Green spaces and cognitive development in
, Mark J. Nieuwenhuijsen
, Mikel Esnaola
, Joan Forns
, Xavier Basagaña
, Ioar Rivas
, Mónica López-Vicente
, Montserrat De Castro Pascual
, Jason Su
, Xavier Querol
, and Jordi Sunyer
Centre for Research in Environmental Epidemiology (CREAL), 08003 Barcelona, Spain;
Experimental and Health Sciences, Pompeu Fabra University, 08003
Barcelona, Catalonia, Spain;
Ciber on Epidemiology and Public Health (CIBERESP), 28029 Madrid, Spain;
Department of Genes and Environment, Division
of Epidemiology, Norwegian Institute of Public Health, 0473, Oslo, Norway;
Department of Geosciences, Institute of Environmental Assessment and Water
Research, Spanish National Research Council (CSIC-IDEA), 08034 Barcelona, Catalonia, Spain;
Environmental Health Sciences, School of Public Health,
University of California, Berkeley, CA 94720-7360;
Department of Environmental Health Sciences, Fielding School of Public Health, University of California,
Los Angeles, CA 90095; and
Hospital del Mar Medical Research Institute (IMIM), 08003 Barcelona, Catalonia, Spain
Edited by Susan Hanson, Clark University, Worcester, MA, and approved May 15, 2015 (received for review February 18, 2015)
Exposure to green space has been associated with better physical
and mental health. Although this exposure could also influence
cognitive development in children, available epidemiological
evidence on such an impact is scarce. This study aimed to assess
the association between exposure to green space and measures of
cognitive development in primary schoolchildren. This study was
based on 2,593 schoolchildren in the second to fourth grades (7–10
y) of 36 primary schools in Barcelona, Spain (2012–2013). Cognitive
development was assessed as 12-mo change in developmental
trajectory of working memory, superior working memory, and in-
attentiveness by using four repeated (every 3 mo) computerized
cognitive tests for each outcome. We assessed exposure to green
space by characterizing outdoor surrounding greenness at home
and school and during commuting by using high-resolution (5 m ×
5 m) satellite data on greenness (normalized difference vegetation
index). Multilevel modeling was used to estimate the associations
between green spaces and cognitive development. We observed
an enhanced 12-mo progress in working memory and superior
working memory and a greater 12-mo reduction in inattentiveness
associated with greenness within and surrounding school bound-
aries and with total surrounding greenness index (including green-
ness surrounding home, commuting route, and school). Adding a
traffic-related air pollutant (elemental carbon) to models explained
20–65% of our estimated associations between school greenness
and 12-mo cognitive development. Our study showed a beneficial
association between exposure to green space and cognitive devel-
opment among schoolchildren that was partly mediated by reduc-
tion in exposure to air pollution.
Contact with nature is thought to play a crucial and irre-
placeable role in brain development (1, 2). Natural envi-
ronments including green spaces provide children with unique
opportunities such as inciting engagement, risk taking, discovery,
creativity, mastery and control, strengthening sense of self, in-
spiring basic emotional states including sense of wonder, and
enhancing psychological restoration, which are suggested to in-
fluence positively different aspects of cognitive development (1–
3). Beneficial effects of green spaces on cognitive development
might accrue from direct influences such as those above, with
green space itself exerting the positive influence or through in-
direct, mediated pathways. The ability of green spaces to miti-
gate traffic-related air pollution (TRAP) (4) could lead to a
beneficial impact of green spaces on cognitive development,
because exposure to TRAP has been negatively associated with
cognitive development in children (5). Further to TRAP, green
spaces can also reduce noise (6), which itself too has been neg-
atively associated with cognitive development (7). Moreover,
proximity to green spaces, particularly parks, has been suggested
to increase physical activity (8), and higher levels of physical
activity are related to improved cognitive development (9).
Outdoor surrounding greenness has also been reported to enrich
microbial input from the environment (10), which may positively
influence cognitive development (10). Through these pathways,
exposure to green space, including outdoor surrounding greenness
and proximity to green spaces, could influence cognitive de-
velopment in children, yet the available population-based evidence
on the association between such exposure and cognitive develop-
ment in children remains scarce.
The brain develops steadily during prenatal and early post-
natal periods, which are considered as the most vulnerable
windows for effects of environmental exposures (11). However,
some cognitive functions closely related with learning and school
achievement—such as working memory and attention—develop
across childhood and adolescence as an essential part of cogni-
tive maturation (12–14). We therefore hypothesized a priori that
exposure to green space in primary schoolchildren could en-
hance cognitive development. Accordingly, our study aimed to
assess the association between indicators of exposure to green
space and measures of cognitive development, including working
memory (the system that holds multiple pieces of transitory in-
formation in the mind where they can be manipulated), superior
working memory (working memory that involves continuous
updating of the working memory buffer), and inattentiveness in
primary schoolchildren. As a secondary aim, we also evaluated
the mediating role of a reduction in air pollution as one of the
potential mechanisms underlying this association.
