Professional Documents
Culture Documents
Effects of Tobacco Smoking On Neuropsychological Function in Schizophrenia in Comparison To Other Psychiatric Disorders and Non Psychiatric Controls
Effects of Tobacco Smoking On Neuropsychological Function in Schizophrenia in Comparison To Other Psychiatric Disorders and Non Psychiatric Controls
Background and Objectives: Compared to the general population the general population in which the rate is approximately
cigarette smoking prevalence is elevated in psychiatric disorders such 20%.1,2 Clinical estimates of smoking prevalence range from
as schizophrenia (SZ), bipolar disorder (BD), and major depressive
disorder (MDD). These disorders are also associated with neuro- 58–88% for individuals with SZ,3 51–70% for those with
cognitive impairments. Cigarette smoking is associated with BD,4–7 and 40–60% for those with MDD.7,8 This pattern has
improved cognition in SZ. The effects of smoking on cognition in been found to persist across cultures, even after controlling for
BD and MDD are less well studied. age, gender, country, and interaction between country and
Methods: We used a cross‐sectional design to study neuropsycho- mental illness.9 Furthermore, both epidemiological1,10 and
logical performance in these disorders as a function of smoking status.
Subjects (N ¼ 108) were SZ smokers (n ¼ 32), SZ non‐smokers
clinical studies11–17 have suggested a reduced likelihood of
(n ¼ 15), BD smokers (n ¼ 10), BD non‐smokers (n ¼ 6), MDD smoking cessation in people with PDs.
smokers (n ¼ 6), MDD non‐smokers (n ¼ 10), control smokers There is evidence for widespread neurocognitive deficits
(n ¼ 12), and control non‐smokers (n ¼ 17). Participants completed (executive functioning, attention and working memory function)
a neuropsychological battery; smokers were non‐deprived. in people with PDs, particularly those with SZ.18 Previous studies
Results: SZ subjects performed significantly worse than controls in
select domains, while BD and MDD subjects did not differ from
have demonstrated that individuals with SZ exhibited significant
controls. Three verbal memory outcomes were improved in SZ deficits in executive function,19 attention, and concentration,20
smokers compared with non‐smokers; smoking status did not alter working memory,21 verbal learning and memory,22 processing
performance in BD or MDD. speed,23 and visuospatial working memory (VSWM).24 Similar-
Conclusions and Scientific Significance: These data suggest that ly, there is evidence for neuropsychological dysfunction in both
smoking is associated with neurocognitive improvements in SZ, but
not BD or MDD. Our data may suggest specificity of cigarette‐
BD and MDD. BD is associated with impairments in processing
smoking modulation of neurocognitive deficits in SZ. (Am J Addict speed, working memory, verbal fluency and episodic memory,25
2013;22:46–53) verbal memory,26 and executive function,27 while MDD is
associated with deficits in episodic memory,28 verbal recall and
recognition,29 executive function,28,30,31 sustained attention, and
visuospatial learning and memory.28,30 In comparison to patients
INTRODUCTION with BD and MDD, however, those with SZ seem to have more
profound neuropsychological impairments.32,33 Recent studies
Cigarette smoking prevalence in psychiatric disorders (PDs) comparing the neuropsychological profiles of persons with SZ
such as schizophrenia (SZ), bipolar disorder (BD), and major and BD (in a euthymic phase) have either reported similar
depressive disorder (MDD) is two to three times higher than in profiles patterns but more severe impairments in SZ34 or more
widespread and severe cognitive deficits in SZ compared
with BD.35,36 In MDD, deficits in executive function are less
Received February 21, 2012; accepted March 21, 2012. severe than SZ,31 while deficits in sustained attention are more
Address correspondence to Dr. George, Department of Psychia- severe and stable in SZ than in BD and MDD, where deficits are
try, University of Toronto, Centre for Addiction and Mental Health more “state‐dependent.”37
(CAMH), 250 College Street, CS 734, Toronto, ON, Canada M5T There is increasing evidence that cigarette smoking and/or
1R8. E‐mail: tony_george@camh.net.
