How much is enough? Optimizing data collection for pediatric functional connectivity research
Poster No:
1572
Submission Type:
Abstract Submission
Authors:
Shefali Rai1, Kirk Graff1, Kate Godfrey1, Shelly Yin1, Daria Merrikh1, Ryann Tansey1, Tamara Vanderwal2, Signe Bray1
Institutions:
1University of Calgary, Calgary, Alberta, 2Department of Psychiatry, University of British Columbia, Vancouver, BC
First Author:
Co-Author(s):
Introduction:
Functional MRI functional connectivity (fMRI-FC) is used to understand how brain network organization changes as children mature, and how networks are impacted in children with neurodevelopmental conditions. However, acquiring sufficient low-motion data presents a challenge to obtaining reliable individual measures of the functional connectome1-3. How much data is 'enough'? And how do design choices – using movies to increase compliance – impact data quantity and quality4-7? Here we used a repeated-sampling approach to investigate the impact of scan duration and viewing condition on test-retest reliability in adults and children.
Methods:
We recruited 25 parent-child pairs (adults 33.8-47.1 years, 12 female; children 6.6-8.9 years, 13 female) that underwent four MRI scans, each including one T1-weighted (TR = 7.29ms, TE = 2.66ms, voxels = 1mm isotropic) and six multi-echo fMRI runs ( TR = 2000ms; TEs = 13ms, 32.33ms, 51.66ms; flip angle = 70°, MB factor=2), each lasting ~6.5 minutes. fMRI data was collected during three viewing conditions: 1) narrative movie (sequential clips from the film Dora and the Lost City of Gold), 2) low-demand videos (similar in concept to Inscapes6) and 3) popular and child-friendly non-narrative video clips (accessed from YouTube.com), totaling ~40 minutes per participant, per session. T1-weighted images were processed via FreeSurfer 6.08, and functional images were preprocessed with Python and MATLAB scripts. Functional preprocessing included slice time correction, multi-echo denoising (Tedana9), strict motion censoring (FD threshold > 0.15mm), bandpass filtering (0.01-0.08), and global signal regression. Surface-projected functional runs were created using Ciftify10, and connectomes were generated by Pearson correlating time series from each parcel, using the Schaefer 200 parcel 17-network atlas. Excluding one child with <30 minutes of censored data, analysis included in 24 children (median = 120 min; range = 72.9 – 160.4 min) and 24 adults (median = 150 min; range = 60.5 – 161.4 min). Connectome split-half reliability was computed using test-retest correlation (FC-TRC11), using ANOVAs to compare across groups and tasks with matched amounts of data.
Results:
Concatenating all conditions, test-retest correlation reached a maximum of r=0.77 for adults and r=0.75 for children (with >50 minutes per split; Figure 1). Up to 30 minutes, adults had significantly higher reliability at 5-minute increments (paired t-tests, all ps<0.001). Adults reached test-retest r>0.7 with 19 minutes of post-censored data and children at 26 minutes. The pre-censored median was 20.7 minutes (range = 20.3 – 54.2 min) for adults and 35.6 minutes (range = 26.6 – 58.5 min) for children, indicating ~72% more data was acquired to achieve comparable reliability in children. In the low-demand condition, both adults and children retained the least data, with significant differences in head motion between viewing conditions and groups (ANOVA, p<0.05). No interaction was found between condition and group (p>0.05). For each condition, adults consistently had higher reliability than children, up to 10 min of post-censored data (ANOVA, p<0.001; Figure 2). At 10 min of post-censored data, there was no significant effect of condition, or condition * group interaction (ANOVA, p>0.05), on reliability.
Conclusions:
Our findings suggest pediatric studies may require >50% longer data collection for comparable reliability to adults and that reliability can match across age groups with sufficient data (e.g., 1h post-censoring). Our study found video type does not impact whole-connectome FC-TRC, but significantly influences data retention. Together our results underscore the need for careful pediatric study design for desired test-retest reliability, aligning with literature supporting movie fMRI in children6,12,13.
Lifespan Development:
Normal Brain Development: Fetus to Adolescence 2
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 1
Methods Development
Novel Imaging Acquisition Methods:
BOLD fMRI
Keywords:
Acquisition
ADULTS
Data analysis
FUNCTIONAL MRI
PEDIATRIC
Other - Reliability
1|2Indicates the priority used for review
Provide references using author date format
Davis, B. R. (2022),'Key considerations for child and adolescent MRI data collection', Frontiers in Neuroimaging, vol. 1, pp. 981947.
DuPre, E. (2021),'TE-dependent analysis of multi-echo fMRI with tedana', The Journal of Open Source Software, vol. 6, no. 66.
Emerson, R. W. (2015),'Network-Level Connectivity Dynamics of Movie Watching in 6-Year-Old Children', Frontiers in Human Neuroscience, vol. 9.
Glasser, M. F. (2013),'The minimal preprocessing pipelines for the Human Connectome Project', NeuroImage, vol. 80, pp. 105–124.
Gordon, E. M. (2017),'Precision Functional Mapping of Individual Human Brains', Neuron, vol. 95, no. 4, pp. 791-807.e7.
Luna, B. (2010),'Methodological approaches in developmental neuroimaging studies', Human Brain Mapping, vol. 31, no. 6, pp. 863–871.
Meer, J. N. V. D. (2020),'Movie viewing elicits rich and reliable brain state dynamics', Nature Communications, vol. 11, no. 1, pp. 5004.
Power, J. D. (2012),'Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion', NeuroImage, vol. 59, no. 3, pp. 2142–2154.
Satterthwaite, T. D. (2012),'Impact of in-scanner head motion on multiple measures of functional connectivity: Relevance for studies of neurodevelopment in youth', NeuroImage, vol. 60, no. 1, pp. 623–632.
Vanderwal, T. (2021),'Stability and similarity of the pediatric connectome as developmental measures', NeuroImage, vol. 226, pp. 117537.
Vanderwal, T. (2015),'Inscapes: A movie paradigm to improve compliance in functional magnetic resonance imaging', NeuroImage, vol. 122, pp. 222–232.
Yeo, B. T. T. (2011),'The organization of the human cerebral cortex estimated by intrinsic functional connectivity', Journal of Neurophysiology, vol. 106, no. 3, pp. 1125–1165.