Green spaces have a range of health benefits, but little is known in
relation to cognitive development in children. This study, based on
comprehensive characterization of outdoor surrounding green-
ness (at home, school, and during commuting) and repeated
computerized cognitive tests in schoolchildren, found an improve-
ment in cognitive development associated with surrounding
greenness, particularly with greenness at schools. This association
was partly mediated by reductions in air pollution. Our findings
provide policymakers with evidence for feasible and achievable
targeted interventions such as improving green spaces at schools
to attain improvements in mental capital at population level.
Author contributions: P.D., M.J.N. , X.Q., and J. Sunyer designed research; M.J.N., J.F.,
M.A.-P., I.R., M.L.-V., M.D.C.P., X.Q., and J. Sunyer performed research; M.E., X.B., J. Su,
and M.J. contributed new reagents/analytic tools; P.D., M.E., and X.B. analyzed data; and
P.D. and J. Sunyer wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
To whom correspondence should be addressed. Email: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
June 30, 2015
Study Setting. We undertook this study in Barcelona, Spain, a port city sit-
uated on the northeastern part of the Iberian Peninsula. It has a Mediter-
ranean climate characterized by hot and dry summers, mild winters, and
maximum precipitation and vegetation during autumn and spring. This study
was conducted in the context of the brain development and air pollution
ultrafine particles in school children (BREATHE) project. Of the 416 schools in
Barcelona, 37 schools were initially selected to obtain maximum contrast in
TRAP levels (i.e., nitrogen dioxide: NO
), of which 36 accepted to participate
and were included in the study (SI Appendix, Fig. S1). Participating schools
were similar to the remaining schools in Barcelona in terms of the neigh-
borhood socioeconomic vulnerability index (0.46 versus 0.50, Kruskal–Wallis test
P=0.57) and NO
levels (51.5 versus 50.9 μg/m
, Kruskal–Wallis test P=0.72).
All schoolchildren (n=4,562) without special needs in the second to
fourth grades (7–10 y) of these schools were invited to participate by letters
or presentations in schools for parents, of which 2,623 (58%) agreed to take
part in BREATHE. All children had been in the school for more than 6 mo
(and 98% more than 1 y) before the beginning of the study. All parents or
guardians signed the informed consent and the study was approved (No.
2010/41221/I) by the Clinical Research Ethical Committee of the Parc de Salut
Outcome:Cognitive Development. Cognitive development was assessedthrough
12-mo change in developmental trajectory of working memory and at-
tention. We selected these functions because they grow steadily during
preadolescence (15, 16). We used computerized n-back test for assessing
working memory (15) and computerized attentional network test (ANT)
(17) for evaluating attention.
From January 2012 to March 2013, children were evaluated every 3 mo
over four repeated visits by using computerized tests in sessions lasting
∼40 min in length. Groups of 10–20 children wearing ear protectors were
assessed together and supervised by one trained examiner per 3–4 children.
For the n-back test, we examined different n-back loads (up to three-back)
and stimuli (colors, numbers, letters, and words). For analysis here, we se-
lected both two-back and three-back loads for number and word stimuli
because they showed a clear age-dependent slope in the four measurements
(4). The two-back predicts general mental abilities (i.e., working memory)
whereas the three-back also predicts superior functions such as fluid in-
telligence (i.e., superior working memory) (18). All sets of n-back tests star-
ted with colors as a training phase to ensure participants’comprehension of
the test. The n-back parameter analyzed was dprime (d′), a measure of
detection subtracting the normalized false alarm rate from the hit rate
[(Z hit rate −Z false alarm rate) ×100]. A higher d′indicates more accurate
test performance. Given that our final findings for numbers and words were
similar, here we only show results for numbers. Among the ANT measures,
we chose hit reaction time standard error (HRT-SE) (SE of RT for correct re-
sponses), a measure of response speed consistency throughout the test (19),
because it showed a clear growth during the 1-y study period. A higher HRT-SE
indicates highly variable reactions related to inattentiveness.
Exposure to Green Space. Our assessment of exposure to green space was
based on a comprehensive characterization of outdoor surrounding green-
ness (photosynthetically active vegetation) encompassing greenness sur-
rounding home, greenness surrounding commuting route between home
and school (hereafter referred to as commuting greenness), and greenness
within and around school boundaries.
To assess outdoor surrounding greenness we applied normalized differ-
ence vegetation index (NDVI) derived from RapidEye data at 5 m ×5m
resolution. NDVI is an indicator of greenness based on land surface re-
flectance of visible (red) and near-infrared parts of spectrum (20). It ranges
between −1 and 1, with higher numbers indicating more greenness. The
RapidEye Imagery is acquired from a constellation of five satellites 630 km
above ground in sun-synchronous orbits. We generated our NDVI map by
using the image obtained on July 23, 2012, that was available for our study
region during our study period (SI Appendix, Fig. S1).
Residential surrounding greenness. Residential surrounding greenness was ab-
stracted as the average of NDVI in a buffer of 250 m (21, 22) around the home
address of each study participant. For 174 children (5.9%) who shared two
homes, we used the address where the child spent most of her/his time.
