CTN 04 00021

Reviews/Mini-Reviews
Clinical & Translational Neuroscience

July-December 2020: 1–21
ª The Author(s) 2020
Neuroplasticity in children and Article reuse guidelines:
sagepub.com/journals-permissions
adolescents in response to treatment DOI: 10.1177/2514183X20974231
journals.sagepub.com/home/ctn
intervention: A systematic review

of the literature
Lisa L Weyandt1, Christine M Clarkin2,3 , Emily Z Holding4,

Shannon E May3, Marisa E Marraccini4,
Bergljot Gyda Gudmundsdottir5, Emily Shepard6,
and Lauren Thompson3
Abstract
The purpose of the present study was to conduct a systematic review of the literature, adhering to PRISMA guidelines,
regarding evidence of neuroplasticity in children and adolescents in response to cognitive or sensory-motor interventions.
Twenty-eight studies employing seven different types of neuroimaging techniques were included in the review. Findings
revealed that significant variability existed across the 28 studies with regard to the clinical populations examined, type of
interventions employed, neuroimaging methods, and the type of neuroimaging data included in the studies. Overall, results
supported that experience-dependent interventions were associated with neuroplastic changes among children and ado-
lescents in both neurotypical and clinical populations. However, it remains unclear whether these molecular neuroplastic
changes, including the degree and direction of those differences, were the direct result of the intervention. Although the
findings are encouraging, methodological limitations of the studies limit clinical utility of the results. Future studies are
warranted that rigorously define the construct of neuroplasticity, establish consistent protocols across measurement
techniques, and have adequate statistical power. Lastly, studies are needed to identify the functional and structural neu-
roplastic mechanisms that correspond with changes in cognition and behavior in child and adolescent samples.
Keywords
Neuroplasticity, neuroimaging, children, adolescents, experience-dependent, sensory-motor, fMRI
Introduction
Broadly speaking, neuroplasticity refers to the ability of the 1
Department of Psychology, Director Interdisciplinary Neuroscience
brain to undergo morphological and neurochemical Program, University of Rhode Island, Kingston, RI, USA
changes as a result of experience. A variety of definitions 2
Physical Therapy Department, University of Rhode Island, Kingston,
on neuroplasticity exist, each emphasizing different ele- RI, USA
3
ments. For example, Cramer et al. defined neuroplasticity Interdisciplinary Neuroscience Program, Graduate School, University of
Rhode Island, Kingston, RI, USA
as “the ability of the nervous system to respond to intrinsic 4
School of Education, University of North Carolina at Chapel Hill, Chapel
or extrinsic stimuli by reorganizing its structure, function Hill, NC, USA
and connections” (p. 1591), while Sarrasin and colleagues’ 5
School of Education, University of Iceland, Reykjavik, Iceland
6
more recent definition focuses on “the capacity of the brain Department of Psychology, University of Rhode Island, Kingston, RI, USA
to modify its neural connections through learning” (p.
Corresponding author:
23).1,2 An abundance of human and other animal research Lisa L Weyandt, Department of Psychology, University of Rhode Island, 142
supports the ability of the brain to change in response to Flagg Road, Kingston, RI 02881, USA.
environmental stimuli and this change can be adaptive Email: [email protected]
Creative Commons Non Commercial CC BY-NC: This article is distributed under the terms of the Creative Commons
Attribution-NonCommercial 4.0 License (https://creativecommons.org/licenses/by-nc/4.0/) which permits non-commercial use,
reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open
Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).
2 Clinical & Translational Neuroscience
(improvement in function) or maladaptive (loss of func- research support the presence of neuroplasticity among
tion).3–8 These underlying cellular, morphological, struc- children and adolescents in the context of experience-
tural, and functional changes are believed to be the result of dependent interventions and, if so, what are those changes
a complex interplay between genetic, biological, psycholo- and how do they relate to functional outcomes?; and (4)
gical, and environmental factors.9,10 What are the implications for future research?
Developmentally, human brain growth is mediated by
genetic and environmental factors from the moment of con-
ception; however, historically it was believed that the brain Methods
underwent little change beyond childhood. Indeed, it was not Search strategies
until the late 1970s that research substantiated that the pre-
frontal cortex undergoes structural changes during adoles- A systematic review of the literature was conducted in
cence.11 Technological advances in brain imaging have January through March 2018, adhering to PRISMA guide-
enabled researchers to demonstrate aspects of brain devel- lines.24 The review process included a comprehensive
opment that continue throughout childhood and adolescence; search of the following databases: Scopus, PubMed, and
however, the ways in which these changes unfold as well as PsychINFO. To identify articles, we used two main search
their effects on behavior and cognition are less clear.12–18 queries: Plasticity þ Neuroimag* þ Child* (P þ N þ
Since the establishment of the Decade of the Brain in the Child*) and Plasticity þ Neuroimag* þ Adolesc* (P þ
1990s, there has been a fivefold increase in neuroplasticity- N þ Adolesc*).
related research, and similarly, an increased interest in the
topic by the media and popular press.19,20 Eligibility criteria
Understanding the mechanisms underlying neuroplasticity
Studies were included in the present review based on the
is important from a basic and an applied perspective.
following criteria: (1) original research; (2) included a
For example, ongoing research in the field of neuroplasticity
treatment intervention that aimed to facilitate neuroplasti-
indicates that negative environmental triggers (e.g. inade-
city and a neuroimaging technique (i.e. functional magnetic
quate care, poverty, neglect, and stress) can result in
resonance imaging (fMRI), diffusion tensor imaging (DTI),
increased vulnerability to psychological disorders or an
and magnetic resonance imaging (MRI)) both with pre- and
impaired stress response later in life, presumably in part
post-measures; (3) was published in English; (4) included
due to morphological and functional brain changes.21,22 On
human participants only (i.e. animal studies were
the other hand, the brain’s ability to adapt both structurally
excluded); (5) included children and/or adolescents (i.e.
and functionally makes childhood the preferred period for
18 or younger); and (6) articles were published between
many surgical interventions (such as hemispherectomies or
January 2008 and March 2018. Articles were excluded if
cochlear implants) that are less effective if delayed until
they met one or more of the following exclusion criteria:
adulthood.23 Understanding neuroplasticity during child-
review article, case study, open study, method or proof-of-
hood and adolescence and investigating methods to foster
concept paper, no treatment intervention utilized, and not
this process across the life span could help to counteract
written in English.
negative outcomes associated with environmental factors
and possibly neurodegeneration later in life.
In response to the body of research that supports that the Data collection and extraction
brain undergoes substantial growth and development dur- The eligibility of the studies was examined by two investi-
ing childhood and adolescence, and the increased interest in gators using a standardized data extraction form, and full
neuroplasticity in the scientific literature and popular press, consensus was reached on the studies included in the review.
the purpose of the present study was to conduct a systema- Information including publication year, sample size, and
tic review of the literature addressing neuroplasticity population characteristics was collected along with two
among children and adolescents using neuroimaging tech- other key components—intervention and imaging data.
niques. A second purpose of the article was to critically A description of the type of intervention including frequency
evaluate the methodological strengths and weaknesses of and duration and pre and post-intervention measures along
these studies in order to provide direction to future with type of neuroimaging performed, regions of interest
research. Specifically, the present systematic review sought (ROIs), and pre and post-measures was extracted from each
to address whether neuroplasticity occurs among children individual study.
and adolescents with and without clinical disorders in
response to experience-dependent intervention based on
neuroimaging findings. The research questions were as fol- Results
lows: (1) What are the key measures and analytical tech-
niques used within this neuroplasticity literature?; (2) What
Search results
are the associated methodological strengths and limitations The search yielded 1122 articles (Scopus ¼ 387, PubMed ¼
from this area of research?; (3) Does current neuroimaging 520, and PsychINFO ¼ 215). Duplicates within each
Weyandt et al. 3
Figure 1. Flow diagram of the selection process of studies on neuroplasticity in children and adolescents (PRISMA, Moher et al.24).
database search were removed (Scopus ¼ 94, PubMed ¼ 12, article review. Upon further review, three additional articles
and PsychINFO ¼ 20) and an additional 7 duplicates were excluded as they did not include a treatment interven-
between the databases were removed. After removing jour- tion with pre- and post-measures, and an additional article
nal articles that did not meet inclusion criteria, 82 article was excluded because it was a case study. This process
titles and abstracts remained. Of the 82 articles, 32 met the resulted in a final total of 28 articles meeting all inclusion
inclusion criteria of the study and were eligible for full criteria (see Figure 1).
Baseline characteristics of studies Typical grouping (N-Typ). 29,31–34,36–39 Studies that