nicotine administration improves selected aspects of cognitive
46
dysfunction in SZ (reviewed in references 38–40). In contrast, following diagnoses: Schizophrenia/Schizoaffective Disorder
the cognitive‐enhancing effects in non‐psychiatric controls are (SZ), Bipolar I Disorder (BD), or MDD. Further, they were
less clear. Some studies have demonstrated enhancements required to be in stable remission from active psychiatric
following acute nicotine administration, while others have symptomatology, as judged by psychiatric evaluation, and
reported no effects or even detrimental effects (for review, see were without changes to psychiatric medication(s) for the past
references 41 and 42). Moreover, recent studies have indicated 3 months. Non‐psychiatric controls were excluded if under
that in non‐psychiatric controls chronic smoking is associated current treatment with psychotropic medications, or if they
with worse cognitive function, particularly in higher‐order showed evidence of a current Axis I PD or substance abuse/
domains,43–45 which may be due to the detrimental effects of dependence on the SCID‐IV (except nicotine or caffeine
acute nicotine or that poor cognitive function acts as a risk dependence). Smokers in the study were required to meet
factor for the initiation of smoking. It has been suggested that DSM‐IV criteria for moderate to heavy nicotine dependence,
cognitive enhancement in SZ as a result of smoking might be with a Fagerström Test for Nicotine Dependence (FTND)67
specific to the domains of working memory, attention, and score of 5, expired breath carbon monoxide (CO) measure-
information processing.46,47 The cognitive‐enhancing effects ment of >10 ppm, and an intake of 15 cigarettes per day, as
of smoking on neurocognitive performance in SZ have been assessed via self‐reported 7‐day Timeline Follow Back.68 Non‐
observed in both cross‐sectional (ie, smokers vs. non‐smokers) smokers were either never‐smokers or those who had abstained
and longitudinal studies (ie, smoking vs. abstinent conditions from cigarettes in the previous 6 months, and who had expired
and nicotine challenge paradigms) and span tasks associated breath CO levels of <10 ppm.
with sustained47–52 and selective attention,53 working memo- After eligibility determination, a total of 108 particip-
ry,24,47,51,54,55 response inhibition,50 processing speed,49,50,54 ants completed neurocognitive testing in a single session:
prepulse inhibition,56–58 and sensorimotor gating.46,59,60 (1) SZ smokers (n ¼ 32), (2) SZ non‐smokers (n ¼ 15),
The high rates of smoking in BD and MDD are often (3) BD smokers (n ¼ 10), (4) BD non‐smokers (n ¼ 6), (5)
attributed to an attempt to self‐medicate affective symptoms2 MDD smokers (n ¼ 6), (6) MDD non‐smokers (n ¼ 10), (7)
and elevated monoamine oxidase (MAO)‐A levels associated control smokers (n ¼ 12), and (8) non‐smoking controls
with the MDD.61–63 However, little is known about whether (n ¼ 17). Written informed consent was obtained from all
smoking modulates cognition in BD or MDD. This is an participants seeking participation as approved by the Human
important issue given that MDD and BD are associated with Investigation Committee at Yale University Medical School.
both high rates of smoking and poor cognitive performance. To All participants were compensated for their participation.
our knowledge, the only study to be conducted to date
demonstrated in a small sample of smokers and non‐smokers Methods and Materials
with bipolar illness (I and II) that smoking status did not All testing was completed at the Neurocognitive Laboratory
significantly alter neuropsychological task performance on of the Program for Smokers with Mental Illness (PRISM) at the
measures of psychomotor speed, attention, memory, learning, Substance Abuse Center of CMHC. The initial screening
and executive function.64 Accordingly, in the present study, we assessment included the following measures: demographics
used a cross‐sectional design to evaluate the effects of smoking questionnaire, SCID‐IV, FTND, urine toxicology screen to rule
status on neuropsychological task performance in subjects with out current substance use, CO measurement, Brief Psychiatric
SZ, BD, and MDD. Rating Scale (BPRS),69 Beck Depression Inventory‐2nd
Edition (BDI‐II),70 Positive and Negative Syndrome Scale
METHODS for Schizophrenia (PANSS)71 (used only with SZ participants),
and the Shipley Full Scale Intelligence Screen.65
Participants
Smokers and non‐smokers with PDs were recruited from the Neuropsychological Assessment
outpatient divisions of The Connecticut Mental Health Center Details of the neuropsychological battery used in the present
(CMHC, New Haven, CT) and its satellite clinics. Healthy study were described previously.47,72 All tests were adminis-
control smokers and non‐smoking controls between the ages of tered according to standardized procedures and supervised by
18 and 65 were recruited from the Greater New Haven the study neuropsychologist (KAS). The order of administra-
community through flyers posted in the community and word‐ tion was randomly alternated in one of three pre‐assigned ways
of‐mouth. Individuals who met any of the following criteria in order to account for potential sequencing bias. Smokers were
were excluded from participation: (1) Full Scale IQ (FSIQ) of given three prespecified smoke breaks during testing to ensure
<70 on a screening measure Shipley Full Scale Intelligence that deprivation from cigarette smoking did not exceed
Screen65; (2) history of a neurological illness/injury or previous 30 minutes47 and therefore minimize any neurocognitive
loss of consciousness; (3) abuse or dependence to alcohol or effects related to smoking withdrawal.