Commuting greenness. Data on the main mode of commute to and from school
was obtained from parents via questionnaires. Approximately 60% of par-
ticipants reported walking as the main mode of commuting, whereas the
38% reported commuting by motor vehicles (private car, bus, motorcycle, or
tram). The remaining 2% reported the underground metro train as the main
mode of transport, for whom we assumed no exposure to greenness during
commuting. For participants reporting walking as the main mode of com-
muting, we identified the shortest walking route to school and for partici-
pants reporting motor vehicles as the main mode of commuting, we
identified the shortest driving route to school, based on street networks
(network distance) by using network analyst extension from ArcGIS software
v10. We defined commuting greenness as the average of NDVI in a 50-m
buffer around the commuting route.
School greenness. To assess greenness within school premises, we first digitized
the school boundaries and then averaged NDVI values within those
boundaries. To assess greenness surrounding schools, we averaged NDVI
values across a 50 m buffer around the school boundaries.
Total surrounding greenness index. We developed a total surrounding greenness
index by averaging residential surrounding greenness (250-m buffer), com-
muting greenness, and greenness within school boundaries weighted by the
daytime (12 h a day) that children were assumed to spend at home (3 h),
commuting (1 h), and school (8 h). To avoid double-counting in developing
this index, we abstracted as the average NDVI over commute corridor beyond
the 250-m home buffer and 50-m school buffer.
Main Analyses. Data on 9,357 tests from 2,593 (99%) children were available
for analysis. Because of the multilevel nature of the data (i.e., multiple visits
for each child within schools), we used linear mixed effects models with the
four repeated cognitive parameters as outcomes (one test at a time), each
measure of exposure to green space (one at a time) as fixed effect predictor,
and child and school as random effects (5). An interaction between age at
each visit and the indicator of exposure to green space was included to
capture changes in 12-mo progress in cognitive trajectory associated with
greenness exposure (5). The main effect of exposure to green space, which
was also included in the model, captured the baseline (visit 1) differences in
cognitive function that were associated with exposure to green space before
the first visit. This model was further adjusted for potential confounders
identified a priori:age (centered at visit 1), sex, and indicators of socioeco-
nomic status (SES) at both individual and area levels. Maternal education (no
or primary/secondary/university) was used as the indicator of individual-level
SES and Urban Vulnerability Index (23), a measure of neighborhood SES at
the census tract (median area of 0.08 km
for the study region) was applied
as the indicator of area-level SES. Linearity of the relation between exposure
to green space and cognitive tests was assumed because generalized addi-
tive mixed models did not show any nonlinearity of associations. We esti-
mated the change in average outcome scores associated with one
interquartile range (IQR) increase (based on all study participants) in average
NDVI. Statistical significance was set at P<0.05. R statistical package was
used to carry out the analyses.
Mediating Role of Traffic-Related Air Pollution. We hypothesized that re-
duction in TRAP levels could be one of the potential mechanisms underlying
the association between greenness exposure and cognitive development. To
quantify such a mediating role, we calculated the percent of the associations
between greenness and cognitive development explained by TRAP as [1 −
)] ×100, where β
was the regression coefficient for the greenness
exposure in a fully adjusted model including the mediator (i.e., TRAP) and β
was the regression coefficient in the fully adjusted model without including
the mediator (24).
We focused on the associations between school greenness and cognitive
development because they were the strongest among our evaluated asso-
ciations (Results) and also because of the availability of data on levels of air
pollutants at BREATHE schools that were monitored as part of the BREATHE
project. Such a high-quality monitored data were not available for TRAP
levels at homes or during commuting. Among the TRAPs monitored in the
BREATHE framework, we chose indoor levels of elemental carbon (EC) for
this mediation analyses. EC is mainly generated by fossil fuel combustion and
is considered as a tracer of road traffic emissions in Barcelona (25). In other
BREATHE analyses, we had observed that indoor EC was associated with
adverse impacts on cognitive development (5) and EC levels were reduced in
schools with higher greenness (4). Detailed description of TRAP sampling
methodology at the BREATHE schools has been published (25, 26).
Children were on average 8.5 y old at baseline and 50% were
girls. Regarding maternal education, 13% of mothers had no or
only primary school, 29% secondary school, and 58% university
education. Further characteristics of the study participants are
presented in SI Appendix, Table S1. Average working memory
increased by 22.8%, superior working memory by 15.2%, and
www.pnas.org/cgi/doi/10.1073/pnas.1503402112 Dadvand et al.
inattentiveness decreased by 18.9% during the follow up (Table 1).
At baseline, higher maternal education was associated with better
cognitive function (SI Appendix,TableS2). For 12-mo progress,
whereas higher maternal education was associated with larger re-
duction in inattentiveness, improvements in working memory and
superior working memory were not associated with maternal edu-
cation (SI Appendix,TableS2). The median (IQR) of our estimated
surrounding greenness for all participants and across strata of ma-
ternal education are presented in Table 2 and SI Appendix,Table
S2, respectively. The Spearman’s correlation coefficient among
residential, school, and commuting surrounding greenness varied
from 0.46 (between surrounding greenness at home and greenness
within school boundaries) to 0.80 (between commuting and school
surrounding greenness) (SI Appendix,TableS3).
