TL;DR
This study reveals that the brain's ability to integrate information over time (intrinsic neural timescales or INTs) shortens significantly during adolescence across species. This developmental refinement, linked to changes in brain circuitry like myelination and E/I balance, is crucial for the maturation of executive functions, particularly working memory.
What Problem Does This Paper Address?
Adolescence is a critical period for the development of executive functions, which are essential for complex cognitive tasks. While we know that brain structure and chemistry mature during this time, how the temporal dynamics of neural processing change to support this maturation remains poorly understood. Specifically, the paper investigates how the brain's "intrinsic neural timescales" (INTs) – the time windows over which neural populations integrate information – evolve during adolescence and how this relates to cognitive development and underlying neurobiological changes.
Methods
The researchers employed a multimodal, cross-species approach:
- Human Scalp EEG: Longitudinal data from 161 neurotypical individuals (ages 10-32) were analyzed to measure INTs using the autocorrelation window (ACW) during both rest and a working memory task.
- Human Intracranial sEEG: Data from 14 epilepsy patients (ages 9-21) provided high-resolution recordings to validate EEG findings.
- Macaque Local Field Potentials (LFPs): Recordings from adolescent macaques were used to examine cross-species conservation of developmental changes in INTs.
- Quantitative MRI (7T): Myelination levels (indexed by R1) were measured in specific cortical depths to investigate structural contributions to INT development.
- Statistical Modeling: Generalized Additive Mixed Models (GAMMs) were extensively used to model developmental trajectories, age-by-task interactions, and relationships between neural measures and behavior.
Datasets
- Human Longitudinal EEG Dataset: Collected from 161 neurotypical participants (ages 10-32) in Pittsburgh, with up to 3 visits per participant.
- Human Intracranial sEEG Dataset: Collected from 14 medication-resistant epilepsy patients (ages 9-21) in Pittsburgh.
- Macaque LFP Dataset: Collected from 8 adolescent macaques (ages 3.2-5.7 years) at Wake Forest University and Vanderbilt University.
- 7T MRI Data: Collected from the same human participants as the EEG dataset.
Key Insights and Ideas
- Adolescence is a "Temporal Refinement" Period: The most striking finding is the consistent shortening of intrinsic neural timescales across adolescence, observed in humans (scalp EEG and sEEG) and macaques. This suggests a fundamental shift in how the brain processes information over time during this developmental stage.
- INTs are Biologically Grounded: The study links shorter INTs to increased excitatory-inhibitory (E/I) balance and reduced aggregate neural spiking, as measured by EEG spectral properties. This provides a mechanistic bridge between local circuit physiology and macroscale temporal dynamics.
- Myelination Plays a Role: Deep layer myelination in frontal and parietal cortices showed a significant age-by-layer interaction, suggesting that structural maturation, particularly in deeper cortical layers supporting long-range connectivity, influences the development of resting-state INTs.
- Shorter INTs = Better Working Memory (in Adolescence): Shorter INTs were associated with improved working memory accuracy and reduced response time variability in adolescents, highlighting the functional significance of this temporal refinement for cognitive development. This advantage shifted in adulthood.
- Cross-Species Conservation: The replication of these findings across humans and macaques strongly suggests that the developmental shortening of INTs is a conserved primate trait, pointing to a fundamental biological process.
Key Results
- Consistent Decrease in ACW: Autocorrelation window (ACW), a measure of INT, significantly decreased with age in frontal, parietal, and occipital regions in human EEG data. This was replicated in human sEEG and macaque LFP data.
- Task vs. Rest Differences: While INTs were generally shorter during rest compared to task states in adults, this state-sensitivity was less pronounced in adolescents.
- E/I Balance and Spiking: A lower aperiodic exponent (indicating increased E/I balance) and a lower aperiodic offset (indicating reduced neural spiking) were both significantly associated with shorter ACWs.
- Working Memory Performance: Shorter ACWs were associated with better working memory accuracy (frontal and parietal) and reduced latency variability (frontal) in adolescents. This advantage was not observed in adults.
- Myelination Link: Significant age-by-layer interactions for deep cortical layers in frontal and parietal regions indicated that developmental changes in deep myelination were linked to shorter resting-state ACWs.
Limitations and Future Work
- EEG's Indirect Nature: The authors acknowledge that EEG measures surface-level dynamics influenced by volume conduction, making it an indirect measure of neural activity. While sEEG and LFP data help address this, the indirect nature of EEG remains a limitation.
- Small sEEG Sample: The sEEG dataset, while valuable, is small and limited to resting-state recordings, restricting the ability to bridge surface-level task dynamics to intracortical dynamics within the same species.
- ACF and PSD Relationship: The authors note the mathematical relationship between ACF and PSD but state their goal was to examine associations with specific spectral components, not to resolve redundancy.
- Future Directions: The study opens avenues for further research into how these temporal refinements support other executive functions, the precise mechanisms by which myelination influences INTs, and how individual differences in INT development might relate to neurodevelopmental disorders. The cross-species framework also allows for deeper investigation into the evolutionary underpinnings of these developmental processes.
Summary updated