Childhood cognitive ability and its relationship withanxiety and depression in adolescence

M. Weeks1, T. C. Wild2, G. B. Ploubidis3, K. Naicker1, J. Cairney4, C. R. North2, and I. Colman1,2*

1 Department of Epidemiology and Community Medicine, University of Ottawa, Canada;

2 School of Public Health, University of Alberta, Canada;

3 London School of Hygiene and Tropical Medicine, UK;

4 Departments of Family Medicine, Psychiatry & Behavioural Neurosciences, and

Clinical Epidemiology & Biostatistics, McMaster University, Canada.

* Address for correspondence: I. Colman, PhD, Department of Epidemiology and Community Medicine, University of Ottawa, 451 Smyth Road, RGN 3230C, Ottawa, ON, K1H 8M5, Canada. (Email: icolman@uottawa.ca)

1

ABSTRACT

Background: Childhood cognitive ability may have protective effects against

internalizing symptoms in adolescence, although this may depend on the time of

symptom assessment and child gender. Also, the effects of childhood stressors on

adolescent internalizing symptoms may be moderated by childhood cognitive ability.

Methods: The sample included 4,405 individuals from the Canadian National

Longitudinal Study of Children and Youth (NLSCY). Between ages 4-5 and 10-11,

children completed a test of verbal ability and scholastic aptitude and a series of

mathematics computation tests. Internalizing symptoms were assessed via self-reports at

ages 12-13 and 14-15.

Results: Greater cognitive ability was generally associated with decreased odds of

internalizing symptoms at age 12-13. However, greater cognitive ability generally

increased, or had no effect on, the odds of internalizing symptoms at age 14-15. Some of

the effects of childhood cognitive ability varied by child gender. Also, childhood

cognitive ability attenuated the effects of family dysfunction and chronic illness

throughout childhood on subsequent internalizing symptoms.

Limitations: These data are largely subject to some degree of reporting bias, the tests of

cognitive ability are limited and may not represent overall cognitive ability, and there

may be intermediary variables that account for the relationship between childhood

cognitive ability and adolescent internalizing symptoms.

Conclusion: Results suggest that programs attempting to increase early cognitive skills

may be particularly beneficial for girls. Also, an increased focus on cognitive skills may

2

attenuate the negative effects of some stressors on subsequent anxious and depressive

symptoms, regardless of child gender.

Key words: Depression; anxiety; cognitive ability; development; epidemiology

3

1. INTRODUCTION

Feelings of anxiety and depression, or internalizing symptoms, are among the

most common psychiatric problems experienced by adolescents (Costello et al., 2005).

Although most individuals will not meet diagnostic criteria for clinical anxiety and mood

disorders, the presence of internalizing symptoms have nevertheless been shown to

increase the risk for subsequent psychiatric disorders (Pine et al., 1999, Ashford et al.,

2008). Moreover, anxious and depressive disorders have been shown to represent a

significant economic burden to society (Greenberg et al., 1999, Greenberg et al., 2003).

The identification of risk and protective factors for internalizing symptoms is therefore an

important first step in the prevention of psychiatric disorders in young people.

One possible protective factor against the development of internalizing symptoms

is childhood cognitive ability, a concept synonymous with general intelligence (Ek et al.,

2013, Woodward et al., 2012, Bruni et al., 2012, Trippas et al., 2013) and involving skills

such as problem solving, memory, and verbal ability (Vernon, 1971, Ashton et al., 1979,

DeFries et al., 1979, Sass et al., 1995). Evidence suggests that child and youth

internalizing symptoms are associated with a deficit in cognitive functioning. For

example, children with social phobia have been shown to have certain types of memory

deficit (Vasa et al., 2007), and may also have a greater amount of neurodevelopmental

delay (Kristensen and Torgersen, 2008) as compared to other children. Children with

greater internalizing symptoms also tend to have more difficulty with various cognitive

tasks such as those involving problem solving (Emerson et al., 2005). Moreover, lower

IQ has been associated with greater depression, especially in young adult men (Zammit et

al., 2004, Rabbitt et al., 1995), and children with anxiety disorders tend to have lower IQ

4

scores than non-anxious children (Hodges and Plow, 1990, Davis et al., 2008, Kristensen

and Torgersen, 2008).

These findings suggest that children experiencing higher levels of anxiety and

depression should have more difficulty in academic settings. There is some evidence of

an educational gradient in depression, possibly resulting from differences in cognitive

ability (Lee, 2011). Indeed, academic performance has been negatively associated with

symptoms of anxiety and depression in children (Cole, 1991, Lin et al., 2011, Rapport et

al., 2001). In addition, both academic competence in adolescence and high school

completion are negatively associated with symptoms of anxiety and depression in young

adulthood (Masten et al., 2005, Topitzes et al., 2009), and early academic achievement

lowers the odds of persistently high levels of depression throughout adolescence and

young adulthood (Stoolmiller et al., 2005). In contrast, other research suggests that

educational attainment in early adulthood is associated with greater internalizing

symptoms at age 53 (Hatch et al., 2007).

Most of the research examining the associations between markers of cognitive

ability and internalizing symptoms to date has been cross-sectional, and thus the direction

of effect in these associations, as well as the role of early cognitive ability on later

internalizing symptoms, remains largely unexamined. Nevertheless, a small number of

longitudinal studies suggest that cognitive ability early in life can have a significant effect

on the later development of internalizing symptoms. For example, higher academic

achievement in grade 1 has been shown to predict lower internalizing symptoms two

years later (Burt and Roisman, 2010). Also, there is evidence that low IQ in adolescence

increases the risk of depression in middle age (Der et al., 2009, Franz et al., 2011) and

5

low IQ in childhood increases the risk of internalizing symptoms in adulthood (Koenen et

al., 2009).

However, the role of intelligence in the development of subsequent internalizing

symptoms may depend on a number of factors, including the age at which these

symptoms are assessed and participant gender. For instance, one study found that IQ at

age 8 was inversely related to depressive symptoms at age 11, but positively related to

these symptoms at ages 13 and 14 (Glaser et al., 2011), while another study found that

childhood cognitive ability was only related to fewer adult internalizing symptoms for

women (Hatch et al., 2007).

