RESEARCH REPORT

doi:10.1111/j.1360-0443.2009.02656.x

Cognitive functioning in substance abuse and

dependence: a population-based study of young adults
2656

add_

1558..15681558..1568

Antti Latvala1,2, Anu E. Castaneda1,2, Jonna Perl1, Samuli I. Saarni1, Terhi Aalto-Setl1,3,
Jouko Lnnqvist1,4, Jaakko Kaprio1,5, Jaana Suvisaari1,6 & Annamari Tuulio-Henriksson1,2
Department of Mental Health and Substance Abuse Services, National Institute for Health and Welfare, Helsinki, Finland,1 Department of Psychology, University
of Helsinki, Finland,2 Department of Child Psychiatry, Hospital for Children and Adolescents, Helsinki University Central Hospital, Finland,3 Department of
Psychiatry, University of Helsinki, Finland,4 Department of Public Health, University of Helsinki, Finland5 and Department of Social Psychiatry,Tampere School of
Public Health, University of Tampere, Finland6

ABSTRACT
Aims To investigate whether substance use disorders (SUDs) are associated with verbal intellectual ability, psychomotor processing speed, verbal and visual working memory, executive function and verbal learning in young adults,
and to study the associations of SUD characteristics with cognitive performance. Participants A population-based
sample (n = 466) of young Finnish adults aged 2135 years. Measurements Diagnostic assessment was based on all
available information from a structured psychiatric interview (SCID-I) and in- and out-patient medical records. Established neuropsychological tests were used in the cognitive assessment. Confounding factors included in the analyses
were comorbid psychiatric disorders and risk factors for SUDs, representing behavioural and affective factors, parental
factors, early initiation of substance use and education-related factors. Findings Adjusted for age and gender, lifetime DSM-IV SUD was associated with poorer verbal intellectual ability, as measured with the Wechsler Adult Intelligence ScaleRevised (WAIS-R) vocabulary subtest, and slower psychomotor processing, as measured with the WAIS-R
digit symbol subtest. Poorer verbal intellectual ability was accounted for by parental and own low basic education,
whereas the association with slower psychomotor processing remained after adjustment for SUD risk factors. Poorer
verbal intellectual ability was related to substance abuse rather than dependence. Other SUD characteristics were not
associated with cognition. Conclusions Poorer verbal intellectual ability and less efficient psychomotor processing are
associated with life-time alcohol and other substance use disorders in young adulthood. Poorer verbal intellectual
ability seems to be related to parental and own low basic education, whereas slower psychomotor processing is
associated with SUD independently of risk factors.
Keywords

Abuse, cognition, dependence, population-based sample, substance use disorders, young adults.

Correspondence to: Antti Latvala, Department of Mental Health and Substance Abuse Services, National Institute for Health and Welfare, Mannerheimintie 166, FIN-00271, Helsinki, Finland. E-mail: antti.latvala@thl.fi
Submitted 23 January 2009; initial review completed 16 April 2009; final version accepted 20 April 2009.

INTRODUCTION
Substance use disorders (SUDs) are characterized by a
maladaptive pattern of substance use leading to clinically
significant impairment or distress. Several studies have
investigated cognitive functioning in people with SUD. In
alcohol use disorders, deficits in executive functions, visuospatial abilities, verbal abilities, learning, memory and
speed of information processing have been observed,
ranging from mild deficits in alcohol abuse and dependence to severe deficits in patients with Korsakoff syn-

drome [1,2]. Impaired cognition has also been reported

in drug use disorders, for example deficits in decision
making and inhibitory cognitive control, reflecting neural
processing in frontal cortical and subcortical areas [3].
Poorer cognitive functioning related to SUDs may
reflect both the effects of long-term heavy substance use
and cognitive differences predating SUD. Heavy use of
alcohol, opioids or stimulants may affect executive and
memory functions [4,5]. On the other hand, poorer general intellectual ability is often already observed in SUDs
in adolescence [6]. Also, findings of lower intellectual

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568

Cognition in substance use disorders

ability in children at elevated genetic risk for SUD [7]

suggest that lower intellectual ability is unlikely to be
caused by substance use. Indeed, prospective studies have
found lower verbal intellectual ability to increase the risk
for later alcohol problems [8]. Lower intellectual ability in
childhood also predicts both lower academic achievement
and illicit drug dependence in adolescence [9].
A necessary strategy for investigating the nature of
the relationship of cognitive deficits with SUD is to take
into account confounding factors that are related to both
SUD and cognitive functioning. Comorbid psychiatric disorders occur commonly with SUDs [10], or precede them
[11]. Psychiatric disorders may be associated with poorer
cognitive functioning and impair later academic and
occupational achievements. Similarly, low educational
level is found consistently in SUDs [12], and educational
achievement is known to overlap with cognitive abilities;
attention and behaviour problems in childhood increase
the risk for SUD [13] but are also associated with poorer
cognitive and intellectual abilities [14]. Moreover, many
other risk factors for SUDs may have associations with
cognitive performance.
Besides SUD risk factors, several characteristics of SUD
might be relevant for cognitive and intellectual functioning. For example, the validity of DSM-IV alcohol dependence and abuse has been under investigation [15], but
whether the type of diagnosis (dependence versus abuse)
is related to cognitive performance is not known. Furthermore, the number of SUD diagnoses during the life-time
and age at SUD onset may reflect the severity of SUD and
thus contribute to the association between SUD and cognition. In addition, comparing people with a current SUD
diagnosis with those in remission sheds light on the possible cognitive deficits related to the state of the disorder.
Both substance use and the incidence of SUDs are
known to peak in young adulthood [16]. Cognitive abilities, in turn, develop through childhood and adolescence,
reaching a stable level by young adulthood [17].
However, the association between SUDs and cognition
among young adults is not well known. Moreover, there
have been very few studies on SUDs and cognition using
general population samples. This paucity may have distorted the current state of knowledge on cognitive functioning in SUDs.
In order to address these issues, we used a representative population-based sample of young Finnish adults to
assess cognitive and verbal intellectual functioning in
alcohol and other substance use disorders. Axis I disorders
and several SUD risk factors, representing behavioural
and affective factors, parental factors, age at substance use
initiation and educational factors, were included in the
analyses, primarily as confounding factors. Associations
of cognitive and intellectual functioning with characteristics related to SUD diagnosis were also studied.

