RESEARCH REPORT

Retinal Response to Light in Young Nonaffected

Offspring at High Genetic Risk of Neuropsychiatric
Brain Disorders
Marc Hébert, Anne-Marie Gagné, Marie-Eve Paradis, Valérie Jomphe, Marc-André Roy, Chantal Mérette,
and Michel Maziade
Background: In neuropsychiatric brain disorders, such as schizophrenia (SZ) and bipolar disorder (BD), the biased effect of chronic drug
therapy and the toxic effect of illness once installed constitute obstacles to the identification of valid biomarkers. Such biomarkers could lie
at the level of retinal function where anomalies have already been reported in adults suffering from neuropsychiatric disorders. Here, we
report a specific electroretinographic (ERG) anomaly in young nonaffected and nonmedicated offspring at high genetic risk (HR) of these
disorders.

Methods: Electroretinography was performed in 29 HR offspring having one parent affected by DSM-IV SZ or BD (mean age: 20.8 years, SD
4.4) and 29 healthy control subjects (mean age: 20.6 years, SD 4.2). The HRs’ parents descended from multigenerational families affected by
SZ or BD.

Results: Rod ERG (b-wave amplitude at Vmax) in HRs was significantly lower than control subjects (p ⬍ .0001; effect size of ⫺1.47), whereas
the cone ERG Vmax showed no difference (p ⫽ .27). No effects of gender, age, and seasons of testing were observed. The anomaly in retinal
response (rod Vmax b-wave amplitude) was observed independently of parents’ diagnosis (SZ; p ⫽ .007, effect size of ⫺1.09; BD: p ⬍ .0001,
effect size of ⫺1.88) and was present in both the younger and older HRs (effect size of ⫺1.6 and ⫺1.8, respectively).

Conclusions: A rod retinal response anomaly before the age of the disease incidence may represent an early and specific biomarker of risk
with meaning for further genetic and prevention research.

Key Words: Biomarker, brain disorders, early detection, electroreti- Despite emerging success in identifying convergent suscepti-
nography, genetics, prevention, risk bility genes and loci for SZ and bipolar disorder (BD) (13–15),
the brain dysfunction underlying neuropsychiatric disorders is
still unknown. Breaking down the diagnostic into simpler phe-

T
he retina, which is part of the central nervous system, may
constitute a powerful site of investigation for understand-
ing brain disorders. Indeed, retinal anomalies, as mea-
sured with the electroretinogram (ERG), have been reported in
disorders such as seasonal affective disorder (SAD) (1–3), schizo-
phrenia (SZ) (4,5), autism (6), and Parkinson’s disease (7). The
ERG is a light-evoked biopotential recorded on the eyes measur-
ing the full dynamic response originating from cones and rods.
With dopamine (DA) being a major neurotransmitter in the
retina, it is believed that the ERG could reflect its direct and
indirect activity in the retina and the brain, respectively (8,9). In
fact, human studies have shown that oral administration of DA
antagonists blocking both D1 and D2 receptors lead to a rod
b-wave decrease (10). In parallel, serotonin (5-HT) activity, as
measured with platelet imipramine-binding sites, was shown to
be negatively correlated (.84) with the rod ERG (11). This is
consistent with research on pigeons showing a decreased ERG
b-wave amplitude after intraocular injection of glutamate analogs
with toxic effects on 5-HT neurons (12). Therefore, the ERG
comes up as a convincing means for the investigation of neuro-
psychiatric disorders in which a DA and/or 5-HT dysfunction is
likely. Unfortunately, in most psychiatric ERG studies, the con-
founding effect of medications exists, leaving questions about
the nature and specificity of the results.
From the Centre de recherche Université Laval Robert-Giffard, Québec,
Canada.
Address correspondence to Marc Hébert, Ph.D., Centre de recherche Univer-
sité Laval Robert-Giffard, 2601 chemin de la Canardière (F-4500), Que-
bec, Quebec G1J 2G3, Canada; E-mail: marc.hebert@crulrg.ulaval.ca.
Received May 21, 2009; revised Aug 2, 2009; accepted Aug 3, 2009.
notypic components is a crucial challenge for pinpointing the
genetic architecture of these disorders (16 –19). But, again, there
are barriers to the dissection of diagnoses into meaningful
intermediate phenotypes (IP) (18,19), such as the confounding
effect of long-lasting psychiatric medication and the “toxic” effect
of psychosis once installed.
Therefore, we have targeted, in well characterized multigen-
erational families of Eastern Quebec (13,14,17), young high-risk
offspring (HR) of parents affected by SZ and/or BD depression.
These young HRs (although at high genetic risk) were not yet
affected by the illness and thus free from psychotropic drugs that
could impact the ERG response.
Under a neurodevelopmental model of major psychosis (20),
our hypothesis was that the young HR offspring would display a
biological endophenotype under the form of ERG anomaly.
Methods and Materials
Sample
Nonaffected HR Offspring. Offspring were targeted from the
most recent generations of kindreds densely affected by SZ and
BD (Supplement 1). The inclusion criterion was having one
parent affected by a DSM-IV SZ or BD who descended from these
kindreds (21). The exclusion criteria were the presence of a
diagnosis of DSM-IV psychotic disorder, BD, or major depres-
sion; brain and metabolic disorders; being pregnant; working on
night shifts; and having traveled two time zones within 1 month
before the experiment. Twenty-nine HR subjects were enrolled
(mean age ⫽ 20.8 years, SD 4.4) among 15 kindreds (6 families
with 1 HR, 6 families with 2 HR, 1 family with 3 HR, and 2 families
with 4 HR, totalling 29 HRs). These kindreds were divided into 21