Although childhood cognitive ability may be a protective factor against

subsequent internalizing symptoms, it is not clear why this would be the case. One

possibility is that children with greater cognitive ability are better equipped to cope with

various sources of stress that may be important in the development and maintenance of

internalizing symptoms. That is, cognitive ability may buffer children and youth against

the effects of stress on anxiety and depression. However, support for this idea is mixed.

One study found that stressful life events and depression in adulthood were positively

associated only among those participants with lower cognitive ability at age 15 (Van Os

and Jones, 1999). Another a study of high-risk adolescents also failed to find evidence

that greater intelligence gives adolescents an advantage in coping with stress (Luthar,

1991).

The first objective of the current study was to examine the relationship between

cognitive ability in childhood and internalizing symptoms in adolescence. We

hypothesized that cognitive ability in childhood would generally be inversely associated

6

with internalizing symptoms in adolescence. However, we acknowledged the possibility

that the effect of cognitive ability depends on the time of assessment of internalizing

symptoms, with more protective effects in younger adolescents than older adolescents.

The second objective was to examine gender differences in the effect of childhood

cognitive ability on adolescent internalizing symptoms. We hypothesized that girls may

benefit more from greater cognitive ability than boys. The third objective was to examine

the potentially moderating role of childhood cognitive ability on the relationships

between childhood stressors and internalizing symptoms in adolescence. Cognitive ability

was expected to attenuate the influence of childhood stressors on adolescent internalizing

symptoms.

2. METHODS

2.1. Study sample

The National Longitudinal Study of Children and Youth (NLSCY) is a nationally

representative prospective cohort study of Canadians that is managed by Statistics

Canada (Canada, 2009). Data collection for the longitudinal sample began in 1994/1995

and continued every two years until 2008/2009. Drawing from 6,908 individuals who

were age 2-3 years in Cycle 1 (1994/1995) or Cycle 2 (1996/1997) of the NLSCY, our

sample included 4,405 individuals with information on cognitive ability in childhood (age

4-5, 6-7, 8-9, and/or 10-11) and internalizing symptoms in adolescence (age 12-13 and/or

14-15). Figure 1 displays the longitudinal response pattern, as well as the level of

response on the main study predictors and outcomes from age 4-5 to age 14-15.

--- Insert Figure 1 ---

7

2.2. Assessments

Childhood cognitive ability. All children aged 4-5 completed the Peabody

Picture Vocabulary Test- Revised (PPVT-R) (Dunn and Dunn, 2006), which assesses

verbal ability and scholastic aptitude. All children in Grade 2 or above completed a

Mathematics Computation Test (MCT), a shortened version of the MCT component of

the Canadian Achievement Tests, Second Edition (CAT/2), which measures

understanding of addition, subtraction, multiplication, and division of whole numbers

(Statistics Canada and Human Resources and Skills Development Canada, 1998). We

thus obtained MCT scores for ages 6-7, 8-9, and 10-11. PPVT-R and MCT scores in the

NLSCY are treated as continuous variables and are standardized based on a norm

samples from across Canada (Statistics Canada and Human Resources and Skills

Development Canada, 1998).

Adolescent internalizing symptoms. Internalizing symptoms were assessed

using a questionnaire that included 7 items from the Ontario Child Health Study (OCHS-

R) (Boyle et al., 1987), assessing symptoms of anxiety and depression. The items were

derived from the Child Behavior Checklist (CBCL) (Achenbach and Edelbrock, 1981)

because they appeared to operationalize DSM-III criteria for emotional disorder

(Statistics Canada and Human Resources and Skills Development Canada, 1998). The

CBCL subscales are derived from DSM-III and DSM-IV criteria (Boyle et al., 1993b)

and have been extensively evaluated in terms of their psychometric properties (Boyle et

al., 1987, Boyle et al., 1993a). The statements were: “I am unhappy or sad,” “I am not as

happy as other people my age,” “I am too fearful or nervous,” “I worry a lot,” “I cry a

lot,” “I am nervous, high-strung, or tense,” and “I have trouble enjoying myself.”

Most at age 6-7 were not in Grade 2 or above and were thus did not complete the MCT (See Figure 1).

8

Response options were: “never or not true,” “sometimes or somewhat true,” or “often or

very true.” The scale has shown good reliability, with a Cronbach alpha of 0.78 reported

for 12-15-year-olds (Statistics Canada, 1992).

In the current study, total scores were grouped into 4 categories representing

degrees of severity to account for floor effects and variation in severity of symptoms (i.e.,

positively skewed distributions), following previously reported procedures (Colman et

al., 2012, Naicker et al., 2012): no symptoms (scores below the 50th percentile), low

symptoms (scores between the 51st and 75th percentile), moderate symptoms (scores

between the 76th and 90th percentile), and severe symptoms (scores above the 90th

percentile).

Covariates. Socioeconomic status (SES) is operationalized in the NLSCY using

household income adjusted for the ‘low-income cut off’ score (i.e., SES/LICO). This

measure takes into account an individual’s income relative to the community and the size

of their family (Cotton, 2001). We examined SES at child age 2-3. This variable was

dichotomized, with scores in the bottom 10th percentile representing ‘low’ SES scores.

Maternal depression was measured with a shortened version of the self-reported

Center for Epidemiologic Studies Depression Scale (CES-D) (Radloff, 1977), which

measures the frequency and severity of symptoms of depression over the past week

through 12 statements (e.g., “I felt depressed”, “I had crying spells”). We examined

maternal depression at child age 2-3. This variable was dichotomized, with a score in the

top 10th percentile representing the presence of maternal depression.

Parents rated a series of 12 statements addressing various aspects of family

functioning, such as problem solving (e.g., “we are able to make decisions about how to

9

solve problems”) and affective involvement (e.g., “we express feelings to each other”).