1559

METHODS
Sample
The present investigation is part of the Mental Health in
Early Adulthood in Finland (MEAF) study [18]. The
sample was assessed initially in 2001 as part of the
nationwide Health 2000 Survey [10], and re-examined
in the period 200305 in the MEAF study investigating
psychiatric disorders among young adults in Finland.
MEAF was a two-phase study. In the first phase, a questionnaire was sent to all 1863 members of the study
population, of whom 1316 (71%) returned the questionnaire. In the second phase, respondents who were
screened positive for mental health or substance use
problems and a random sample of people who screened
negative were invited to participate in a mental health
interview and neuropsychological assessment.
The MEAF questionnaire contained scales that
assessed mental health and substance use. A positive
screen for substance use entailed scoring at least three
in the Cut-down, Annoyed, Guilt, Eye-opener (CAGE)
questionnaire [19], or the self-reported use of any illicit
drug at least six times. The CAGE questionnaire, a
widely used screening instrument for alcohol problems,
contains four dichotomous questions assessing problems
related to drinking (need to cut down, annoyed by criticism, feeling guilty, need for an eye-opener). In addition
to screen-positive persons, individuals with hospital
treatment due to any mental or substance use disorder
(ICD chapter V: mental and behavioural disorders)
during the life-time according to the Finnish Hospital
Discharge Register were asked to participate. Details of
the sampling and screening procedures have been
reported previously [18]. Participants provided written
informed consent, and the study protocol was approved
by the ethics committees of the National Public Health
Institute and the Hospital District of Helsinki and
Uusimaa.

Diagnostic assessment
Of the 982 individuals invited for psychiatric and neuropsychological assessment, 546 (55.6%) participated. Previous analyses indicated that attrition depended on age,
sex and education, but not on mental disorders, psychological symptoms or substance-use-related problems
reported in the MEAF questionnaire [18]. The psychiatric
interview was conducted by experienced psychiatric
research nurses or psychologists using the Research
Version of the Structured Clinical Interview for DSMIV-TR [20]. The Global Assessment of Functioning (GAF)
and the Social and Occupational Functioning Assessment Scale (SOFAS) were also included. All interviews
were reviewed jointly by a psychiatrist and the inter-

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568

1560

Antti Latvala et al.

viewer. In addition, case-notes from hospital and outpatient treatments during the life-time were obtained,
and diagnostic assessment was based on all available
information from the interview and case records. Two
psychiatrists and two residents in psychiatry made the
final best-estimate diagnoses. All SUDs except for nicotine
dependence were assessed. Reliability of the diagnoses
was tested on 40 cases rated by all four clinicians. For
alcohol disorders, the unweighted pairwise kappa values
ranged from 0.94 to 1.

Neuropsychological assessment
Neuropsychological testing preceded the psychiatric
interview in the same session. Validated measures of
verbal intellectual ability, psychomotor performance and
processing speed, executive function, working memory
and verbal learning were used. Verbal intellectual ability
was assessed with the vocabulary subtest of the Wechsler Adult Intelligence ScaleRevised (WAIS-R) [21], and
the sum score of correct answers (1 or 2 points each)
was included in the analyses. Psychomotor performance
and processing speed was examined with the digit
symbol subtest of WAIS-R [21], and the number of correctly filled items was used in the analyses. The Trail
Making Test (TMT) [22] was administered to assess
executive functioning. The TMT contains two parts. In
part A, consecutively numbered circles on a test sheet
must be connected by lines to obtain the correct
sequence (123 . . . ). In part B, the same number of
consecutively numbered and lettered circles must be
connected by alternating between the two sequences
(1A2B3C . . . ). In both parts, the participant is
urged to perform as fast as possible. Time to complete
parts A and B and the difference score BA were analysed. Scores in the digit span forward and backward subtests, and the visual span forward and backward subtests
of the Wechsler Memory ScaleRevised (WMS-R) [23]
were used as variables for verbal and visual working
memory, respectively, and the letternumber sequencing
subtest of WAIS-III [24] was used as an additional
measure of verbal working memory. Verbal learning was
assessed with the California Verbal Learning Test (CVLT)
[25]. Total recall from trials 15 (learning performance),
short-delay recall and long-delay recall were included in
the analyses. Higher scores indicate superior performance in all the neuropsychological measures, except in
TMT. Valid neuropsychological data were available for
466 individuals aged 2135 years. Reasons for exclusion
included alcohol or other substance use during the
testing day (excluding tobacco), disturbances in the
testing situation, being a psychologist or student of psychology, native language other than Finnish, neurological disorders and life-time psychotic disorders.

SUD risk factors

Measures of risk factors for SUD were gathered from
questionnaire and interview data from both the baseline
Health 2000 study and the MEAF study. Behavioural
and affective risk factors included self-reported attention
or behaviour problems at school, an eight-item aggression measure based on selected items from the four subscales of the BussPerry Aggression Questionnaire [26]
(theoretical range 840, Cronbachs alpha = 0.82) classified further into low, moderate and high, and a fivepoint item assessing trait anxiousness, classified further
into low, moderate and high. Risk factors related to
parents were self-reported parental alcohol problems,
experienced before the age of 16, and parental basic education, measured as the highest secondary education of
either of the parents. Measures of substance use initiation were self-reported age at initiation of daily smoking
and age at initiation of drinking to intoxication.
Education-related risk factors were self-reported learning
difficulties at school and lower basic education. The SUD
risk factor measures are described in more detail in the
supplementary table which accompanies the online
version of this paper (see Supporting Information at the
end).
Statistical analysis
The initial cluster sampling design of the Health 2000
Survey [10] was taken into account using survey settings
in Stata 9 [27], and post-stratification weights were
applied to adjust for non-response and to correct the
survey distributions in order to correspond to the population distributions. Further, the two-phase screening for
the MEAF interview and neuropsychological tests was
taken into account using expansion weights calculated
for screen-positives (M) by dividing their total number by
the number interviewed (M1), i.e. M/M1, and for the
screen-negatives in the same way, N/N1 [28]. The final
weights used in the statistical analyses were calculated
by multiplying the expansion weights by the poststratification weights. The weighting procedure has been
described in more detail previously [18].
Associations between SUD diagnosis, cognitive measures and confounding factors were studied with t-tests,
c2 tests and linear regression models. Assumptions of
linear regression were tested with the ShapiroWilk test
for the normality of the residuals, and plots of residuals
versus fitted values were made to check the homoscedasticity of the residuals. Logarithmic, square, and 1/square
root transformations were used for the cognitive measures to approximate normality, when needed. Standardized cognitive variables were used in the regression
models to enable comparisons of predictor variable effects
for different cognitive outcomes. In visual span forward,

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568

Cognition in substance use disorders

1561

Table 1 Psychiatric diagnoses in the sample.

