Developmental Associations between Cognition and Adaptive Behavior in Intellectual and Developmental Disability

Background. Intellectual and developmental disabilities (IDDs) are associated with both cognitive challenges and difficulties in conceptual, social, and practical areas of living (DSM–5). Individuals with IDD often present with an intellectual disability in addition to a developmental disability such as autism or Down syndrome. Those with IDD may present with deficits in intellectual functioning as well as adaptive functioning that interfere with independence and living skills. The present study sought to examine associations of longitudinal developmental change in domains of cognition (NIH Toolbox Cognition Battery, NIHTB-CB) and adaptive behavior domains (Vineland Adaptive Behavior Scales-3; VABS-3) including Socialization, Communication, and Daily Living Skills (DLS) over a two-year period. Methods. Eligible participants for this multisite longitudinal study included those who were between 6 and 26 years at Visit 1, and who had a diagnosis of, or suspected intellectual disability (ID), including borderline ID. Three groups were recruited, including those with fragile X syndrome, Down syndrome, and other/idiopathic intellectual disability. In order to examine the association of developmental change between cognitive and adaptive behavior domains, bivariate latent change score (BLCS) models were fit to compare change in the three cognitive domains measured by the NIHTB-CB (Fluid, Crystallized, Composite) and the three adaptive behavior domains measured by the VABS-3 (Communication, DLS, and Socialization). Results. Over a two-year period, change in cognition (both Crystalized and Composite) was significantly and positively associated with change in daily living skills. Also, baseline cognition level predicted growth in adaptive behavior, however baseline adaptive behavior did not predict growth in cognition in any model. Conclusions. The present study demonstrated that developmental improvements in cognition and adaptive behavior are associated in children and young adults with IDD, indicating the potential for cross-domain effects of intervention. Notably, improvements in Daily Living Skills on the VABS-3 emerged as a primary area of adaptive behavior that positively related to improvements in cognition. This work provides evidence for the clinical, “real life” meaningfulness of the NIHTB-CB in IDD, and important empirical support for the NIHTB-CB as a fit-for-purpose performance-based outcome measure for this population.


Background
Intellectual and developmental disabilities (IDDs) are associated with both cognitive challenges and di culties in conceptual, social, and practical areas of living (DSM-5 1 ).Those with IDD typically present with de cits in intellectual functioning including reasoning, problem solving, planning, abstract thinking, judgment, academic learning, and learning from experience, as well as de cits in adaptive functioning that interfere with independence and living skills within the context of social and cultural developmental norms and expectations 1 .Adaptive functioning refers to one's ability to function independently across home and community contexts throughout the lifespan 2 and in many individuals with IDD, is an important measure of well-being both at a given moment 3,4 and over time 5,6 .There is evidence that adaptive behaviors may also serve as a better marker of overall functioning within one's environment than intellectual ability alone for this population 3,4 .
There is abundant and well characterized cross-sectional 7,8 and longitudinal [9][10][11][12] evidence of de cits in cognition [13][14][15][16][17] and adaptive behavior 9,[18][19][20] among speci c etiologies within IDD (i.e., fragile X syndrome [FXS], Down syndrome [DS], and Williams syndrome [WS]) in development.This strong foundation in the literature has identi ed phenotypic patterns of relative strengths and weaknesses in both cognitive and adaptive domains across IDD etiologies over time.However, less is known about how cognition and adaptive behavior develop together in individuals with IDD.Some have assessed both adaptive behavior and cognition longitudinally, but separately.For example, in a sample of children aged 3-12 years with autism spectrum disorder compared to those with FXS, differential patterns of longitudinal change were observed in both intelligence and adaptive behavior across groups, yet no direct associations of longitudinal change were made between domains of cognitive and adaptive function 21,22 , which can limit a full understanding of the ways in which speci c aspects of cognition may impact certain domains of daily functioning.Also, recently the validity of these results and others has been more closely scrutinized in terms of the pro le and performance of scores (e.g., standard scores, raw scores, age-equivalent scores) used in analyses 23 .These critiques center around issues of oor-effects and scaling of scores for individuals -often those with IDD -who perform near or below the oor, and whether standard scores or age-adjusted scores adequately re ect individuals' performance in such cases.