Items are rated on a scale from 0 (“strongly agree”) to 3 (“strongly disagree”), with

scores ranging from 0-36 and higher scores indicating a greater degree of family

dysfunction. We examined family dysfunction at child ages 4-5, 6-7, 8-9, and 10-11. This

variable was dichotomized, with scores in the top 10th percentile representing the

presence of family dysfunction.

Parents reported on the presence of any chronic illness in their child (e.g.,

diabetes) as well as the presence of 15 stressful life events (SLEs; e.g., “death of a family

member”) that children had experienced in the previous 2 years. We examined chronic

illness SLEs in the previous 2 years as measured at child ages 4-5, 6-7, 8-9, and 10-11.

We dichotomized chronic illness into ‘yes’ (i.e., presence of one or more chronic illness)

and ‘no chronic illness’. We also dichotomized SLEs, where more than one SLE in the

previous 2 years was considered to represent a ‘high’ degree of stressful life events.

2.2. Statistical analysis

A series of multinomial logistic regressions was used to examine the relationship

between cognitive ability in childhood (PPVT-R at age 4-5, and MCT scores at ages 6-7,

8-9, and 10-11) and severity of internalizing symptoms in adolescence (ages 12-13 and

14-15). Odds ratios (OR) were obtained for unadjusted models, as well as controlling for

child gender, SES, family dysfunction, SLEs, maternal depression. We also used

multinomial logistic regression to examine gender differences, by examining the

interaction between child gender and cognitive ability in childhood in the prediction of We examined the change in odds associated with a one standard deviation increase in PPVT-R and MCT scores, with ‘no symptoms’ as the reference category. In these analyses, family dysfunction and SLEs were measured at the same age as the predictor, whereas SES and maternal depression were measured at age 2-3.

10

adolescent internalizing symptoms while controlling for SES. Lastly, multinomial logistic

regression was used to examine the moderating role of childhood cognitive ability in the

relationship between childhood stress (chronic illness, family dysfunction, SLEs) and

adolescent internalizing symptoms, while controlling for child gender and SES.

All analyses were performed using SAS 9.3 (SAS Institute Inc, Cary, North

Carolina, USA). Survey responses were weighted using the appropriate longitudinal

sampling weights provided by Statistics Canada. These weights are based on the inverse

of each child’s probability of selection into the survey, and are adjusted to account for

non-response at each cycle (Statistics Canada and Human Resources and Skills

Development Canada, 1998).

3. RESULTS

3.1. Overall regression analyses Results from the multinomial regression analyses are shown in Table 2. Overall,

cognitive ability appeared to be negatively associated with internalizing problems

measured at age 12-13, most consistently so for severe symptoms. Higher PPVT-R scores

at age 4-5 decreased the odds of moderate (OR=0.83; CI: 0.75, 0.93) and severe

(OR=0.80; CI: 0.70, 0.92) internalizing symptoms while controlling for covariates.

Higher MCT scores at age 6-7 decreased the odds of severe symptoms in the unadjusted

model (OR=0.74; CI: 0.58, 0.95) and increased the odds of mild symptoms after

controlling for covariates (OR=1.21; CI: 1.05, 1.39). After controlling for covariates,

higher MCT scores at age 8-9 decreased the odds of moderate (OR=0.83; CI: 0.75, 0.93)

11

and severe (OR=0.85; CI: 0.74, 0.98) symptoms, and higher MCT scores at age 10-11

decreased the odds of severe (OR=0.85; CI: 0.75, 0.98) symptoms.

In contrast, when internalizing symptoms were measured at age 14-15, childhood

cognitive ability was either not associated, or positively associated, with these symptoms.

Although higher PPVT-R scores decreased the odds of mild symptoms (OR=0.84; CI:

0.76, 0.92), they were not associated with moderate or severe symptoms. Interestingly,

higher MCT scores at age 6-7 increased the odds of mild (OR=1.44; CI: 1.21, 1.72) and

severe (OR=1.50; CI: 1.18, 1.92) internalizing symptoms, and higher MCT scores at age

8-9 increased the odds of moderate (OR=1.14; CI: 1.02, 1.29) and severe (OR=1.20; CI:

1.05, 1.37) symptoms, after controlling for covariates.

3.2. Moderating role of child gender

Several significant interactions between child gender and cognitive ability

emerged (Table 3). Higher PPVT-R scores decreased the odds of moderate internalizing

symptoms at age 12-13 for girls (OR=0.70; CI: 0.61, 0.80) but not boys (OR=1.10; CI:

0.95, 1.28), and decreased the odds of mild internalizing symptoms at age 14-15 for girls

(OR=0.76; CI: 0.67, 0.87) but not boys (OR=0.92; CI: 0.81, 1.04). Higher MCT at age 6-

7 increased the odds of mild internalizing symptoms at age 14-15 for boys (OR=1.82; CI:

1.43, 2.32) but not girls (OR=1.14; CI: 0.89, 1.46), whereas higher MCT scores at age 8-

9 increased the odds of moderate internalizing symptoms at age 14-15 for girls

(OR=1.30; CI: 1.13, 1.50) but not boys (OR=0.83; CI: 0.68, 1.02).

3.3. Moderating role of childhood stressors

12

Significant interactions between childhood cognitive ability and childhood

stressors were also observed (see Table 4). Overall, the pattern of results suggests that

greater cognitive ability attenuates the role of stress in the development of subsequent

internalizing. First, there was evidence that cognitive ability attenuates the effect of

family dysfunction on adolescent internalizing symptoms. Among the findings, at lower

PPVT-R scores, family dysfunction at age 8-9 increased the odds of severe internalizing

symptoms at age 12-13 (OR=1.81; CI: 1.21, 2.71), but had no effect on these symptoms

at higher PPVT-R scores (OR=0.82; CI: 0.45, 1.48). Similarly, at lower MCT scores,

greater odds of severe internalizing symptoms at age 14-15 were predicted by family

dysfunction at age 8-9 (OR=2.61; CI: 1.67, 4.06) and age 10-11 (OR=2.90; CI: 1.84,

4.56). At higher MCT scores, these symptoms were not predicted by family dysfunction

at age 8-9 (OR=0.53; CI: 0.26, 1.09) or age 10-11 (OR=1.36; CI: 0.87, 2.11).