Main Analyses. We observed an enhanced 12-mo progress in
working memory and superior working memory and a greater
12-mo reduction in inattentiveness associated with greenness within
and surrounding school boundaries and with the total surrounding
greennessindex(Table2,Fig.1,andSI Appendix,Fig.S2). Com-
muting greenness was also associated with improved 12-mo progress
in working memory and superior working memory, although the
association for superior working memory was only marginally sta-
tistically significant. We did not observe any association between
residential surrounding greenness and cognitive measurements
(Table 2). None of the indicators of outdoor greenness were asso-
ciated with baseline cognitive measurements (Table 2).
The findings for n-back tests with “word”stimuli were con-
sistent with the aforementioned results for “number”stimuli (SI
Appendix,TableS4). The association between commuting green-
ness and 12-mo progress in superior working memory, which had
borderline statistical significance for the three-back test using
number stimuli, was statistically significant for the test using
To explore the possibility of an impact of green space exposure
on other ANT measures than inattentiveness, we repeated the
main analyses by using alerting, orienting, and executive pro-
cessing (one at a time) abstracted from ANT as outcome. We did
not observe any statistically significant association for these
outcomes with any of indicators of green space exposure (SI
Appendix, Table S5), which was consistent with our observation
that these measures did not show any clear growth during the
We conducted a number of sensitivity analyses as described in
SI Appendix that showed the robustness of our findings to al-
ternative definition of total surrounding greenness index and
commuting greenness and to including a range of relevant
covariates in models (e.g., socioeconomic indicators and condi-
tion of venue at the time of cognitive tests).
Mediating Role of Traffic-Related Air Pollution. The Spearman’s
correlation coefficients between school EC levels and greenness
within and surrounding school boundaries were −0.62 and −0.66
(P<0.01), respectively. Adding EC to models explained 20–65%
of associations between school greenness and 12-mo progress in
cognitive functions (Table 3). Including EC reduced effect sizes
in all models. EC made the associations between school surrounding
greenness and superior working memory and between greenness
within and surrounding school boundaries and inattentiveness much
smaller and statistically nonsignificant (Table 3).
To our knowledge, this is the first epidemiological study to re-
port on the impact of exposure to green space on cognitive de-
velopment in schoolchildren. School and total surrounding
greenness index were associated with enhanced 12-mo progress
in indicators of working memory and superior working memory
and greater 12-mo reduction in inattentiveness. Commuting
greenness was also associated with better 12-mo progress in
working memory. Adding EC to our models explained 20–65%
of our estimated associations between green spaces and 12-mo
Interpretation of Results. Over a 12-mo period, we observed that
an IQR exposure increment in total surrounding greenness index
was associated with a 5% increase in the progress of working
memory, a 6% increase in the progress of the superior working
memory, and a 1% reduction of inattentiveness. Among our
assessed exposure measures, we observed the strongest associa-
tions for greenness within or surrounding school boundaries.
Children spend a considerable part of their active daily time at
schools and “green exercise”has been related to better mental
health (27). Furthermore, the combination of physical activity in
school with daily peaks of TRAPs in urban areas that often coincide
with school time could result in a considerable inhaled dose of air
pollutants at school. Consistently, in our other BREATHE analysis
oftheimpactofTRAPsoncognitive development using the same
measures of cognitive development as in this study, we also ob-
served stronger associations for levels at school compared with
those at home (5). Therefore, the ability of school greenness in
reducing pollutant levels (4) might explain, in part, why we observed
the strongest associations for school greenness.
We found some indications for an enhanced 12-mo progress in
working memory associated with commuting greenness. Because
of the strong correlation between greenness surrounding school
boundaries and commuting greenness, it was not possible to
determine the independent impact of commuting greenness (i.e.,
whether commuting greenness is a surrogate for school sur-
rounding greenness). Therefore, our findings for commuting
greenness should be interpreted with caution. To the best of our
knowledge, this study is the first reporting on the potential im-
pact of commuting greenness on health in general and on cog-
nitive development in particular. We hypothesize that green
exercise and visual access to greenness might underlie such an
association, if any.
The beneficial associations for 12-mo progress in cognitive
functions were stronger than those at baseline. Baseline esti-
mates reflected the association between cognitive test scores at
the first visit and the cumulative green space exposure preceding
the study period, whereas our exposure assessment was based
Table 1. Description of the cognitive outcomes in children [median (25th–75th %)]
Working memory (WM)
(two-back numbers), d′*
(three-back numbers), d′*
First visit 2,278 8.5 206 (129, 360) 112 (53, 171) 271 (205, 338)
Second visit 2,425 8.7 221 (129, 392) 112 (59, 190) 250 (186, 321)
Third visit 2,347 9.1 234 (129, 392) 128 (59, 190) 247 (183, 317)
Fourth visit 2,307 9.4 253 (152, 392) 129 (64, 210) 228 (165, 294)
*The n-back d′is a measure of detection subtracting the normalized false alarm rate from the hit rate [(Z hit rate −Z false alarm rate) ×100].