included individuals with a diagnosed developmental dis-
Significant variability existed across the 28 studies with
ability (n ¼ 8) were assigned into the Clinical Neurodeve-
regard to the type of clinical population studied, the type
lopmental (C-Dev) group. These studies included
of experience-dependent intervention employed, interven-
individuals that were born preterm or were diagnosed with
tion frequency and delivery method, timing of assess-
attention deficit hyperactivity disorder (ADHD), learning
ments, neuroimaging method, and type of data extracted
disabilities, obsessive compulsive disorder (OCD), math
from that imaging. Table 1 provides a summary of the
anxiety, or autism spectrum disorder.25,30,40–45 Finally,
study characteristics for the 28 included studies. The
studies that assessed individuals with a neurological injury
majority of the studies (57%) were conducted in the
diagnosis were assigned into the Clinical Neurological
United States (n ¼ 16), three in Canada, one in the United
Injury (C-Inj) group. This included diagnoses of apraxia,
Kingdom, two in the Netherlands, one in Germany, one in
traumatic brain injury, and cerebral palsy including spastic
Switzerland, one in Spain, one in Iran, one in Israel, and
hemiplegia.26–28,35,46–52
one in Japan. Most of the studies conducted in the United
States were conducted in California and Massachusetts
(four studies in each state), two in Alabama, and one study Description of interventions
each in New Jersey, New York, Ohio, Pennsylvania, All 28 studies examined the effects of an experience-
Texas, and Wisconsin. Studies were published between dependent intervention employing a variety of behavioral
2008 and 2018; the median publication year was 2015. outcome measures. These interventions were either cogni-
Eighty-two percent of the studies (n ¼ 23) were published tive (n ¼ 13) or sensory-motor focused (n ¼ 15) (see
with in the last 5 years. Table 1). Four studies in the N-Typ category,29,31,33,39 eight
studies in the C-Dev category,25,30,40–45 and one study in
Plasticity terminology and measurement the C-Inj category28 used cognitive training interventions.
Five studies in the N-Typ category32,34,36–38 implemented
Among the 28 reviewed studies, broad and more specific
sensory-motor interventions using either music or auditory-
terms for neuroplasticity were used. Most of the studies
related interventions. There were no studies in the C-Dev
used the term plasticity (n ¼ 13) or neuroplasticity (n ¼ 8).
category that examined experience-dependent plasticity
Alternative terms included neural change, neuronal change,
among children and adolescents using a sensory-motor inter-
neural plasticity, functional connectivity, brain connectivity,
vention. Ten studies in the C-Inj category examined
experience-dependent structural change, experience-
experience-dependent plasticity among children and adoles-
dependent plasticity, and developmental plasticity.25–34
cents using a sensory-motor intervention.26,27,35,46–52
Furthermore, even broader terms for plasticity were used
The type of intervention appeared to be closely corre-
in some studies, such as neuromodulation.35 The terminol-
lated with the clinical population that was being studied
ogy of plasticity in studies correspondingly influenced the
(see Figure 2). For example, children and adolescent parti-
selection of imaging modality.
cipants that had a motor impairment such as speech apraxia
were treated with a sensory-motor intervention which
Description of samples focused on improved motor function, while individuals
Sample sizes ranged from 4 to 235 (M ¼ 34.36, SD ¼ with a learning disability were treated with a cognitive-
41.63) resulting in a total of 962 participants. There was based intervention focused on those skills (i.e. reading or
one outlier study with 235 participants. Excluding that math). Nonclinical populations assessed both types of inter-
study is more reflective of the remaining 27 studies with ventions approximately equally.
a sample size ranging from 4 to 65 (M ¼ 26.92, SD ¼
13.94) resulting in a total of 727 participants. Gender (bin- Description of intervention delivery, frequency,
ary only) was reported in 25 of the studies for a total male/ and duration
female ratio of 463/408. Age was reported in all 28 studies;
mean age ranged from 0.59 to 17.3 years. Across all stud- The studies in this review demonstrated inconsistent report-
ies, the mean of mean ages was 9.43 years. ing regarding the intensity, volume, and frequency of the
interventions employed. Six studies did not report any of
this information while the remaining studies reported signif-
Description of study populations icant variability ranging from 20 min one time a week to 6 h
The high degree of heterogeneity of sample participants a day for 5 days a week. Duration and mode of delivery of
across all 28 studies led the authors to categorize each study interventions was reported in all 28 studies. The median
into one of three groups based on the clinical or nonclinical duration was 8 weeks; however, the majority of interven-
diagnoses (see Figure 2). tions (79%) were 16 weeks or less (see Table 2). Five of the
All studies assessing nonclinical, that is, neurotypical 28 studies, or 18% of the articles, investigated the effects of
populations (n ¼ 9) were assigned into the Nonclinical long-duration interventions that lasted longer than 16 weeks
Table 1. Review of studies assessing neuroplasticity in children and adolescents: Summary of study characteristics.
Comparison Imaging
Author Location Population N group Ages (years) Gender (M/F) Race/SES Intervention Imaging method category
Alves-Pinto et al.26 Germany C–Inj 16 Yes 8 to 16 and 15 to 17 10/6 NR Sensory-motor Task-based MRI ACT
Music
Amad et al.36 England N–Typ 31 Yes 16.8 and 17.8 15/16 NR Sensory-motor Resting-state fMRI CON
Drumming
Azizi et al.47 Iran C–Inj 4 Yes 9, 11, 12, and 4 NR NR Sensory-motor DTI STR
Alter G treadmill Structural MRI
Bakhtiari et al.48 Canada C–Inj 16 Yes 11.6 10/6 NR Sensory-motor Task-based fMRI CON
LSV T LOUD
Benasich et al.34 USA (NJ) N–Typ 49 Yes 0.59 (7.1 months) 26/16 Yes Sensory-motor EEG ACT
Passive acoustic
Cao et al.49 USA (TX) C–Inj 11 Yes 10.2 and 9.8 7/7 NR Sensory-motor fNIRS ACT
CIMT
Carlson et al.35 Canada C–Inj 34 Yes 12.1 and 12.72 21/14 NR Sensory-motor fMRI ACT
CIMT and HABIT MR spectroscopy
Transcranial direct current
stimulation
Cope et al.27 USA (WI) C–Inj 15 Yes 11 6/9 NR Sensory-motor Task-based fMRI ACT
CIMT and HABIT
Everts et al.25 Switzerland C–Dev 23 Yes 10 and 10.7 and 9.8 11/12 Yes Cognitive Task-based fMRI ACT
Memory
Friel et al.50 USA (NY) C–Inj 20 Yes 9.2 and 8.2 12/9 Yes Sensory-motor Structural MRI STR
HABIT Single-pulse transcranial
magnetic stimulation
Motor evoked potentials
(MEP)
Habibi et al.37 USA (CA) N–Typ 35 Yes 6.68 and 7.15 and 7.16 26/12 Yes Sensory-motor EEG ACT
Music vs. Sports
Hoekzema et al.40 Spain C–Dev 18 Yes 11.22 and 11.33 15/3 NR Cognitive Structural MRI STR
Huyser et al.44 Netherlands C–Dev 58 Yes 13.78 and 13.6 22/36 NR Cognitive Structural MRI STR
CBT
Hyde et al.38 USA (MA) N–Typ 31 Yes 6.32 and 5.9 15/16 Yes Sensory-motor Structural MRI STR
Music
Iuculano et al.41 USA (CA) C–Dev 30 Yes 7.5 to 9.6 NR NR Cognitive Structural MRI ACT
Math Wise fMRI
Jolles et al.29 Netherlands N–Typ 24 Yes 12.24 and 22.04 Children ¼ 4/5; NR Cognitive Resting state ACT
adults ¼ 7/8 Memory Event-related fMRI
Jolles et al.39 USA (CA) N–Typ 18 No 7.7 to 9.1 7/11 NR Cognitive DTI CON
Math tutoring Task-related fMRI
Kadis et al.46 Canada C–Inj 28 Yes 4.5 and 4.1 17/11 NR Sensory-Motor Vertex-based MRI STR
PROMPTs
(continued)
5
6
Table 1. (continued)
Comparison Imaging
Author Location Population N group Ages (years) Gender (M/F) Race/SES Intervention Imaging method category
Matsudaira et al.33 Japan N–Typ 235 No 10.65 116/109 Yes Cognitive Structural MRI STR
Parental Praise
Maximo et al.30 USA (AL) C–Dev 28 Yes 10 and 11 24/4 NR Cognitive Task-based fMRI CON
Language
Comprehension
Meyler et al.42 USA (PA) C–Dev 28 Yes 10.8 and 10.8 6/22 NR Cognitive Task-based fMRI ACT
Reading Program
Romeo et al.43 USA (MA) C–Dev 65 Yes 7.75 43/22 Yes Cognitive Structural MRI STR
Reading program
Rosenberg-Lee et al.31 USA (CA) N–Typ 34 Yes 8.5 and 8.8 13/21 NR Cognitive Task-based fMRI ACT
Math Wise and Galaxy
Math
Schlaug et al.32 USA (MA) N–Typ 33 Yes 6.49 Not provided NR Sensory-motor Task-based fMRI STR
Music
Sterling et al.51 USA (AL) C–Inj 10 No 3.4 6/4 No Sensory-motor Structural MRI STR
CIMT
Supekar et al.45 USA (CA) C–Dev 28 Yes 8.51 and 8.68 HMA: 6/8; LMA: NR Cognitive Task-based fMRI ACT
6/8 Math Tutoring
Weinstein et al.52 Israel C–Inj 12 No 10.3 and 12.45 6/6 No Sensory-motor fMRI ACT
Magic HABIT DTI
Yuan et al.28 USA (OH) C–Inj 28 Yes 13.72 and 13.37 13/15 Yes Cognitive DTI CON
AIM MRI
ACT: Activation; C-Dev: Clinical Developmental; C-Inj: Clinical Injury; CON: Connectivity; DTI: diffusion tensor imaging; fMRI: functional magnetic resonance imaging; fNIRS: functional near-infrared spectroscopy; MRI:
magnetic resonance imaging; NR: not reported; N-Typ: Nonclinical-Neuro Typical; STR: Structural.
Weyandt et al. 7
Figure 2. Distribution of studies by clinical population category. Key: Non-italicized ¼ sensory-motor intervention; italicized ¼
cognitive-based intervention.
on neuroplasticity of children or adolescents.26,32,33,37,38 The MRI and DTI; and fMRI and DTI.28,35,39,41,50,52 Based
majority of the studies (75%) employed interventions that on imaging modality and measurement of plasticity, stud-
were delivered 1:1 while 11% were delivered as a group and ies were grouped into three categories. These groups
14% compared 1:1 to group. reflect how the study’s authors interpreted and measured
neuroplasticity as (a) changes in brain structure (STR)
(n ¼ 10),32,33,38,40,43,44,46,47,50,51 (b) activation of regions
Description of neuroimaging of interest (ACT) (n ¼ 13),25–27,29,31,34,35,37,41,42,45,49,52 or
There was substantial heterogeneity in imaging modality (c) functional connectivity (CON) (n ¼ 5)28,30,36,39,40 (see
employed, processing pipelines, and analytical approaches Table 1; Figure 3).
across studies (see Table 2). The neuroimaging modality or The studies identified in this systematic review measured
combination of modalities differed among the studies, neuroplasticity using single or multimodal imaging strate-
demonstrating the flexibility of neuroimaging to identify and gies based on the authors’ working definition of plasticity
localize changes in plasticity in response to an intervention. and the specific research questions. Neuroimaging modal-
MRI (n ¼ 8), fMRI (n ¼ 10), DTI (n ¼ 1), functional near- ities differ in their spatial resolution, temporal resolution,
infrared spectroscopy (fNIRS) (n ¼ 1), and electroencepha- and signal of interest that approximates or directly measures
lography (EEG) (n ¼ 2) were used by the studies reviewed neural activity. fMRI relies on the detection of blood-
here. Furthermore, several studies used multimodal imaging oxygen-level-dependent (BOLD) signal to indirectly
to investigate neural changes or brain structure, combining measure neuronal activity.53 Task-based fMRI provides
fMRI and structural MRI; DTI and task-based fMRI; fMRI, information about regions of activation in response to a cog-
magnetic resonance spectroscopy, and transcranial direct nitive, motor, or sensory task. By contrast, resting-state
magnetic stimulation (TMS); structural MRI and TMS; fMRI detects spontaneous changes in the BOLD signal
8
Table 2. Review of studies assessing neuroplasticity in children and adolescents: Summary of behavioral interventions and imaging.
Behavior
Intervention frequency/ Intervention intervention Imaging
Author Population duration delivery Intervention outcome Imaging method ROI Imaging outcome measure outcome
Alves-Pinto C-Inj 78 weeks (18 months): 1:1 Sesnory-motor Ø Task-based MRI Primary motor cortex and Activation—Neuronal dynamics
j
et al.26 30–45 min 2 a week Music cerebellum (changes in neuronal activity)
of motor network;
Connectivity—then quantified
by dynamic causal modeling
(DCM)
Amad et al.36 N-Typ 8 weeks: 30 min 3 Group Sensory-motor
j Resting-state fMRI Whole-brain connectivity Connectivity: Changes resting-
j
a week Drumming state functional connectivity
47
Azizi et al. C-Inj 8 weeks: 45 min 3 1:1 Sensory-motor
j DTI Cortical spinal tract (CST) in Structural: Percentage signal
j
a week Alter G treadmill Structural MRI brain stem and the corona change (PSC) in DTI
radiate parameters in affected and
unaffected sides of brain;
functional angioscopy (FA)
measurement of entire
corticospinal tract white
matter
Bakhtiari C-Inj 16 weeks: 1 h a day 4 days 1:1 Sensory-motor
j Task-based fMRI Anterior cingulate gyrus, Connectivity: Changes in strength
j
et al.48 a week for 4 weeks LSV T LOUD inferior frontal gyrus, of interactions between brain
cerebellum, middle regions; PCS
temporal gyrus, secondary
motor area, superior
temporal gyrus, superior
marginal gyrus, and
posterior cingulate gyrus
Benasich N-Typ 6 weeks: 20 min 1 1:1 vs. group Sensory-motor
j EEG Frontal, frontocentral, and Activation: EEG/event-related
j
et al.34 a week 6 weeks Passive acoustic central channels potentials (ERPs) latency and
amplitude
Cao et al.49 C-Inj 2 weeks: 6 h a day 1:1 Sensory-motor Ø fNIRS Premotor cortex, Activation: Cortical activation
j
5 days a week Constraint-induced supplementary motor patterns, as measured by
movement therapy area, and primary motor changes in BOLD signal using
(CIMT) and sensory cortex the fNIRS metrics—laterality
index (global measure of
hemodynamics) and time-to-
peak/duration (more localized
measure of hemodynamics)
Carlson C-Inj 2 weeks: 5 a week and 1:1 Sensory-motor
j fMRI Primary motor (M1) cortex Activation: MR spectroscopy was Ø
et al.35 Cathodal tDSC or CIMT and HABIT MR spectroscopy used to measure metabolite
sham was administered Transcranial direct levels at sites of transcranial
daily M-F for 20 min to current direct stimulation
M1 Area stimulation
Cope et al.27 C-Inj 2 weeks: 4 h a day 5 days a 1:1 Sensory-motor
j Task-based fMRI Premotor area Activation—Brain reorganization
j
week CIMT measured by PSC
(continued)
Behavior
Everts et al.25 C-Dev 4 weeks: training 60 min 1:1 Cognitive