There was also evidence that cognitive ability attenuated the effect of chronic

illness on adolescent internalizing symptoms. For instance, at lower MCT scores at age 6-

7, chronic illness increased the odds of severe internalizing symptoms at age14-15

(OR=2.24; CI: 2.09, 5.60), whereas this was not true at (OR=0.54; CI: 0.25, 1.19).

Lastly, there was some evidence that greater cognitive ability may actually

exacerbate the effect of recent SLEs on internalizing symptoms. At high levels of MCT

scores at age 10-11, the presence of 1 or more SLE increased the odds of moderate

internalizing symptoms at age 12-13 (OR=3.30; CI: 1.50, 7.25) but had no effect on

symptoms at low levels of MCT scores (OR=0.43; CI: 0.15, 1.19).

13

4. DISCUSSION

The primary goal of the current study was to clarify the role of childhood

cognitive ability in the development of internalizing symptoms in adolescence. Some

previous research suggests that greater cognitive ability may protect children against

subsequent internalizing difficulties (Burt and Roisman, 2010, Koenen et al., 2009).

Consistent with this previous research, we found that cognitive ability was consistently

associated with decreased odds of moderate and severe internalizing symptoms at age 12-

13, even while controlling for potential confounders. Children’s cognitive ability in

childhood therefore seems to be a protective factor against the development of

internalizing symptoms in early adolescence. Moreover, parents and educators could flag

academic underperformance as an early warning sign of future emotional difficulties in

children.

However, the protective role of cognitive ability appeared to be dependent on the

age at which internalizing symptoms were assessed. When internalizing symptoms were

measured at age 14-15, greater childhood cognitive ability had either no effect or

increased the odds of mild to severe symptoms (see Table 2). Although these findings do

not provide support for the idea of cognitive ability as a protective factor, they

nevertheless mirror recent work examining the role of IQ at age 8 on depressive

symptoms at ages 11 through 14 (Glaser et al., 2011). In fact, the apparent reversal of

effect in the current study was most evident in the case of age 8-9 MCT scores, which

lends strong support to this previous study. Although the mechanism responsible for this

effect remains unclear, the authors suggest that the reversal of effect may be tied to

hormonal changes associated with the onset of puberty.

14

An implication of these findings is that, although there seem to be important

associations between childhood cognitive ability and adolescent internalizing, the role of

cognitive ability is most likely influenced by other factors. If the reversal of effect is tied

to factors related to a period of adolescence, such as hormonal or social changes, middle

adolescence might represent a temporary period where other factors overshadow the

otherwise beneficial effects of greater cognitive ability. It may also be that the protective

impact of greater cognitive ability would return in later adolescence and adulthood, as has

been observed in studies linking higher childhood cognitive ability to lower rates of

depression in adulthood (Der et al., 2009, Franz et al., 2011, Koenen et al., 2009).

The next goal of the current study was to examine gender differences in the role

of childhood cognitive ability in predicting adolescent internalizing symptoms. Again,

these findings seem to depend on the age of assessment of internalizing symptoms, as

well as cognitive ability. Higher PPVT-R scores decreased the odds of moderate and mild

symptoms at age 12-13 and 14-15, respectively, for females only. Higher MCT scores at

age 6-7 increased the odds of mild symptoms at age 14-15 for males only. However,

higher MCT scores at age 8-9 increased the odds of moderate symptoms at age 14-15 for

females only. Thus, depending on age of assessment, cognitive ability in the current study

can be seen as having gender-specific protective effects, as well as gender-specific

detrimental effects. Although the reasons for these gender differences are not clear, they

are potentially important, especially in the context of the widely acknowledged higher

rate of depression for females (Kessler et al., 1993), and the emergence of associated

gender differences in adolescence (Wade et al., 2002). Our observed gender differences,

in conjunction with the reversal of the effect of cognitive ability on internalizing

15

symptoms at 12-13 versus 14-15, lends support to a previously proposed idea that

hormonal mechanisms may be responsible for this reversal of effect (Glaser et al., 2011).

Finally, we examined the moderating role of childhood cognitive ability in the

relationship between childhood stress and adolescent internalizing symptoms. Overall,

our findings support our hypothesis that greater cognitive ability in childhood attenuates

the effect of stress on subsequent internalizing symptoms. We found evidence that

childhood family dysfunction and chronic illness had significantly less negative effect on

subsequent internalizing symptoms for children with higher cognitive ability. Children

with greater cognitive ability in childhood may be better at coping with the stress of

family dysfunction and chronic illness in childhood, perhaps because they are more

capable of problem-solving and adapting to more stressful environments.

In contrast, there was also some evidence that SLEs in late childhood actually had

a stronger effect on subsequent internalizing symptoms in early adolescence for children

with higher cognitive ability. Although the reason for this finding is unclear, it suggests

that in some cases there may be a detrimental effect of increased cognitive ability. One

possible explanation relies upon the distinction between acute versus chronic stress.

Participants in the current study were asked to report on a list of SLEs in the previous two

years, whereas family dysfunction and chronic illnesses may have represented chronic

stressors in many cases. Indeed, we found that greater cognitive ability attenuated the

effects of family dysfunction and chronic illness at different time-points in childhood. In

contrast, greater cognitive ability only exacerbated the effects of SLEs measured at age

10-11, suggesting that only the most recent SLEs had differential effects on children with

high versus low cognitive ability. This is consistent with a study showing that recent (i.e.,

16

within past two years) SLEs have a stronger effect on adolescent internalizing than

previous (more than two years prior) SLEs (Colman et al., 2012).

