Advancing PET Surface-based Analysis - Integration in a Multimodal Imaging Pipeline
Poster No:
2273
Submission Type:
Abstract Submission
Authors:
Clara Fonteneau1, Mika Naganawa1, Sophie Holmes1, Markus Helmer2, Jie Lisa Ji2, Irina Esterlis1, William Martin3, Richard Carson1, Alan Anticevic1
Institutions:
1Yale University, New Haven, CT, 2Manifest Technologies, N/A, N/A, 3Johnson & Johnson Innovative Medicine, N/A, N/A
First Author:
Co-Author(s):
Introduction:
Investigating patient-specific differences remains a challenge when using neuroimaging approaches. Current studies have focused independently on pharmacological effects either using broad cortical network functional connectivity (i.e. resting state networks) acquired with fMRI or receptor quantification acquired with positron emission tomography (PET), but rarely integrating both modalities. PET analyses are traditionally implemented in a 3-dimensional volumetric space, which limits multi-modal comparisons across cortical areas that inherently follow a 2-dimensional geometry. We hypothesize superior alignment of modalities by representing the cerebral cortex as a two-dimensional geometry, i.e. surface-based analyses. Specifically, we aim to test the hypothesis that mapping PET data to single-subject cortical surfaces will reduce inter-subject variance in kinetic modeling as compared to a traditional volume-based workflow. This workflow will allow us to leverage state-of-the-art cortical and subcortical parcellations.
Methods:
We set the framework to optimize the integration of PET with BOLD fMRI data across both cortex and subcortex to enable multi-modal comparisons of receptor distribution in relation to clinical dysconnectivity. All neuroimaging data was processed using the Quantitative Neuroimaging Environment and Toolbox (QuNex), which integrates HCP Pipeline workflows. PET data was analyzed using the novel surface-based workflow developed (Figure 1). We applied surface-based analytics either after (traditional volume-based workflow) or before (surface-based workflow) PET BPND quantification to assess impact on alignment and inter-subject variance.
Results:
We demonstrate the feasibility of our proposed framework. We have successfully integrated PET surface-based analysis workflow within QuNex, a multimodal imaging pipeline. This allows us to integrate PET analytics with other neuroimaging modalities (Figure 1). We show that using an optimized surface-based workflow yields a more precise alignment on the cortical surface where quantification is done before compared to after mapping to the surface.
·Figure 1 Workflow schematic
Conclusions:
This optimized framework is a key step in multimodal neuroimaging integration (e.g. PET and fMRI) aimed at investigating the specific molecular effects of pharmacological treatments (e.g. ketamine), which would be critical in advancing the mechanistic understanding of neuropsychiatric disorders and their treatment. This would provide a more complete picture and help identify networks where pharmacological treatment effects relate to symptom response, enabling a more individually-targeted treatment plan.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia)
Modeling and Analysis Methods:
Methods Development
PET Modeling and Analysis 2
Neuroinformatics and Data Sharing:
Workflows 1
Keywords:
MRI
Positron Emission Tomography (PET)
Workflows
Other - Multimodal
1|2Indicates the priority used for review