#58114 | AsPredicted

'Longitudinal predictors of cognitive well-being in children below poverty'
(AsPredicted #58114)


Created:       02/11/2021 02:39 PM (PT)

This is an anonymized version of the pre-registration.  It was created by the author(s) to use during peer-review.
A non-anonymized version (containing author names) should be made available by the authors when the work it supports is made public.

1) Have any data been collected for this study already?
It's complicated. We have already collected some data but explain in Question 8 why readers may consider this a valid pre-registration nevertheless.

2) What's the main question being asked or hypothesis being tested in this study?
1) Concurrent and longitudinal associations with academic performance:
Relevance of cognitive test score for academic performance:
a. We predict children’s performance on cognitive tests at baseline will be associated with their grades in school (H1a). However, we hypothesize that the concurrent association between cognitive performance and grades may be weaker for children in poverty, who face many potential barriers to academic success (H1b). If this is the case, we expect to find an interaction for the relation between concurrent cognitive performance and grades as a function of poverty status.
b. Given Cattell's classic hypothesis that current cognitive functioning supports the acquisition of crystallized knowledge, we will test whether cognitive test performance at T0 predicts grades at T2, controlling for grades at T0. We hypothesize that the longitudinal association between cognitive performance and grades may be weaker for children in poverty (H2). If this is the case, we expect to find an interaction for the relation between cognitive performance and grades as a function of poverty status.
Relations between T0 brain connectivity and academic performance:
c. We hypothesize that LFPN-DMN connectivity is differentially associated with academic performance for children above and below poverty. Our primary prediction is a concurrent relation (H3): higher LFPN-DMN connectivity will be associated with higher grades for children in poverty and lower grades for children above poverty, mirroring our findings for test scores.
If, in our preliminary analyses (part a.), we find a strong concurrent association between cognitive test scores and academic performance for children below poverty, this outcome would hardly be surprising. If not, however, a positive association between LFPN-DMN connectivity and academic performance for these children would indicate that LFPN-DMN connectivity is relevant for two at least partially dissociable metrics of performance.
d. Based on prior research showing longitudinal but not concurrent brain-behavior relations, we also seek to test the hypothesis that LFPN-DMN connectivity supports knowledge acquisition (H4). We predict that LFPN-DMN connectivity will be longitudinally associated with grades, controlling for grades at T0.

2) Longitudinal associations with attention problems:
a. We seek to replicate a prior finding that stronger LFPN-DMN connectivity is associated with greater attention problems longitudinally for children above poverty, controlling for attention problems at T0. If this is a general phenomenon, we would expect children below poverty to show the same relation (H5a). However, given our prior findings regarding differential brain-behavior relations for cognitive performance, children below poverty could conceivably show the opposite pattern (H5b).

3) Describe the key dependent variable(s) specifying how they will be measured.
We will use a combination of data from ABCD’s baseline assessment (T0, ages 9-10) and the two-year follow-up (T2, ages 11-12).
- Cognitive test performance: NIH toolbox fluid ability composite, as provided by the ABCD data curators.
- LFPN-DMN connectivity: Resting state connectivity (average strength of correlation across nodes within/between specific brain networks/nodes) between LFPN and DMN, as identified using the ABCD in-house processing at T0 and T2.
- Academic performance: CBCL mean of reported grades at T0 and T2, separately (Failing = 1…Above average = 4).
- Attention problems: CBCL attention subscale at T0 and T2, separately.
- Poverty level: Binary above- or below-poverty measure, considering annual household income and number of people in home. Children will be considered to be in poverty if they are living in a household of 4 of less with a total income of less than $25,000, or a household of 5 or more with a total income of less than $35,000.
- DMN-CON connectivity: Resting state connectivity (average strength of correlation across nodes within/between specific brain networks/nodes) between DMN and CON, as identified using the ABCD in-house processing at T0 and T2.
- LFPN-CON connectivity: Resting state connectivity (average strength of correlation across nodes within/between specific brain networks/nodes) between LFPN and CON, as identified using the ABCD in-house processing at T0 and T2.

4) How many and which conditions will participants be assigned to?
None.

5) Specify exactly which analyses you will conduct to examine the main question/hypothesis.
For each set of analyses, we will run linear mixed effects models. We will start by including each of the following covariates: fixed effects of age, motion (mean framewise displacement), and sex; random intercepts for study site and family. We will drop covariates that do not contribute significantly to model fit (assessing significance using the Anova() function from the “car” package in R, for a Type 2 Anova test).