However, the relative effects of recent versus previous SLEs do not explain why

recent stressors may be particularly problematic for adolescents with greater cognitive

ability. There has been some research suggesting that more cognitively advanced children

may be more prone to worrying in response to cognitive tasks (Muris et al., 2002). It is

therefore possible that, in response to acute stress, children with greater cognitive ability

are less capable of coping with stressful situations. For example, it has been argued that

more intelligent children may view intelligence as fixed or unchangeable, and can

therefore show a reduction in perseverance when faced with difficult problems (Burhans

and Dweck, 1995). Indeed, one study showed that, in the presence of stressful life events,

adolescents with high grades in school did not differ from those with low grades in terms

of internalizing symptoms (Luthar, 1991); another showed that gifted children

adolescents display stronger physiological reactions to stressful cognitive tasks than non-

gifted children (Roberts and Lovett, 1994).

There are some important limitations to the current study that should be

addressed. First, the data used for this project are largely self- or parent- reported, and are

therefore subject to some degree of reporting bias. Second, there are multiple ways of

assessing cognitive ability (including IQ tests, school grades, and performance on

specific cognitive tasks); the NLSCY relies on limited tests of verbal ability, scholastic

aptitude (PPVT-R) and mathematical computation (MCT). Although the concordance of

our findings with previous studies (Der et al., 2009, Franz et al., 2011, Koenen et al.,

2009) suggests that these measures are capturing constructs similar to IQ, it is possible

17

that other methods of assessment would yield different results. In particular, it is difficult

to separate cognitive potential from cognitive performance, the latter of which can be

affected by factors such as test anxiety, as well as subject-specific anxiety (e.g., math

anxiety) (Ashcraft and Faust, 1994, Gass and Curiel, 2011).

Third, childhood cognitive ability may not be directly causally related to

adolescent internalizing symptoms, but rather may be indirectly associated through

intermediary variables that modify mental health risk, such as academic achievement

(Hatch et al., 2007). However, internalizing symptoms could also lead to problems with

attention and memory, and thus have an effect on cognitive ability. Indeed, anxiety and

depression have both been conceptualized as leading to lower academic performance (Da

Fonseca et al., 2009, Owens et al., 2008), and inducing sad mood in adults has been

shown to affect performance on cognitive tasks (Chepenik et al., 2007).

A number of biological substrates, as well as early childhood factors, likely

influence cognitive ability. For example, research suggests that infant developmental

delay influences adolescent depression (Colman et al., 2007, North et al., 2013) and

cognitive impairment may be part of a causal pathway from infant malnutrition to

adolescent depression (Waber et al., 2011). Also, there is good evidence that impaired

cognition plays a role in the development of psychosis, through direct insults to the brain

as well as interactions with environmental factors (Morgan et al., 2010, Murray and

Fearon, 1999). In future studies, the role of cognitive ability can be more fully explored

as part of a pathway towards internalizing symptoms.

This study also had a number of strengths. The NLSCY is a nationally

representative cohort of Canadian children and youth and our results are therefore highly

18

generalizable. Few studies have examined the role of childhood cognitive ability in the

development of internalizing symptoms in adolescence, and our study adds substantially

to the relatively few studies that examine this question. To our knowledge, this was the

first study to examine the moderating role of childhood cognitive ability on the

relationship between childhood stress and adolescent internalizing. Our results have shed

light on the important and complex role that childhood cognitive ability may play in the

development of these often debilitating symptoms.

19

REFERENCES

ACHENBACH, T. M. & EDELBROCK, C. S. 1981. Behavioral-Problems and Competencies Reported by Parents of Normal and Disturbed-Children Aged 4 through 16. Monographs of the Society for Research in Child Development, 46, 1-82.

ASHCRAFT, M. H. & FAUST, M. W. 1994. Mathematics Anxiety and Mental Arithmetic Performance - an Exploratory Investigation. Cogn Emot, 8, 97-125.

ASHFORD, J., SMIT, F., VAN LIER, P. A. C., CUIJPERS, P. & KOOT, H. M. 2008. Early risk indicators of internalizing problems in late childhood: a 9-year longitudinal study. Journal of Child Psychology and Psychiatry, 49, 774-780.

ASHTON, G. C., POLOVINA, J. J. & VANDENBERG, S. G. 1979. Segregation analysis of family data for 15 tests of cognitive ability. Behavior Genetics, 9, 329-47.

BOYLE, M. H., OFFORD, D. R., HOFMANN, H. G., CATLIN, G. P., BYLES, J. A., CADMAN, D. T., CRAWFORD, J. W., LINKS, P. S., RAEGRANT, N. I. & SZATMARI, P. 1987. Ontario Child Health Study .1. Methodology. Archives of General Psychiatry, 44, 826-831.

BOYLE, M. H., OFFORD, D. R., RACINE, Y., FLEMING, J. E., SZATMARI, P. & SANFORD, M. 1993a. Evaluation of the revised Ontario Child Health Study scales. Journal of Child Psychology and Psychiatry and Allied Disciplines, 34, 189-213.

BOYLE, M. H., OFFORD, D. R., RACINE, Y., SANFORD, M., SZATMARI, P. & FLEMING, J. E. 1993b. Evaluation of the Original Ontario-Child-Health-Study Scales. Canadian Journal of Psychiatry-Revue Canadienne De Psychiatrie, 38, 397-405.

BRUNI, O., KOHLER, M., NOVELLI, L., KENNEDY, D., LUSHINGTON, K., MARTIN, J. & FERRI, R. 2012. The Role of NREM Sleep Instability in Child Cognitive Performance. Sleep, 35, 649-656.

BURHANS, K. K. & DWECK, C. S. 1995. Helplessness in early childhood: the role of contingent worth. Child Development, 66, 1719-38.

BURT, K. B. & ROISMAN, G. I. 2010. Competence and psychopathology: cascade effects in the NICHD Study of Early Child Care and Youth Development. Development and Psychopathology, 22, 557-67.

CANADA, S. C. A. H. R. A. S. D. 2009. National Longitudinal Survey of Children and Youth: Survey Overview for the 2008/2009 Data Collection, Cycle 8. Statistics Canada and Human Resources and Skills Development Canada. Ottawa, ON: Statistics Canada.

CHEPENIK, L. G., CORNEW, L. A. & FARAH, M. J. 2007. The influence of sad mood on cognition. Emotion, 7, 802-11.