0006-3223/10/$36.00 BIOL PSYCHIATRY 2010;67:270 –274

doi:10.1016/j.biopsych.2009.08.016 © 2010 Society of Biological Psychiatry
M. Hébert et al. BIOL PSYCHIATRY 2010;67:270 –274 271

sibships, among which 14 comprised a single HR subject, 6

comprised 2 HR subjects, and 1 comprised 3 HR subjects.
Healthy Control Subjects. A sample of 29 unrelated healthy
control subjects balanced for age and gender (mean age: 20.6
years, SD 4.2) were selected from the same population through
local advertising. The exclusion criteria were the same as for the
offspring, with the addition of any Axis I DSM-IV diagnosis and
a positive family history of SZ or BD spectrum disorders. Season
of measurement was also balanced with 14 HRs and 12 control
subjects tested in winter and 15 HRs and 17 control subjects
tested in summer (␹2 ⫽ .28, p ⫽ .60). The study was approved by
the Institutional Ethics Committee and signed consent was
obtained.
Electroretinography
Procedure. Electroretinography technique and protocol
used in this study were shown to provide reliable retinal
measures of both cones and rods (22). Recordings were obtained
in both eyes (averaged for analysis) with Dawson-Trick-Litzkow
electrodes (Shieldex 33/9 Thread, Statex, Bremen, Germany)
secured deep in the conjunctival sac. Ground and reference
electrodes (Grass Gold Cup Electrodes filled with Grass EC2 Figure 1. Examples in one subject of typical electroretinographic responses
Electrode Cream; Astro-Med Inc., Brossard, QC, Canada) were obtained at Vmax in both photopic (A) and scotopic (B) conditions. The
photopic Vmax was achieved at 7.5 cd/m2 and scotopic at .056 cd/m2. The
secured to the forehead and external canthi, as previously
a-wave amplitude is measured from baseline to trough of the a-wave, and
described (22). the b-wave from trough of the a-wave to peak of the b-wave (A). The arrows
Photopic ERG was used to assess cone function (day vision). indicate flash onset.
The subject was light-adapted for 15 minutes to a light back-
ground set at 80 candela (cd)/m2 provided by a Ganzfeld Color
eliminate the presence of a diagnosis of major psychiatric
Dome (Espion, DiagnoSys LLC, Littleton, Massachusetts) in
disorder (SZ, BD, major depression) (21). In summary, the
which an integrated camera allowed continuous monitoring of
methods consisted of a direct interview with the parents and the
the eyes during testing, as well as flash stimulation by means of
children using the Kiddie-Schedule for Affective Disorders and
a xenon strobe (white flash) or light-emitting diodes (LEDs)
Schizophrenia (K-SADS) (23) for subjects under 18 or the Struc-
(colored flash). A cone luminance-response function (LRF) was
tured Clinical Interview for DSM-V (SCID) (24) for subjects over
provided using 18 increasing white flash intensities ranging from
18. All available medical records were also reviewed and a
.076 to 3200 cd/m2 per second with an interstimulus interval set
lifetime best estimate DSM-IV diagnosis used all available infor-
at 1.5 seconds. Subjects were then dark-adapted for 30 minutes
mation concerning the offspring. A diagnosis of major psychosis
before the scotopic ERG was performed to assess rod function
was eliminated in this HR sample, but a nonpsychotic diagnosis
(night vision). A rod LRF was obtained using 15 increasing green
