Intrinsic timescale evolves along a sensorimotor-association cortical axis in neurodevelopment

1223 

Abstract Submission 

Golia Shafiei¹, Valerie Sydnor², Matthew Cieslak³, Arielle Keller¹, Audrey Luo¹, Kahini Mehta¹, Taylor Salo¹, Dani Bassett³, Russell Shinohara¹, Theodore Satterthwaite³

¹University of Pennsylvania, Philadelphia, PA, ²University of Pittsburgh, Pittsburgh, PA, ³UPenn, Philadelphia, PA

Golia Shafiei  
University of Pennsylvania
Philadelphia, PA

Valerie Sydnor  
University of Pittsburgh
Pittsburgh, PA

Matthew Cieslak  
UPenn
Philadelphia, PA

Arielle Keller  
University of Pennsylvania
Philadelphia, PA

Audrey Luo  
University of Pennsylvania
Philadelphia, PA

Kahini Mehta  
University of Pennsylvania
Philadelphia, PA

Taylor Salo  
University of Pennsylvania
Philadelphia, PA

Dani Bassett  
UPenn
Philadelphia, PA

Russell Shinohara  
University of Pennsylvania
Philadelphia, PA

Theodore Satterthwaite  
UPenn
Philadelphia, PA

Intrinsic timescale is a commonly used measure of spontaneous neural dynamics that quantifies the time window of information processing of neuronal populations and is associated with cognition and complex behavior [1]. Intrinsic timescale displays a hierarchical cortical organization across multiple species and modalities, such that neural activity at the sensory cortex demonstrates a shorter timescale–adapting to rapid changes in the environment–and association cortex exhibits longer timescales–sustaining neural activity for a longer time to process higher order functions. A similar organizational pattern is observed across a range of neurobiological properties that largely align to an archetypical axis spanning from sensorimotor to association cortices (S-A axis) [2]. Recent work in human neurodevelopment has demonstrated that the S-A axis also constrains developmental patterns of functional brain organization in youth [3,4]. However, less is known about age-related changes in intrinsic timescale during development and whether its cortical maturation patterns follow the same large-scale organizational principles. Here we estimate intrinsic timescale in youth and investigate its neurodevelopmental trajectories with respect to the S-A axis of brain organization.

We used functional Magnetic Resonance Imaging (fMRI) data from the Human Connectome Project: Development (HCP-D [5]; N = 610, age range 8-21 years, 327 female). Functional MRI data were preprocessed using fMRIPrep and XCP-D. The autocorrelation function (ACF) of the preprocessed time-series was used to estimate an intrinsic timescale for each region as the sum of its positive autocorrelation values, indicating the timescale at which the ACF of a given region decays [6]. A Generalized Additive Model (GAM) was first applied to assess neurodevelopmental variations in the whole-cortex average timescale across individuals, including age as the smooth variable and in-scanner motion and sex as covariates. Region-wise GAMs were then used to estimate neurodevelopmental trajectories of intrinsic timescale at each cortical region. Age effects on regional timescales were estimated as partial R2, calculated as differences in the model fit with and without including the smooth term (i.e., age). Regional neurodevelopmental trajectories were compared with an archetypical organizational map of the cortex that spans the S-A axis, obtained from the neuromaps toolbox [1,7].

Applying GAM on whole-cortex data, we found that the average timescale increases during development in youth (Fig 1a). Region-wise GAM analysis demonstrated that regional neurodevelopmental trajectories of intrinsic timescale were heterogeneous across the cortex, with smaller age effects in sensorimotor cortex and larger positive age effects in association cortex (Fig. 1b). Comparing regional age effects with S-A axis rankings, we found that intrinsic timescales mature along the cortical hierarchy captured by the S-A axis (Fig. 1c; r_s = 0.38, p_spin = 0.0024). Further inspection of regional GAM fits confirmed that intrinsic timescale remains relatively stable in sensorimotor regions during child and adolescent development whereas timescale increases in association regions (Fig. 1d). Finally, we compared intrinsic timescale derived from task fMRI (Go/NoGo Task) with resting-state data to assess the extent to which cognitive demand influences intrinsic timescale in a neurodevelopmental cohort. Consistent with previous reports in adults, we found that the average whole-cortex intrinsic timescale significantly increases with higher task demand (Fig. 1e; paired t-test: t = 30.95, p < 0.05).

Intrinsic timescale increases during development, in particular in the association cortex at the top of the cortical hierarchy. Developmental effects follow a hierarchical pattern and recapitulate the S-A axis of cortical organization, revealing convergence between major axes of cortical organization and development.

Early life, Adolescence, Aging ¹

Task-Independent and Resting-State Analysis ²

Normal Development

BOLD fMRI

Data analysis

Development

FUNCTIONAL MRI

Modeling

Other - Intrinsic timescale