SUD (n = 58)

No SUD (n = 408)

Diagnosis

Alcohol abuse/dependence
Cannabis abuse/dependence
Amphetamine abuse/dependence
Opioid abuse/dependence
Other substance abuse/dependence
Polysubstance dependence
Any substance abuse/dependence
Abuse
Dependence
Current
In remission
Number of life-time SUD diagnoses
1
2 or more
Age at SUD onseta
19
18
Axis Ib or personality disorder***
Number of Axis I diagnosesb***
0
1
2
3 or more
Mood disorders*
Depressive disorders*
Bipolar disorders*
Anxiety disorders***
Eating disorders
Personality disorder***

56
8
6
2
2
4

97
14
10
3
3
7

31
27
27
31

53
47
47
53

46
12

79
21

31
25
38

55
45
66

149

37

21
19
8
10
22
20
2
18
2
12

36
33
14
17
38
35
3
31
3
21

264
78
48
18
92
89
2
50
24
10

65
19
12
4
23
22
5
12
6
2

SUD: substance use disorder. aInformation was not available for two cases. bOther than SUD. *P < 0.05, ***P < 0.001.

visual span backward, CVLT short delay recall and CVLT

long delay recall, modest ceiling effects were detected
(512% of the observations). Therefore, we used tobit
regression in addition to linear regression to study the
associations between SUD diagnosis and these measures;
no significant changes in the results occurred (data not
shown).

RESULTS
Description of the sample
Both Axis I and personality disorders were more common
in people with SUD (Table 1), while all the selected risk
factors were associated with life-time SUD diagnosis
(Table 2).
Intellectual and cognitive function in SUD
In the first phase of the analyses, the means of cognitive
measures in individuals from the SUD group were found

to be lower (reflecting poorer performance) than in the

no-SUD group in six tests: vocabulary, digit symbol,
letternumber sequencing, CVLT total learning and
CVLT short delay recall. Adjusting for age and gender,
differences in vocabulary and digit symbol remained
statistically significant, whereas differences in digit span
forward, letternumber sequencing, TMT part A and
CVLT total learning were bordering on being significant
(P < 0.10) (Table 3).
Based on these results, vocabulary, digit symbol, digit
span forward, letternumber sequencing, TMT part A
and CVLT total learning were selected for the next phase
of analyses. Associations of SUD risk factors with these
cognitive measures were studied in separate linear regression models, adjusting for age and gender (Table 4). Both
parental and own low basic education was associated
strongly with all cognitive measures. Other SUD risk
factors had associations with some of the cognitive measures (Table 4), with the exception of the Axis I disorder
diagnosis, anxiousness and age at initiation of drinking;

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568

1562

Antti Latvala et al.

Table 2 Distributions of age, gender and the substance use disorder (SUD) risk factors in the sample.
SUD (n = 58)

No SUD (n = 408)
%

t or c2

df

(3.7)

-2.81*

464

251
157

61.5
38.5

23.90***

43.1
34.5
22.4

362
24
22

88.7
5.9
5.4

76.31***

8
22
28

13.8
37.9
48.3

88
243
77

21.6
59.6
18.9

25.15***

35
16
7

60.3
27.6
12.1

320
73
15

78.4
17.9
3.7

12.26**

15
17
26

25.9
29.3
44.8

269
94
45

65.9
23.0
11.0

52.29***

7
36
15

12.1
62.1
25.9

116
261
31

28.4
64.0
7.6

22.34***

9
8
20
21

15.5
13.8
34.5
36.2

194
58
104
52

47.6
14.2
25.5
12.8

31.35***

9
26
23

15.5
44.8
39.7

129
207
72

31.6
50.7
17.7

16.86***

34
11
13

58.6
19.0
22.4

356
30
22

87.3
7.4
5.4

32.35***

13
45

22.4
77.6

251
157

61.5
38.5

31.62***

Variable

Age: mean (SD)

Gender
Female
Male
Attention or behaviour problems at school
No
Yes
Missing
Aggression
Low
Moderate
High
Anxiousness
Low
Moderate
High
Parental alcohol problems
No
Yes
Missing
Parental basic education
Some high school
Less than high school
Missing
Age at initiation of daily smoking
Non-smoker
>17 years
1517 years
<15 years
Age at initiation of drinking to intoxication
>17 years or never
1517 years
<15 years
Learning difficulties at school
No
Yes
Missing
Basic education
High school degree
Less than high school

29.3

(3.7)

16
42

27.6
72.4

25
20
13

27.9

SD: standard deviation; df: degrees of freedom. *P < 0.05, **P < 0.01, ***P < 0.001.

these variables were not carried over to the next phase of

analysis.
Multiple regression models were then used to study
the association of a SUD diagnosis with the cognitive
measures, adjusting for selected SUD risk factors and
covariates (Table 5). SUD diagnosis had an independent
association with poorer performance on digit symbol, but
not on vocabulary, digit span forward, letternumber
sequencing, TMT part A and CVLT total learning. Statistically significant predictors of lower vocabulary score

were male gender and both parental and own low basic
education, whereas older age at testing was related to
better vocabulary score. Besides SUD diagnosis, poorer
performance on digit symbol was related to male gender,
learning difficulties at school and both parental and own
low basic education. Low education also predicted poorer
performance on all the other cognitive measures,
whereas low parental education and learning difficulties
at school predicted poorer performance on letternumber
sequencing. Besides vocabulary and digit symbol, male

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568

Cognition in substance use disorders

1563

Table 3 Means and 95% confidence intervals (CIs) of cognitive measures in substance use disorder (SUD) and no-SUD groups, and
regression coefficients of cognitive measures on SUD diagnosis.