(anxiety disorders, disruptive disorders, personality and learning
flash (wavelength peak: 509 nm) intensities ranging from .000178
disorders, alcohol and substance abuse disorders) was present in
to 1 cd/m2 per second with an interstimulus interval set at 2
about 60% of HRs, as reported before by our group (21), which
seconds (first 11 intensities) and 5 seconds (last 4 intensities). For
is consistent with rates reported in other HR samples (25).
all recordings, at least 10 responses were averaged for each
In control subjects, a shortened version of the SCID or
intensity to achieve a low signal-to-noise ratio.
K-SADS was used to ensure that they were exempt from any
ERG Analyses. Waveform analysis was performed offline
diagnoses. In the analysis, the presence of nonpsychotic diagno-
with the experimenter being blind to the status (HR or control
sis was controlled for when comparing HR subjects with control
subject) of the participant’s recording. The ERG response is
subjects on ERG.
usually composed, first, of a negative component named the
For the diagnoses of the affected parents, we administered a
a-wave followed by a more prominent positive component, the
lifetime best estimate diagnosis based on several sources of
b-wave (see example in Figure 1). The a-wave originates mainly
information that have been the object of several previous reports
from the photoreceptors, whereas the b-wave is generated by the
(26). The parents had either a DSM-IV lifetime diagnosis of SZ
bipolar-Müller cells complex, postsynaptic to photoreceptors.
(n ⫽ 9) or BD (n ⫽ 20).
The a-wave amplitude is measured from baseline to trough, and
the b-wave amplitude is measured from the trough of the a-wave
to the peak of the b-wave. At lower intensities, when only rods Statistical Analysis
are responding, no a-wave is observed, and the b-wave is We compared the HR offspring (n ⫽ 29) to control subjects
measured relative to baseline. Each system (rods and cones) (n ⫽ 29) on each of the eight ERG parameters by means of
demonstrates a specific LRF from which parameters such as analyses of variance (ANOVAs). To address the possible effect of
saturation amplitude (Vmax) and retinal sensitivity (log K) are lack of independence among observations due to some of the
derived. In the current study, Vmax was used as the reference offspring who were sibs, multilevel modeling (27) was carried
point for analysis. For details, see Figure S1 in Supplement 1. out using the group assessment (HRs vs. control subjects) as a
first level and the sibships nested in the group as a second
Psychiatric Assessment random level. In this multilevel model, we further adjusted for
In the present sample of 29 HRs, a best estimate diagnosis potential confounders such as gender, age, season of ERG
procedure providing a DSM-IV diagnosis was administered to assessment, or the presence of a nonpsychotic diagnosis. Fur-

www.sobp.org/journal
272 BIOL PSYCHIATRY 2010;67:270 –274 M. Hébert et al.

Table 1. Comparison of Offspring ERG Parameters to Normal Control Subjects

Multilevel Modela Adjusted Multilevel Modela Adjusted

Unadjusted Mean (SD) for Age and Gender for Age, Gender, and Season
ERG Parameters Offspring Control Subjects Effect Sizeb p Value Effect Sizeb p Value