In another study, 58 adults with DS (ages 31-55 years) were seen up to seven times over a 10-year period, and change in both adaptive behavior (using the Inventory for Client and Agency Planning) and cognitive skills (using the Woodcock-Johnson Early Developmental Battery) was assessed 24 .Results revealed a modest decline in all areas of adaptive function (i.e., social communication, community living, motor skills, personal living skills, broad independence) over time, however a differential pattern of growth and decline emerged for cognitive skills.Comprehension, knowledge, and auditory processing continued to improve into the seventh decade of life, however measures of short-term memory, long-term memory, and visual processing revealed maximum performance around 50 years, followed by decline in these skills 24 .Again, associations between adaptive behavior and cognitive skills were not made directly, yet these results indicate that in adulthood, speci c areas of cognitive and adaptive functioning may track together in people with DS 24 .
More recently, Hahn et al., (2015) assessed the effect of non-verbal cognition (measured by the Mullen Scales of Early Learning taken at time 1) on the rate of growth (slopes) and starting level (intercepts) of adaptive behavior subscales in children with FXS.They found that non-verbal cognition signi cantly predicted the rate of growth in daily living skills and motor domains, and signi cantly predicted the starting level in socialization, communication, and motor domains 11 .In a similar study, intelligence (Using the Kaufman Brief Intelligence Test, 2nd Edition) and adaptive behavior (using the Vineland Adaptive Behavior Scales, Second Edition) were examined longitudinally in a sample of children with WS between the ages of 14 and 49 years.Intelligence remained stable while adaptive behavior decreased over time 10 .Baseline intelligence was correlated with a higher intercept (i.e., starting level ) of adaptive behavior across all domains (i.e., communication, daily living, socialization), but higher baseline intelligence also predicted greater decreases in adaptive behavior composite and communication scores over time 10 .
To our knowledge, no study to-date has investigated the association between developmental changes in cognitive skills and adaptive behavior in children or adults with IDD.It is important to characterize how these two broad domains track within people with IDD over time as they are the primary areas of de cit in this population across the lifespan.Moreover, understanding both the direction and patterns of association between adaptive behavior and cognition is especially relevant to educational, behavioral, and pharmacological interventions.Speci cally, there is great utility in investigating whether improvements in particular cognitive skills -measured by performance-based tests -that may be targeted in treatment, track with clinically meaningful improvements in adaptive behavior.As cognitive tests are increasingly utilized as key performance-based clinical outcome assessments (what the Food and Drug Administration [FDA] refer to as a "PerfO" 25 ) in clinical trials and other treatments for people with IDD, it is critical to understand how changes in cognition as measured by such instruments may associate with and perhaps impact adaptive or functional changes in the individual's daily life.
Although there is no cure or signi cant disease modifying treatment for any form of IDD, there have been promising results in animal models, and strong translational research based in FXS and DS for targeted treatments 15,[26][27][28] .Human trials utilizing various classes of medications in both FXS (see Berry-Kravis et al., 2018; for review) and DS [29][30][31][32] have not been met with the same successes as these treatments in animal models, possibly due to challenges in translation from animal models to humans, maintaining safe and adequate dosages, or inappropriate, insensitive, or invalid outcome measures 27,33 .Despite these setbacks, recent studies have made adjustments to outcome measures that have led to exciting advances.In a double-blind, placebo-controlled trial, adults with DS made signi cant improvements in recognition memory, inhibitory control, and caregiver-reported functional academics over a one-year period when treated with green tea extract containing epigallocatechin-3-gallate 30 .There is also evidence that components of the NIH Toolbox Cognition Battery (NIHTB-CB) 34 detected treatment effects in a 24week phase 2 randomized, placebo-controlled, crossover trial of a phosphodiesterase-4D allosteric inhibitor (BPN14770) in 30 adult males with FXS 35 .In this study, compared to placebo, cognitive improvement in BPN14470-treated patients was detected by the language-based NIHTB-CB Crystallized Cognition Composite score (comprised of the Oral Reading Recognition and Picture Vocabulary tests).