j Task-based fMRI Left and right parietal Activation: ROI measured by PCS
j
1 a week; intensive Memory Left and right frontal and lateralization index
practice 12 min 5
a week 4 weeks
Friel et al.50 C-Inj 3 weeks: 6 h a day 1:1 Sensory-motor
j Structural MRI Motor cortex and CST Structural and Activation—
j
5 days a week Hand-arm intensive Single-pulse Changes in motor map
bimanual therapy transcranial representations in the motor
(HABIT) magnetic cortex of participants as
stimulation detected by structural MRI.
Motor evoked Neural excitability of motor
potentials (MEP) cortex was measured by
single-pulse transcranial
magnetic stimulation.
Habibi et al.37 N-Typ 104 weeks (2 years): 6 h 1:1 vs. group Sensory-motor
j EEG Central auditory system Activation: Difference in cortical
j
weekly Music vs. Sports including associative potentials/ERPs
auditory temporal regions,
anterior cingulate cortex,
Heschl’s gyrus, primary
and secondary auditory
cortices
Hoekzema C-Dev 2 weeks: 45 min 5 1:1 Cognitive Ø Structural MRI Frontal lobes, striatum, and Structural: Measure of regional
j
et al.40 a week 2 weeks cerebellum longitudinal volume changes
(voxel-based morphometry
with tensor-based
morphometric elements)
Huyser et al.44 C-Dev 16 weeks; one session 1:1 Cognitive
j Structural MRI Frontostriatal circuitry— Structural: Voxel-based
j
weekly CBT orbitofrontal cortex, morphometry—global and
lateral and medial regional gray matter and white
prefrontal cortex, anterior matter differences between
cingulate cortex, and time points
striatum
Hyde et al.38 N-Typ 64 weeks (15 months): 1:1 vs. group Sensory-motor
j Structural MRI Whole brain analysis Structural: Voxel-based
j
30 to 40 min weekly Instrumental vs. morphometry
Noninstrumental/
Singing
Iuculano C-Dev 8 weeks 1:1 Cognitive
j Structural MRI Posterior parietal, prefrontal, Structural: Voxel-based
j
et al.41 Math Wise fMRI ventral temporal-occipital morphometry in ROI
cortices (regions Activation—Support Vector
associated with Machine (SVM)—functional
mathematical learning brain activity patterns; Brain
disability compared to Plasticity Index (BPI)
typically developing
populations)
(continued)
9
10
Behavior
Jolles et al.29 N-Typ 6 weeks: 3 blocks of 1:1 Cognitive

j Resting-state MRI Frontoparietal network and Activation: Neural activation as Ø Adolescents
30 trials Memory Event-related fMRI the default network measured by BOLD signal
j Adults
Connectivity: Seed-correlation
approach to identify functional
connectivity within and
between the two networks
Jolles et al.39 N-Typ 8 weeks: 40–50 min 3 1:1 Cognitive
j DTI Left superior longitudinal Connectivity: FA Ø
a week Math tutoring Task-related fMRI fasciculus (SLF) connecting
the inferior parietal,
frontal, and temporal
regions
Kadis et al.46 C-Inj 8 weeks 1:1 Sensory-motor
j Vertex-based MRI Broca’s area, Wernicke’s Structural—Cortical thickness Ø
PROMPT ¼ Prompts area, and the posterior
for Restructuring supramarginal gyrus
Oral Motor (regions associated with
Phonetics Targets language, speech, and
voluntary oral-motor
control)
Matsudaira N-Typ 156 weeks (3 years) 1:1 Cognitive
j Structural MRI Amygdala and hippocampus Structural: Regional gray matter
j
et al.33 Parental Praise volume
Maximo C-Dev 10 weeks: 4 h sessions 1:1 Cognitive
j Task-based fMRI Left fusiform gyrus, inferior Connectivity: Local functional
j
et al.30 5 a week Visualizing and frontal gyrus, inferior connectivity assessed using
Verbalizing for occipital gyrus, superior local connection density
Language parietal lobule, middle
Comprehension and frontal gyrus, precentral
Thinking V/V gyrus, supplementary
Intervention motor area, superior
temporal gyrus, and
thalamus (reading
network)
Local connectivity was
calculated separately for
each ROI
Meyler et al.42 C-Dev 100 h of instruction; Group Cognitive
j Task-based fMRI Cortical activation (whole Activation: Brain activation ROI
j
duration of remedial Reading Program brain) measured using a
training not indicated hemodynamic response
function estimate for each
individual at each time point
(continued)
Behavior
Romeo et al.43 C-Dev 6 weeks: 4 h a day Group Cognitive

j Structural MRI Inferior frontal and posterior Structural: Cortical thickness
j
5 a week Reading program temporal regions that are using symmetrized percent
associated with language change (SPC) which is the rate
and phonological of change at each location with
processing respect to the average
thickness across both time
points
Rosenberg- N-Typ 8 weeks: 45–50 min 1:1 Cognitive
j Math Task Task-based fMRI Hippocampus Activation: Functional
j
Lee et al.31 3 a week Math Wise and Galaxy Ø Math connectivity based on cortical
Math Programs Reasoning activation patterns during task
based math solving activity
Schlaug et al.32 N-Typ 120 weeks (30 months): 1:1 vs. group Sensory-motor
j Task-based fMRI Midsagittal callosal and seven Structural: Voxel-based
j
30 min or 60 min Music midsagittal subregions morphological changes
weekly
Sterling et al.51 C-Inj 3 weeks: 3 h a day 1:1 Sensory-motor
j Structural MRI Sensorimotor cortex, Structural—gray matter voxel-
j
5 days a week CIMT anterior motor areas, and based morphological changes
hippocampus
Supekar C-Dev 8 weeks: 40–50 min 1:1 Cognitive
j Task-based fMRI Amygdala circuits Activation: Changes in amygdala
j
et al.45 3 per week Math Tutoring circuits
Weinstein C-Inj 2 weeks: 6 h a day 1:1 Sensory-motor
j fMRI Corticospinal tracts Activation—fMRI: Lateralization
j
et al.52 5 days a week Magic HABIT DTI index was calculated based on
or the total number of activated
6 weeks: 2 h a day voxels in each ROI
5 days a week
Yuan et al.28 C-Inj 10 weeks: 60–90 min 1:1 Cognitive
j DTI Superior frontal gyrus, Connectivity—global network
j
sessions weekly plus AIM (Attention MRI angular gyrus, fusiform connectivity and regional
20–40 min in home intervention and gyrus, paracentral lobule, network connectivity
practice sessions 2–3 management) supramarginal gyrus,
times per week superior temporal gyrus,
and thalamus
C-Dev: Clinical Developmental; C-Inj: Clinical Injury; DTI: diffusion tensor imaging; fMRI: functional magnetic resonance imaging; fNIRS: functional near-infrared spectroscopy; MRI: magnetic resonance imaging; NR: not
reported; N-Typ: Nonclinical-Neuro Typical; BOLD: blood-oxygen-level-dependent; ROI: region of interest.
11
The order of preprocessing steps is subject to researcher

discretion based on the imaging protocol and research
questions. To date, there is a lack of agreement in the
neuroimaging field about the order of preprocessing
steps.64,65 The multiple steps of preprocessing and pro-
cessing steps are referred to as “pipelines” for preproces-
sing and processing. Below, we briefly discuss the
processing pipelines of the reviewed studies.
Preprocessing and data quality control