Cones Vmax
b-Wave amplitude 100.1 (16.7) 96.9 (13.3) .27 .314 .35 .188
a-Wave amplitude ⫺20.26 (6.57) ⫺22.79 (5.22) .43 .116 .40 .134
b-Wave implicit time 29.62 (1.59) 29.57 (1.46) .01 .981 .01 .979
a-Wave implicit time 14.19 (1.06) 14.40 (0.75) ⫺.23 .385 ⫺.22 .415
Log K .65 (.05) .64 (.07) .22 .421 .22 .424
Rods Vmax
b-Wave amplitude 122.9 (16.6) 158.8 (23.6) ⫺1.47 ⬍.0001 ⫺1.40 ⬍.0001
b-Wave implicit time 60.21 (10.64) 54.17 (7.37) .65 .017 .63 .020
Log K ⫺2.23 (.19) ⫺2.24 (.13) .08 .774 .04 .891
ERG, electroretinography; LSM, least squares mean.
a
To account for possible correlation among subjects within the same sibship, a multilevel model was carried out using the MIXED procedure of SAS, version
9.1.3. Sibships nested in the group were used as the second level and modeled according to a random effect. Degrees of freedom were obtained by the
method of Kenward-Roger (31) that is available with the option DDFM ⫽ KR in the MODEL statement of the MIXED procedure.
b
Effect sizes were calculated using the difference of LSM between the two groups standardized by a pooled standard deviation according to the formula
provided by Kelly (28).

thermore, we verified whether the group effect (belonging to HR .007; effect size ⫽ ⫺1.09) and HRBD (t ⫽ ⫺6.47; p ⬍ .0001; effect
or to control group) depended on gender or age by testing the size ⫽ ⫺1.88). The rod Vmax amplitude was lower in both the
interaction terms group ⫻ gender and group ⫻ age separately. younger (ⱕ17 years; effect size of ⫺1.6) and older HRs (ⱖ18
The effect of age was treated as a dichotomous variable in HRs years; effect size of ⫺1.8).
and control subjects: ⱕ17 and ⱖ18 years. Finally, we compared
the ERG parameters in the HRs of a SZ parent (HRSZ) with those Discussion
of a BD parent (HRBD) versus control subjects to verify if the
type of parental illness had an effect on ERG. These present HRs at extreme genetic risk of SZ or mood
Our multiple testing, due to analyses performed on each of disorder displayed lowered b-wave maximal amplitude (about
the five cone and three rod ERG measures, made us fix the 23%) only at the rod level. When compared with the only
threshold of significance (p value) at .006 (i.e., .05/8). A p value electrophysiological study that assessed the full dynamic re-
greater than .006 but below .02 (i.e., .15/8) was interpreted as a sponse of both cones and rods in nonmedicated patients, namely
trend. All analyses were performed using the MIXED procedure SAD (1), the present findings showed an important difference.
of SAS (version 9.1.3; SAS Institute, Inc., Cary, North Carolina). Whereas in the SAD study a decreased cone response (17%) and
Effect sizes were calculated using Cohen’s statistics (28). rod sensitivity (log K) were observed in winter, in the current
study, only a substantial rod Vmax decrease was observed.
Results Indeed, one has to be cautious in interpreting these data as
Group analyses (Table 1) revealed no significant difference
between HRs and normal control subjects for any of the cone
ERG parameters (b-wave amplitude and implicit time, a-wave
amplitude and implicit time, log K). For the rod ERG parameters
(Figure 2, Table 1), ANOVAs revealed a significant effect with
Vmax b-wave amplitude (F ⫽ 31,29; p ⬍ .0001; effect size ⫽
⫺1.47) and a trend with Vmax b-wave implicit time (F ⫽ 6,12; p ⫽
.017; effect size ⫽ .65) with differences remaining significant
after adjusting for age, gender, and season of testing (Table 1).
When the effect of group (rod Vmax) was adjusted for the
presence of nonpsychotic diagnoses, the corresponding adjusted
p value remained significant and as low as p ⫽ .0002, indicating
no effect of the nonpsychotic diagnoses in HRs.
The group effect on rod Vmax amplitude did not depend on
gender (p ⫽ .766) or on age (p ⫽ .806). Indeed, the effect size
illustrating the group effect on rod Vmax amplitude was very
similar between the two strata based on gender or age. Of
interest, 72.4% (21/29) of HR offspring demonstrated a rod Vmax Figure 2. The rod Vmax b-wave amplitude values for each subject in off-
amplitude that was at least one standard deviation lower than spring and controls obtained in scotopic condition. The scotopic Vmax was
normal control subjects, whereas 31% (9/29) showed a response selected at intensity ⫺1.25 log cd/m2 based on the break point observed on
the group averaged luminance response function. Ordinate: Vmax amplitude
two standard deviations below normal control subjects (Figure 2). (␮V). The mean and standard deviation of each group are represented by a
When the HRs were divided according to the illness of their long and a short horizontal line respectively. High risk offspring subjects are
parent, the differences in rod Vmax amplitude when compared represented by filled circles, and control subjects by filled triangles. Graph
with control subjects were present both in HRSZ (t ⫽ ⫺2.86; p ⫽ produced with GraphPad Prism 5.01 (La Jolla, California).