Cognitive measure
Verbal intellectual ability
WAIS-R: vocabulary (a)
Psychomotor processing speed
WAIS-R: digit symbol (a)
Verbal working memory
WMS-R: digit span forward (a)
WMS-R: digit span backward (a)
WAIS-III: letternumber sequencing (b)
Visual working memory
WMS-R: visual span forward (c)
WMS-R: visual span backward (d)
Executive function
TMT: part A (a)
TMT: part B (e)
TMT: part Bpart A (e)
Verbal learning
CVLT: total learning, trials 15 (c)
CVLT: free recall, short delay (c)
CVLT: free recall, long delay (c)

SUD (n = 58)

No SUD (n = 408)

Meana

(95% CI)

Meana

(95% CI)

41.3

(39.043.6)

45.1

(44.046.2)

0.035

-0.32

0.022

54.8

(52.057.7)

63.9

(62.565.3)

<0.001

-0.65

<0.001

7.1
6.5
10.1

(6.67.6)
(6.16.8)
(9.410.8)

7.6
6.8
10.9

(7.47.9)
(6.67.0)
(10.511.2)

0.075
0.114
0.043

-0.29
-0.20
-0.32

0.071
0.149
0.078

9.1
8.9

(8.69.6)
(8.49.4)

9.4
9.0

(9.19.6)
(8.89.2)

0.401
0.896

-0.12
-0.08

0.455
0.681

28.3
65.1
36.2

(24.931.7)
(59.371.0)
(32.340.1)

24.8
59.6
35.0

(23.825.8)
(56.462.8)
(32.337.7)

0.056
0.111
0.637

0.37
-0.24
0.02

0.081
0.139
0.887

50.7
11.5
12.0

(48.053.3)
(10.912.2)
(11.412.8)

54.9
12.3
12.7

(53.855.9)
(12.012.6)
(12.413.1)

0.004
0.025
0.075

-0.29
-0.23
-0.17

0.058
0.122
0.295

bb

TMT: Trail Making Test; CVLT: California Verbal Learning Test; WMSR: Wechsler Memory ScaleRevised; WAIS: Wechsler Adult Intelligence Scale.
a
Means and 95% CIs are estimated without covariates. bRegression coefficient of cognitive measures (standardized variables) on SUD diagnosis adjusting
for age and gender. For regression models, the following variables were transformed to approximate normality: TMT part A and part Bpart A
(logarithmic transformation), TMT part B (1/square root transformation), CVLT measures (square transformation). Among the no-SUD group, data were
missing for (a) two participants, (b) seven participants, (c) three participants, (d) four participants and (e) 20 participants. Estimated using survey settings
and weights.

gender was related to poorer performance on CVLT total

learning.
Vocabulary and digit symbol performance:
SUD characteristics
Restricting analyses to the SUD group (n = 58), we
assessed the effects of diagnosis type (abuse versus dependence), current disorder (current versus in remission),
early onset of SUD (age at SUD onset: 18 years versus
19 years), number of life-time SUD diagnoses (1 versus
at least 2) and comorbid Axis I disorder, and personality
disorder on vocabulary and digit symbol performance. In
vocabulary, participants with a substance abuse diagnosis performed poorer than those with a substance dependence diagnosis (38.8 versus 45.0, t = -2.49, df = 56,
P < 0.05), whereas current phase of the disorder, age at
SUD onset, number of life-time SUD diagnoses, comorbid
Axis I disorder or personality disorder were not related to
vocabulary score. None of the SUD characteristics or
comorbid disorders were related to performance on digit
symbol. Of the covariates, comorbid psychiatric diagnoses and SUD risk factors, abuse and dependence groups
differed only in aggression, with the dependence group
having higher aggression scores (15.7 versus 19.7,

t = -2.54, df = 56, P < 0.05). Adjusting for aggression

did not affect the results on cognitive functioning. Compared to people with substance abuse, those with substance dependence also had lower SOFAS scores (77.2
versus 68.1, t = 2.46, df = 56, P < 0.05) and almost statistically significantly lower GAF scores (73.2 versus
65.9, t = 1.95, df = 56, P = 0.056). Adjusting for these
measures did not affect the cognitive functioning results
either.

DISCUSSION
Using a representative sample of young Finnish adults,
and a comprehensive diagnostic and neuropsychological
assessment, we found poorer verbal intellectual ability, as
assessed with the WAIS-R vocabulary subtest, and less
efficient psychomotor processing, as assessed with the
WAIS-R digit symbol subtest, to be associated with SUD
diagnosis. Only borderline associations (P < 0.10) with
executive function, verbal working memory and verbal
learning processes were observed, whereas the functioning of visual working memory showed no association
with SUD. Further analyses suggested that the association with verbal intellectual ability was accounted for and

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568

 2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

-0.02
-0.23
-0.67***
-0.65***

-0.17
-0.23
-0.42*
-0.89***

(-0.27; 0.23)
0.02
(-0.54; 0.08)
0.00
(-0.97; -0.37) -0.47*
(-0.87; -0.44) -0.47***

(-0.60; 0.01) -0.29*

(-0.60; -0.04) -0.39**
(-0.81; -0.14) -0.17

0.07
0.23

(-0.25; 0.29)
(-0.30; 0.31)
(-0.84; -0.10)
(-0.67; -0.27)

-0.05
-0.02
-0.51*
-0.49***

(-0.35; 0.26) -0.14

(-0.37; 0.33)
0.15
(-0.90; -0.12) 0.23
(-0.74; -0.25) 0.42***

0.21
0.22
0.23

(-0.26; 0.46)
(-0.46; 0.10)
(-0.36; 0.33)

-0.05
0.10

(-0.57; -0.00) 0.10

(-0.68; -0.10) -0.18
(-0.49; 0.15) -0.02

(-0.11; 0.52)
(-0.40; 0.31)

(-0.13; 0.36)
(-0.61; 0.19)

(95% CI)

(-0.07; 0.51) -0.03

(-0.50; 0.51)
0.30
(-0.67; -0.04) 0.35*
(-0.70; -0.26) 0.28*

0.21
-0.04

0.12
-0.21

-0.01
-0.16
-0.12
-0.34**

(-0.41; 0.13) 0.09

(-0.15; 0.46) -0.04
(-0.06; 0.51) -0.36*
(0.20; 0.63) -0.53***

(-0.11; 0.53) 0.21

(-0.06; 0.50) -0.12
(-0.08; 0.54) -0.05

(-0.28; 0.22)
(-0.25; 0.86)
(0.07; 0.62)
(0.01; 0.56)