Hit reaction time SE (HRT-SE), SE of reaction time for correct responses as a measure of response speed consistency throughout the test.
Dadvand et al. PNAS
June 30, 2015
on the home address of participants and the school they were
attending during the study period, not including potential prior
different addresses or schools to their current ones. Part of our
observed larger estimates for 12-mo progress might therefore
reflect better characterization of exposure, but it could also be
due to the window of vulnerability for these high executive func-
tions that develop significantly during the primary school age
(12–14). This window of vulnerability might also explain why we
observed the strongest associations for 12-mo progress in superior
working memory that develops considerably during this period.
We did not observe any statistically significant difference in
12-mo progress in working memory and superior working
memory (for which we found associations with green space ex-
posure) between strata of maternal education. Moreover, further
adjustment of our analyses for other indicators of SES like pa-
rental employment, marital status, and ethnicity (SI Appendix, SI
Methods) did not change the interpretation of our findings no-
tably. Furthermore, removing SES indicators (maternal educa-
tion and neighborhood SES) from our fully adjusted models did
not result in a considerable change in the interpretation of our
findings (SI Appendix, Table S6). Additionally, we did not ob-
serve any statistically significant effect modification by maternal
education or neighborhood SES for our associations (P>0.1).
These observations might suggest that our results were unlikely
to have been affected by residual SES confounding.
Available Evidence and Potential Underlying Mechanisms. We are
not aware of previous epidemiological studies on the impact of
green space exposure on cognitive development in schoolchildren;
therefore, it is not possible to compare our findings with those of
others. Our findings, however, are consistent with several previous
observations. Residential surrounding greenness has been related
to better mental health including lower risk of depression and
anxiety in children (28). Higher school greenness has been asso-
ciated with better student performance at schools (29). Experi-
mental studies have shown walking in nature or watching photos
of nature could improve directed-attention abilities in adults (30)
and have “therapeutic effects”on attention deficit hyperactivity
disorder symptoms in children (31–34). Our previous cross-sec-
tional analysis of BREATHE participants showed a protective
impact of home and school greenness on behavioral problems
including hyperactivity and inattention (35). That analysis was
based on behavioral screening questionnaires rated by teachers
and parents. In those questionnaires behavioral aspects that
characterized hyperactivity/inattention were modestly correlated
(Spearman’s correlation coefficients ranging between 0.18 and 0.23)
with the ANT inattentiveness score (at baseline) used in this study.
A study by Wells (2000) reported that relocation to residences with
higher “naturalness”improved cognitive function in a sample of 17
children (36). In an analysis of BREATHE schools, we observed
that higher greenness inside and surrounding school boundaries was
associated with lower TRAPs levels at schools (5), in line with our
other study showing lower levels of personal exposure to TRAPs
(based on personal monitors) associated with higher residential
surrounding greenness in Barcelona (22). Another BREATHE
analysis, using the same cognitive measures as the current study,
demonstrated that higher levels of TRAPs at school were associated
with diminished 12-mo cognitive progress (5). Thus, reduction of
exposure to TRAPs associated with higher greenness could have
partly underlain our observed associations. Consistently, in the
current analysis we observed that including a TRAP (EC) in our
models could explain one-fifth to two-thirds of the associations,
suggesting that our observed beneficial associations between green-
ness exposure and cognitive development could have been partly
also suggest that other mechanisms may account for 35–80% of our
observed associations that was not explained by reduction in TRAP
exposure. Higher ambient noise has been related with adverse im-
pacts on cognitive development (7). The ability of green spaces to
Fig. 1. Twelve-month progress (with 95% confidence bands) in superior
working memory for participants with the first (low greenness) and third
(high greenness) tertiles of greenness within the school boundaries.
Table 2. Adjusted difference (95% confidence interval) in baseline and 12-mo progress of working memory, superior working
memory, and inattentiveness per one interquartile range (IQR) change in greenness
(2-back number stimuli, d′)
Superior working memory
(3-back number stimuli, d′)
Median (IQR) Baseline Progress Baseline Progress Baseline Progress
Home 0.091 (0.053) 0.2 (-3.8, 4.2) 0.7 (-2.6, 4.1) 0.6 (-2.5, 3.7) −0.1 (-2.7, 2.6) 2.0 (-1.4, 5.4) −0.7 (-3.1, 1.7)
Within 0.094 (0.085) 0.3 (−6.8, 7.4) 9.8 (5.2, 14.0)* 0.9 (−5.0, 6.8) 6.9 (3.4, 10.0)* −4.0 (−12.0, 4.0) −3.4 (−6.6, −0.2)*
0.100 (0.120) 3.2 (−4.3, 11) 9.5 (4.5, 15.0)* 1.5 (−4.8, 7.8) 6.3 (2.3, 10.0)* −5.1 (−14.0, 3.6) −3.7 (−7.3, −0.1)*
Commuting 0.100 (0.062) 1.5 (−3.5, 6.6) 4.9 (1.0, 8.8) * 3.5 (−0.6, 7.5) 3.1 (0.0, 6.1) 0.2 (−4.5, 4.9) −1.2 (−4.0, 1.7)
0.094 (0.073) 0.0 (−6.9, 6.5) 9.8 (5.0, 15.0)* 1.7 (−4.4, 7.8) 6.7 (2.8, 11.0)* −2.4 (−9.8, 4.9) −3.9 (−7.4, −0.4)*
Difference adjusted for age, sex, maternal education, and residential neighborhood socioeconomic status with school and subject as nested random effects.