Before analyzing and interpreting neuroimaging data, it is
important to check the quality of the acquired images.
Researchers can assess data quality using a variety of
approaches such as visual inspection of data, removal of
scans with excessive noise, band-pass filtering, shim cor-
rection, and algorithms.64 Data quality control may also be
assessed by calculating signal-to-noise ratios and contrast-
to-noise ratios.66 Of the 28 articles reviewed, four studies
Figure 3. Overall distribution of neuroplasticity studies by type did not report data quality control measures in their pre-
of imaging including population category, type of imaging, and processing methods. 28,32,33,40 Several studies reported
sample size. Note: Number in circle indicates reference article. visual inspection of data quality or signal prior to data
processing. Four studies, or 14.3% of the articles reviewed,
inspected data quality for quality before analysis.34,37,43,44
without task presentation and resting-state fMRI is com- Two studies reported quality control strategies for EEG
monly used to approximate the functional connectivity of recordings.34,37
brain regions.53 Acquiring fMRI images from children is Capturing accurate, physiologically relevant signals and
particularly challenging given the environment of the scan- quality data is important for downstream analysis. A com-
ner and the duration of acquisition.54 mon strategy for stabilizing the signal of fMRI time series
Similar to fMRI, the neuroimaging method fNIRS is to remove data from the first several runs or volumes for
approximates neuronal activity as a function of metabolic each scan to improve quality and signal detection. Five
demand and cardiac perfusion, which is referred to as neu- studies reported using this strategy.25,41,45,48,52 Signal can
rovascular coupling.55 fNIRS measures changes in near- be additionally be improved by applying a linear shim cor-
infrared light absorption to indirectly measure changes in rection to each slice during reconstruction.67 Four studies
the concentration of oxyhemoglobin and deoxyhemoglo- reported using this technique during MRI acquisition to
bin.56 In comparison to fMRI, fNIRS has a shorter acqui- optimize magnetic field homogeneity, reduce blurring, and
sition period and is both portable and robust.49 Unlike improve signal.31,41,43,45 It is important to note that a spe-
fNIRS and fMRI, which indirectly measure brain activity, cific issue for fMRI data is the acquisition time required to
structural MRI measures changes in structure, such as sample multiple slices and stack them into a three-
alterations in volume or thickness of gray matter. DTI was dimensional image, which can result in temporal errors in
also used in some of the studies. The principle of DTI fMRI time series data, or “offset delays, which can be
image contrast relies on the anisotropic diffusion of water corrected for using slice timing correction methods before
molecules along axonal tracts in white matter when a gra- or after motion correction.”64 Only 4 of the total 14 fMRI
dient is applied that increases magnetic field strength in one studies reviewed here carried out slice-time correction on
direction.57 After fiber tract reconstruction by tractography acquired functional images prior to processing the
or fMRI-guided fiber tracking, valuable information can be data.31,36,41,45 Additionally, applying a frequency filter to
obtained about white matter microstructure and axonal neuroimaging data is a widely used method for removing
tract orientation.58,59 nonneuronal physiological artifacts and improving signal
detection.68 Five studies used band-pass filtering to limit
data collection to relevant signals.30,34,36,37,49
Methodological heterogeneity and
limitations of selected studies
Neuroimaging results are dependent on instrument para-
Motion detection and correction methods
meters of the imaging modality employed and on strategies Motion artifacts in neuroimaging data can cause systematic
used to define and detect low-quality data, remove noise and disruptions in image resolution, such as blurring and ghost-
artifacts, and prepare the data for statistical analyses.60–63 ing artifacts, which occur more frequently in pediatric
Weyandt et al. 13
populations.69–71 Motion during imaging is time-locked to provided participants with a tour of the MRI center and
the acquired images, so it is necessary to effectively correct offered a simulation of the MRI setup by allowing the
for motion before analyzing and interpreting data.72–74 participants to lie in a tunnel while wearing a baseball
Furthermore, motion during diffusion-weighted imaging catcher’s mask to simulate a head coil while listening to
can affect mean diffusivity and fractional anisotropy MRI sounds.50 During the actual scanning session, partici-
metrics whereas movement during task-based fMRI can pants watched a movie of their choice and the child’s
result in spurious activations, especially for movement guardian was present in the room.50 Head restraints and
tasks.75 Given that motion during image acquisition affects head molds are additional methods that can be used to deter
the quality of neuroimaging data, we discuss several impor- motion during imaging.50,80 Of the 28 reviewed articles,
tant methods used by the reviewed studies to address this two studies reported using foam pads or pillows to reduce
concern. Improving image quality by increasing the signal- movement or muscle tension during imaging.44,50 Addi-
to-noise ratio affects the accuracy of later processing steps, tionally, sedation during pediatric neuroimaging helps to
such as segmentation.66 Therefore, it is crucial that motion control for motion during data acquisition as previously
artifacts in neuroimaging data are corrected for before data described in the literature.81,82 Of the 28 articles reviewed
analysis and interpretation.76,77 Five of the 28 articles, or here, one study reported using propofol to sedate partici-
17.8% of the studies reviewed, did not correct for motion in pants before acquiring structural MRI images.51
their neuroimaging data or report motion correction proce-
dures in the methods section. Four of the reviewed studies,
or 14.3% of the studies, reported using default software
Statistical analyses
settings to remove motion artifacts from the imaging data In the studies reviewed, missing data were a prevalent issue
as a preprocessing step.29,36,42,47 Reporting motion thresh- in studies with multiple neuroimaging time points. To cor-
olds in the methods section of an article provides transpar- rect for missing data, some studies removed scans from the
ency about how the authors quantified motion artifacts participant that missed a session. Removing participants’
during pre- or post-processing. 66 Of the 28 articles scans due to a missed imaging session is an example of a
reviewed, 6 studies listed the motion threshold for images deletion method for handling missing data. Deletion meth-
acquired using MRI and DTI.25,26,41,45,48,52 Three out of the ods may introduce bias and reduce statistical power, espe-
28 articles used de-spiking procedures to correct for devi- cially if there is a small sample size to begin with.83,84
ant volumes that arose from spikes in movement.31,41,45 Existing methods for correcting for missing data include
Several studies monitored movement during image likelihood-based methods, multiple imputation, and
acquisition to detect head motion or eye movements.34,35,37 weighting.84 Two studies used strategies to address missing
EEG recordings are susceptible to ocular artifacts, such as data in their analyses. Friel et al. interpolated missing data
eye movement and blink artifacts.78 Two out of the 28 based on the group average from 6-month time point and
articles included in this review used EEG to measure neu- Hyde et al. replaced missing data with the series’ mean.38,50
roplasticity.34,37 Benasich et al. and Habibi et al. monitored Second, although many studies reported p values for
eye movements from EEG recordings using electrodes statistical tests, few reported effect size estimates. Provid-
located above and lateral to the eyes.34,37 Moreover, Bena- ing estimates of effect size, such as beta-values from
sich et al. played movies or conducted silent puppet shows regression models, eta squared values, and Cohen’s d val-
to hold infants’ attention during EEG acquisition.34,37,78 ues, provide meaningful information and increase interpret-
Two studies removed motion artifacts from neuroimaging ability of results.85 Five of the 28 reviewed studies reported
data after image acquisition by modeling motion artifacts estimates of effect size to describe the magnitude of an
as nuisance regressions in a general linear statistical model observed outcome.28,31,32,35,51 Yuan et al. standardized
of the physiological signal of interest.27,30 continuous variables and calculated parameter estimates
Several studies described familiarizing and acclimating using a mixed model for dependent variables; the coeffi-
participants with the scanning environment prior to neuroi- cients were subsequently treated as mean differences.28
maging to reduce motion during actual data acquisition. Furthermore, Rosenberg-Lee et al. and Sterling et al.
Weinstein et al. noted that participants in their study prac- reported Cohen’s d values as estimates of effect size
ticed lying in a mock scanner before the MRI scan and a whereas Schlaug et al. included eta squared values in the
movie was played during the scanning session while the results.31,32,51
child’s guardian was present in the room to reduce move- Lastly, when multiple hypothesis tests are conducted,
ment and anxiety.52 Additionally, during the fMRI motor there is an increased risk of incorrectly rejecting a null
task, videos were recorded to monitor mirror movements hypothesis, which can result in type I error, or false posi-
that were rated retrospectively by the authors using the tives.86 This is particularly relevant to neuroimaging ana-
Woods and Teuber scale.79 Jolles et al. also familiarized lytical methods.86–88 Of the 28 papers reviewed, 7 studies
participants with the MRI scanner environment before did not report corrections for multiple compari-
scanning sessions and provided detailed instructions to par- sons.25,27,29,35,36,46,47 Two articles reported that multiple
ticipants prior to acquiring fMRI images.29 Friel et al. comparison tests were not carried out due to the small
sample size of the study.27,46 Similarly, Carlson et al. did heterogeneity in statistical methods across the studies,
not correct for multiple comparisons because of the low 46%, or 13 of the 28 reviewed articles, reported using
number of comparisons made during analysis.35 The 20 analysis of variance models.25,32,34–37,39,42,44,45,48–50 As
articles that reported multiple correction methods used var- discussed in the next section, these different approaches
ious statistical approaches, including family-wise error rate affect the methodological strengths and limitations of the