Prospective clinical trials continue to build upon this important work; however, a critical question remains unanswered: do improvements made in cognitive skills, as measured by such tests, extend to adaptive improvements in the everyday lives of people with ID? Though newly deployed outcome measures such as the NIHTB-CB have demonstrated reliability, validity, and treatment sensitivity in a clinical trial 35 , it is imperative that we determine whether different degrees of improvement on clinic-based cognitive tests are associated with changes in the daily functioning of people with IDD.Whether or not improvements in outcomes (i.e., cognitive skills from the NIHTB-CB) translate into identi able and clinically signi cant improvements in downstream areas of functioning (such as academic skills, activities of daily living, communication, or social skills) will likely be a key determinant for clinical trials moving forward, and ultimate FDA acceptance of key outcome measures and eventual approval of targeted medications.
Our psychometric studies of the NIHTB-CB in children, adolescents, and young adults with IDD have demonstrated its feasibility and validity 7,36 , as well as its sensitivity to developmental change 12 among this population.Therefore, the present study sought to build upon this previous work with a data-driven approach, to explore associations of longitudinal change in domains of cognition (i.e., NIHTB-CB subtests) and adaptive behavior domains (i.e., VABS-3 Socialization, Communication and Daily Living Skills) over a two-year period.Our previous work has identi ed some limitations using the NIHTB-CB in individuals with IDDs, particularly for those with lower mental ages (i.e., < 5 years) 7 .A speci c limitation relates to composite scores in the NIHTB-CB.For instance, the Fluid Cognition Composite is comprised of ve subtests, all of which must be administered, valid, and have a "completed" status in order to produce the composite score.Many individuals in our longitudinal sample are unable to pass practice and thus complete all ve subtests.Therefore, they do not have composite scores available, thus limiting our power to test hypotheses at the construct level.Structural equation modeling (SEM) provides an analytic framework to help combat this particular issue, given that this modeling approach is robust to missing data, aiding our ability to retain the full sample in our models -even if a single participant was able to provide only one valid NIHTB-CB subtest score.In the present study, bivariate latent change score (BLCS) models provided two-year estimates of both cognitive and adaptive behavior change in individuals with FXS, DS, and other intellectual disability (OID).This modeling framework allowed us to examine the association between latent change for cognition and adaptive behavior across construct levels of each assessment.

Participants
Eligible participants for this multisite longitudinal study included those who were between 6 and 26 years at Visit 1, and who had a diagnosis of, or suspected IDD.During Visit 1, ID or borderline ID criteria were based on the DSM-5 1 , with adaptive behavior de cits measured by the Vineland Adaptive Behavior Scales, Third Edition (VABS-3) 2 and IQ < 80 on the Stanford-Binet Intelligence Scales, 5th Edition (SB5).Three groups were recruited: FXS (full mutation, with genetic con rmation), DS (with genetic con rmation if possible), and OID (with genetic con rmation of negative fragile X mutation).A mental age equivalent of at least 3.0 years as measured by the SB5 was required, in concordance with NIHTB-CB age limits.Participants were required to be stable with usual treatment for at least 4 weeks before each visit.
Exclusion criteria consisted of uncorrectable or uncorrected vision impairment, signi cant motor impairment preventing touch screen or keypad responses, or history of head injury, brain infection, stroke, or other neurological problems such as uncontrolled daily seizures or excessive sedation from medication.Recruitment sources consisted of research registries, yers at local clinics, announcements through parent support foundation websites, and mailings to families registered with state departments that provide services to individuals with IDD.A total of 318 participants with IDD were recruited at Visit 1, and of those recruited, 54 individuals were ineligible: 20 with IQ > 79 and 34 with mental age below 3 years, leaving a nal sample of 264.Full protocol, details of the NIHTB-CB, and its performance at baseline in the present ID samples has been reported previously 7,12,36 .