(-0.37; 0.28) -0.06

(-0.26; 0.45) -0.17

(-0.14; 0.33)
(-0.35; 0.26)
(-0.70; -0.02)
(-0.75; -0.31)

(-0.08; 0.51)
(-0.36; 0.13)
(-0.36; 0.26)

(-0.26; 0.23)
(-0.63; 0.31)
(-0.38; 0.20)
(-0.60; -0.09)

(-0.33; 0.21)
(-0.47; 0.14)

(-0.17; 0.28)
(-0.67; -0.01)

(95% CI)

CVLT: totalb

(-0.15; 0.28) 0.06

(-0.16; 0.63) -0.34*

(95% CI)

TMT: part Aa

0.22
(-0.09; 0.49)
(-0.66; 0.47)
0.03
(-0.34; 0.18) -0.35*
(-0.49; -0.07) -0.48***

(-0.05; 0.50)
(-0.55; 0.13)

(-0.22; 0.22)
(-0.69; 0.00)

(95% CI)

Letternumber

CI: confidence interval; TMT: Trail Making Test; CVLT: California Verbal Learning Test. aLogarithmic transformation; bsquare root transformation; cother than SUD. (a) Compared to low; (b) less than high school, compared to some
high school; (c) compared to non-smokers; (d) compared to >17 years or never; (e) less than high school, compared to high school degree. *P < 0.05, **P < 0.01, ***P < 0.001. Estimated using survey settings and weights.

(-0.40; 0.06)
(-0.52; 0.06)
(-0.73; -0.10)
(-1.10; -0.67)

(-0.41; 0.12) -0.29

(-0.69; -0.17) -0.32*
(-0.64; -0.08) -0.48**

-0.15
-0.43**
-0.36*

(-0.29; 0.27)
0.20
(-0.66; 0.25) -0.09
(-0.67; -0.15) -0.08
(-0.87; -0.37) -0.28*

-0.01
-0.21
-0.41**
-0.62***

(-0.07; 0.47)
(-0.66; 0.24)
(-0.48; 0.08)
(-0.80; -0.30)

(-0.47; 0.11)
0.23
(-0.65; -0.03) -0.21

0.20
-0.21
-0.20
-0.55***

(-0.36; 0.12)
0.00
(-0.80; -0.21) -0.34

(95% CI)

Digit span forward

-0.18
-0.34*

(-0.10; 0.32) -0.12

(-0.81; -0.15) -0.51**

(-0.41; 0.20)
(-0.69; 0.02)

0.11
-0.48**

Axis I disorderc
Attention or behaviour problems
at school
Aggression (a)
Moderate
High
Anxiousness (a)
Moderate
High
Parental alcohol problems
Parental basic education (b)
Age at initiation of daily smoking (c)
>17 years
1517 years
<15 years
Age at initiation of drinking to
intoxication (d)
1517 years
<15 years
Learning difficulties at school
Low education (e)

(95% CI)

Digit symbol

-0.11
-0.33

Risk factor

Vocabulary

Table 4 Separate regression models of cognitive measures (standardized variables) on substance use disorder (SUD) risk factors, adjusting for age and gender.

1564
Antti Latvala et al.

Addiction, 104, 15581568

 2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

-0.22
-0.25
0.01
-0.36
-0.43***
0.12

0.00
-0.08
0.29
-0.15

(-0.21; 0.49)
(-0.29; 0.28)
(-0.16; 0.54)
(-0.95; -0.05)
(-0.72; -0.16)

0.20
0.11
0.07
0.11
0.34**
0.08

(0.05; 0.65)
-0.16
(-0.07; 0.67) -0.20
(-0.51; 0.11)
0.25
(-0.61; -0.14) 0.17

(-0.03; 0.03)
(-0.27; 0.20)
(-0.57; 0.15)
(-0.10; 0.75)

(95% CI)

(95% CI)

CVLT: totalb

(-0.11; 0.51) 0.28

(-0.18; 0.40) 0.05
(-0.23; 0.37) 0.18
(-0.22; 0.44) -0.27
(0.09; 0.59) -0.50**
0.19

(-0.45; 0.14) 0.08

(-0.52; 0.13) 0.11
(-0.03; 0.53) 0.05
(-0.10; 0.44) -0.21

(-0.04; 0.61)
(-0.20; 0.31)
(-0.15; 0.50)
(-0.69; 0.16)
(-0.78; -0.22)

(-0.17; 0.33)
(-0.18; 0.39)
(-0.20; 0.30)
(-0.49; 0.06)

(-0.03; 0.03) 0.00

(-0.03; 0.03)
(-0.30; 0.15) -0.52*** (-0.73; -0.31)
(-0.16; 0.74) -0.11
(-0.45; 0.22)
(-0.60; 0.30) -0.02
(-0.41; 0.37)

(95% CI)

TMT: part Aa

CI: confidence interval; TMT: Trail Making Test; CVLT: California Verbal Learning Test. aLogarithmic transformation; bsquare root transformation. (a) Compared to female; (b) compared to low; (c) less than high school, compared to
some high school; (d) compared to non-smokers; (e) less than high school, compared to high school degree. *P < 0.05, **P < 0.01, ***P < 0.001. Estimated using survey settings and weights.