any illicit substance within the past 3 months; or (4) inability to
provide written informed consent. Digit Span: Forward and Backward. Digit span (5–
All patient participants had to meet Structured Clinical 10 min) is a two‐part subtest of the Wechsler intelligence and
Interview for DSM‐IV (SCID‐IV)66 criteria for one of the memory batteries that is used for measuring span of immediate
Trail Making Test, Parts A and B (TMT). The TMT (5– Demographic and Clinical Characteristics of
10 min) is administered in two parts: Part A, which primarily the Sample
measures psychomotor speed, requires participants to draw Significant main effects of smoking status were observed
lines connecting consecutively numbered circles; Part B is used for sex, years of education and IQ, and main effects of
to assess cognitive set‐shifting (a type of executive function), diagnosis were found on most baseline variables except CO
and requires participants to alternate connecting lines between levels (p ¼ .08; Table 1).
letters and numbers in a more complex sequence.75
Comparison of Neuropsychological Performance
Wisconsin Card Sorting Test (WCST). The WCST across Diagnoses and Smoking Status
(25 min) assesses aspects of executive functioning including CVLT‐II
cognitive flexibility in response to feedback; it also involves A significant main effect of diagnosis was found across all
learning and memory, attention, and concentration. Commonly outcomes (Total, Trial 1, Trial 5, List B, short‐term free recall
reported outcomes are percent of total errors, number of [STFR], short‐term cued recall [STCR], long‐term free recall
categories completed, and percent of perseverative errors.76 [LTFR], long‐term cued recall [LTCR], and recognition/
discriminability [R/D]). There were also significant interaction
Visuospatial Working Memory (VSWM). The VSWM effects of diagnosis by smoking status on List B, LTCR, and
(20 min) is a computerized delayed‐response spatial working R/D scores, and trends towards an interaction effect on STCR
memory task, which assesses working memory for nonverbal (Table 2).
(object) visuospatial stimuli.24 Bonferroni‐adjusted pair‐wise comparisons were conducted
to analyze specific group differences on all significant
California Verbal Learning Test—Second Edition (CVLT‐ outcomes. Individuals with SZ scored significantly lower
II). The CVLT‐II (30 min) is a verbal learning and memory than non‐psychiatric controls on CVLT Total (p < .01), Trial 1
test that measures encoding, immediate and delayed recall, and (p < .05), Trial 5 (p < .05), STFR (p < .0001), STCR
recognition of new verbal information with the presentation of (p < .0001), LTFR (p < .0001), and LTCR (p < .0001); while
a 16‐item list.77 BD and MDD participants did not differ significantly from
non‐psychiatric controls on any of these outcomes (p ¼ 1.0 in
Statistical Analyses all cases). Adjusted pair‐wise comparisons on List B scores
Group demographic characteristics were explored via a revealed that none of the PD groups differed significantly from
series of ANOVAs and chi‐square analyses (see Table 1). The controls (SZ: p ¼ .58, BD: p ¼ .96, MDD: p ¼ 1.0). Within
diagnostic/smoking groups differed significantly on demo- the control group, however, smokers had significantly lower
graphic characteristics (ie, age, years of education, and gender); List B scores than non‐smokers (p < .05); no effect of smoking
therefore, subsequent testing included these variables as status was found within the SZ (p ¼ 1.0), BD (p ¼ .42), or
covariates. Although the groups also differed with regard to MDD (p ¼ 1.0) groups. On LTCR, among individuals with
distribution of race, the overwhelming majority of the sample SZ, scores were significantly higher among smokers versus
was Caucasian and the other races were too sparsely populated non‐smokers (p < .05). This was not the case in the other
across groups for race to be meaningfully included in subsequent diagnostic groups, where no smoking effect was observed
analyses. Consequently, race was dropped as a potential (BD: p ¼ .70, controls: p ¼ 1.0, MDD: p ¼ .37). On R/D
covariate. Homogeneity of variance and normality of data scores, adjusted comparisons revealed that none of the PD
distributions were evaluated on all dependent variables. For groups differed significantly from controls (p ¼ .19, p ¼ 1.0,
some outcome measures (noted as appropriate), due to the p ¼ 1.0, respectively), although among individuals with SZ,
distributional properties of the data, a natural‐logarithm smokers scored significantly higher than non‐smokers (p < .01).