Protocol
The NIHTB-CB, VABS-3 interview and SB5 were completed at Visit 1.For some participants, assessments were conducted over two days.After completion of the SB5, participants completed the NIHTB-CB while their parent/caregiver completed the VABS-3 with a psychologist or trained personnel.The same procedure was conducted again approximately 2-years later at Visit 2.

Measures
The NIHTB-CB 37  For the present study, VABS-3 growth scale values (GSVs) were used for all analyses as they have been shown to be sensitive in individuals with IDD, particularly given their robust performance longitudinally and being less susceptible to oor effects 39,40 .
The Stanford-Binet Intelligence Scales, 5th ed.(SB5), which is standardized for individuals between 2-85 years, provides an overall index of intellectual ability reported as the Full-Scale IQ (FSIQ).In part, due to its broad developmental range, the SB5 has performed well in our prior studies of IDD 7,36,41 .Our protocol utilizes mental (rather than chronologic) age to select NIHTB-CB test versions and VABS-3 start points, which were derived from the SB5 FSIQ 38 for each participant.

Statistical Analyses
In order to examine the association of developmental change between cognitive and adaptive behavior domains, permutations of bivariate latent change score (BLCS) models were used to compare change in the three cognitive domains measured by the NIHTB-CB (Fluid, Crystallized, Composite) and the three AB domains measured by the VABS-3 (Communication, DLS, and Socialization), resulting in the evaluation of nine models plus one full model (including all cognitive domains and all AB domains).Latent change score models are a type of structural equation modeling that provide estimates of change as latent variables based on two or more time points.In the BLCS framework, each model can assess the association between the latent change estimated for two constructs of interest 42 .We have utilized latent change scores previously to characterize developmental change in this sample across individual NIHTB-CB subtests 12 .Missing data were handled with full information maximum likelihood estimation, which is a standard recommendation to provide accurate parameter estimates in the presence of missing data 43 .
Generally, each model contained latent scores for cognition and AB at Visit 1 and Visit 2 that were derived from the observed scores from each domain's respective subtests 42,[44][45][46][47] .Latent change scores for cognition (ΔCognition) and AB (ΔAB) were included to model change from Visit 1 to Visit 2. Furthermore, we included an estimate of the correlated change between ΔAB and ΔCognition in each model to assess cross-domain coupling of adaptive behavior and cognition.Time between visits and participant age were each used as a covariates at the latent level in all models to control for any differences in cognitive and AB change due to variations in timing between Visit 1 and Visit 2, as well as age-related changesmodeled as those between 6 and 16 years at Visit 1, and those 16 years or older at Visit 1, which we have previously demonstrated in this population 12 .Analyses included all participants with a valid NIHTB-CB score, even without completion of visit 2 as BLCS models are robust to missingness 42 .Supplementary Fig. 1 graphically presents a generic representation of these models, and Table 3 provides details of each model's speci cation.For model t we rst speci ed base models in which nothing was correlated, and each variable received an equated intercept and variance across time 48 .We then assessed each model's t by comparing to its base model (utilizing robust t parameters including CFI, TLI, and RMSEA) using methods from Savalei 49 , and indices of t based on Little 50 (i.e., CFI > 0.85; TLI > 0.85; RMSEA < .08).

Descriptive statistics
A sample of 264 individuals were included in the present analyses.Descriptive statistics for sex assigned at birth, race, ethnicity, and diagnostic group are provided in Table 1 and for cognitive and adaptive behavior scores at Visit 1 in Table 2.The rst six models (Models 1-6) assessed ΔAB, where AB was modeled as one of three VABS-3 subscales: Communication (Comm.),Socialization (Soc.), and Daily Living Skills (DLS), and ΔCOG, where COG was modeled as one of two NIHTB-CB composites (Fluid and Crystallized).Of these models, Model 1 and Model 2 demonstrated relatively poor model t.These Models were followed up with Models 1.1 and 2.1, which omitted written communication from the AB Communication domain, and were subsequently found to have good t.Model 6 did not demonstrate adequate model t, and was not further evaluated.Models 3, 4, and 5 were deemed to have good t.