(-0.50; 0.07)
(-0.40; 0.15)
(-0.47; 0.17)
(-0.90; -0.19)
(-0.66; -0.15)

0.35*
0.30
-0.20
-0.38**

0.00
-0.03
-0.21
0.32

Letter-number

(-0.52; 0.09)
0.14
(-0.54; 0.05)
0.00
(-0.29; 0.31)
0.19
(-0.71; 0.00) -0.50*
(-0.66; -0.21) -0.44**
0.13

-0.21
-0.13
-0.15
-0.55**
-0.41**
0.29

-0.06
-0.20
-0.09
-0.22
-0.80***
0.28
(-0.32; 0.19)
(-0.47; 0.08)
(-0.34; 0.17)
(-0.52; 0.07)
(-1.06; -0.53)

(0.08; 0.63)
(-0.29; 0.41)
(-0.16; 0.33)
(-0.37; 0.08)

(-0.25; 0.22)
0.36*
0.06
(-0.11; 0.46)
(-0.48; 0.04)
0.09
(-0.74; -0.22) -0.15

(-0.14; 0.36) -0.02

(-0.24; 0.38)
0.18
(-0.21; 0.29) -0.22
(-0.57; -0.09) -0.48***

0.11
0.07
0.04
-0.33**

(95% CI)
(-0.02; 0.03)
(-0.03; 0.39)
(-0.47; 0.28)
(-0.21; 0.50)

(0.01; 0.06)
0.00
(-0.03; 0.03)
0.00
(-0.41; -0.04) -0.57*** (-0.76; -0.37) 0.18
(-0.29; 0.29) -0.34*
(-0.64; -0.04) -0.10
(-0.33; 0.39)
0.15
(-0.18; 0.48)
0.15

(95% CI)

0.04**
-0.23*
0.00
0.03

Age
Gender: male (a)
Any substance abuse/dependence
Attention or behaviour problems
at school
Aggression (b)
Moderate
High
Parental alcohol problems
Parental basic education (c)
Age at initiation of daily smoking (d)
>17 years
1517 years
<15 years
Learning difficulties at school
Basic education (e)
R2

(95% CI)

Digit span forward

Digit symbol

Variable

Vocabulary

Table 5 Multiple regression models of cognitive measures (standardized variables) on substance use disorder (SUD) diagnosis and risk factors, adjusting for age and gender.

Cognition in substance use disorders

1565

Addiction, 104, 15581568

1566

Antti Latvala et al.

possibly mediated by parental and own low basic education. In contrast, the association with psychomotor processing speed remained statistically significant after
adjusting for low basic education and other SUD risk
factors. Among participants with SUD, the diagnosis of
substance abuse, rather than dependence, was found to
be associated with lower verbal intellectual ability,
whereas type of diagnosis and other SUD characteristics,
comorbid Axis I disorders and personality disorders were
not related to psychomotor processing speed.
Several studies have found lower verbal intellectual
ability in relation to alcohol and other substance disorders [1,6]. Our results suggest that low basic education,
a common risk factor for SUD, is one factor accounting
for this association. It is well known that education contributes strongly to performance in tests assessing verbal
intellectual ability [29], and previous studies have found
education to also correlate with verbal measures such as
vocabulary and abstraction in people with SUDs [1].
Genetically informative studies have suggested that
genetic factors explain a significant proportion of the
variance of both intellectual ability and educational
achievement, and there are also shared genetic factors
influencing both intellectual and educational phenomena [30]. Interestingly, both parental and own low basic
education associated with poorer verbal intellectual
ability and with SUD in the present study. Both genetic
and environmental factors contribute to the association
between parental education and verbal intellectual
ability in the offspring [31], but the relationships
between these phenomena and SUDs are not well
known.
Similar to cognitive ability and educational achievement, individual differences in SUD liability are to a large
extent influenced by genetic factors, with heritability estimates ranging from 30% to 70% [32]. In the present
study, parental alcohol problems were associated with
both SUD and cognitive performance in young adulthood,
possibly reflecting both genetic and environmental
transmission.
Neural correlates of verbal intellectual ability are not
well known, but general intelligence is believed to reflect
the functioning and interconnections of a widely distributed network of brain areas [33]. A recent magnetic
resonance imaging study suggested that lower verbal
intellectual ability is more likely to predate heavy substance use than reflect impairment caused by use. Schottenbauer et al. [34] found that in alcohol-dependent
patients, the WAIS-R vocabulary score was not related to
brain shrinkage caused by alcohol use and ageing, but
correlated with pre-morbid brain size as measured by
intracranial volume. In contrast, performance on a
measure of fluid intelligence was predicted by brain
shrinkage, as measured by the ratio of cerebrum and

intracranial volumes [34]. This finding, in conjunction

with studies of intellectual abilities predicting later substance use and problems [8,9], suggests that our finding
of poorer verbal intellectual ability in young adults with a
life-time SUD diagnosis may reflect cognitive differences
predating SUD. This interpretation is supported by the
finding that the age of SUD onset and number of life-time
SUD diagnoses, which are likely to be related to heaviness
of substance use during the life-time, were not associated
with verbal intellectual ability.
The finding that poorer verbal intellectual ability was
associated with substance abuse rather than dependence
was unexpected, as dependence is generally considered a
more severe form of SUD than abuse. The validity of the
DSM-IV diagnostic criteria for alcohol abuse has often
been questioned [15], but some support for the diagnostic
utility of alcohol abuse has been found in prospective
studies showing different developmental courses for
abuse and dependence [35]. The present findings of
poorer verbal intellectual ability, but not processing
speed, in substance abuse compared to dependence
support the division of abuse and dependence as separate
diagnostic categories.
Less efficient psychomotor processing in SUDs has also
been reported previously [1]. We found that parental and
own low basic education were associated strongly with
slower psychomotor processing, but the association with
SUD remained statistically significant after adjustment
for these and other risk factors. One study found less efficient psychomotor processing to correlate with greater
estimated life-time alcohol consumption but not with
years of life-time alcohol dependence or length of sobriety [36]. Beatty et al. also failed to find influences of SUD
chronicity on digit symbol performance [1]. Interestingly,
we found a relationship between psychomotor slowing
and SUDs among young adults with a relatively short
history of these disorders. On the other hand, white
matter atrophy in alcohol dependence reflects differences
in exposure to alcohol [37], and there is evidence that the
digit symbol task is related to volume and microstructure
of white matter pathways [38].
We found no associations between psychomotor processing and characteristics of SUD: type or current phase
of the disorder, age of onset and number of life-time SUD
diagnoses. These findings suggest that less efficient psychomotor processing in young adults with SUD might not
be related to severity of the disorder. Interestingly, deficits
in performance on digit symbol and similar coding tasks
are often observed also in other psychiatric disorders
besides SUDs, particularly in schizophrenia and bipolar
disorder [39,40]. Unfortunately, estimates of life-time
consumption of alcohol and other substances were not
available in the present study, making it difficult to interpret the findings on psychomotor processing. It should