www.sobp.org/journal
M. Hébert et al.
suggestive of different pathophysiologies, since some SAD pa-
tients (up to 20%) may demonstrate manic episodes in the spring
or summer and therefore receive a diagnosis of bipolar I or II
disorder (29). Since in the SAD study (1) only patients with
atypical symptoms, mild depression, and no history of spring or
summer manic episode were selected, the latter ERG results may
be applicable only to this specific sample of SAD. In any case, the
magnitude of the selective rod difference between the HRs and
normal control subjects is intriguing. Although there is some
substantial overlap, no HR showed a rod Vmax amplitude beyond
the average value of control subjects, as shown in Figure 2.
Confounding factors such as age, gender, and season of testing
were controlled, but we cannot eliminate the possibility of an
uncontrolled factor yet to be discovered that could have magni-
fied the group difference at the rod level only.
Our result also contrasts with previous ERGs obtained in
medicated patients with schizophrenia, where a decreased cone
a-wave was reported (4,5). In our study, although not statistically
significant, considering the effect size of .4 for the a-wave
amplitude and the small sample size, one cannot exclude a type
II error. Of interest, Balogh et al. (4) observed that the a-wave
reduction was present in patients with schizophrenia but not
those with bipolar disorder. Considering that our HR offspring
descended from parents affected by either schizophrenia or
bipolar disorder, this may have dampened the a-wave effect in
our total HR sample.
Even though a definite neurobiological explanation for the
observed rod function reduction in these HR offspring is not easy
to formulate, an anomaly in neurotransmitter production or
receptor sensitivity may be involved. Based on animal research
effects of dopamine (10) and serotonin (12) on b-wave ampli-
tude, a specific rod response reduction might be caused by either
a DA receptor downregulation or a hyposensitivy (D1/D2) or
lower serotonin activity. Another possibility would be a reduced
number of rod photoreceptors or interneuron architecture. Con-
sidering that the eye is part of the central nervous system, our
data might indeed be a reflection of a dysfunctional neurodevel-
opmental process occurring at the brain level, which would be
compatible with the prevalent neurodevelopmental hypothesis
for SZ (20).
Since the offspring of a parent having major psychosis display
early in life nonpsychotic behavior disorders in our sample (21)
and in others (25), we adjusted our analysis for the presence of
psychiatric diagnosis and found that the retinal response effect
remained. Also, the retinal response anomaly was present in the
youngest HRs as it was in the eldest, diminishing the likelihood
of a sole relationship with a prodromal state.
We observed that the type of family history, either SZ or BD,
was not associated with the HRs’ ERG anomaly, suggesting that
this potential indicator of a neurodevelopmental anomaly was
shared by SZ and BD. This finding is not surprising, given the
growing evidence that patients affected by SZ or BD and their
relatives share phenotypic, genetic, and physiological features
possibly underlain by a common pathological pathway
(13,14,30,31).
A limitation of our study is that the HRs’ sampling frame based
on densely affected kindreds in the Eastern Québec population
(17) may have accentuated the genetic effect to explain the
disease and facilitated the observation of anomalies by magnify-
ing the findings. Moreover, the very high genetic liability of the