COLE, D. A. 1991. Preliminary support for a competency-based model of depression in children. Journal of Abnormal Psychology, 100, 181-90.

COLMAN, I., ATAULLAHJAN, A., NAICKER, K. & VAN LIESHOUT, R. J. 2012. Birth Weight, Stress, and Symptoms of Depression in Adolescence: Evidence of Fetal Programming in a National Canadian Cohort. Canadian Journal of Psychiatry-Revue Canadienne De Psychiatrie, 57, 422-428.

20

COLMAN, I., PLOUBIDIS, G. B., WADSWORTH, M. E. J., JONES, P. B. & CROUDACE, T. J. 2007. A longitudinal typology of symptoms of depression and anxiety over the life course. Biological Psychiatry, 62, 1265-1271.

COSTELLO, E. J., EGGER, H. & ANGOLD, A. 2005. 10-year research update review: the epidemiology of child and adolescent psychiatric disorders: I. Methods and public health burden. Journal of the American Academy of Child and Adolescent Psychiatry, 44, 972-86.

COTTON, C. 2001. Developments in the Low Income Cutoffs. Cotton, C. Ottawa, ON: Statistics Canada.

DA FONSECA, D., CURY, F., SANTOS, A., PAYEN, V., BOUNOUA, L., BRISSWALTER, J., RUFO, M., POINSO, F. & DERUELLE, C. 2009. When depression mediates the relationship between entity beliefs and performance. Child Psychiatry and Human Development, .40, pp.

DAVIS, T. E., OLLENDICK, T. H. & NEBEL-SCHWALM, M. 2008. Intellectual ability and achievement in anxiety-disordered children: A clarification and extension of the literature. Journal of Psychopathology and Behavioral Assessment, 30, 43-51.

DEFRIES, J. C., JOHNSON, R. C., KUSE, A. R., MCCLEARN, G. E., POLOVINA, J., VANDENBERG, S. G. & WILSON, J. R. 1979. Familial resemblance for specific cognitive abilities. Behavior Genetics, 9, 23-43.

DER, G., BATTY, G. D. & DEARY, I. J. 2009. The association between IQ in adolescence and a range of health outcomes at 40 in the 1979 US National Longitudinal Study of Youth. Intelligence, 37, 573-580.

DUNN, L. & DUNN, L. 2006. Peabody Picture Vocabulary Test. Pearson.EK, U., WESTERLUND, J. & FERNELL, E. 2013. General versus executive cognitive

ability in pupils with ADHD and with milder attention problems. Neuropsychiatric Disease and Treatment, 9, 163-168.

EMERSON, C. S., MOLLET, G. A. & HARRISON, D. W. 2005. Anxious-depression in boys: an evaluation of executive functioning. Archives of Clinical Neuropsychology, 20, 539-546.

FRANZ, C. E., LYONS, M. J., O'BRIEN, R., PANIZZON, M. S., KIM, K., BHAT, R., GRANT, M. D., TOOMEY, R., EISEN, S., XIAN, H. & KREMEN, W. S. 2011. A 35-year longitudinal assessment of cognition and midlife depression symptoms: The Vietnam Era Twin Study of Aging. The American Journal of Geriatric Psychiatry, .19, pp.

GASS, C. & CURIEL, R. 2011. Test Anxiety in Relation to Measures of Cognitive and Intellectual Functioning. Archives of Clinical Neuropsychology, 26, 553-553.

GLASER, B., GUNNELL, D., TIMPSON, N. J., JOINSON, C., ZAMMIT, S., SMITH, G. D. & LEWIS, G. 2011. Age- and puberty-dependent association between IQ score in early childhood and depressive symptoms in adolescence. Psychological Medicine, 41, 333-43.

GREENBERG, P. E., KESSLER, R. C., BIRNBAUM, H. G., LEONG, S. A., LOWE, S. W., BERGLUND, P. A. & COREY-LISLE, P. K. 2003. The economic burden of depression in the United States: how did it change between 1990 and 2000? Journal of Clinical Psychiatry, 64, 1465-75.

21

GREENBERG, P. E., SISITSKY, T., KESSLER, R. C., FINKELSTEIN, S. N., BERNDT, E. R., DAVIDSON, J. R., BALLENGER, J. C. & FYER, A. J. 1999. The economic burden of anxiety disorders in the 1990s. Journal of Clinical Psychiatry, 60, 427-35.

HATCH, S. L., JONES, P. B., KUH, D., HARDY, R., WADSWORTH, M. E. & RICHARDS, M. 2007. Childhood cognitive ability and adult mental health in the British 1946 birth cohort. Social Science and Medicine, 64, 2285-96.

HODGES, K. & PLOW, J. 1990. Intellectual Ability and Achievement in Psychiatrically Hospitalized Children with Conduct, Anxiety, and Affective-Disorders. Journal of Consulting and Clinical Psychology, 58, 589-595.

KESSLER, R. C., MCGONAGLE, K. A., SWARTZ, M., BLAZER, D. G. & NELSON, C. B. 1993. Sex and Depression in the National Comorbidity Survey .1. Lifetime Prevalence, Chronicity and Recurrence. Journal of Affective Disorders, 29, 85-96.

KOENEN, K. C., MOFFITT, T. E., ROBERTS, A. L., MARTIN, L. T., KUBZANSKY, L., HARRINGTON, H., POULTON, R. & CASPI, A. 2009. Childhood IQ and adult mental disorders: a test of the cognitive reserve hypothesis. American Journal of Psychiatry, 166, 50-7.

KRISTENSEN, H. & TORGERSEN, S. 2008. Is social anxiety disorder in childhood associated with developmental deficit/delay? European Child and Adolescent Psychiatry, 17, 99-107.

LEE, J. 2011. Pathways from education to depression. J Cross Cult Gerontol, 26, 121-35.

LIN, F.-G., TUNG, H.-J., HSIEH, Y.-H. & LIN, J.-D. 2011. Interactive influences of family and school ecologies on the depression status among children in marital immigrant families. Research in Developmental Disabilities, .32, pp.