Fifty-meter buffer around school boundaries.
www.pnas.org/cgi/doi/10.1073/pnas.1503402112 Dadvand et al.
reduce noise (6) might therefore explain a part of our observed
associations (37). Moreover, proximity to green spaces has been
reported to increase physical activity (38), and physical activity has
been associated with better cognitivefunctioninchildren(9).Fur-
thermore, parental psychological stress and depression have been
reported to be adversely associated with cognitive development in
their children (39) and exposure to green space has been associated
with evidence of stress restorative effects and reduced depression in
adults (3, 28). A growing body of evidence also suggests that a failure
of the immunoregulatory pathways due to a reduced exposure to
macroorganisms and microorganisms in Westernized populations
might play a role in impairment of brain development (10, 40) with
childhood as a particular window of vulnerability (41). Therefore,
the ability of outdoor surrounding greenness to enhance immuno-
regulation-inducing microbial input from the environment (10)
could have been another mechanism underlying our observed as-
sociation between greenness exposure and cognitive development.
Implications for Policymakers. Approximately one-half of the world
population lives in cities, and it is projected that by 2030, three of
every five persons will live in urban areas worldwide (42). Urban
areas are characterized by a network of nonnatural built-up in-
frastructures with increased pollutant levels and less green en-
vironments (43). Children’s exposure to these pollutants such as
air pollution and noise has been associated with detrimental
impacts on their cognitive development. Our findings suggest for
a beneficial impact of green space exposure on cognitive de-
velopment, with part of this effect resulting from buffering
against such urban environmental pollutants. This impact was
more evident for surrounding greenness at school and for
working memory and superior working memory, which are pre-
dictors of learning and academic attainment (44). Schoolchildren
with a superior working memory progress of less than one-10th
of a percentile (45) of the distribution can be classified as im-
paired superior working memory progress. Our results suggest that
if schools increased greenness within their boundaries by the ob-
served IQR (Fig. 1), then 8.8% of children with impaired superior
working memory progress would move out of this category. Our
findings, therefore, hold importance for policymakers when trans-
lating evidence into feasible and achievable targeted interventions
such as improving greenness at schools, given that improved cog-
nitive development in children attending schools with more green-
ness could result in an advantage in mental capital, which, in turn,
would have lasting effects through the life-course.
Strengths and Limitations of Study. This study was based on re-
peated computerized tests of cognitive development to quantify
different aspects of cognitive development in study participants.
These tests have been reported to have acceptable internal
consistency, reasonable factorial structure, and good criterion
validity and statistical dependencies for use in general population
(46). We applied one of the most comprehensive approaches to
date to assess exposure to green space by characterizing the
outdoor surrounding greenness at home and school and during
commuting by using high-resolution (5 m ×5 m) satellite data
on greenness, enabling us to account for small-area green spaces
(e.g., home gardens, street trees, and green verges) in a stan-
Our study also faced some limitations. The generalizability of
our findings might have been affected by selection bias in that
those participants participated in BREATHE were different
from those not participated with respect to SES. Approximately
58% of mothers in our study population had a university degree,
which was higher that the regional average of 50% among
women between 25 and 39 y old living in Barcelona (47). We did
not, however, observe any indication of effect modification by
maternal education in our associations. Moreover, the Urban
Vulnerability Index of the schools was not associated with school
participation rate (Spearman’s correlation coefficient =−0.09, P=
0.61); these observations might suggest that the socioeconomic
status was less likely to be a major predictor of participating in
the study. Similarly, school greenness was not associated with
participation rate at schools (Spearman’s correlation coefficients
of −0.06 with Pvalue =0.72 for greenness within school
boundaries and 0.13 with Pvalue =0.43 for greenness sur-
rounding schools). Our exposure assessment focused on expo-
sure during the school age, overlooking other potential windows
of susceptibility such as prenatal and preschool periods. In-
vestigating these windows of susceptibility presents an opportu-
nity for future studies. By using an NDVI map obtained at a
single point in time (2012), we effectively assumed that the
spatial distribution of NDVI across our study region remained
constant over the study period (2012). The findings of our pre-
vious studies support the stability of the NDVI spatial contrast
over seasons and years (21, 48). Finally, data were not available
for some potentially relevant confounders, such as parental
mental health status.
Exposure to outdoor surrounding greenness was associated with
a beneficial impact on cognitive development in schoolchildren.
