Poster No:
2268
Submission Type:
Abstract Submission
Authors:
Jelle Dalenberg1, Débora Peretti2, Ramesh Marapin1, A.M. Van Der Stouwe1, Alma Torres-Torres1, Remco Renken1, Marina Tijssen1
Institutions:
1University Medical Center Groningen, Groningen, Groningen, 2University of Geneva, Geneva, Geneva
First Author:
Co-Author(s):
Ramesh Marapin
University Medical Center Groningen
Groningen, Groningen
Introduction:
The Next Move in Movement Disorders (NEMO) project is dedicated to characterizing hyperkinetic movement disorders; i.e. including the phenotypes tremor, myoclonus, dystonia, and myoclonus-dystonia [1]. Precise phenotypic classification is essential for effective diagnosis and treatment planning. The project seeks to develop computer-aided tools to enhance diagnostic accuracy, assess disease progression, and individualize treatment approaches. For the NEMO project, we collected data from 140 patients using movement registration (as described in [1]) and neuroimaging measurements. These neuroimaging measurements include fMRI and 18F-fluorodeoxyglucos (FDG) PET scans to explore the relation between the phenotypes and brain function. Here, we outline our efforts from the past four years to improve data quality and to address challenges such as head movements, protocol design, and data preprocessing. Our standardized protocols enable comparative analyses between movement disorders, contributing to our understanding of each disorder's distinctive attributes.
Methods:
Hyperkinetic movement disorders, often mild at rest, intensify during tasks [2]. As body movements increase the risk of movement induced artifacts, we carefully selected neuroimaging protocol designs [3].
Two BOLD fMRI protocols were created: (1) a protocol with relatively high spatial precision and a short TR: Full brain T2*-weighted EPI with 2mm isotropic voxels, TR 1600ms, and (2) a multi-echo protocol that is more resilient to movement induced artifacts: Full brain T2*-weighted EPI with 3.5mm isotropic voxels, TR 1101ms.
Preprocessing involved fMRIPrep for single-echo data and a Nipype pipeline for multi-echo data, including fMRIPrep, tedana, and ANTs [4]–[7].
Spatial and temporal quality of the protocols were assessed using signal to noise ratios (sSNR/tSNR), framewise displacement (FD), and DVARS in 18 participants. 3 conditions were tested using these two protocols: a) resting-state (protocol 1,2), b) hand movement task (1,2), and a hand-posture task (2).
For FDG PET imaging, dynamic acquisition of PET images was implemented to track head movements, and correction was performed using a frame-based image-registration (FIR) methodology [8].
Results:
fMRI Protocols Optimization: Task scans increased artifacts (FD +37%, p<0.001) and reduced tSNR (-22%, p=0.002), while protocol 1 had 189% higher sSNR (p<0.001). After preprocessing, tSNR increased by 39% for protocol 1 (p<0.001), with resting state surpassing movement task tSNR (+26%, p=0.003). Protocol 2 showed no tSNR differences across tasks (p=0.47), but tSNR was 287% greater (p<0.001) than protocol 1 after preprocessing. Little head movements during rest led to selecting protocol 1 for the rest scan. No distinct tSNR differences in multi-echo task scans led to choosing the kinetic hand movement task to evoke more movement disorders.
PET Protocol: To address artifacts associated with motion, a dynamic acquisition protocol for FDG PET was implemented without restricting head movement. The FIR approach corrected motion in the dynamic images, providing a single static image for analysis.
PET Preprocessing: An in-house preprocessing pipeline was developed, combining fMRIPrep and Nipype. The robust procedure included HD-BET [9] for brain extraction, two-step coregistration using ANTs, and transformation into subject and MNI spaces. The NEMO FDG PET preprocessing pipeline is accessible at https://github.com/jrdalenberg/PETBrainPreprocessing, offering an open-source, Nipype workflow for PET BIDS [10] preprocessing.
Conclusions:
The NEMO project represents one of the most extensive studies into rare movement disorders. Its distinctive contribution lies in the integration of PET and fMRI alongside movement registration measurements. Notably, this study is the first to systematically apply these modalities across multiple hyperkinetic movement disorders, contributing to a more standardized approach to compare these rare movement disorders in future studies.
Modeling and Analysis Methods:
Methods Development
Motion Correction and Preprocessing
Motor Behavior:
Motor Behavior Other 2
Neuroinformatics and Data Sharing:
Workflows 1
Keywords:
Design and Analysis
FUNCTIONAL MRI
Motor
Movement Disorder
Positron Emission Tomography (PET)
Workflows
1|2Indicates the priority used for review
Provide references using author date format
[1] A. M. M. Van Der Stouwe et al., “Next move in movement disorders (NEMO): Developing a computer-aided classification tool for hyperkinetic movement disorders,” BMJ Open, vol. 11, no. 10, pp. 1–7, 2021.
[2] W. F. Abdo, B. P. C. van de Warrenburg, D. J. Burn, N. P. Quinn, and B. R. Bloem, “The clinical approach to movement disorders,” Nat. Rev. Neurol., vol. 6, no. 1, pp. 29–37, Jan. 2010.
[3] J. D. Power, M. Plitt, S. J. Gotts, P. Kundu, V. Voon, and P. A. Bandettini, “Ridding fMRI data of motion-related influences : Removal of signals with distinct spatial and physical bases in multiecho data,” Proc. Natl. Acad. Sci., vol. 115, no. 9, 2018.
[4] B. B. Avants, C. L. Epstein, M. Grossman, and J. C. Gee, “Symmetric diffeomorphic image registration with cross-correlation: Evaluating automated labeling of elderly and neurodegenerative brain,” Med. Image Anal., vol. 12, no. 1, pp. 26–41, 2008.
[5] E. DuPre et al., “TE-dependent analysis of multi-echo fMRI with tedana,” J. Open Source Softw., vol. 6, no. 66, p. 3669, 2021.
[6] O. Esteban et al., “fMRIPrep: a robust preprocessing pipeline for functional MRI,” Nat. Methods, vol. 16, no. 1, pp. 111–116, 2019.
[7] K. Gorgolewski et al., “Nipype: A flexible, lightweight and extensible neuroimaging data processing framework in Python,” Front. Neuroinform., vol. 5, no. August, 2011.
[8] N. Costes, A. Dagher, K. Larcher, A. C. Evans, D. L. Collins, and A. Reilhac, “Motion correction of multi-frame PET data in neuroreceptor mapping: Simulation based validation,” Neuroimage, vol. 47, no. 4, pp. 1496–1505, 2009.
[9] F. Isensee et al., “Automated brain extraction of multisequence MRI using artificial neural networks,” Hum. Brain Mapp., vol. 40, no. 17, pp. 4952–4964, 2019.
[10] M. Norgaard et al., “PET-BIDS, an extension to the brain imaging data structure for positron emission tomography,” Sci. Data, vol. 9, no. 1, pp. 1–7, 2022.