transformation was applied prior to testing. This was not the case within the other diagnostic groups, where
A series of two‐way analyses of covariance (ANCOVAs) no difference was found between smokers and non‐smokers
were conducted to assess the interactive effects of diagnosis (BD: p ¼ 1.0, controls, p ¼ .52, MDD: p ¼ 1.0).
(SZ, MDD, BD, or control) and smoking (smokers vs. non‐
smokers) on neurocognitive test scores. Three covariates CPT
(age, gender, and years of education) were included in all No evidence of a smoking or diagnosis by smoking effect
analyses. When significant ANCOVA effects were observed, was found within any of the CPT tests. A significant diagnosis
a series of Bonferroni‐adjusted pair‐wise comparisons were effect was found in the CPT variability analysis, but all
Main
effect of
Control Control SZ SZ BD BD MDD MDD smoking Main
smokers non‐smokers smokers non‐smokers smokers non‐smokers smokers non‐smokers status* effect of
(n ¼ 12) (n ¼ 17) (n ¼ 32) (n ¼ 15) (n ¼ 10) (n ¼ 6) (n ¼ 6) (n ¼ 10) (p) diagnosis (p)*
Age M (SD) 38.0 (9.3) 38.5 (15.3) 41.3 (6.9) 39.9 (9.2) 46.9 (7.2) 41.3 (13.3) 44.8 (5.1) 46.9 (10.7) .648 <.05
Sex (% male) 83.3 (10) 70.6 (12) 81.3 (26) 53.3 (8) 60.0 (6) 0.0 (0) 33.3 (2) 50.0 (5) <.05 <.05
Race
Caucasian 10 14 15 9 10 5 5 10 .27
Morisano et al.
<.0001
African American 2 3 16 6 0 0 1 0
Hispanic 0 0 1 0 0 1 0 0
Years of education 13.5 (2.7) 15.9 (2.4) 11.6 (2.4) 13.4 (2.4) 13.4 (2.0) 15.7 (2.3) 13.3 (1.8) 15.0 (3.3) <.001 <.001
Estimated FSIQ 97.5 (11.7) 111.4 (13.4) 88.4 (11.0) 90.7 (11.2) 103.5 (10.5) 106.0 (8.4) 100.8 (11.9) 108.3 (9.3) <.01 <.001
# Cigs per day 21.1 (8.3) 0 23.7 (10.9) 0 21.4 (8.7) 0 15.8 (12.5) 0 – <.05
Baseline carbon 20.4 (5.2) 1.0 (.6) 24.8 (14.7) 1.1 (.7) 16.5 (6.1) 5 (.5) 20.8 (13.5) 1.2 (.6) – .08
monoxide
level (ppm)
FTND 6.1 (1.2) .0 (.0) 6.8 (1.6) .0 (.0) 6.7 (1.8) .0 (.0) 7.2 (1.2) .0 (.0) – <.05
January–February 2013
BDI 3.9 (5.8) 1.3 (1.4) 9.9 (9.5) 7.0 (6.0) 12.6 (6.9) 10.2 (11.7) 17.8 (9.0) 13.7 (9.2) .07 <.001
BPRS 17.1 (8.7) 19.1 (1.8) 35.4 (6.8) 38.9 (5.0) 24.9 (3.2) 30.0 (1.9) 28.6 (4.2) 27.3 (6.7) .53 <.001
PANSS positive – – 13.7 (3.7) 15.3 (2.2) – – – – .13 –
PANSS negative – – 14.9 (3.2) 14.3 (2.7) – – – – .54 –
PANSS general – – 28.8 (5.5) 30.5 (4.3) – – – – .32 –
PANSS total – – 57.2 (10.5) 60.1 (8.4) – – – – .36 –
*
ANOVA results are reported for continuous variables, Chi‐square results are reported for categorical variables.