(FWER) correction, false discovery rate (FDR), Monte studies, as well as the interpretation of the findings.
Carlo simulations, Tukey tests, and Bonferroni
tests.26,28,30–34,37,38,40–45,51,52 One study conducted post
hoc Tukey tests and reduced the degrees of freedom using Methodological strengths and limitations
Greenhous-Geisser epsilon to reduce the risk of type I In response to the second aim of the review, findings
error.37 revealed important methodological limitations among the
studies. Specifically, differences existed in terms of varia-
tions of imaging modality, sample sizes, order and number
Discussion of steps of preprocessing and data processing pipelines,
leading to a multiplicity of analytical methods. It is under-
Key measures and analytical techniques standable that different imaging protocols and experimen-
A main aim of the review was to identify key measures and tal designs may require specialized pipelines for
analytical techniques used within the neuroplasticity liter- preprocessing and processing of neuroimaging data; how-
ature pertaining to children and adolescents. Results ever, it is important that researchers are transparent in
revealed substantial variability among studies with respect reporting the preprocessing and processing steps and sta-
to types of cognitive and sensory motor interventions as tistical analyses used to increase the validity and reprodu-
well as delivery protocol, frequency, and duration of inter- cibility of experiments. MRI and fMRI data obtained from
ventions. For example, duration of interventions varied pediatric populations typically has more motion artifacts
from 20 min once per day to 6 h per day for 5 days per and lower quality imaging data compared to adult popula-
week. Delivery protocol also varied; however, 75% of stud- tions.69,70 Similarly, segmentation and quantification of
ies used one-to-one delivery while the others used group brain regions taken from high-resolution MR images can
methods. Given the heterogeneity across studies, results have a significant impact on imaging findings.89 Eleven of
remain equivocal; however, findings do suggest that inter- the 28 studies (39%) included in the present review were
vention studies of longer duration likely have a more robust volumetric neuroimaging studies. Brain size and shape dif-
influence on neuroplastic changes than those of shorter fers substantially among children and it is therefore espe-
duration. With regard to specific type of intervention, the cially important to correct for intracranial volume (ICV).90
present review was unable to determine whether sensory- In the current review, however, only 7% of the studies
motor interventions or cognitive interventions were more reported ICV correction methods. Consequently, the
impactful as several neuroimaging findings supported impact on the findings from the remaining studies that did
alterations in functional connectivity, volumetric changes, not employ correction procedures is suspect but
changes in neural activation, or other parameters of neuro- unknown.33,43 Additionally, only two studies reported that
plasticity; however, two cognitive interventions and two blind observers checked neuroimaging data for motion arti-
sensory-motor interventions had inconclusive imaging facts, accurate segmentation, or correct identification of
findings (Table 2).29,35,39,46 Similarly, findings across stud- structures of interest.32,43 Sterling et al. reported that a
ies were inconsistent with respect to whether cognitive or second investigator inspected the accuracy of lesion
sensory motor interventions had a larger impact on neuro- masks.51 Including post-processing inspections and com-
plastic changes across age and type of disability. Replica- bining manual data inspection with automated software
tion of designs using larger samples, varying ages, and processing improve validity. Likewise, reporting and shar-
different types of disability along with effect size informa- ing methods for neuroimaging studies that provide specific
tion is vital for understanding these potential relationships. guidelines for future studies will be critical in improving
In terms of analytical techniques, there was substantial quality and validity.64,91–93
heterogeneity in imaging modality employed, processing Motion detection is a major methodological concern of
pipelines, and approaches across studies. The majority of this review as MRI data obtained from pediatric popula-
studies employed MRI or fMRI. In terms of analyses, most tions has increased motion artifacts compared to older
studies used multivariate or cluster-level analytical tech- populations. The studies included in the present review that
niques and/or multiple analyses using the general linear addressed artifacts (not all did) used a variety of methods to
model. Of the 28 articles reviewed, 5 studies, or 17.9% of detect, limit, and remove movement from neuroimaging
the articles, described using nonparametric statistical tests data. It is possible, however, to prospectively correct for
such as the Wilcoxon signed rank test and Spearman rank artifacts in neuroimaging data caused by head motion dur-
correlations when data violated normality assumptions of ing scanning acquisition to reduce motion artifacts and
the general linear model.25,27,28,35,49 Although there was improve data quality.94 Examples of prospective motion
Weyandt et al. 15
correction techniques include BLADE/PROPELLER MRI is a more liberal correction method compared to Bonferroni
and PROMO.95,96 PROMO utilizes interleaved spiral navi- adjustment that provides more sensitivity and power for
gator scans and image-based tracking to ensure that the detecting statistical differences while correcting for type I
relationship between the imaged object, subject to motion, errors.103 RFT is commonly used for cluster-level correc-
is constant with the imaged volume by adjusting MR pulse tions of neuroimaging data. 86 The specific methods
sequences as the pose (orientation) of the object changes employed to account for false positive findings included
during scanning.94,97 It is critical that future neuroimaging FDR, FWER, and Monte Carlo simulations, as well as
studies report methods for motion correction, including the broadly termed whole brain methods. Only three of the
parameters for motion threshold, so that effective methods studies reviewed here controlled for FDR, which typically
can be adopted to improve the quality of images and involves rejecting a number of hypotheses in order to main-
increase the validity of statistical analyses based on ima- tain the FDR below a predetermined level, or specified a
ging results.69 more sophisticated known approach for controlling for
A third methodological problem of many of the studies FDR (i.e. a hidden Markov Random Field
reviewed pertains to statistical power and risk of type I or II model).28,38,52,87,104 FWER with a Bonferroni correction
error. Power analysis requires an estimate of the effect was employed by one study and two studies used Monte
size.98 In the present review, only five of the studies Carlo simulations to establish more stringent thresholds to
(17.8%) acknowledged statistical power or effect account for assumptions of independence.32,105,106 We
size.27,39,41,49,50 Adequate sample size is necessary for a direct readers to reviews of methods for analysis of neuroi-
study to have sufficient power. In the present review, maging data for further information about statistical
35.7% of studies (10 out of 28 articles) had total sample approaches.107,108
sizes of 20 participants and under.26–28,40,47–52 Only 4 of Lastly, an important methodological consideration per-
the 28 studies (14%) conducted analyses with total sample tains to data registration. Due to the global and local neu-
sizes greater than 50.30,33,43,44 To increase transparency roanatomical changes that occur during development in
and interpretation of research findings, future studies are children and adolescents, it is necessary that researchers
encouraged to report power analyses conducted to deter- use age-specific MRI templates rather than adult templates
mine sample size and effect sizes of research outcomes in for accurate registration, spatial normalization, and seg-
conjunction with p values. mentation.66 A commonly used practice for registering
Further, many of the studies used multiple hypothesis pediatric neuroimaging data to a reference template is reg-
tests comparing brain voxel activations to zero for each istration to adult brain atlases MNI305 and MNI152 fol-
voxel of the brain. Functional neuroimaging data can con- lowed by corrections for pediatric brain anatomy.109,110
sist of 100,000 voxels, each voxel corresponding to specific The disparities in shape, morphology, and size of an adult
spatial locations in the brain with different intensity values brain compared to a brain of a child brain result in errors in
and the more hypothesis tests conducted, the higher like- segmentation and increased deformations in nonlinear
lihood of a false positive.86–88 Future studies are encour- transformations when adult templates are used as a refer-
aged to employ statistical methods to reduce the likelihood ence for pediatric neuroimaging, generating variabilities
of false positives; use built-in automated processes for and less robust registration.111 Warping or applying non-
image segmentation; use theory to select apriori brain linear deformations to fit linear brain atlases to neuroima-
regions of interest for analysis; and employ multiple com- ging data from individuals with morphometric differences
parison corrections.99–101 Notably, of the 28 studies in this results in more accurate registration.112 Furthermore, regis-
review, less than 50% used a priori ROIs based on previous tering pediatric brain scans to adult brain atlases or tem-
literature to guide neuroimaging data collection and anal- plates can result in inaccurate classification of brain
ysis.25,29,30–33,35,37,39,40,41,43–46,49,51 tissue. 113 Researchers have the option of using open-
Additional methods that are often used for whole brain access brain atlases constructed from pediatric datasets
analysis to quantify the likelihood and thus control for false from Cincinnati Children’s Hospital Medical Center, IDEA
positives, include the family-wise error rate (FWER) which group, and Johns Hopkins University, among other
may include Bonferroni correction, random field theory resources for normalization and segmentation of neuroima-
(RFT), and permutation tests, as well as the false discovery ging acquired from pediatric populations.66 Furthermore,
rate (FDR). FWER Bonferroni correction is a conservative Richards et al. generated a Neurodevelopmental MRI Data-
method of controlling for type I error by determining the base of age-specific MRI reference templates ranging in
rate of false positives among all statistical tests.102 Due to age from 2 weeks to 89 years old.113
the stringency of Bonferroni adjustment for multiple tests In addition to the methodological problems characteriz-