The next set of models (Models A-C) assessed ΔAB across the three subscales of the VABS-3 (Comm., Soc., and DLS) and ΔCOG across the full NIHTB-CB (comprised of its seven subtests).Of these models, Models B and C were deemed to have acceptable t.See Fig. 1 for a graphical representation of the SEM for Model B.
A nal model (Full Model) assessed ΔAB across the VABS-3 (all domains) and ΔCOG across the full NIHTB-CB.The model t was poor and not examined further.Notably, none of the models where AB was modeled using the VABS-3 Comm.subscales were shown to have good t until removing the written communication subdomain.However, for the other two AB domains (DLS & Soc.), models wherein COG was de ned as either Crystallized, Fluid, or the full NIHTB-CB were shown to have good t.

Correlation between cognitive and adaptive behavior change
In two of the seven models of good t, a signi cant relationship was observed between the change in cognition (ΔCOG) and change in adaptive behavior (ΔAB).A positive relationship between the ΔCOG and ΔAB was observed in Model 3, where COG was a variable comprised of the NIHTB-CB subscales in the Crystallized domain (PV and ORR) and AB was comprised of the VABS-3 subscales in the DLS domain (Personal, Domestic, and Community).Similarly, a positive relationship between ΔCOG and ΔAB was observed in Model B, where AB was again comprised of the VABS-3 subscales in the DLS domain, but COG was comprised of all of the NIHTB-CB subscales (Fig. 2).These ndings indicate that change in cognition, speci cally in the Crystallized domain, relates to change in daily living skills over time in our sample.A summary of the parameters of interest from these models is provided in Table 4.

Discussion
The present study sought to examine the relationship between developmental change in cognition and adaptive behavior in children and young adults with IDD.We modeled this association using bivariate latent change score models.We found that developmental improvements in language-based crystalized cognition as measured by the NIHTB-CB were related to improvement in daily living skills, and that improvement in overall cognition was also related to improvements in daily living skills.Models that included the VABS-3 Communication domain (i.e., Models 1, 2, A, and the Full Model) did not have adequate model t for analysis.Follow-up analyses indicated that the measurement model for VABS-3 Communication did not t, with the written communication domain demonstrating a high degree of covariance with the other areas comprising VABS-3 Communication (i.e., receptive communication and expressive communication).Models excluding written communication were subsequently t (Model The present study is the rst, to our knowledge, to assess the relation between developmental change in cognition and adaptive behavior in individuals with IDD, two particularly important areas of functioning for this population, as de cits in each of these areas constitute core featiures of IDD.The present study demonstrates that developmental improvements in cognition and adaptive behavior are associated in children and young adults with IDD, indicating the potential for cross-domain effects of intervention.Notably, improvements in Daily Living Skills on the VABS-3 emerged as a primary area of adaptive behavior that positively related to improvements in cognition.Developmental change in all domains of adaptive behavior were predicted by cognitive skills at Visit 1; speci cally Fluid Cognition at Visit 1 predicted improvements in DLS and Communication, and the Cognition Composite at Visit 1 predicted improvement in DLS and Socialization.
These ndings demonstrate some of the "real-life" improvements that are associated with cognitive growth in a longitudinal study of youth with IDD.Previous work has shown that cognitive ability, as measured via IQ testing, is correlated with adaptive functioning 51 at the population level.With the strength of this relationship in mind, it begs the question whether changes in one of these domains will result in change in the other, particularly in individuals with lower IQ.Establishing such reciprocity has implications for how treatment trials in IDD are designed and implemented.However, until now, no previous studies have looked at how change in these abilities over time might relate to one another in IDD; rather, only limited work characterizing the natural history of cognitive ability and/or adaptive functioning in IDD with speci c etiologies, such as WS 10 , FXS 23 and DS 27 .Our ndings indicate that changes in cognitive ability, as measured via the NIHTB-CB, are related to changes in adaptive functioning, as measured by the VABS-3, over only a two-year period.