LUTHAR, S. S. 1991. Vulnerability and resilience: a study of high-risk adolescents. Child Development, 62, 600-16.

MASTEN, A. S., ROISMAN, G. I., LONG, J. D., BURT, K. B., OBRADOVIC, J., RILEY, J. R., BOELCKE-STENNES, K. & TELLEGEN, A. 2005. Developmental cascades: linking academic achievement and externalizing and internalizing symptoms over 20 years. Developmental Psychology, 41, 733-46.

MORGAN, C., CHARALAMBIDES, M., HUTCHINSON, G. & MURRAY, R. M. 2010. Migration, Ethnicity, and Psychosis: Toward a Sociodevelopmental Model. Schizophrenia Bulletin, 36, 655-664.

MURIS, P., MERCKELBACH, H., MEESTERS, C. & VAN DEN BRAND, K. 2002. Cognitive development and worry in normal children. Cognitive Therapy and Research, 26, 775-787.

MURRAY, R. M. & FEARON, P. 1999. The developmental 'risk factor' model of schizophrenia. Journal of Psychiatric Research, 33, 497-499.

NAICKER, K., WICKHAM, M. & COLMAN, I. 2012. Timing of first exposure to maternal depression and adolescent emotional disorder in a national Canadian cohort. PLoS One, 7, e33422.

NORTH, C. R., WILD, T. C., ZWAIGENBAUM, L. & COLMAN, I. 2013. Early neurodevelopment and self-reported adolescent symptoms of depression and anxiety in a national canadian cohort study. PLoS One, 8, e56804.

22

OWENS, M., STEVENSON, J., NORGATE, R. & HADWIN, J. A. 2008. Processing efficiency theory in children: Working memory as a mediator between trait anxiety and academic performance. Anxiety Stress and Coping, 21, 417-430.

PINE, D. S., COHEN, E., COHEN, P. & BROOK, J. 1999. Adolescent depressive symptoms as predictors of adult depression: Moodiness or mood disorder? American Journal of Psychiatry, 156, 133-135.

RABBITT, P., DONLAN, C., WATSON, P., MCINNES, L. & BENT, N. 1995. Unique and interactive effects of depression, age, socioeconomic advantage, and gender on cognitive performance of normal healthy older people. Psychology and Aging, 10, 307-13.

RADLOFF, L. S. 1977. The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401.

RAPPORT, M. D., DENNEY, C. B., CHUNG, K. M. & HUSTACE, K. 2001. Internalizing behavior problems and scholastic achievement in children: Cognitive and behavioral pathways as mediators of outcome. Journal of Clinical Child Psychology, 30, 536-551.

ROBERTS, S. M. & LOVETT, S. B. 1994. Examining the F in Gifted - Academically Gifted Adolescents Physiological and Affective Responses to Scholastic Failure. Journal for the Education of the Gifted, 17, 241-259.

SASS, K. J., BUCHANAN, C. P., WESTERVELD, M., MAREK, K. L., FARHI, A., ROBBINS, R. J., NAFTOLIN, F., VOLLMER, T. L., LERANTH, C., ROTH, R. H. & ET AL. 1995. General cognitive ability following unilateral and bilateral fetal ventral mesencephalic tissue transplantation for treatment of Parkinson's disease. Archives of Neurology, 52, 680-6.

STATISTICS CANADA 1992. Growing up in Canada: national longitudinal study of children and youth. Ottawa, ON: Statistics Canada.

STATISTICS CANADA AND HUMAN RESOURCES AND SKILLS DEVELOPMENT CANADA 1998. National Longitudinal Survey of Children and Youth - Cycle 1 Data Dictionary. Statistics Canada and Human Resources and Skills Development Canada. Ottawa, ON: Statistics Canada.

STOOLMILLER, M., KIM, H. K. & CAPALDI, D. M. 2005. The course of depressive symptoms in men from early adolescence to young adulthood: identifying latent trajectories and early predictors. Journal of Abnormal Psychology, 114, 331-45.

TOPITZES, J., GODES, O., MERSKY, J. P., CEGLAREK, S. & REYNOLDS, A. J. 2009. Educational success and adult health: Findings from the Chicago Longitudinal Study. Prevention Science, .10, pp.

TRIPPAS, D., HANDLEY, S. J. & VERDE, M. F. 2013. The SDT Model of Belief Bias: Complexity, Time, and Cognitive Ability Mediate the Effects of Believability. Journal of Experimental Psychology. Learning, Memory, and Cognition.

VAN OS, J. & JONES, P. B. 1999. Early risk factors and adult person--environment relationships in affective disorder. Psychological Medicine, 29, 1055-67.

VASA, R. A., ROBERSON-NAY, R., KLEIN, R. G., MANNUZZA, S., MOULTON, J. L., GUARDINO, M., MERIKANGAS, A., CARLINO, A. R. & PINE, D. S. 2007. Memory deficits in children with and at risk for anxiety disorders. Depression and Anxiety, 24, 85-94.

23

VERNON, P. E. 1971. Analysis of cognitive ability. British Medical Bulletin, 27, 222-6.WABER, D. P., EAGLESFIELD, D., FITZMAURICE, G. M., BRYCE, C., HARRISON, R. H. &

GALLER, J. R. 2011. Cognitive Impairment as a Mediator in the Developmental Pathway From Infant Malnutrition to Adolescent Depressive Symptoms in Barbadian Youth. Journal of Developmental and Behavioral Pediatrics, 32, 225-232.

WADE, T. J., CAIRNEY, J. & PEVALIN, D. J. 2002. Emergence of gender differences in depression during adolescence: national panel results from three countries. Journal of the American Academy of Child and Adolescent Psychiatry, 41, 190-8.

WOODWARD, L. J., CLARK, C. A. C., BORA, S. & INDER, T. E. 2012. Neonatal White Matter Abnormalities an Important Predictor of Neurocognitive Outcome for Very Preterm Children. PLoS One, 7.