and strict significance levels, it is possible that researchers ing many of the studies discussed herein, there are several
may not detect statistically significant differences.86 By limitations of the review that are important to acknowl-
contrast, FDR controls for false positives among statisti- edge. It is possible that studies not meeting inclusion cri-
cally significant tests where the null hypothesis is teria for the present review may have yielded different
rejected.86 For exploratory studies and health studies, FDR findings. Available research may be biased toward positive
results, that is, data not supporting neuroplastic effects are limitation across all studies with only 7 of the 21 studies
less likely to be published. reporting diversity information (e.g. race or SES).43
A major shortcoming is that this review was unable to Collectively, these studies support that children and ado-
produce findings concerning potential age effects within lescents are capable of responding to cognitively to inter-
the child and adolescent samples, and information was ventions, and these cognitive changes correspond with
lacking in the studies concerning racial and ethnic diversity neuroplastic changes—as measured by neural connectivity,
as well as other forms of diversity (e.g. socioeconomic alterations in neuronal activation across multiple areas, and
status [SES], gender, and disabilities). This shortcoming increased cortical thickness in regions that are functionally
is not unique to neuroplasticity studies and is unfortunately related to the focus of the cognitive intervention
common within the field of neuroimaging.10,114 employed.25,30,41–43 The studies tentatively support the
hypothesis that the capacity for training-induced changes
related to activation, connectivity, or structure may also
serve to mitigate or counteract local gray matter volumetric
Neuroimaging research supports the presence of
decreases seen in neurodevelopmental disorders such as
neuroplasticity among children and adolescents in the ADHD or aberrant network connectivity as seen in individ-
context of experience-dependent interventions uals with MLD.115,116
Recent neuroimaging technology has allowed for an In addition to neuroplastic changes, environmental
enhanced understanding of experience-dependent changes interventions were typically associated with corresponding
in the human brain in relation to various environmental cognitive and behavioral changes such as increased atten-
experiences (e.g. musical training, cognitive interventions, tion, planning, and memory performance, as well as
reading and math interventions, and motor interventions). improvements in motor functioning, math and reading per-
Although other animal and human adult literature provides formance.25–31,35,39,40–43,46–52 Cognitive behavior therapy
evidence of experience-dependent neuroplasticity, a com- for children with OCD and anxiety was also associated with
prehensive account of these processes in the developing neuroplastic changes in both grey and white matter relative
brain of children and adolescents is currently lacking. The to a control group and these changes were positively asso-
purpose of the present study, therefore, was to conduct a ciated with symptom severity.44,45 As a whole, findings
systematic review of neuroimaging studies that have exam- across the 28 studies support previous animal and adult
ined neuroplasticity among children and adolescents in human studies documenting the neuroplastic effects of
response to a treatment intervention and to evaluate the experience-dependent activities and response to
methodological strengths and weakness of those studies. injury.117–120
In response to the first aim, results of this review support Integrating principles of neuroplasticity in the imple-
the presence of neuroplastic changes among children and mentation of experience-dependent activities, whether they
adolescents in the context of experience-dependent are cognitive or sensory-motor, may be a critical compo-
interventions. nent to the success of an intervention. Principles such as
Three clinical categories such as Neurotypical (N-Typ), specificity, repetition, intensity, and salience are just a few
Neurodevelopmental (N-Dev), and Neurological Injury (N- key principles outlined by Kleim and Jones and these prin-
Inj) emerged from the review. Results supported that a ciples were represented among several of the studies
variety of experience-dependent interventions that we reviewed.121 A critical limitation, however, that needs to
defined as cognitive-based (i.e. social skills training, beha- be addressed in future studies concerns the length of time
vioral, or academic intervention) or sensory-motor training (intensity and duration) of the studies and timing of neu-
(i.e. music or motor-based training) were associated with roimaging studies. All studies integrated pre- and post-
neuroplastic changes across all three clinical categories. intervention measures. However, the overall length of the
Collectively, 22 of the 28 studies included in the present intervention varied from fairly short duration at 2 to 4
review provided evidence that both structural and func- weeks (n ¼ 9), moderately longer at 6 to 16 weeks (n ¼
tional neuroplastic changes occur among children and ado- 13), and relatively few longitudinal studies that were 1 to 3
lescents as a result of experience-dependent intervention years (n ¼ 6) (see Table 2). Critical to understanding and
(see Table 2). Three of the remaining studies identified interpreting neuroplastic changes in brain structures, gray
significant functional improvement across the nonclinical/ and/or white matter volumes, and other definitions of neu-
clinical and nonclinical comparison groups, but no signif- roplasticity is the inclusion of a follow-up imaging session
icant neuroplastic changes.29,35,46 The other three identi- in addition to a post-intervention imaging session. Includ-
fied significant neuroplastic changes but no significant ing a long-term follow-up imaging session (third time
functional improvements.26,40,49 Although there was diver- point) allows researchers to assess the retention of changes
sity across the populations studied regarding nonclinical observed at the imaging session that took place directly
versus clinical and a mix of comparison and control groups, after an intervention or therapy program. Of the 28 articles
racial, economic, and other forms of diversity remain a that we reviewed, 25% of the studies reported a follow-up
Weyandt et al. 17
imaging session post-intervention, ranging from 12-weeks and direction of these differences, and the clinical implica-
post-intervention to 1-year post-intervention.34,42,48–52 tions of these findings. In order to draw meaningful con-
Research in adult populations supports a wide range of clusions about neuroplasticity in children and adolescents
temporal-related neuroplasticity from transient structural as a result of experience-dependent interventions, future
gray matter changes to more persistent changes which may studies should (a) explicitly and rigorously define the con-
be delayed and require several weeks to months to generate struct of neuroplasticity; (b) clearly describe guidelines for
the necessary structural adaptations that might be revealed measurement of neuroplasticity; (c) examine questions of
on imaging.122,123 Integrating adequate time durations for whether and how neuroplasticity occurs based on experi-
both delivery and assessment of interventions along with mental and longitudinal designs; (d) identify the functional
neuroimaging will be important to more accurately captur- and structural neuroplastic mechanisms that correspond
ing these temporal components of neuroplasticity. with changes in cognition and behavior; and (e) provide
information about software, preprocessing of data, and ana-
Implications for future research lytical procedures to increase reproducibility and transpar-
ency of studies.
It is clear that there are many challenges related to studying
the complexities of brain neuroplasticity in children and Author contributions
adolescents. However, there are also many opportunities LW and MM conceived and designed the study. CC and SM
to minimize and mitigate certain limitations, including the researched the literature. All authors reviewed and edited the
development of clearer guidelines for defining and measur- manuscript and approved the final version of the manuscript.
ing neuroplasticity. Based on the methodological limita-
tions identified in the present review, future studies are Data availability statement
encouraged to (1) increase and report racial, ethnic, and Data sharing is not applicable to this article as no datasets were
additional forms of diversity within the populations stud- generated or analyzed during the current study.
ied; (2) assess the potential age effects within the child and
adolescent samples; (3) integrate defined principles of neu- Declaration of conflicting interests
roplasticity (i.e. intensity, duration, and saliency) and ana-
The author(s) declared no potential conflicts of interest with
lyze the impact of those variables on intervention and respect to the research, authorship, and/or publication of this
imaging outcome measures; (4) report the brain registration article.
and extraction method and use those appropriate to pedia-
tric samples; (5) pursue longitudinal studies with multi- Funding
point neuroimaging and behavioral and/or motor outcome The author(s) disclosed receipt of the following financial support
assessment periods; and (6) describe the analysis workflow for the research, authorship, and/or publication of this article:
followed in each study. We also refer readers to the thor- Research reported in this publication was supported by the
ough review of methods in neuroimaging research by Pol- National Center for Research Resources of the National Institutes
drack et al. for a comprehensive list of suggestions for of Health under Award Number G20RR030883. The content is
increasing transparency in reporting methods.64 Inadequate solely the responsibility of the authors and does not necessarily
reporting of analysis workflows reduces reproducibility of represent the official views of the National Institutes of Health.
studies, as there are a multitude of possible pipelines for
processing and analysis of neuroimaging data that can lead ORCID iD
to variable results.124 Implementation of these recommen- Christine M Clarkin https://orcid.org/0000-0002-9082-4083
dations would facilitate greater understanding of the under-
lying factors involved in neuroplastic changes and the References
degree to which these morphological and neurochemical 1. Cramer SC, Sur M, Dobkin BH, et al. Harnessing neuroplas-
changes correspond with cognitive and behavioral change. ticity for clinical applications. Brain 2011; 134(6):
1591–1609.
2. Sarrasin JB, Nenciovici L, Foisy LMB, et al. Effects of teach-
Conclusion ing the concept of neuroplasticity to induce a growth mindset