In addition to the relationship between changes in cognitive ability and changes in adaptive behavior, an individual's ability in these domains at the rst visit also predicted change within and between these domains.This was particularly evident within the adaptive behavior domain, as starting with higher adaptive behavior skills at the rst visit was associated with less growth in adaptive behavior over the following 2 years in all models.Such a pattern could be indicative of a regression to the mean, or evidence that as an individual approaches a skill ceiling, their growth in that domain will begin to slow.Interestingly, this effect was not apparent within the cognitive domain, as cognitive ability at Visit 1 was not signi cantly associated with change in cognitive ability, only with change in adaptive behavior.The models revealed a positive relationship such that higher cognitive ability at Visit 1 was related to increased growth in adaptive behavior.Taken with the cross-domain ndings indicating that changes in cognition were associated with changes in adaptive behavior, these ndings indicate that targeting improvement in cognitive skills in ID may result in positive adaptive behavior change.However, the present study cannot draw any causal inferences as it was correlational.Future experiments could verify whether changes in one domain cause change in the other, and innovative cognitive interventions studied experimentally with controlled, randomized clinical trials 52,53 , examining the impact on adaptive behaviors, could be fruitful.
The importance of developing endpoint measures in the eld of ID is evident, 54 with many existing measures of the concept of interest (i.e., cognition) deemed inadequate or not fully " t for purpose".
Measures commonly used in clinical trials have been critiqued for their limitations in validation and sensitivity to change in individuals with IDD 55 .When evaluating clinical outcome assessments (COAs) for ID, PerfO's (i.e., direct assessments) of cognition are limited because they largely assess general cognition (e.g., IQ tests) and are less likely to show short-term change; thus, COAs for ID often consist of observer reports or clinician reported outcomes 56 .The NIHTB-CB was developed, in part, with the express purpose of lling the performance outcome gap for cognition in intervention studies.However, it was not created, validated, or normed with consideration of IDD, a population currently undergoing clinical trials targeting cognition and in urgent need of suitable primary outcome measures.Nonetheless, validity evidence for the NIHTB-CB has been collected in IDD 7 and it shows sensitivity to change in this population 12 .Evaluating the clinical meaningfulness of the NIHTB-CB was an important next step.The present study established its clinical meaningfulness through its relation to adaptive behaviors.The VABS-3 is often used as an outcome measure in clinical trials for ID; however, the VABS-3 is not a direct assessment (it is a combination of an observer report outcome and clinician-reported outcome), perhaps limiting its sensitivity, nor does it measure the concept often being targeted in many current or planned clinical trials for IDD (e.g., cognition RMSEA: <0.01 = great, 0.05 − 0.01 = good, 0.08 − 0.05 = acceptable, 0.10 − 0.08 = mediocre, > 0.10 = poor CFI: >0.99 = great, 0.95-0.99= good, 0.90-0.95= acceptable, 0.85-0.90= mediocre, < 0.85 = poor TLI: >0.99 = great, 0.95-0.99= good, 0.90-0.95= acceptable, 0.85-0.90= mediocre, < 0.85 = poor ; and Asuragen Inc to develop testing standards for FMR1 testing; D. Hessl has received funding from the following, all of which are directed to the UC Davis, in support of fragile X treatment programs, and he receives no personal funds and has no relevant nancial interest in any of the commercial entities listed: Autifony, Ovid, Tetra/Shionogi, Healx, and Zynerba pharmaceutical companies to consult on outcome measures and clinical trial design.D. Hessl and EBK are members of the Clinical Trials Committee of the National Fragile X Foundation.Funding.This study was funded by the NICHD (R01HD076189), the Health and Human Services Administration of Developmental Disabilities (90DD0596), and the MIND Institute Intellectual and Developmental Disabilities Research Center (P50HD103526).

