Biol Psychiatry 40: 181–188.
- 60. Light GA, Braff DL (1999) Human and animal studies of schizophrenia-related gating deficits. Curr Psychiatry Rep 1: 31–40.
- 61. Adler LE, Hoffer LD, Wiser A, Freedman R (1993) Normalization of auditory physiology by cigarette smoking in schizophrenic patients. Am J Psychiatry 150: 1856–1861.
- 62. Swerdlow NR, Auerbach P, Monroe SM, Hartston H, Geyer MA, et al. (1993) Men are more inhibited than women by weak prepulses. Biol Psychiatry 34: 253–260.
- 63. Hetrick WP, Sandman CA, Bunney WE, Jin Y, Potkin SG, et al. (1996) Gender differences in gating of the auditory evoked potential in normal subjects. Biol Psychiatry 39: 51–58.
- 64. Rentzsch J, Jockers-Scherubl MC, Boutros NN, Gallinat J (2008) Test-retest reliability of P50, N100 and P200 auditory sensory gating in healthy subjects. Int J Psychophysiol 67: 81–90.
- 65. Fuerst DR, Gallinat J, Boutros NN (2007) Range of sensory gating values and test-retest reliability in normal subjects. Psychophysiology 44: 620–626.
- 66. Smith DA, Boutros NN, Schwarzkopf SB (1994) Reliability of P50 auditory event-related potential indices of sensory gating. Psychophysiology 31: 495–502.
- 67. Boutros NN, Overall J, Zouridakis G (1991) Test-retest reliability of the P50 mid-latency auditory evoked response. Psychiatry Res 39: 181–192.
- 68. Clementz BA, Geyer MA, Braff DL (1997) P50 suppression among schizophrenia and normal comparison subjects: a methodological analysis. Biol Psychiatry 41: 1035–1044.
- 69. Green MF, Kern RS, Heaton RK (2004) Longitudinal studies of cognition and functional outcome in schizophrenia: implications for MATRICS. Schizophr Res 72: 41–51.
- 70. Marder SR, Fenton W (2004) Measurement and Treatment Research to Improve Cognition in Schizophrenia: NIMH MATRICS initiative to support the development of agents for improving cognition in schizophrenia. Schizophr Res 72: 5–9.
- 71. Green MF, Nuechterlein KH (2004) The MATRICS initiative: developing a consensus cognitive battery for clinical trials. Schizophr Res 72: 1–3.
- 72. Twamley EW, Savla GN, Zurhellen CH, Heaton RK, Jeste DV (2008) Development and Pilot Testing of a Novel Compensatory Cognitive Training Intervention for People with Psychosis. Am J Psychiatr Rehabil 11: 144–163.
- 73. McGurk SR, Twamley EW, Sitzer DI, McHugo GJ, Mueser KT (2007) A meta-analysis of cognitive remediation in schizophrenia. Am J Psychiatry 164: 1791–1802.
- 74. Kawakubo Y, Kamio S, Nose T, Iwanami A, Nakagome K, et al. (2007) Phonetic mismatch negativity predicts social skills acquisition in schizophrenia. Psychiatry Res 152: 261–265.
- 75. Green MF, Nuechterlein KH, Kern RS, Baade LE, Fenton WS, et al. (2008) Functional co-primary measures for clinical trials in schizophrenia: results from the MATRICS Psychometric and Standardization Study. Am J Psychiatry 165: 221–228.
- 76. Leifker FR, Patterson TL, Bowie CR, Mausbach BT, Harvey PD (2010) Psychometric properties of performance-based measurements of functional capacity: test-retest reliability, practice effects, and potential sensitivity to change. Schizophr Res 119: 246–252.
- 77. Thompson A, Nelson B, Yung A (2011) Predictive validity of clinical variables in the “at risk” for psychosis population: international comparison with results from the North American Prodrome Longitudinal Study. Schizophr Res 126: 51–57.
- 78. Seidman LJ, Giuliano AJ, Meyer EC, Addington J, Cadenhead KS, et al. (2010) Neuropsychology of the prodrome to psychosis in the NAPLS consortium: relationship to family history and conversion to psychosis. Arch Gen Psychiatry 67: 578–588.
- 79. Jahshan C, Heaton RK, Golshan S, Cadenhead KS (2010) Course of neurocognitive deficits in the prodrome and first episode of schizophrenia. Neuropsychology 24: 109–120.
- 80. Jahshan C, Cadenhead KS, Rissling AJ, Kirihara K, Braff DL, et al. (2012) Automatic sensory information processing abnormalities across the illness course of schizophrenia. Psychol Med 42: 85–97.
- 81. Addington J, Cadenhead KS, Cannon TD, Cornblatt B, McGlashan TH, et al. (2007) North American Prodrome Longitudinal Study: a collaborative multisite approach to prodromal schizophrenia research. Schizophr Bull 33: 665–672.
- 82. Turetsky B, Colbath EA, Gur RE (1998) P300 subcomponent abnormalities in schizophrenia: II. Longitudinal stability and relationship to symptom change. Biol Psychiatry 43: 31–39.
- 83. Green MF, Butler PD, Chen Y, Geyer MA, Silverstein S, et al. (2009) Perception measurement in clinical trials of schizophrenia: promising paradigms from CNTRICS. Schizophr Bull 35: 163–181.
- 84. Swerdlow NR (2011) Are we studying and treating schizophrenia correctly? Schizophr Res 130: 1–10.