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568

Cognition in substance use disorders

also be noted that the statistical power to detect differences in the subsample of participants with a SUD diagnosis (n = 58) was limited.
Using a subsample of the present MEAF data, Castaneda et al. found no evidence of cognitive dysfunction
in unipolar depressive disorders among young adults
[41]. In a similar vein, our analyses suggested no association between cognitive functioning and life-time Axis I
disorders besides SUDs. In addition, among people with
SUD neither comorbid Axis I disorders nor personality
disorders were associated with verbal intellectual ability
or psychomotor processing speed. The results concerning
personality disorders must be interpreted with caution,
however, because a diagnostic interview for personality
disorders was not conducted, and the diagnostic assessment was thus based mainly on medical case records,
probably resulting in diagnosing only people with severe
personality disorders [18].
The present negative findings on other cognitive
domains besides verbal ability and psychomotor processing speed might reflect the fact that our sample was
comprised of young adults, and the possible deficits in
executive function and working memory related to heavy
use might not have thus had enough time to develop. On
the other hand, our representative sample and exceptionally comprehensive diagnostic assessment probably
resulted in improved validity of the present cognitive
results in this age group.
Some limitations of the present investigation are
noteworthy. First, as the study setting was crosssectional, causal and temporal relations between the
studied phenomena could not be firmly established.
Optimally, prospective longitudinal assessments of substance use and cognitive developmental trajectories,
starting in childhood, would help to shed light on causal
relations between these phenomena. Secondly, measures
of SUD risk factors were self-reported. However, these
measures come from a general health survey, not profiled as focusing specifically on SUDs, which should
serve to reduce reporting bias. Thirdly, TMT was the
only executive function measure included in our neuropsychological test battery. Finally, attrition might have
affected the results, as people with the most severe SUDs
may have been less likely to participate. Due to the twophase study design, there were non-respondents in both
of the study phases. However, non-response was not
related to self-reported mental health or alcohol use
problems [18]. To correct statistically for non-response,
post-stratification and expansion weights were used.
These limitations notwithstanding, the present results
on a representative sample highlight poorer verbal intellectual function and less efficient psychomotor processing related to life-time alcohol and other substance
disorders in young adults.

1567

Declarations of interest
None.
Acknowledgements
This study was supported by grants from the Academy
of Finland (Dr Suvisaari, grant 210714; Dr TuulioHenriksson, grant 117159; Dr Kaprio, grant 118555),
the Academy of Finland Centre of Excellence in Complex
Disease Genetics (Dr Kaprio), the Yrj Jahnsson Foundation (Dr Suvisaari, Dr Aalto-Setl) and the Jalmari and
Rauha Ahokas Foundation (Dr Aalto-Setl). We thank
Merja Blom, Margit Keinnen-Guillaume, Helena Kurru,
Maija Lindgren MPsych, Taina Laajasalo PhD, Marko
Manninen MPsych, Tuula Mononen and Sebastian
Therman MPsych, for skilfully conducted interviews, and
Tuula Koski, Kirsi Niinist and Satu Vierti MSc, for
administrative work. Mark Phillips BA, is thanked for
revising the language. We also thank Arpo Aromaa MD,
PhD and all the other collaborators in the Health 2000
study group, as well as all participants.
References
1. Beatty W. W., Tivis R., Stott H. D., Nixon S. J., Parsons O. A.
Neuropsychological deficits in sober alcoholics: influences of
chronicity and recent alcohol consumption. Alcohol Clin Exp
Res 2000; 24: 14954.
2. Parsons O. A. Neurocognitive deficits in alcoholics and
social drinkers: a continuum? Alcohol Clin Exp Res 1998;
22: 95461.
3. Goldstein R. Z., Volkow N. D. Drug addiction and its underlying neurobiological basis: neuroimaging evidence for the
involvement of the frontal cortex. Am J Psychiatry 2002;
159: 164252.
4. Oscar-Berman M., Marinkovic K. Alcohol: effects on
neurobehavioral functions and the brain. Neuropsychol Rev
2007; 17: 23957.
5. Rogers R. D., Robbins T. W. Investigating the neurocognitive
deficits associated with chronic drug misuse. Curr Opin
Neurobiol 2001; 11: 25057.
6. Tarter R. E., Mezzich A. C., Hsieh Y. C., Parks S. M. Cognitive
capacity in female adolescent substance abusers. Drug
Alcohol Depend 1995; 39: 1521.
7. Tarter R. E., Kirisci L., Mezzich A., Cornelius J. R., Pajer K.,
Vanyukov M. et al. Neurobehavioral disinhibition in childhood predicts early age at onset of substance use disorder.
Am J Psychiatry 2003; 160: 107885.
8. Windle M., Blane H. T. Cognitive ability and drinking behavior in a national sample of young adults. Alcohol Clin Exp
Res 1989; 13: 438.
9. Fergusson D. M., Horwood L. J., Ridder E. M. Show me the
child at seven II: childhood intelligence and later outcomes
in adolescence and young adulthood. J Child Psychol Psychiatry 2005; 46: 8508.
10. Pirkola S. P., Isomets E., Suvisaari J., Aro H., Joukamaa M.,
Poikolainen K. et al. DSM-IV mood-, anxiety- and alcohol
use disorders and their comorbidity in the Finnish general
population: results from the Health 2000 Study. Soc Psychiatry Psychiatr Epidemiol 2005; 40: 110.

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568

1568

Antti Latvala et al.