Figures Figure 1
Figures

Table 1
is an iPad-based assessment that provides information about uid cognition, crystalized cognition, and a cognition composite through 7 tests.Flanker Inhibitory Control and Attention (FICA), Dimensional Change Card Sort (DCCS), List Sorting Working Memory (LSWM), Pattern Comparison Processing Speed (PCPS), and Picture Sequence Memory (PSM) comprise the Fluid Cognition Composite, and Picture Vocabulary (PV) and Oral Reading Recognition (ORR) comprise the Crystalized Cognition Composite.The two composites combine for the Total Cognition Composite score.A published manual of standardized NIHTB-CB administration procedures for IDD can be found in Ref.
38.In the present study, the NIHTB-CB Version 2.0 was used.Unadjusted standard scores (USSs; non-age adjusted standard scores) were used for all Toolbox tests.USSs have a mean of 100 and SD of 15.The USSs are recommended for longitudinal measurement because, like change sensitive or growth scale scores, they are not adjusted based on age-related growth of normative peers.The Vineland Adaptive Behavior Scales 3rd ed.(VABS-3) 2 interview form was used to measure adaptive behavior (AB) domains including Communication (consisting of Expressive Language, Receptive Language, and Written Language), Daily Living Skills (DLS; consisting of Personal, Home, and Community skills), and Socialization (consisting of Interpersonal Relationships, Play and Leisure, and Coping Skills).

Table 2
Descriptive statistics for study variables (Visit 1) latent change score model t evaluation Twelve bivariate latent change score models were conducted assessing the relationship between change in the adaptive behavior (ΔAB) domains and change in the cognition (ΔCOG) domains from Visit 1 to Visit 2. The interval between Visit 1 and Visit 2 in years (M = 2.44, SD = 0.81) was included as a covariate at the latent level, and age, split into those between 6 and 16 years, and 16 years or older at Visit 1 was included as a covariate at the Visit 1 level, as well as the latent level.A model t statistic summary is presented in Table3.
*Non-normal distribution, median and IQR are reported. 1dicates NIHTB-CB Fluid Composite subtest; 2 indicates NIHTB-CB Crystalized composite subtest.DCCS = Dimensional change card sorting; FICA = Flanker inhibitory control and attention; PVT = Picture vocabulary test; PSM = Picture sequence memory; PCPS = Pattern comparison processing speed; ORRT = Oral reading recognition test; and LSWM = List sorting working memory.BivariateNote: Shaded models determined to have adequate t.Goodness of t according to Little (2013):

Table 4 .
Correlation of latent change (ΔAB and ΔCOG) COG at Visit 1 signi cantly predicted increased developmental change in AB for Models 4, and 2.1, as well as Models B and C, however COG at Visit 1 did not predict ΔCOG in any model.This pattern of results indicates that both Crystalized Cognition, Fluid Cognition and the Total Cognition Composites are good Note: Signi cant paths are denoted by *p < 0.05, **p < 0.01, ***p < 0.001.Shaded rows denote models with good t.Visit 1 Cross-domain coupling of AB and COG Tables 5 and 6 present regression parameters for COG at Visit 1, and AB at Visit 1 respectively predicting ΔCOG and ΔAB.Regression components of the models with good t (i.e., Models 3, 4, 5, B, C, 1.1 and 2.1) were evaluated to examine the in uence of COG at Visit 1 on ΔAB and ΔCOG (Table5), and AB at Visit 1 on ΔAB and ΔCOG (Table6).