sample suggests caution before generalizing the findings to all
children of SZ and BD parents of the population.
BIOL PSYCHIATRY 2010;67:270 –274 273
This research was supported in part by a Canada Research
Chair (#950 –200810) in psychiatric genetics of which MM is the
Chair and by a Canadian Institutes of Health Research (CIHR)
grant (#MOP-74430).
We are grateful to our professional research assistants, Linda
René, Claudie Poirier, Lisette Gagnon, Louise Bélanger, and
Nicole LeClerc, and to the family members, adults, and children
who participated in this study. We also thank Institut Inter-
Universitaire de Recherche sur les Populations for their collabo-
ration regarding the Balsac database. MH is supported by Fonds
de la Recherche en Santé du Québec and CIHR (#MOP-82707).
MM and M-AR have been consultants for GlaxoSmithKline
(GSK) and Eli Lilly and have received research funding from
GSK, Eli Lilly, and AstraZeneca that are not related to the
material of this study. MH, A-MG, M-EP, CM, and VJ reported no
biomedical financial interests or potential conflicts of interest.
Supplementary material cited in this article is available online.
1. Lavoie MP, Lam RW, Bouchard G, Sasseville A, Charron MC, Gagné AM, et
al. (2009): Evidence of a biological effect of light therapy on the retina of
patients with seasonal affective disorder. Biol Psychiatry 66:253–258.
2. Hébert M, Dumont M, Lachapelle P (2002): Electrophysiological evi-
dence suggesting a seasonal modulation of retinal sensitivity in subsyn-
dromal winter depression. J Affect Disord 68:191–202.
3. Hébert M, Beattie C, Tam E, Yatham L, Lam R (2004): Electroretinography
in patients with winter seasonal affective disorder. Psychiatry Res 127:
27–34.
4. Balogh Z, Benedek G, Keri S (2008): Retinal dysfunctions in schizophre-
nia. Prog Neuropsychopharmacol Biol Psychiatry 32:297–300.
5. Warner R, Laugharne J, Peet M, Brown L, Rogers N (1999): Retinal func-
tion as a marker for cell membrane omega-3 fatty acid depletion in
schizophrenia: A pilot study. Biol Psychiatry 45:1138 –1142.
6. Ritvo ER, Creel D, Realmuto G, Crandall AS, Freeman BJ, Bateman JB, et al.
(1988): Electroretinograms in autism: A pilot study of b-wave ampli-
tudes. Am J Psychiatry 145:229 –232.
7. Jaffe MJ, Bruno G, Campbell G, Lavine RA, Karson CN, Weinberger DR
(1987): Ganzfeld electroretinographic findings in Parkinsonism: Un-
treated patients and the effect of levodopa intravenous infusion. J Neu-
rol Neurosurg Psychiatry 50:847– 852.
8. Filip V, Balik J (1978): Possible indication of dopaminergic blockade in
man by electroretinography. Int Pharmacopsychiatry 13:151–156.
9. Fornaro P, Dell’Osso L, Perossini M, Placidi GF, Castrogiovanni P (1984):
Effects of nomifensine on dopaminergic transmission in healthy volun-
teers evaluated by means of the electroretinographic technique. Acta
Neurol (Napoli) 6:5–10.
10. Holopigian K, Clewner L, Seiple W, Kupersmith MJ (1994): The effects of
dopamine blockade on the human flash electroretinogram. Doc Oph-
thalmol 86:1–10.
11. Castrogiovanni P, Di Muro A, Maremmani I, Perossini M, Marazziti D
(1989): Interaction between the serotonin and dopamine systems in
humans: Preliminary findings. Brain Res 504:129 –131.
12. Porciatti V, Bagnoli P, Alesci R, Fontanesi G (1987): Pharmacological
dissociation of the b-wave and pattern electroretinogram. Doc Ophthal-
mol 65:377–383.
13. Maziade M, Roy M-A, Chagnon Y, Cliche D, Fournier J, Montgrain N, et al.
(2005): Shared and specific susceptibility loci for schizophrenia and
bipolar disorder: A dense genome scan in Eastern Quebec families. Mol
Psychiatry 10:486 – 499.