SZ ¼ Schizophrenia; BD ¼ Bipolar Disorder; MDD ¼ Major Depressive Disorder; FTND ¼ Fagerstrom Test for Nicotine Dependence; BDI ¼ Beck Depression Inventory; PANSS ¼ Positive and Negative
Symptoms Scale for Schizophrenia.
49
50
TABLE 2. Neurocognitive outcomes across subgroups
Diagnosis Main Main
Control Control SZ SZ BD BD MDD MDD X smoking effect of effect of
smokers non‐smokers smokers non‐smokers smokers non‐smokers smokers non‐smokers interaction* smoking status* diagnosis*
CPT, hit rate (%)† n ¼ 12, 98.5 (1.1) n ¼ 17, 98.6 (1.8) n ¼ 31, 97.8 (2.7) n ¼ 15, 97.2 (3.7) n ¼ 10, 98.8 (.95) n ¼ 6, 98.7 (1.6) n ¼ 6, 99.0 (.49) n ¼ 10, 98.8 (1.7) F(3,96) ¼ .08, p ¼ .97 F(1,96) ¼ .97, p ¼ .97 F(3,96) ¼ 1.69, p ¼ .33
CPT, omissions† n ¼ 12, 1.5 (1.1) n ¼ 17, 1.4 (1.8) n ¼ 31, 2.2 (2.7) n ¼ 15, 2.8 (3.8) n ¼ 10, 1.2 (1.0) n ¼ 6, 1.3 (1.6) n ¼ 6, 1.0 (.49) n ¼ 10, 1.2 (1.7) F(3,96) ¼ .14, p ¼ .94 F(1,96) ¼ .28, p ¼ .60 F(3,96) ¼ 2.13, p ¼ .10
CPT, comissions† n ¼ 12, 37.3 (19.9) n ¼ 17, 21.6 (20.1) n ¼ 31, 33.4 (22.8) n ¼ 15, 28.7 (20.6) n ¼ 10, 24.4 (12.6) n ¼ 6, 44.5 (14.9) n ¼ 6, 21.0 (10.1) n ¼ 10, 25.6 (11.7) F(3,96) ¼ 2.63, p ¼ .06 F(1,96) ¼ .01, p ¼ .92 F(3,96) ¼ .39, p ¼ .76
CPT, variability† n ¼ 12, 8.4 (3.9) n ¼ 17, 9.4 (8.5) n ¼ 31, 13.5 (8.7) n ¼ 15, 15.4 (9.1) n ¼ 10, 8.6 (5.5) n ¼ 6, 9.7 (3.0) n ¼ 6, 13.9 (16.7) n ¼ 10, 7.6 (4.1) F(3,96) ¼ .18, p ¼ .91 F(1,96) ¼ .08, p ¼ .78 F(3,96) ¼ 3.46, p < .05
CPT, D′ n ¼ 12, 2.6 (.8) n ¼ 17, 3.8 (2.1) n ¼ 31, 2.7 (1.4) n ¼ 15, 2.6 (1.0) n ¼ 10, 2.9 (1.0) n ¼ 6, 2.1 (.79) n ¼ 6, 2.8 (1.1) n ¼ 10, 2.9 (.79) F(3,96) ¼ 1.95, p ¼ .13 F(1,96) ¼ .18, p ¼ .67 F(3,96) ¼ .58, p ¼ .63
VSWM, 30‐second n ¼ 12, 1.3 (.6) n ¼ 17, 1.1 (.5) n ¼ 32, 1.9 (.9) n ¼ 14, 2.6 (1.6) n ¼ 10, 1.6 (.7) n ¼ 6, 2.2 (2.1) n ¼ 6, 1.0 (.3) n ¼ 10, 1.5 (1.3) F(3,96) ¼ .69, p ¼ .56 F(1,96) ¼ 2.01, p ¼ .16 F(3,96) ¼ 5.36, p < .01
delay (cm)
VSWM, 60‐second n ¼ 12, 1.4 (.7) n ¼ 17, 1.0 (.6) n ¼ 32, 2.2 (1.2) n ¼ 14, 2.3 (1.7) n ¼ 10, 1.8 (.8) n ¼ 6, 2.6 (1.7) n ¼ 6, 1.3 (.6) n ¼ 10, 1.5 (.8) F(3,96) ¼ 1.30, p ¼ .28 F(1,96) ¼ .39, p ¼ .53 F(3,96) ¼ 4.97, p < .01