Table 3. Difference (95% confidence interval) in 12-mo cognitive trajectory per one
interquartile range change in greenness estimated by main analyses and models further
including school indoor elemental carbon (EC) interaction with age
Outcomes/exposures Main analyses
Further adjusted for EC
Within school 9.8 (5.2, 14.0)* 8.7 (2.5, 15.0)* 20.4
Surrounding school 9.5 (4.5, 15.0)* 6.9 (0.9, 13.0)* 27.4
Superior working memory
Within school 6.9 (3.4, 10.0)* 4.9 (0.1, 9.8)* 29.0
6.3 (2.3, 10.0)* 3.3 (-1.5, 8.1) 47.6
Within school −3.4 (-6.6, -0.2)* −1.2 (-5.6, 3.2) 64.7
Surrounding school −3.7 (-7.3, -0.1)* −1.8 (-6.1, 2.5) 51.4
Adjusted for age, sex, maternal education, and residential neighborhood socioeconomic status with school and
subject as nested random effects.
Estimates per 0.085 and 0.120 change respectively in greenness within and surrounding school boundaries (i.e.,
Fifty-meter buffer around school boundaries.
Dadvand et al. PNAS
June 30, 2015
These associations were only partly mediated by reduction in
TRAP levels, suggesting that other mechanisms likely underlie
this association. Our observed beneficial associations were
consistent for working memory, superior working memory, and
inattentiveness and were more evident for greenness at school.
Further studies are warranted to replicate our findings in
other settings with different climates and to investigate other
cognitive functions with different windows of susceptibility
such as prenatal and preschool periods.
ACKNOWLEDGMENTS. We thank all the families and schools participating in
the study for their altruism and their collaboration; Xavier Mayoral for the
technical support of the n-back test; and Cecilia Persavento, Judit Gonzalez,
Laura Bouso, and Pere Figueras for conducting the field work. The research
leading to these results has received funding from the European Research
Council (ERC) under ERC Grant Agreement 268479—the BREATHE project.
The research (PHENOTYPE) leading to the methodology applied for the ex-
posure assessment in this study has received funding from the European
Community’s Seventh Framework Program (FP7/2007-2013) under Grant
Agreement 282996. P.D. is funded by Ramón y Cajal Fellowship RYC-2012-
10995 awarded by the Spanish Ministry of Economy and Competitiveness.
1. Kahn PH, Kellert SR (2002) Children and Nature: Psychological, Sociocultural, and
Evolutionary Investigations (MIT Press, Cambridge, MA).
2. Kellert SR (2005) Building for Life: Designing and Understanding the Human-Nature
Connection (Island, Washington).
3. Bowler DE, Buyung-Ali LM, Knight TM, Pullin AS (2010) A systematic review of evi-
dence for the added benefits to health of exposure to natural environments. BMC
Public Health 10:456.
4. Dadvand P, et al. (2015) The association between greenness and traffic-related air
pollution at schools. Sci Total Environ 523:59–63.
5. Sunyer J, et al. (2015) Association between traffic-related air pollution in schools and
cognitive development in primary school children: A prospective cohort study. PLoS
6. Gidlöf-Gunnarsson A, Öhrström E (2007) Noise and well-being in urban residential
environments: The potential role of perceived availability to nearby green areas.
Landsc Urban Plan 83(2):115–126.
7. Klatte M, Bergström K, Lachmann T (2013) Does noise affect learning? A short review
on noise effects on cognitive performance in children. Front Psychol 4:578.
8. James P, Banay RF, Hart JE, Laden F (2015) A review of the health benefits of
greenness. Curr Epidemiol Reports 2(2):131–142.
9. Fedewa AL, Ahn S (2011) The effects of physical activity and physical fitness on
children’s achievement and cognitive outcomes: A meta-analysis. Res Q Exerc Sport
10. Rook GA (2013) Regulation of the immune system by biodiversity from the natural
environment: An ecosystem service essential to health. Proc Natl Acad Sci USA
11. Grandjean P, Landrigan PJ (2014) Neurobehavioural effects of developmental toxic-
ity. Lancet Neurol 13(3):330–338.
12. Anderson P (2002) Assessment and development of executive function (EF) during
childhood. Child Neuropsychol 8(2):71–82.
13. Ullman H, Almeida R, Klingberg T (2014) Structural maturation and brain activity
predict future working memory capacity during childhood development. J Neurosci
14. Østby Y, Tamnes CK, Fjell AM, Walhovd KB (2011) Morphometry and connectivity of
the fronto-parietal verbal working memory network in development. Neuro-
15. Jaeggi SM, Buschkuehl M, Perrig WJ, Meier B (2010) The concurrent validity of the
N-back task as a working memory measure. Memory 18(4):394–412.
16. Rueda MR, Rothbart MK, McCandliss BD, Saccomanno L, Posner MI (2005) Training,
maturation, and genetic influences on the development of executive attention. Proc
Natl Acad Sci USA 102(41):14931–14936.
17. Rueda MR, et al. (2004) Development of attentional networks in childhood. Neuro-
18. Shelton JT, Elliott EM, Matthews RA, Hill BD, Gouvier WD (2010) The relationships of
working memory, secondary memory, and general fluid intelligence: Working
memory is special. J Exp Psychol Learn Mem Cogn 36(3):813–820.