11. Sihvola E., Rose R. J., Dick D. M., Pulkkinen L., Marttunen
M., Kaprio J. Early-onset depressive disorders predict the use
of addictive substances in adolescence: a prospective study
of adolescent Finnish twins. Addiction 2008; 103: 2045
53.
12. Hasin D. S., Stinson F. S., Ogburn E., Grant B. F. Prevalence,
correlates, disability, and comorbidity of DSM-IV alcohol
abuse and dependence in the United States: results from the
National Epidemiologic Survey on Alcohol and Related
Conditions. Arch Gen Psychiatry 2007; 64: 83042.
13. Elkins I. J., McGue M., Iacono W. G. Prospective effects of
attention-deficit/hyperactivity disorder, conduct disorder,
and sex on adolescent substance use and abuse. Arch Gen
Psychiatry 2007; 64: 114552.
14. Plomin R., Price T. S., Eley T. C., Dale P. S., Stevenson J.
Associations between behaviour problems and verbal and
nonverbal cognitive abilities and disabilities in early
childhood. J Child Psychol Psychiatry 2002; 43: 619
33.
15. Grant B. F., Harford T. C., Muthen B. O., Yi H. Y., Hasin D. S.,
Stinson F. S. DSM-IV alcohol dependence and abuse: further
evidence of validity in the general population. Drug Alcohol
Depend 2007; 86: 15466.
16. Perkonigg A., Pfister H., Hofler M., Frohlich C., Zimmermann P., Lieb R. et al. Substance use and substance use
disorders in a community sample of adolescents and young
adults: incidence, age effects and patterns of use. Eur Addict
Res 2006; 12: 18796.
17. Waber D. P., De Moor C., Forbes P. W., Almli C. R., Botteron
K. N., Leonard G. et al. The NTH MRI study of normal brain
development: performance of a population based sample of
healthy children aged 618 years on a neuropsychological
battery. J Int Neuropsychol Soc 2007; 13: 72946.
18. Suvisaari J., Aalto-Setl T., Tuulio-Henriksson A.,
Hrknen T., Saarni S. I., Perl J. et al. Mental disorders in
young adulthood. Psychol Med 2009; 39: 28799.
19. Mayfield D., McLeod G., Hall P. The CAGE questionnaire:
validation of a new alcoholism screening instrument. Am J
Psychiatry 1974; 131: 11213.
20. First M. B., Spitzer R. L., Gibbon M., Williams J. B. W. Structured Clinical Interview for DSM-IV-TR Axis I Disorders
Patient Edition (SCID I/P, 2/2001 Revision). New York:
Biometrics Research, New York State Psychiatric Institute;
2001.
21. Wechsler D. Wechsler Adult Intelligence Scale, Revised. San
Antonio, TX: The Psychological Corporation; 1981.
22. Reitan R. M., Wolfson D. The HalsteadReitan Neuropsychological Test Battery: Theory and Clinical Interpretation. Tuscon,
AZ: Neuropsychology Press; 1993.
23. Wechsler D. Wechsler Memory Scale, Revised. San Antonio,
TX: The Psychological Corporation; 1987.
24. Wechsler D. Wechsler Adult Intelligence Scale, 3rd edn. San
Antonio, TX: The Psychological Corporation; 1997.
25. Delis D. C., Kramer J. H., Kaplan E., Ober B. A. California
Verbal Learning Test: Adult Version. San Antonio, TX: The
Psychological Corporation; 1987.
26. Buss A. H., Perry M. The Aggression Questionnaire. J Pers
Soc Psychol 1992; 63: 4529.
27. StataCorp. Stata Statistical Software: Release 9. College
Station, TX: StataCorp LP; 2005.
28. Pickles A., Dunn G., Vazquez-Barquero J. L. Screening for
stratification in two-phase (two-stage) epidemiological
surveys. Stat Methods Med Res 1995; 4: 7389.
29. Lezak M. D., Howieson D. B., Loring D. W. Neuropsychological

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

Assessment, 4th edn. New York, NY: Oxford University Press;

2004.
Bartels M., Rietveld M. J. H., Van Baal G. C. M., Boomsma D.
I. Heritability of educational achievement in 12-year-olds
and the overlap with cognitive ability. Twin Res 2002; 5:
54453.
Neiss M., Rowe D. C. Parental education and childs verbal
IQ in adoptive and biological families in the National Longitudinal Study of Adolescent Health. Behav Genet 2000; 30:
48795.
Agrawal A., Lynskey M. T. Are there genetic influences on
addiction: evidence from family, adoption and twin studies.
Addiction 2008; 103: 106981.
Song M., Zhou Y., Li J., Liu Y., Tian L. X., Yu C. S. et al. Brain
spontaneous functional connectivity and intelligence. Neuroimage 2008; 41: 116876.
Schottenbauer M. A., Momenan R., Kerick M., Hommer D.
W. Relationships among aging, IQ, and intracranial volume
in alcoholics and control subjects. Neuropsychology 2007;
21: 33745.
Schuckit M. A., Smith T. L., Landi N. A. The 5-year clinical
course of high-functioning men with DSM-IV alcohol abuse
or dependence. Am J Psychiatry 2000; 157: 202835.
Sassoon S. A., Fama R., Rosenbloom M. J., OReilly A., Pfefferbaum A., Sullivan E. V. Component cognitive and motor
processes of the digit symbol test: differential deficits in alcoholism, HIV infection, and their comorbidity. Alcohol Clin
Exp Res 2007; 31: 131524.
Bjork J. M., Grant S. J., Hommer D. W. Cross-sectional volumetric analysis of brain atrophy in alcohol dependence:
effects of drinking history and comorbid substance use disorder. Am J Psychiatry 2003; 160: 203845.
Turken A. U., Whitfield-Gabrieli S., Bammer R., Baldo J. V.,
Dronkers N. F., Gabrieli J. D. E. Cognitive processing speed
and the structure of white matter pathways: convergent
evidence from normal variation and lesion studies. Neuroimage 2008; 42: 103244.
Frantom L. V., Allen D. N., Cross C. L. Neurocognitive
endophenotypes for bipolar disorder. Bipolar Disord 2008;
10: 38799.
Dickinson D., Ramsey M. E., Gold J. M. Overlooking the
obvious: a meta-analytic comparison of digit symbol coding
tasks and other cognitive measures in schizophrenia. Arch
Gen Psychiatry 2007; 64: 53242.
Castaneda A. E., Suvisaari J., Marttunen M., Perl J.,
Saarni S. I., Aalto-Setl T. et al. Cognitive functioning in a
population-based sample of young adults with a history of
non-psychotic unipolar depressive disorders without psychiatric comorbidity. J Affect Disord 2008; 110: 3645.

Supporting information
Additional Supporting Information may be found in the
online version of this article:
Table S1 Origins and contents of substance use disorder
(SUD) risk factor variables.
Please note: Wiley-Blackwell is not responsible for the
content or functionality of any supporting materials supplied by the authors. Any queries (other than missing
material) should be directed to the corresponding author
for the article.

2009 The Authors. Journal compilation 2009 Society for the Study of Addiction

Addiction, 104, 15581568