delay (cm)
WCST, % errors† n ¼ 12, 27.9 (16.7) n ¼ 15, 18.6 (13.1) n ¼ 32, 32.2 (17.3) n ¼ 15, 37.0 (21.3) n ¼ 10, 30.0 (19.8) n ¼ 6, 36.2 (13.2) n ¼ 6, 21.7 (15.1) n ¼ 10, 17.1 (6.1) F(3,95) ¼ 1.99, p ¼ .12 F(1,95) ¼ .81, p ¼ .37 F(3,95) ¼ 4.57, p < .01
WCST, % perseverative n ¼ 12, 12.6 (6.8) n ¼ 14, 9.3 (5.3) n ¼ 32, 14.6 (8.9) n ¼ 15, 22.3 (15.9) n ¼ 10, 20.4 (17.8) n ¼ 6, 15.3 (6.5) n ¼ 6, 11.8 (10.6) n ¼ 10, 9.0 (3.7) F(3,94) ¼ 1.58, p ¼ .20 F(1,94) ¼ .46, p ¼ .50 F(3,94) ¼ 3.55, p < .05
errors†
WCST, categories completed‡ n ¼ 12, 4.5 (2.1) n ¼ 14, 5.7 (1.1) n ¼ 32, 3.8 (2.2) n ¼ 15, 3.3 (2.6) n ¼ 10, 5.0 (2.2) n ¼ 6, 4.2 (2.1) n ¼ 6, 4.5 (2.3) n ¼ 10, 5.9 (.3)
x2(3, n ¼ 105) ¼ 7.68, x2(1, n ¼ 105) ¼ .89, x2(3, n ¼ 105) ¼ 3.84,
p ¼ .0531 p ¼ .35 p ¼ .28
TMT Part A (scaled score) n ¼ 12, 11.7 (2.1) n ¼ 17, 10.7 (3.3) n ¼ 30, 8.2 (3.1) n ¼ 15, 8.7 (2.8) n ¼ 10, 10.0 (2.4) n ¼ 6, 8.2 (3.2) n ¼ 6, 11.5 (1.4) n ¼ 10, 10.5 (2.5) F(3,94) ¼ .99, p ¼ .40 F(1,94) ¼ 4.56, p < .05 F(3,94) ¼ 4.97, p < .01
TMT Part B (scaled score) n ¼ 12, 11.1 (2.2) n ¼ 17, 11.9 (3.5) n ¼ 30, 7.8 (2.9) n ¼ 14, 8.3 (2.6) n ¼ 10, 9.6 (1.9) n ¼ 6, 8.0 (3.0) n ¼ 6, 10.5 (2.9) n ¼ 10, 11.8 (3.9) F(3,93) ¼ 1.50, p ¼ .22 F(1,93) ¼ .53, p ¼ .47 F(3,93) ¼ 7.01, p < .001
CVLT, total score n ¼ 12, 50.8 (11.6) n ¼ 16, 43.7 (9.8) n ¼ 31, 36.1 (11.1) n ¼ 15, 34.9 (8.9) n ¼ 10, 47.9 (11.3) n ¼ 6, 41.5 (7.2) n ¼ 6, 43.0 (17.4) n ¼ 10, 49.8 (10.2) F(3,95) ¼ 1.79, p ¼ .15 F(1,95) < .01, p ¼ .98 F(3,95) ¼ 5.46, p < .01
CVLT, Trial 1 n ¼ 12, 5.3 (2.0) n ¼ 16, 6.6 (1.7) n ¼ 32, 4.7 (1.6) n ¼ 15, 3.8 (1.2) n ¼ 10, 5.7 (2.0) n ¼ 6, 4.8 (1.3) n ¼ 6, 6.0 (1.7) n ¼ 10, 6.2 (2.3) F(3,96) ¼ 2.32, p ¼ .08 F(1,96) ¼ .83, p ¼ .36 F(3,96) ¼ 4.83, p < .01
CVLT, Trial 5 n ¼ 12, 10.8 (1.9) n ¼ 16, 12.9 (2.7) n ¼ 32, 9.0 (3.2) n ¼ 15, 8.9 (2.7) n ¼ 10, 11.4 (2.2) n ¼ 6, 10.3 (2.7) n ¼ 6, 9.7 (5.5) n ¼ 10, 11.7 (2.8) F(3,96) ¼ 1.47, p ¼ .23 F(1,96) ¼ .18, p ¼ .67 F(3,96) ¼ 3.47, p < .05
CVLT, List B n ¼ 12, 4.1 (2.1) n ¼ 16, 6.2 (1.4) n ¼ 32, 3.8 (1.7) n ¼ 15, 4.5 (2.4) n ¼ 10, 4.7 (1.3) n ¼ 6, 3.7 (1.2) n ¼ 6, 5.5 (1.2) n ¼ 10, 5.7 (1.5) F(3,96) ¼ 3.20, p < .05 F(1,96) ¼ .33, p ¼ .57 F(3,96) ¼ 3.22, p < .05