57The NIHTB-CB remedies these concerns, and based on the ndings of the present study, also characterizes change that relates to established clinically meaningful outcomes, such as adaptive behavior.Latent change score models were used to resolve issues with missing data and the use of growth scale values (GSVs) in the present study.Missing data were particularly problematic for NIHTB-CB tests in the Fluid domain, notably Flanker and DCCS.These tasks are challenging for individuals with IDD 7 as well as individuals of young mental ages, including young typically developing children57.Unfortunately, missing scores on any individual NIHTB-CB test prohibits the generation of Fluid, Crystallized, and Composite scores.For this reason, large portions of our sample would have been excluded from the analyses if a latent variable approach had not been used to model the cognitive domains.Additionally, VABS-3 GSVs are only available at the subdomain level.GSVs cannot be averaged across subdomains to create composites for Communication, Socialization, and DLS domains because they are "a unitless measure and therefore cannot be compared or combined across subdomains" 58 .By using latent change score models, we were able to use our data in full to examine the broader constructs of cognition and adaptive functioning.However, this approach presents practical challenges.Namely, the latent variable "scores" in these models thus do not match the composite or domain scores generated by the NIHTB-CB or the VABS-3.For this reason, the relationships between latent variables cannot be translated into practical terms regarding the standard output that are generated by these tests (e.g., "An x point change in the NIHTB-CB Crystallized Composite is associated with a y point change in the VABS-3 ABC).The present study instead provides evidence that change in certain domains of cognitive function and adaptive behavior are related at the latent level.These ndings provide direction for future work at the measurement level as the domains of cognition and adaptive behavior can potentially be more narrowly speci ed in future studies.Another study caveat to emphasize pertains to the time span between assessments (approximately two years), and the number of observations across development (maximum of two), factors that likely reduced power to detect associations of cognition and adaptive behavior growth.More observations over a longer period of development would likely produce better and stronger estimates of these associations.Related to the data missingness of the NIHTB-CB tests, future development of the NIHTB-CB should involve individuals with IDD to improve the probability that the measure can be used in clinical trials targeting cognition with this population.The development of the NIH Infant and Toddler (Baby) Toolbox (NBT), including domains of cognition and executive function, language, numeracy/early mathematics, motor, and social functioning, is currently underway.The NBT aims to capture neurodevelopment at younger ages (1-42 months old) for both research and clinical use.Individuals with IDD represent a clear clinical population of interest for this measure, particularly due to the much lower mental ages often seen in this population, and the limited feasibility we have observed for some uid reasoning tests in individuals with these lower mental ages.This work provides evidence for the clinical, "real life" meaningfulness of the NIHTB-CB in IDD, and important empirical support for the NIHTB-CB as a t-for-purpose performance-based outcome measure for this population.Ethics approval and consent to participate.Institutional review board approval was obtained at each site before study initiation.Written consent was obtained from each guardian (or adult participants in the case of individuals who were capable to provide their own consent).Consent for publication.Not applicable.Availability of data and materials.Data are available from the NIMH Data Archive (nda.nih.gov/)-IDC3738.Competing interests.AJK, JK, AD, EC, ES, D. Harvey and KR: no relevant disclosures; EBK has received funding from the following, all of which are directed to Rush University Medical Center in support of rare disease programs, and she receives no personal funds and has no relevant nancial interest in any of the commercial entities listed: Acadia, Alcobra, Anavex, Biogen, BioMarin, Cydan, Fulcrum, GeneTx, GW, Ionis, Lumos, Marinus, Neuren, Neurotrope, Novartis, Orphazyme, Ovid, Roche, Seaside Therapeutics, Tetra, Ultragenyx, Yamo, and Zynerba to consult on trial design and development strategies and/or to conduct clinical studies in FXS or other NNDs or neurodegenerative disorders; Vtesse/Sucampo/Mallinckrodt Pharmaceuticals to conduct clinical trials in Nieman Pick ConclusionsIn summary, the present study demonstrated that cognitive level, as well as change in cognition over a two-year period of development, as measured by the NIHTB-CB, are associated with growth in adaptive behavior, especially daily living skills, among youth with intellectual and developmental disabilities.