14. Maziade M, Chagnon YC, Roy MA, Bureau A, Fournier A, Mérette C
(2009): Chromosome 13q13-q14 locus overlaps mood and psychotic
disorders: The relevance for redefining phenotype. Eur J Hum Genet
17:1034 –1042.
15. Mérette C, Roy MA, Bureau A, Fournier A, Emond C, Cliche D, et al. (2008):
Replication of linkage with bipolar disorder on chromosome 16p in the
Eastern Quebec population. Am J Med Genet B Neuropsychiatr Genet
147B:737–744.
16. Braff DL, Freedman R, Schork NJ, Gottesman I (2007): Deconstructing
schizophrenia: An overview of the use of endophenotypes in order to
understand a complex disorder. Schizophr Bull 33:21–32.
www.sobp.org/journal
274 BIOL PSYCHIATRY 2010;67:270 –274
17. Maziade M, Rouleau N, Gingras N, Boutin P, Paradis M-E, Jomphe V, et al.
2009): Shared neurocognitive dysfunctions in young offspring at ex-
treme risk for schizophrenia or bipolar disorder in Eastern Quebec mul-
tigenerational families. Schizophr Bull 35:919 –930.
18. Szatmari P, Maziade M, Zwaigenbaum LZ, Mérette C, Roy MA, Joober R,
Palmour R (2007): Informative phenotypes for genetic studies of psychi-
atric disorders. Am J Med Genet B Neuropsychiatr Genet 144:581–588.
19. Maziade, M, Palmour R, Phaneuf D, Mérette C, Roy M-A (2003): Schizo-
phrenia and bipolar disorder: Linkage on chromosomes 5 and 11. In:
Cooper DN, NTE editor. Nature Encyclopedia of the Human Genome, Vol 1.
London: Macmillan Publishers Ltd, Nature Publishing Group, 178 –181.
20. Murray RM, O’Callaghan E, Castle DJ, Lewis SW (1992): A neurodevelop-
mental approach to the classification of schizophrenia. Schizophr Bull
18:319 –331.
21. Maziade M, Gingras N, Rouleau N, Poulin S, Jomphe V, Paradis ME, et al.
(2008): Clinical diagnoses in offspring from Eastern Québec multigen-
erational families densely affected by schizophrenia or bipolar disorder.
Acta Psychiatr Scand 117:118 –126.
22. Hébert M, Lachapelle P, Dumont M (1995): Reproducibility of electroretino-
grams recorded with DTL electrodes. Doc Ophthalmol 91:333–342.
23. Kaufman J, Birmaher B, Brent D, Rao U, Flynn C, Moreci P, et al. (1997):
Schedule for Affective Disorders and Schizophrenia for School-Age Chil-
www.sobp.org/journal
M. Hébert et al.
dren-Present and Lifetime Version (K-SADS-PL): Initial reliability and
validity data. J Am Acad Child Adolesc Psychiatry 36:980 –988.
24. Spitzer RL, Williams JB, Gibbon M, First MB (1992): The Structured Clini-
cal Interview for DSM-III-R (SCID) I: History, rationale and description.
Arch Gen Psychiatry 49:624 – 629.
25. Hans SL, Auerbach JG, Styr B, Marcus J (2004): Offspring of parents with
schizophrenia: Mental disorders during childhood and adolescence.
Schizophr Bull 30:303–315.
26. Roy M-A, Lanctôt G, Mérette C, Cliche D, Fournier J-P, Boutin P, et al.
(1997): Clinical and methodological factors related to reliability of the
best-estimate diagnostic procedure. Am J Psychiatry 154:1726 –1733.
27. Goldstein H (1998): Multilevel models. In: Armitage P, Colton T, editors.
Encyclopedia of Biostatistics, Vol 4. Chichester, United Kingdom: John
Wiley and Sons Ltd., 2725–2731.
28. Kelley K (2007): Confidence intervals for standardized effect sizes: The-
ory, application, and implementation. J Stat Softw 20:1–24.
29. Westrin A, Lam RW (2007): Seasonal affective disorder: A clinical update.
Ann Clin Psychiatry 19:239 –246.
30. Schretlen DJ, Cascella NG, Meyer SM, Kingery LR, Testa SM, Munro CA, et
al. (2007): Neuropsychological functioning in bipolar disorder and
schizophrenia. Biol Psychiatry 62:179 –186.
31. Berrettini W (2003): Evidence for shared susceptibility in bipolar disor-
der and schizophrenia. Am J Med Genet C Semin Med Genet 123C:59 – 64.