CVLT, short‐term free recall n ¼ 12, 9.2 (2.1) n ¼ 16, 10.8 (3.4) n ¼ 32, 6.6 (2.6) n ¼ 15, 5.9 (3.2) n ¼ 10, 9.7 (2.2) n ¼ 6, 9.0 (3.1) n ¼ 6, 8.2 (3.6) n ¼ 10, 10.0 (2.7) F(3,96) ¼ 1.43, p ¼ .24 F(1,96) ¼ .30, p ¼ .58 F(3,96) ¼ 9.20, p < .0001
CVLT, short‐term cued recall n ¼ 12, 10.6 (1.1) n ¼ 16, 11.8 (2.3) n ¼ 32, 8.3 (2.8) n ¼ 15, 6.3 (3.4) n ¼ 10, 10.7 (2.1) n ¼ 6, 10.8 (2.1) n ¼ 6, 10.0 (3.3) n ¼ 10, 11.8 (2.0) F(3,96) ¼ 2.61, p ¼ .06 F(1,96) ¼ .27, p ¼ .61 F(3,96) ¼ 14.25, p < .0001
CVLT, long‐term free recall n ¼ 12, 9.3 (1.9) n ¼ 15, 11.0 (3.2) n ¼ 32, 6.7 (3.0) n ¼ 15, 5.4 (3.7) n ¼ 10, 10.3 (2.4) n ¼ 6, 9.7 (2.3) n ¼ 6, 8.8 (4.2) n ¼ 10, 10.7 (3.1) F(3,95) ¼ 1.74, p ¼ .16 F(1,95) ¼ .02, p ¼ .88 F(3,95) ¼ 9.77, p < .0001
CVLT, long‐term cued recall n ¼ 12, 10.3 (1.2) n ¼ 15, 11.5 (2.6) n ¼ 32, 8.3 (2.7) n ¼ 15, 6.1 (3.5) n ¼ 10, 11.9 (1.9) n ¼ 6, 10.2 (2.7) n ¼ 6, 9.3 (4.0) n ¼ 10, 12.1 (2.3) F(3,95) ¼ 3.79, p < .05 F(1,95) ¼ .21, p ¼ .65 F(3,95) ¼ 10.91, p < .0001
CVLT recognition/ n ¼ 12, 12.1 (3.7) n ¼ 15, 13.9 (2.0) n ¼ 32, 13.6 (1.8) n ¼ 15, 11.2 (3.4) n ¼ 10, 13.8 (1.9) n ¼ 6, 13.0 (2.0) n ¼ 6, 14.5 (1.4) n ¼ 10, 15.2 (1.2) F(3,95) ¼ 4.27, p < .01 F(1,95) ¼ .59, p ¼ .44 F(3,95) ¼ 2.73, p < .05
discriminability
*
ANCOVA results (covariates: age, gender, and years of education); †Due to the distributional properties of the data, a natural‐logarithm transformation was applied to the outcome measure prior to testing; ‡Over
50% of the subjects had completed all six categories. This variable was dichotomized as <6 (41%) versus 6 (58%) and a logistic regression analysis was performed instead of an ANCOVA.
SZ ¼ Schizophrenia; BD ¼ Bipolar Disorder; MDD ¼ Major Depressive Disorder; CPT ¼ Continuous Performance Test; VSWM ¼ Visuospatial Working Memory; WCST ¼ Wisconsin Card Sorting Test;
TMT ¼ Trial Making Test; CVLT ¼ California Verbal Learning Test.
TMT
WCST
VSWM
DISCUSSION
years of education.
January–February 2013
forward or backward performance (data not shown).