19. Conners CK, Staff MHS (2000) Conners’Continuous Performance Test II: Computer
Program for Windows Technical Guide and Software Manual (Mutli-Health Systems,
North Tonwanda, NY).
20. Weier J, Herring D (2011) Measuring Vegetation (NDVI & EVI) (Natl Aeronaut Space
Admin, Greenbelt, MD).
21. Dadvand P, et al. (2012) Surrounding greenness and pregnancy outcomes in four
Spanish birth cohorts. Environ Health Perspect 120(10):1481–1487.
22. Dadvand P, et al. (2012) Surrounding greenness and exposure to air pollution during
pregnancy: An analysis of personal monitoring data. Environ Health Perspect 120(9):
23. Spanish Ministry of Public Works (2012) Atlas of Urban Vulnerability in Spain.
Methodology and Contents, ed Aja AH (Spanish Ministry of Public Works, Madrid).
24. Preacher KJ, Kelley K (2011) Effect size measures for mediation models: Quantitative
strategies for communicating indirect effects. Psychol Methods 16(2):93–115.
25. Amato F, et al. (2014) Sources of indoor and outdoor PM2.5 concentrations in primary
schools. Sci Total Environ 490:757–765.
26. Rivas I, et al. (2014) Child exposure to indoor and outdoor air pollutants in schools in
Barcelona, Spain. Environ Int 69:200–212.
27. Thompson Coon J, et al. (2011) Does participating in physical activity in outdoor
natural environments have a greater effect on physical and mental wellbeing than
physical activity indoors? A systematic review. Environ Sci Technol 45(5):1761–1772.
28. Maas J, et al. (2009) Morbidity is related to a green living environment. J Epidemiol
Community Health 63(12):967–973.
29. Wu C-D, et al. (2014) Linking student performance in Massachusetts elementary
schools with the “greenness”of school surroundings using remote sensing. PLoS ONE
30. Berman MG, Jonides J, Kaplan S (2008) The cognitive benefits of interacting with
nature. Psychol Sci 19(12):1207–1212.
31. van den Berg AE, van den Berg CG (2011) A comparison of children with ADHD in a
natural and built setting. Child Care Health Dev 37(3):430–439.
32. Taylor AF, Kuo FE (2009) Children with attention deficits concentrate better after
walk in the park. J Atten Disord 12(5):402–409.
33. Taylor AF, Kuo FE, Sullivan WC (2001) Coping with ADD: The surprising connection to
green play settings. Environ Behav 33(1):54–77.
34. Kuo FE, Taylor AF (2004) A potential natural treatment for attention-deficit/hyper-
activity disorder: Evidence from a national study. Am J Public Health 94(9):1580–1586.
35. Amoly E, et al. (2014) Green and blue spaces and behavioral development in Barce-
lona schoolchildren: The BREATHE project. Environ Health Perspect 122(12):
36. Wells NM (2000) At home with nature effects of greenness on children’s cognitive
functioning. Environ Behav 32(6):775–795.
37. Stansfeld SA, et al.; RANCH study team (2005) Aircraft and road traffic noise and
children’s cognition and health: A cross-national study. Lancet 365(9475):1942–1949.
38. Lee AC, Maheswaran R (2011) The health benefits of urban green spaces: A review of
the evidence. J Public Health (Oxf) 33(2):212–222.
39. Ramchandani P, Psychogiou L (2009) Paternal psychiatric disorders and children’s
psychosocial development. Lancet 374(9690):646–653.
40. Rook GAW, Lowry CA, Raison CL (2013) Microbial ’Old Friends’, immunoregulation
and stress resilience. Evol Med Public Health 2013(1):46–64.
41. Rook GAW, Lowry CA, Raison CL (April 13, 2014) Hygiene and other early childhood
influences on the subsequent function of the immune system. Brain Res, 10.1016/j.
42. Martine G, Marshall A (2007) State of World Population 2007: Unleashing the Po-
tential of Urban Growth (United Nations Popul Fund, New York).
43. Escobedo FJ, Kroeger T, Wagner JE (2011) Urban forests and pollution mitigation:
Analyzing ecosystem services and disservices. Environ Pollut 159(8-9):2078–2087.
44. Alloway TP, Alloway RG (2010) Investigating the predictive roles of working memory
and IQ in academic attainment. J Exp Child Psychol 106(1):20–29.
45. Lezak MD, Howieson DB, Loring DW, Hannay HJ, Fischer JS (2004) Neuropsychological
Assessment (Oxford Univ Press, New York).
46. Forns J, et al. (2014) The n-back test and the attentional network task as measures of
child neuropsychological development in epidemiological studies. Neuropsychology
47. Barcelona City Council (2013) Statistical Yearbook of Barcelona City. Year 2013.
(Barcelona City Council, Barcelona).
48. Dadvand P, et al. (2014) Inequality, green spaces, and pregnant women: Roles of
ethnicity and individual and neighbourhood socioeconomic status. Environ Int 71:
www.pnas.org/cgi/doi/10.1073/pnas.1503402112 Dadvand et al.