Findings of the present systematic review provide evidence on motivation, achievement, and brain activity: a meta-anal-
of neuroplasticity in children and adolescents in response to ysis. Trends Neurosci Educ 2018; 12: 22–31.
experience-dependent interventions. However, clearer 3. Hamaide J, Lukacova K, Van Audekerke J, et al. Neuroplas-
guidelines for defining and measuring neuroplasticity are ticity in the cerebello-thalamo-basal ganglia pathway: a long-
sorely needed. Although neuroimaging findings shed light itudinal in vivo MRI study in male songbirds. NeuroImage
on structural and functional changes following a wide vari- 2018; 181: 190–202.
ety of cognitive and sensory-motor interventions, it 4. Kennard MA. Age and other factors in motor recovery from
remains unclear whether these differences are a direct con- precentral lesions in monkeys. Am J Physiol 1936; 115:
sequence of molecular neuroplastic changes, the degree 138–146.
5. Kolb B and Whishaw IQ. Plasticity in the neocortex: mechan- 23. Rapoport JL and Gogtay N.Brain neuroplasticity in healthy,
isms underlying recovery from early brain damage. Prog hyperactive and psychotic children: insights from neuroima-
Neurobiol 1989; 32(4): 235–276. ging. Neuropsychopharmacology 2008; 33(1): 181–197.
6. Kühlmann AYR, De Rooij A, Hunink MGM, et al. Music 24. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting
affects rodents: a systematic review of experimental research. items for systematic reviews and meta-analyses: the PRISMA
Front Behav Neurosci 2018; 12: 301. statement. BMJ; 2009; 339: 332–336.
7. Rosenzweig MR, Bennett EL and Diamond MC. Brain 25. Everts R, Mürner-Lavanchy I, Schroth G, et al. Neural change
changes in response to experience. Sci Am 1972; 226(2): following different memory training approaches in very pre-
22–29. term born children—a pilot study. Dev Neurorehabil 2017;
8. Rosenzweig MR, Krech D, Bennett EL, et al. Effects of envi- 20(1): 14–24.
ronmental complexity and training on brain chemistry and 26. Alves-Pinto A, Turova V, Blumenstein T, et al. fMRI assess-
anatomy: a replication and extension. J Comp Physiol Psy- ment of neuroplasticity in youths with neurodevelopmental-
chol 1962; 55(4): 429–437. associated motor disorders after piano training. Eur J Paediatr
9. Killion BE and Weyandt LL. Brain structure in childhood Neurol 2015; 19(1): 15–28.
maltreatment-related PTSD across the lifespan: a systematic 27. Cope SM, Liu XC, Verber MD, et al. Upper limb function and
review. Appl Neuropsychol 2020; 9(1): 68–82. brain reorganization after constraint-induced movement ther-
10. Weyandt LL.Clinical neuroscience: foundations of psycholo- apy in children with hemiplegia. Dev Neurorehabil 2010;
gical and neurodegenerative disorders. New York, NY: Rou- 13(1): 19–30.
tledge, 2019. 28. Yuan W, Treble-Barna A, Sohlberg MM, et al. Changes in
11. Huttenlocher PR. Synaptic density in human frontal cortex— structural connectivity following a cognitive intervention in
developmental changes and effects of aging. Brain Res 1979;
children with traumatic brain injury. Neurorehabil Neural
163(2): 195–205.
Rep 2017; 31(2): 190–201.
12. Fuchs E and Flügge G.Adult neuroplasticity: more than 40
29. Jolles DD, Van Buchem MA, Crone EA, et al. Functional
years of research. Neural Plast 2014; 2014.
brain connectivity at rest changes after working memory
13. Hadanny A, Bechor Y, Catalogna M, et al. Hyperbaric oxy-
training. Hum Brain Mapp 2013; 34(2): 396–406.
gen therapy can induce neuroplasticity and significant clin-
30. Maximo JO, Murdaugh DL, O’Kelley S, et al. Changes in
ical improvement in patients suffering from fibromyalgia
intrinsic local connectivity after reading intervention in chil-
with a history of childhood sexual abuse-randomized con-
dren with autism. Brain Lang 2017; 175: 11–17.
trolled trial. Front Psychol 2018; 9: 2495.
31. Rosenberg-Lee M, Iuculano T, Bae SR, et al. Short-term
14. Oberman L and Pascual-Leone A.Changes in plasticity across
cognitive training recapitulates hippocampal functional
the lifespan: cause of disease and target for intervention. Prog
changes associated with one year of longitudinal skill devel-
Brain Res 2013; 207: 91–120.
opment. Trends Neurosci Educ 2018; 10: 19–29.
15. Pascual-Leone A, Amedi A, Fregni F, et al. The plastic
32. Schlaug G, Forgeard M, Zhu L, et al. Training-induced neu-
human brain cortex. Annu Rev Neurosci 2005; 28(1):
roplasticity in young children. Ann NY Acad Sci 2009; 1169:
377–401.
16. Konrad K, Firk C and Uhlhaas PJ. Brain development during 205–208.
adolescence. Deutsches Arzteblatt International 2013; 33. Matsudaira I, Yokota S, Hashimoto T, et al. Parental praise
110(25): 425–431. correlates with posterior insular cortex gray matter volume in
17. Selemon LD. A role for synaptic plasticity in the adolescent children and adolescents. PLoS One 2016; 11(4): e0154220.
development of executive function. Transl Psychiatry 2013; 34. Benasich AA, Choudhury NA, Realpe-Bonilla T, et al. Plas-
3(3): e238. ticity in developing brain: active auditory exposure impacts
18. Silk TJ and Wood AG. Lessons about neurodevelopment prelinguistic acoustic mapping. J Neurosci 2014; 34(40):
from anatomical magnetic resonance imaging. J Dev Behav 13349–13363.
Pediatr 2011; 32(2): 158–168. 35. Carlson HL, Ciechanski P, Harris AD, et al. Changes in spec-
19. Celebrating a decade of progress. Nat Neurosci 1999; 2(6): troscopic biomarkers after transcranial direct current stimula-
487. tion in children with perinatal stroke. Brain Stimul 2018;
20. O’Connor C and Joffe H. How has neuroscience affected lay 11(1): 94–103.
understandings of personhood? A review of the evidence. 36. Amad A, Seidman J, Draper SB, et al. Motor learning induces
Public Underst Sci 2013; 22(3): 254–268. plasticity in the resting brain—drumming up a connection.
21. Center on the Developing Child at Harvard University. The Cereb Cortex 2017; 27(3): 2010–2021.
science of early childhood development, https://developing 37. Habibi A, Cahn BR, Damasio A, et al. Neural correlates of
child.harvard.edu/resources/inbrief-science-of-ecd/ (2008, accelerated auditory processing in children engaged in music
accessed 10 May 2020). training. Dev Cogn Neurosci 2016; 21: 1–14.
22. Weaver ICG, Cervoni N, Champagne FA, et al. Epigenetic 38. Hyde KL, Lerch J, Norton A, et al. Musical training shapes
programming by maternal behavior. Nat Neurosci 2004; 7(8): structural brain development. J Neurosci 2009; 29(10):
847–854. 3019–3025.
Weyandt et al. 19
39. Jolles D, Wassermann D, Chokhani R, et al. Plasticity of left 53. Roalf DR and Gur RC. Functional brain imaging in neurop-
perisylvian white-matter tracts is associated with individual sychology over the past 25 years. Neuropsychology 2017;
differences in math learning. Brain Struct Funct 2016; 31(8): 954–971.
221(3): 1337–1351. 54. Wilke M, Holland SK, Myseros JS, et al. Functional magnetic
40. Hoekzema E, Carmona S, Ramos-Quiroga JA, et al. Training- resonance imaging in pediatrics. Neuropediatrics 2003; 34:
induced neuroanatomical plasticity in ADHD: a tensor-based 225–233.
morphometric study. Hum Brain Mapp 2011; 32(10): 55. Scholkmann F, Kleiser S, Metz AJ, et al. A review on con-
1741–1749. tinuous wave functional near-infrared spectroscopy and ima-
41. Iuculano T, Rosenberg-Lee M, Richardson J, et al. Cognitive ging instrumentation and methodology. NeuroImage 2014;
tutoring induces widespread neuroplasticity and remediates 85: 6–27.
brain function in children with mathematical learning disabil- 56. Brigadoi S, Ceccherini L, Cutini S, et al. Motion artifacts in
ities. Nat Commun 2015; 6: 8453. functional near-infrared spectroscopy: a comparison of
42. Meyler A, Keller TA, Cherkassky VL, et al. Modifying the motion correction techniques applied to real cognitive data.
brain activation of poor readers during sentence comprehen- NeuroImage 2014; 85: 181–191.
sion with extended remedial instruction: a longitudinal study 57. Feldman HM, Yeatman JD, Lee ES, et al. Diffusion tensor
of neuroplasticity. Neuropsychologia 2008; 46(10): imaging: a review for pediatric researchers and clinicians. J
2580–2592. Dev Behav Pediatr 2010; 31: 346–356.
43. Romeo RR, Christodoulou JA, Halverson KK, et al. Socio- 58. Mori S, Oishi K and Faria AV. White matter atlases based on
economic status and reading disability: neuroanatomy and diffusion tensor imaging. Curr Opin Neurol 2009; 22:
plasticity in response to intervention. Cereb Cortex 2018; 362–369.
59. Zhu D, Zhang T, Jiang X, et al. Fusing DTI and fMRI data: a
28(7): 2297–2312.
survey of methods and applications. NeuroImage 2014; 102:
44. Huyser C, Van Den Heuvel OA, Wolters LH, et al. Increased
184–191.
orbital frontal gray matter volume after cognitive behavioural
60. Carp J.The secret lives of experiments: methods reporting in
therapy in paediatric obsessive compulsive disorder. World J
the fMRI literature. NeuroImage 2012; 63(1): 289–300.
Biol Psychiatry 2013; 14(4): 319–331.
61. Churchill NW, Oder A, Abdi H, et al. Optimizing preproces-
45. Supekar K, Iuculano T, Chen L, et al. Remediation of child-
sing and analysis pipelines for single-subject fMRI. I. Stan-
hood math anxiety and associated neural circuits through
dard temporal motion and physiological noise correction
cognitive tutoring. J Neurosci 2015; 35(36): 12574–12583.
methods. Hum Brain Mapp 2012; 33(3): 609–627.
46. Kadis DS, Goshulak D, Namasivayam A, et al. Cortical thick-
62. Churchill NW, Spring R, Afshin-Pour B, et al. An automated,
ness in children receiving intensive therapy for idiopathic
adaptive framework for optimizing preprocessing pipelines in
apraxia of speech. Brain Topogr 2014; 27(2): 240–247.
task-based functional MRI. PLoS One 2015; 10(7):
47. Azizi S, Marzbani H, Raminfard S, et al. The impact of an
e0131520.
anti-gravity treadmill (AlterG) training on walking capacity
63. Guo Q, Parlar M, Truong W, et al. The reporting of observa-
and corticospinal tract structure in children with cerebral
tional clinical functional magnetic resonance imaging stud-
palsy. Proceedings of the Annual International Conference ies: a systematic review. PLoS One; 2014; 9(4): e94412.
of the IEEE Engineering in Medicine and Biology Society 64. Poldrack RA, Baker CI, Durnez J, et al. Scanning the horizon:
2017; 2017: 1150–1153. towards transparent and reproducible neuroimaging research.
48. Bakhtiari R, Cummine J, Reed A, et al. Changes in brain Nat Rev Neurosci 2017; 18(2): 115–126.
activity following intensive voice treatment in children with 65. Power JD, Plitt M, Kundu P, et al. Temporal interpolation
cerebral palsy. Hum Brain Mapp 2017; 38(9): 4413–4429. alters motion in fMRI scans: magnitudes and consequences
49. Cao J, Khan B, Hervey N, et al. Evaluation of cortical plas- for artifact detection. PLoS One 2017; 12(9): e0182939.
ticity in children with cerebral palsy undergoing constraint- 66. Phan TV, Smeets D, Talcott JB and Vandermosten M.Pro-
induced movement therapy based on functional near-infrared cessing of structural neuroimaging data in young children:
