Fingerprinting individual differences in lesion impact through imaging: The FIDELITI Dashboard
Poster No:
337
Submission Type:
Abstract Submission
Authors:
Helen Carlson1, Jordan Hassett1, Brandon Craig1, Alicia Hilderley1, Keith Yeates1, Melanie Noel1, Jillian Miller1, Frank MacMaster2, Signe Bray1, Karen Barlow3, Brian Brooks1, Catherine Lebel1, Nils Forkert1, Adam Kirton1
Institutions:
1University of Calgary, Calgary, Alberta, 2Dalhousie School of Medicine, Halifax, Nova Scotia, 3University of Queensland, Brisbane, Queensland
First Author:
Co-Author(s):
Introduction:
Childhood and adolescence are periods of massive developmental change continuing into early adulthood. Traditionally used developmental growth charts measuring age-related trajectories for height and weight have provided insights into normal variation around reference data. Availability of very large open-source neuroimaging databases have recently afforded the creation of similar developmental trajectories in brain biomarkers. For neurodevelopmental diseases or brain injuries, deviations from typical developmental trajectories are of particular interest and may explain disabilities while predicting long-term development across the lifespan. Here we introduce the FIDELITI Dashboard (Fingerprinting Individual Differences in Lesion Impact Through Imaging), a patient-centered dashboard that visualizes multimodal brain neuroimaging biomarkers "at a glance". We present several cases illustrating clinical utility of the dashboard by capturing personalized neuroimaging profiles of children with perinatal stroke, the leading cause of hemiparetic cerebral palsy, a non-progressive but lifelong motor disability. In addition to hemiparesis, individuals may also show deficits in attention, executive function, language, and vision. Identifying areas of concern early may facilitate personalized therapeutic interventions.
Methods:
The FIDELITI Dashboard was developed in Python. Reference neuroimaging biomarkers were extracted from 828 typically developing volunteers aged between 6.5-24.0 years (mean age (SD) = 14.50 (3.75) years, 47% male) from either the Human Connectome Project Development (n=609, five 3T Siemens Prisma scanners) or the Alberta Children's Hospital imaging collaboration (n=219, one 3T GE MR750w scanner). For additional clinical validation, scans of six children with perinatal stroke (mean age (SD) = 11.2 (2.4) years, 5 males, 3T GE MR750w scanner) were processed and visualized using the FIDELITI Dashboard. Cortical morphometry metrics (cortical thickness, region volumes) were extracted from T1-weighted images using CAT12. Functional connectivity between Harvard-Oxford atlas regions was extracted from resting state functional scans using CONN. White matter microstructure metrics (fractional anisotropy, mean diffusivity) were extracted for 21 major white matter bundles reconstructed using diffusion scans processed in MRtrix3. These biomarkers have previously been shown to be associated with clinical function (Craig, 2021). ComBat (Fortin, 2017) was applied to biomarkers individually to harmonize, specifying sex and age as covariates to preserve. Extensive visualization and harmonization options are provided for users to customize their dashboards. Fundamental components within the dashboard are organized into six domains based on previous literature regarding functional circuits: Sensorimotor, Language, Vision, Attention/Executive function, Memory, and Audition. Over 200 parameters across these six functional domains and four primary imaging modalities are summarized in a dashboard format providing a fully customizable, at-a-glance summary of brain imaging biomarkers.
Results:
For children with stroke, deviations from the reference cohort (Figure 1) were seen for cortical thickness and volume of lesioned precentral gyrus as well as interhemispheric functional connectivity between primary motor cortices. Metrics extracted from the non-lesioned hemisphere often fell within the normal variation of the reference cohort. Additional, non-motor domains also showed deviations in some stroke participants, such as functional connectivity for language and executive function networks identifying domains that could potentially be treated with intensive cognitive therapy.
Conclusions:
The FIDELITI Dashboard is patient-centered, fully customizable, and has potential applications for many other neurodevelopmental conditions or early life brain injuries. FIDELITI is available online (https://cumming.ucalgary.ca/labs/carlson-imaging/projects/fideliti-dashboard).
Disorders of the Nervous System:
Neurodevelopmental/ Early Life (eg. ADHD, autism) 1
Lifespan Development:
Early life, Adolescence, Aging 2
Modeling and Analysis Methods:
Other Methods
Novel Imaging Acquisition Methods:
Multi-Modal Imaging
Keywords:
Development
FUNCTIONAL MRI
Morphometrics
Movement Disorder
Neurological
Open-Source Software
Pediatric Disorders
Plasticity
Tractography
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
1|2Indicates the priority used for review
