Assessing neurocognitive maturation in early adolescence based on baby and adult functional brain landscapes

Omid Kardan,Natasha Jones,Muriah D. Wheelock,Cleanthis Michael,Mike Angstadt,M. Fiona Molloy,Lora M. Cope,Meghan M. Martz,Katherine L. McCurry,Jillian E. Hardee,Monica D. Rosenberg,Alexander Weigard,Luke W. Hyde,Chandra Sripada,Mary M. Heitzeg

DOI: https://doi.org/10.1101/2024.09.26.615215

2024-09-26

Abstract:Adolescence is a period of growth in cognitive performance and functioning. Recently, data-driven measures of brain-age gap, which can index cognitive decline in older populations, have been utilized in adolescent data with mixed findings. Instead of using a data-driven approach, here we assess the maturation status of the brain functional landscape in early adolescence by directly comparing an individual′s resting-state functional connectivity (rsFC) to the canonical early-life and adulthood communities. Specifically, we hypothesized that the degree to which a youth′s connectome is better captured by adult networks compared to infant/toddler networks is predictive of their cognitive development. To test this hypothesis across individuals and longitudinally, we utilized the Adolescent Brain Cognitive Development (ABCD) Study at baseline (9-10 years; n = 6,489) and 2-year-follow-up (Y2: 11-12 years; n = 5,089). Adjusted for demographic factors, our anchored rsFC score (AFC) was associated with better task performance both across and within participants. AFC was related to age and aging across youth, and change in AFC statistically mediated the age-related change in task performance. In conclusion, we showed that a model-fitting-free index of the brain at rest that is anchored to both adult and baby connectivity landscapes predicts cognitive performance and development in youth.

Neuroscience

What problem does this paper attempt to address?

### Problems the Paper Aims to Solve

This paper aims to assess the brain functional maturity of early adolescents (9-12 years old). Specifically, the researchers attempt to evaluate the maturity state of the brain's functional landscape by comparing individuals' resting-state functional connectivity (rsFC) with typical functional networks of infants and adults. They hypothesize that the closer the adolescent's brain connectivity pattern is to the adult network rather than the infant network, the higher their cognitive development level.

### Main Research Background

1. **Adolescent Brain Development**:
- Adolescence is a critical period for brain development, during which functional connectivity patterns (i.e., "functional connectome") undergo significant changes.
- Understanding the relationship between individual variability in functional connectomes and cognitive performance has been a focus of many recent studies.

2. **Limitations of Existing Methods**:
- In recent years, data-driven methods (such as machine learning) have been widely used in brain-behavior association studies, but these methods have shown limited effectiveness in predicting adolescent cognitive development.
- For example, the "brain age gap" method performs well in older populations, but its application results in adolescents have been inconsistent.

3. **Research Objectives**:
- This paper proposes a new, theoretically grounded, and computationally simple summary metric of the functional connectome, called the Anchored rsFC Maturity Score (AFC), to assess adolescents' cognitive performance and development.
- This metric quantifies the degree of brain functional maturity by comparing individuals' rsFC with typical networks of infants and adults.

### Research Methods

1. **Data Source**:
- The study used data from the Adolescent Brain Cognitive Development (ABCD) study, including baseline (9-10 years old, n=6,489) and two-year follow-up (11-12 years old, n=5,089) data.

2. **Functional Connectivity Matrix Construction**:
- Functional connectivity matrices for each participant were constructed by downsampling resting-state fMRI data to 333 cortical regions.
- Each connectivity matrix was arranged according to typical networks of infants and adults.

3. **AFC Score Calculation**:
- For each participant, the [Conn within – Conn between] value was calculated under the infant and adult network definitions, which is the average value of within-network connections minus the average value of between-network connections.
- The AFC score is defined as the [Conn within – Conn between] value of the adult network minus the [Conn within – Conn between] value of the infant network.

4. **Statistical Analysis**:
- Regression analysis was conducted using linear mixed-effects regression models (lmer), adjusting for factors such as data collection site, head motion, scan count, gender, and race.
- Confidence intervals for mediation effects were estimated using the Bootstrap method.

### Main Findings

1. **Infant and Adult Network Definitions Outperform Random and Distance Baseline Networks**:
- In adolescents aged 9-10 and 11-12, the infant and adult network definitions better captured brain functional connectivity patterns compared to random and distance baseline networks.

2. **AFC Score Correlates with Cognitive Performance**:
- Higher AFC scores were associated with better performance on the N-back task, and this relationship held at both baseline and two-year follow-up.
- AFC scores significantly increased over two years, and this increase was related to improved performance on the N-back task.

3. **Sensitivity Analysis**:
- The results remained robust after excluding participants with N-back task accuracy below 60% or retaining only one sibling per family.

### Conclusion

This paper proposes a new functional connectome-based metric (AFC score) to assess early adolescents' cognitive development. By comparing individuals' rsFC with typical networks of infants and adults, this metric can effectively predict cognitive performance and its development. This method provides a new tool for tracking neurocognitive maturity in adolescents.