spectroscopy. J Biomed Opt 2015; 20(4): 046009. bridging the gap between current practice and state-of-the-
50. Friel KM, Kuo HC, Fuller J, et al. Skilled bimanual training art methods. Dev Cogn Neurosci 2018; 33: 206–223.
drives motor cortex plasticity in children with unilateral cer- 67. Glover GH and Lai S. Self-navigated spiral fMRI: interleaved
ebral palsy. Neurorehabil Neural Repair 2016; 30(9): versus single-shot. Magn Reson Med 1998; 39(3): 361–368.
834–844. 68. Jiang X, Bian GB and Tian Z. Removal of artifacts from EEG
51. Sterling C, Taub E, Davis D, et al. Structural neuroplastic signals: a review. Sensors (Switzerland) 2019; 19(5): 987.
change after constraint-induced movement therapy in chil- 69. Brown TT, Kuperman JM, Erhart M, et al. Prospective
dren with cerebral palsy. Pediatrics 2013; 131(5): motion correction of high-resolution magnetic resonance
e1664–e1669. imaging data in children. NeuroImage 2010; 53(1): 139–145.
52. Weinstein M, Myers V, Green D, et al. Brain plasticity fol- 70. Davidson MC, Thomas KM and Casey BJ. Imaging the
lowing intensive bimanual therapy in children with hemipar- developing brain with fMRI. Ment Retard Dev Disabil Res
esis: preliminary evidence. Neural Plast 2015; 2015: 798481. Rev 2003; 9: 161–167.
71. Greene DJ, Koller JM, Hampton JM, et al. Behavioral inter- 88. Vul E, Harris C, Winkielman P, et al. Puzzlingly high corre-
ventions for reducing head motion during MRI scans in chil- lations in fMRI studies of emotion, personality, and social
dren. NeuroImage 2018; 171: 234–245. cognition. Perspect Psychol Sci 2009; 4(3): 274–290.
72. Friston KJ, Worsley KJ, Frackowiak RSJ, et al. Assessing the 89. Morey RA, Petty CM, Xu Y, et al. A comparison of auto-
significance of focal activations using their spatial extent. mated segmentation and manual tracing for quantifying hip-
Hum Brain Mapp 1994; 1(3): 210–220. pocampal and amygdala volumes. NeuroImage 2009; 45(3):
73. Hajnal JV, Bydder GM and Young IR. Use of fluid attenuated 855–866.
inversion recovery sequence in magnetic resonance imaging. 90. Giedd JN, Blumenthal J, Jeffries NO, et al. Brain develop-
The Lancet 1994; 344: 1783. ment during childhood and adolescence: a longitudinal MRI
74. Zaitsev M, Maclaren J and Herbst M. Motion artifacts in study. Nat Neurosci 1999; 2: 61–63.
MRI: a complex problem with many partial solutions. J Mag 91. Carp J. Better living through transparency: improving the
Reson Imaging 2015; 42: 887–901. reproducibility of fMRI results through comprehensive meth-
75. McLaren DG, Ries ML, Xu G, et al. A generalized form of ods reporting. Cogn Affect Behav Neurosci 2013; 13(3):
context-dependent psychophysiological interactions (gPPI): a 660–666.
comparison to standard approaches. NeuroImage; 2012; 92. Elliott ML, Knodt AR, Ireland D, et al. What is the test-retest
61(4): 1277–1286. reliability of common task-functional MRI measures? New
76. Makowski C, Lepage M and Evans AC. Head motion: the empirical evidence and a meta-analysis. Psychol Sci 2020;
dirty little secret of neuroimaging in psychiatry. J Psychiatry 31(7): 792–806.
Neurosci 2019; 44: 62–68. 93. Nichols TE, Das S, Eickhoff SB, et al. Best practices in data
77. Reuter M, Tisdall MD, Qureshi A, et al. Head motion during analysis and sharing in neuroimaging using MRI. Nat Neu-
MRI acquisition reduces gray matter volume and thickness rosci 2017; 20: 299–303.
estimates. NeuroImage 2015; 107: 107–115. 94. Maclaren J, Armstrong BSR, Barrows RT, et al. Measure-
78. Ille N, Beucker R and Scherg M. Spatially constrained inde- ment and correction of microscopic head motion during mag-
pendent component analysis for artifact correction in EEG netic resonance imaging of the brain. PLoS One 2012; 7(11):
and MEG. NeuroImage 2001; 13(6): 159. e48088.
79. Woods BT and Teuber HL. Mirror movements after child- 95. Pipe J. Motion correction with PROPELLER MRI: applica-
hood hemiparesis. Neurology 1978; 28(11): 1152–1158. tion to head motion and free-breathing cardiac imaging.
80. Power JD, Barnes KA, Snyder AZ, et al. Spurious but sys- Magn Reson Med 1999; 42(5): 963–969.
tematic correlations in functional connectivity MRI networks 96. White N, Roddey C, Shankaranarayanan A, et al. PROMO:
arise from subject motion. NeuroImage 2012; 59(3): real-time prospective motion correction in MRI using image-
2142–2154. based tracking. Magn Reson Med 2010; 63(1): 91–105.
81. Ahmad R, Hu HH, Krishnamurthy R, et al. Reducing sedation 97. Watanabe K, Kakeda S, Igata N, et al. Utility of real-time
for pediatric body MRI using accelerated and abbreviated prospective motion correction (PROMO) on 3D T1-weighted
imaging protocols. Pediatric Radiol 2018; 48: 37–49. imaging in automated brain structure measurements. Sci Rep
82. Kirindoga R, Thurm AE, Hirschtritt ME, et al. Risks of 2016; 6: 38366.
propofol sedation/anesthesia for imaging studies in pedia- 98. Rossi JS. Statistical power of psychological research: What
tric research: eight years of experience in a clinical have we gained in 20 years? J Consult Clin Psychol 1990;
research center. Arch Pediatr Adolesc Med 2010; 164(6): 58(5): 646–656.
554–560. 99. Ashburner J and Friston KJ. Voxel-based morphometry—the
83. Jeliĉić H, Phelps E and Lerner RM. Use of missing data methods. NeuroImage 2000; 11(6): 805–821.
methods in longitudinal studies: the persistence of bad prac- 100. Costafreda SG. Parametric coordinate-based meta-analysis:
tices in developmental psychology. Dev Psychol 2009; 45(4): valid effect size meta-analysis of studies with differing sta-
1195–1199. tistical thresholds. J Neurosci Methods 2012; 210(2):
84. Matta TH, Flournoy JC and Byrne ML. Making an 291–300.
unknown unknown a known unknown: missing data in 101. Weyandt LL. The physiological bases of cognitive and
longitudinal neuroimaging studies. Dev Cogn Neurosci behavioral disorders. Mahwah, NJ: Lawrence Erlbaum
2018; 33: 83–98. Associates, 2006.
85. Chen G, Taylor PA and Cox RW. Is the statistic value all we 102. Genovese CR, Lazar NA and Nichols T. Thresholding of
should care about in neuroimaging? NeuroImage 2017; 147: statistical maps in functional neuroimaging using the false
952–959. discovery rate. NeuroImage 2002; 15(4): 870–878.
86. Lindquist MA and Mejia A. Zen and the art of multiple com- 103. Glickman ME, Rao SR and Schultz MR. False discovery
parisons. Psychosom Med 2015; 77(2): 114–125. rate control is a recommended alternative to Bonferroni-
87. Nguyen HD, McLachlan GJ, Cherbuin N, et al. False discov- type adjustments in health studies. J Clin Epidemiol 2014;
ery rate control in magnetic resonance imaging studies via 67: 850–857.
Markov random fields. IEEE Trans Med Imaging 2014; 104. Tzarouchi LC, Xydis V, Zikou AK, et al. Diffuse periven-
33(8): 1735–1748. tricular leukomalacia in preterm children: assessment of
Weyandt et al. 21
grey matter changes by MRI. Pediatr Radiol 2011; 41(12): 114. Jones LM, Ginier E, Debbs J, et al. Exploring representation
1545–1551. of diverse samples in fMRI studies conducted in patients
105. Park HJ, Kim CH, Park ES, et al. Increased GABA-A receptor with cardiac-related chronic illness: a focused systematic
binding and reduced connectivity at the motor cortex in chil- review. Front Hum Neurosci 2020; 14: 108.
dren with hemiplegic cerebral palsy: a multimodal investiga- 115. Carmona S, Proal E, Hoekzema EA, et al. Ventro-striatal
tion using 18F-fluoroflumazenil PET, immunohistochemistry, reductions underpin symptoms of hyperactivity and impul-
and MR imaging. J Nucl Med 2013; 54(8): 1263–1269. sivity in attention-deficit/hyperactivity disorder. Biol Psy-
106. Schafer RJ, Lacadie C, Vohr B, et al. Alterations in func- chiatry 2009; 66(10): 972–977.
tional connectivity for language in prematurely born adoles- 116. Butterworth B and Walsh V. Neural basis of mathematical
cents. Brain 2009; 132: 661–670. cognition. Curr Biol 2011; 21: R618–R621.
107. King KM, Littlefield AK, McCabe CJ, et al. Longitudinal 117. Borges LR, Fernandes AB, Melo LP, et al. Action observa-
modeling in developmental neuroimaging research: com- tion for upper limb rehabilitation after stroke. Cochrane
mon challenges, and solutions from developmental psychol- Database Syst Rev 2018; 2018(10): CD011887.
ogy. Dev Cogn Neurosci 2018; 33: 54–72. 118. Kempermann G. Environmental enrichment, new neurons
108. Madhyastha T, Peverill M, Koh N, et al. Current methods and the neurobiology of individuality. Nat Rev Neurosci
and limitations for longitudinal fMRI analysis across devel- 2019; 20: 235–245.
opment. Dev Cogn Neurosci 2018; 33: 118–128. 119. Gulyaeva NV. Molecular mechanisms of neuroplasticity: an
109. Ghosh SS, Kakunoori S, Augustinack J, et al. Evaluating the expanding universe. Biochemistry 2017; 82(3): 237–242.
validity of volume-based and surface-based brain image reg- 120. Leggio MG, Mandolesi L, Federico F, et al. Environmental
istration for developmental cognitive neuroscience studies in enrichment promotes improved spatial abilities and
children 4 to 11years of age. NeuroImage 2010; 53(1): 85–93. enhanced dendritic growth in the rat. Behav Brain Res
110. Hoeksma MR, Kenemans JL, Kemner C, et al. Variability in 2005; 163(1): 78–90.
spatial normalization of pediatric and adult brain images. 121. Kleim J and Jones T. Principles of experience-dependent
Clin Neurophysiol 2005; 116(5): 1188–1194. neural plasticity: implications for rehabilitation after brain
111. Sanchez CE, Richards JE and Almli CR. Neurodevelopmen- damage. J Speech Lang Hear Res 2008; 51(1): S225–S239.
tal MRI brain templates for children from 2 weeks to 4 years 122. Draganski B, Gaser C, Busch V, et al. Changes in grey,
of age. Dev Psychobiol 2012; 54(1): 77–91. matter induced by training. Nature 2004; 427(6972):
112. Evans AC, Dai W, Collins DL, et al. Warping of a compu- 311–312.
terized 3D atlas to match brain images for quantitative neu- 123. Draganski B, Gaser C, Kempermann G, et al. Temporal and
roanatomical and functional analysis. SPIE Medical spatial dynamics of brain structure changes during extensive
Imaging 1991; 1445: 236–246. learning. J Neurosci 2006; 26(23): 6314–6317.
113. Richards JE, Sanchez C, Phillips-Meek M, et al. A database 124. Carp J. On the plurality of (methodological) worlds: estimat-
of age-appropriate average MRI templates. NeuroImage ing the analytic flexibility of fMRI experiments. Front Neu-
2016; 124: 1254–1259. rosci 2012; 6: 149.

CTN 04 00021

Uploaded by

Copyright:

Available Formats

CTN 04 00021

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CTN 04 00021

Uploaded by

Copyright:

Available Formats

Reviews/Mini-Reviews

Clinical & Translational Neuroscience

intervention: A systematic review

Lisa L Weyandt1, Christine M Clarkin2,3 , Emily Z Holding4,

Baseline characteristics of studies Typical grouping (N-Typ). 29,31–34,36–39 Studies that

Everts et al.25 C-Dev 4 weeks: training 60 min 1:1 Cognitive

Jolles et al.29 N-Typ 6 weeks: 3 blocks of 1:1 Cognitive

Romeo et al.43 C-Dev 6 weeks: 4 h a day Group Cognitive

The order of preprocessing steps is subject to researcher

Preprocessing and data quality control

You might also like