ZAMMIT, S., ALLEBECK, P., DAVID, A. S., DALMAN, C., HEMMINGSSON, T., LUNDBERG, I. & LEWIS, G. 2004. A longitudinal study of premorbid IQ Score and risk of developing schizophrenia, bipolar disorder, severe depression, and other nonaffective psychoses. Archives of General Psychiatry, 61, 354-60.

24

Age 2-3 Age 4-5 Age 6-7 Age 8-9 Age 10-11 Age 12-13 Age 14-15

Female 50.1

PPVT-R Score 100.2 (14.2)

MCT Score 291.6 (41.7) 358.8 (53.0) 426.0 (55.2)

Maternal Depression 10.4

Family Dysfunction 15.3 8.4 8.5 13.3

≥1 Stressful Life Event 2.7 2.5 2.5 2.9

Chronic Illness 18.4 21.8 26.0 29.2

Internalizing Symptoms

Low 58.4 57.2

Mild 21.1 20.1

Moderate 12.6 12.9

Severe 8.0 9.9

Table 1. Percent prevalence or Mean (SD) for study variables

25

Table 2. Cognitive ability predicting internalizing symptoms at ages 12-13 and 14-15

Cognitive Ability Measure

Internalizing Symptom

Severity (‘low’ as reference)

Symptoms at age 12-13

Unadjusted OR (CI) Adjusted OR (CI)

Symptoms at age 14-15

Unadjusted OR (CI) Adjusted OR (CI)4

PPVT-R n=3402 n=3232 n=3078 n=2928(age 4-5) Mild *0.91 [0.83, 0.98] 0.92 [0.85, 1.01] *0.83 [0.76, 0.91] *0.84 [0.76, 0.92]

Moderate *0.83 [0.75, 0.92] *0.83 [0.75, 0.93] 0.94 [0.84, 1.04] 0.94 [0.84, 1.05]Severe *0.82 [0.73, 0.94] *0.80 [0.70, 0.92] 0.98 [0.87, 1.07] 0.97 [0.86, 1.11]

MCT n=1211 n=1127 n=1058 n=990(age 6-7) Mild *1.19 [1.05, 1.36] *1.21 [1.05, 1.39] *1.35 [1.15, 1.60] *1.44 [1.21, 1.72]

Moderate 0.86 [0.72, 1.04] 0.90 [0.74, 1.09] 1.12 [0.92, 1.35] 1.20 [0.98, 1.47]Severe *0.74 [0.58, 0.95] 0.79 [0.60, 1.04] *1.44 [1.17, 1.78] *1.50 [1.18, 1.92]

MCT n=3307 n=3096 n=3002 n=2809(age 8-9) Mild *0.91 [0.84, 0.99] 0.91 [0.83, 1.00] 1.09 [0.99, 1.20] 1.09 [0.99, 1.21]

Moderate *0.83 [0.74, 0.92] *0.83 [0.75, 0.93] 1.10 [0.98, 1.23] *1.14 [1.02, 1.29]Severe *0.81 [0.71, 0.93] *0.85 [0.74, 0.98] *1.18 [1.04, 1.34] *1.20 [1.05, 1.37]

MCT n=3535 n=3372 n=3240 n=3093(age 10-11) Mild 1.00 [0.92, 1.09] 1.04 [0.95, 1.13] 1.00 [0.92, 1.09] 1.02 [0.93, 1.12]

Moderate *0.85 [0.77, 0.94] 0.91 [0.82, 1.01] 1.07 [0.96, 1.19] 1.12 [1.00, 1.25]Severe *0.83 [0.73, 0.95] *0.85 [0.75, 0.98] 1.01 [0.90, 1.14] 1.07 [0.94, 1.22]

Adjusted for Gender, SES, current family dysfunction, ≥ 1 stressful life events in previous two years, and PMK depression at age 2-3. See footnote in Methods section regarding reduced sample sizes.

26

Table 3. Gender differences in the relationship between childhood cognitive ability and adolescent internalizing symptoms

Interaction Internalizing Outcome

Internalizing Symptom Level

Interaction significance

OR (CI)for Males

OR (CI)for Females

Sex x PPVT-R age 4-5 Age 12-13 Moderate p< .0001 1.10 [0.95, 1.28] 0.70 [0.61, 0.80]Age 14-15 Mild p= .045 0.92 [0.81, 1.04] 0.76 [0.67, 0.87]

Sex x MCT age 6-7 Age 14-15 Mild p= .009 1.82 [1.43, 2.32] 1.14 [0.89, 1.46]

Sex x MCT age 8-9 Age 14-15 Moderate p< .001 0.83 [0.68, 1.02] 1.30 [1.13, 1.50]

27

Table 4. Moderating role of cognitive ability in the relationship between childhood stress and adolescent internalizing

Interaction Internalizing outcome

Internalizing symptom level

Interaction significance

OR (CI) at lowCognitive Ability

OR (CI) at highCognitive Ability

PPVT-R x Family Dysfunction at age 4-5 Age 12-13 Mild p= .002 0.53 [0.36, 0.77] 1.27 [0.86, 1.89]Severe p= .021 1.81 [1.21, 2.71] 0.82 [0.45, 1.48]

MCT x Family Dysfunction at age 8-9 Age 14-15 Severe p= .001 2.42 [1.30, 4.53] 0.53 [0.26, 1.09]

MCT x Family Dysfunction at age 10-11 Age 12-13 Mild p= .023 1.84 [1.28, 2.65] 1.00 [0.69, 1.46]

MCT x Family Dysfunction at age 10-11 Age 14-15 Mild p= .044 1.41 [0.97, 2.06] 0.83 [0.58, 1.20]Severe p= .018 2.90 [1.84, 4.56] 1.36 [0.87, 2.11]

PPVT-R x Chronic Illness at age 4-5 Age 14-15 Mild p= .008 1.71 [1.28, 2.27] 0.94 [0.65, 1.36]

MCT x Chronic Illness at age 6-7 Age 14-15 Severe p= .001 3.24 [1.49, 7.04] 0.44 [0.19, 0.98]

MCT x SLE (> 1) at age 10-11 Age 12-13 Moderate p= .004 0.43 [0.15, 1.19] 3.30 [1.50, 7.25]

28