Core time – The rich club’s role in shaping the intrinsic cortical timescales

2019 

Abstract Submission 

Falko Mecklenbrauck^1,2, Ricarda Schubotz^1,2

¹University of Muenster, Muenster, NRW, Germany, ²Otto Creutzfeldt Center for Cognitive and Behavioral Neuroscience, Muenster, NRW, Germany

Falko Mecklenbrauck  
University of Muenster|Otto Creutzfeldt Center for Cognitive and Behavioral Neuroscience
Muenster, NRW, Germany|Muenster, NRW, Germany

Ricarda Schubotz  
University of Muenster|Otto Creutzfeldt Center for Cognitive and Behavioral Neuroscience
Muenster, NRW, Germany|Muenster, NRW, Germany

Different brain areas express different intrinsic timescales [9] that are crucial to process stimuli on various levels of temporal persistency [4]. The reason for the association of brain regions and specific intrinsic timescales is still unknown. Previous studies suggested that the timescales follow a gradient from unimodal to multimodal fields [6] or from sparsely to strongly connected network nodes [7]. However, another prominent organization has not been investigated yet: At the network level, the structural rich club (RC) architecture has been identified to synchronize the connectome through its slow intrinsic timescale, thus influencing the temporal layout of the cortex [1,5,10]. This study investigates the RC's role in shaping the timescales in direct comparison to other organizational schemes and gradients.

We collected resting state functional magnetic resonance imaging (rs-fMRI) and diffusion weighted imaging (DWI) data of 46 healthy right-handed participants (32 female, age = 21.76 ± 2.75 years). Rs-fMRI was recorded with a 3T Siemens MAGNETOM Prisma MR scanner (TR/TE = 1000/34 ms, FA = 57°, FOV = 210 x 210 mm², 66 slices, slice thickness = 2.2 mm, acceleration factor = 6). DWI data (TR/TE = 7300/90 ms, initial rotation = -90°, FOV = 320 x 320 mm², 56 slices, slice thickness = 2.5 mm, acceleration factor = 2) was collected in the same session. Functional preprocessing included the removal of the first five volumes, slice-timing and motion correction. After registering the individual functional and structural datasets we performed a nuisance regression [6]. Intrinsic cortical timescales were calculated using a model-free approach [9]. Next, we defined the different organizational schemes of interest: (1) The RC identification followed our previous methods [8]. (2) We also tested the diverse club (DC) as an alternative to the RC [2]. (3) For investigating of the anterior-posterior gradient we used the stereotactic coordinates of each area. (4) We extracted the following structural graph measures: nodal degree, betweenness centrality, closeness centrality, participation coefficient, and within-module degree z-score as well as (5) the surface area and gray matter thickness for each area of each participant.

Using likelihood ratio tests, the comparisons to a null model revealed that the RC classification, χ²(2, N = 9495) = 59.07, p < .001, V = 0.06, DC classification, χ²(2, N = 9495) = 37.05, p < .001, V = 0.04, coordinates, χ²(3, N = 9495) = 193.89, p < .001, V = 0.08, structural graph measures, χ²(5, N = 9495) = 63.72, p < .001, V = 0.04, and surface area and gray matter thickness, χ²(2, N = 9495) = 192.63, p < .001, V = 0.10, significantly explained variance of the intrinsic cortical timescales. Furthermore, adding the RC classification to the other individual models significantly explained additional variance beyond the DC classification, χ²(2, N = 9495) = 32.06, p < .001, V = 0.04, coordinates, χ²(2, N = 9495) = 6.65, p = .036, V = 0.02, graph measures, χ²(2, N = 9495) = 45.69, p < .001, V = 0.05, and surface area and gray matter thickness, χ²(2, N = 9495) = 41.98, p < .001, V = 0.05. Multicollinearity checks revealed no significant dependence between the organizational schemes, all VIF < 10.

The results demonstrate that the differences in intrinsic cortical timescales can be explained by the membership of areas to the RC and DC, the position of the area in the cortex, the variation in different graph-theoretical measures and the surface area and gray matter thickness of the respective area. Focusing on the RC, we found that this classification explains additional variance beyond the other organizational schemes, which reiterates the importance of the RC organization for the temporal layout of the cortex.

Connectivity (eg. functional, effective, structural)

Diffusion MRI Modeling and Analysis

Multivariate Approaches

Task-Independent and Resting-State Analysis ¹

Anatomy and Functional Systems ²

FUNCTIONAL MRI

Morphometrics

STRUCTURAL MRI

Tractography

WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC

Other - Connectivity