1) Concurrent and longitudinal associations with academic performance.
a. Linear mixed effects model associating cognitive test performance at T0 with academic performance at T0, with an interaction between T0 academic performance and poverty level.
b. Linear mixed effects model associating cognitive test performance at T0 with academic performance at T2, with an interaction between T2 academic performance and poverty level, and a fixed effect of T0 academic performance.
c. Linear mixed effects model associating academic performance at T0 with LFPN-DMN connectivity at T0, with an interaction between LFPN-DMN connectivity and poverty level.
d. Linear mixed effects model associating academic performance at T2 with LFPN-DMN connectivity at T0, with an interaction between connectivity and poverty level, controlling for academic performance at T0.

2) Concurrent and longitudinal associations with attention problems.
a. Linear mixed effects model associating attention problems at T0 with LFPN-DMN connectivity at T0, with an interaction between LFPN-DMN connectivity and poverty level.
b. Linear mixed effects model associating attention problems at T2 with LFPN-DMN connectivity at T0, with an interaction between LFPN-DMN connectivity and poverty level.

6) Describe exactly how outliers will be defined and handled, and your precise rule(s) for excluding observations.
For each of the relevant analyses, we will exclude children who are missing data for one or more of the variables. In addition, we will use ABCD’s usability guidelines to determine usability of fMRI data, as in our prior analyses.

7) How many observations will be collected or what will determine sample size?
No need to justify decision, but be precise about exactly how the number will be determined.

As many as usable at T0 and T2 according to the criteria outlined above.

8) Anything else you would like to pre-register?
(e.g., secondary analyses, variables collected for exploratory purposes, unusual analyses planned?)

1) CON-LFPN associations with academic performance. CON has been heavily implicated in alerting the LFPN to external challenges, including difficult cognitive tasks. We hypothesize that stronger CON-LFPN connectivity is associated with worse academic performance in both children above and below poverty (secondary H1). We will test this hypothesis concurrently: Linear mixed effects model associating academic performance at T0 with CON-LFPN connectivity at T0, with an interaction between CON-LFPN and poverty level.

We will also test whether academic performance is overall higher for children in poverty: Linear mixed effects model associating academic performance at T0 with poverty level.

2) Interplay between three brain networks and attention problems. Though immature at age 9, CON is theorized to serve as an intermediary the DMN and LFPN, enabling switching attention between internally and externally guided mental states. Thus, we will test whether these patterns of CON connectivity are differentially associated attention problems for children above and below poverty. We predict that better-performing children below (but not above) poverty will exhibit weaker CON connectivity with both of the other networks (secondary H2). We will test this hypothesis concurrently: Linear mixed effects model associating attention problems at T0 with an interaction between LFPN-DMN connectivity and poverty level, an interaction between DMN-CON and poverty level, and an interaction between LFPN-CON and poverty level.

If better attention is associated with lower DMN-CON connectivity for children below poverty but higher connectivity for children above poverty, we would perform a longitudinal analysis. Specifically, we would test the hypothesis that children in poverty show accelerated development of CON connectivity at age 8-9 relative to their higher-income peers (secondary H3). Here, we predict that the slope of the association between DMN-CON connectivity and attention for children above poverty would flip signs (negative at T2, similar to what was observed at T0 for children below poverty).

We will also test whether cognitive performance is negatively related, concurrently, to children’s attention problems: Linear mixed effects model associating cognitive test performance at T0 with attention problems at T0, with an interaction between T0 attention problems and poverty level.

Data have already been collected but we have not yet accessed data from the CBCL, or any data from T2. (We have previously worked with LFPN-DMN connectivity at T0).

We are pre-registering a distinct set of questions with the same dataset concurrently; see pre-registration titled, “Vigilance as a potential mechanism of cognitive resilience.”

####BUNDLE####
This pre-registration is part of a bundle of similar and/or related pre-registrations sharing at least one author. When a pre-registration in a bundle is shared with reviewers or made public, all of them are. Links to all other pre-registrations in the bundle are listed below:

#58109 - https://aspredicted.org/QWQ_C5N - Title: 'ABCD 2.0: Vigilance as a potential mechanism of cognitive resilience'