Poster No:
264
Submission Type:
Abstract Submission
Authors:
Daniel Feldman1,2,3, Molly Prigge2,3,4, Nick Lange5, Doug Dean III4,6, Jubel Morgan2,3, Carolyn King2,3, Erin Bigler7,8,9,10, Andrew Alexander4,6,11, Brandon Zielinski2,8,12,13, Janet Lainhart4,13, Jace King1,2,3
Institutions:
1Department of Biomedical Engineering, University of Utah, Salt Lake City, UT, 2Department of Radiology & Imaging Sciences, University of Utah, Salt Lake City, UT, 3Utah Center for Advanced Imaging Research, University of Utah, Salt Lake City, UT, 4Waisman Center, University of Wisconsin-Madison, Madison, WI, 5Department of Psychiatry, Harvard Medical School, Cambridge , MA, 6Departments of Medical Physics, University of Wisconsin-Madison, Madison, WI, 7Department of Psychology and Neuroscience Center,Brigham Young University, Provo, UT, 8Department of Neurology, University of Utah, Salt Lake City, UT, 9Department of Psychiatry, University of Utah, Salt Lake City, UT, 10Department of Neurology, University of California-Davis, Davis, CA, 11Department of Psychiatry, University of Wisconsin-Madison, Madison, WI, 12Departments of Pediatrics, Neurology, and Neuroscience, University of Florida, Gainesville, FL, 13Department of Pediatrics, University of Utah, Salt Lake City, UT
First Author:
Daniel Feldman
Department of Biomedical Engineering, University of Utah|Department of Radiology & Imaging Sciences, University of Utah|Utah Center for Advanced Imaging Research, University of Utah
Salt Lake City, UT|Salt Lake City, UT|Salt Lake City, UT
Co-Author(s):
Molly Prigge
Department of Radiology & Imaging Sciences, University of Utah|Utah Center for Advanced Imaging Research, University of Utah|Waisman Center, University of Wisconsin-Madison
Salt Lake City, UT|Salt Lake City, UT|Madison, WI
Nick Lange
Department of Psychiatry, Harvard Medical School
Cambridge , MA
Doug Dean III
Waisman Center, University of Wisconsin-Madison|Departments of Medical Physics, University of Wisconsin-Madison
Madison, WI|Madison, WI
Jubel Morgan
Department of Radiology & Imaging Sciences, University of Utah|Utah Center for Advanced Imaging Research, University of Utah
Salt Lake City, UT|Salt Lake City, UT
Carolyn King
Department of Radiology & Imaging Sciences, University of Utah|Utah Center for Advanced Imaging Research, University of Utah
Salt Lake City, UT|Salt Lake City, UT
Erin Bigler
Department of Psychology and Neuroscience Center,Brigham Young University|Department of Neurology, University of Utah|Department of Psychiatry, University of Utah|Department of Neurology, University of California-Davis
Provo, UT|Salt Lake City, UT|Salt Lake City, UT|Davis, CA
Andrew Alexander
Waisman Center, University of Wisconsin-Madison|Departments of Medical Physics, University of Wisconsin-Madison|Department of Psychiatry, University of Wisconsin-Madison
Madison, WI|Madison, WI|Madison, WI
Brandon Zielinski
Department of Radiology & Imaging Sciences, University of Utah|Department of Neurology, University of Utah|Departments of Pediatrics, Neurology, and Neuroscience, University of Florida|Department of Pediatrics, University of Utah
Salt Lake City, UT|Salt Lake City, UT|Gainesville, FL|Salt Lake City, UT
Janet Lainhart
Waisman Center, University of Wisconsin-Madison|Department of Pediatrics, University of Utah
Madison, WI|Salt Lake City, UT
Jace King
Department of Biomedical Engineering, University of Utah|Department of Radiology & Imaging Sciences, University of Utah|Utah Center for Advanced Imaging Research, University of Utah
Salt Lake City, UT|Salt Lake City, UT|Salt Lake City, UT
Introduction:
The global prevalence of autism spectrum disorder (ASD) is estimated to be 1 in 100 individuals (Ziedan et al., 2022). Autistic individuals are at higher risk for co-occurring conditions such as depression, anxiety, and epilepsy (Kuhlthau et al., 2010). Further, ASD is known to impact caregivers and lifetime care can constitute a significant financial burden (Lindly et al., 2022; Roggee et al., 2019). Despite this prevalence and health impact, significant barriers exist to effective and efficient diagnosis of ASD, specifically in adults (Lewis, 2017). Autism diagnosis requires a qualified team to assess behavioral, historical, and parental report information (Horlin et al., 2014). As such, an alternative efficient, cost-effective biological method for ASD diagnosis is needed. Multimodal MRI may be an apt tool to fill this need (Schielen et al., 2024).
Methods:
120 individuals [65 typically developing (TD), age 27.7 years +/- 6.73 | 55 ASD, age 27.3 +/- 8.07], participated in a multimodal MRI study using a Siemens 3T Prisma-Fit at the University of Utah. The MRI contained a structural MP2RAGE (sMRI), multi-shell spin-echo echo-planar pulse diffusion-weighted image (dMRI), and multi-band multi-echo echo-planar resting-state functional scan (fMRI). sMRI was processed using FreeSurfer v6.0.0 to extract cortical and subcortical thickness, surface area, and volumes (Knussmann et al., 2022). dMRI was processed with an in-house pipeline to extract quantitative maps of fractional anisotropy (FA), radial diffusivity (RD), axial diffusivity (AD), and orientation dispersion index (ODI) (DiPiero et al., 2023). fMRI was processed using the AFNI ME-ICA pipeline to quantify functional connectivity values between time-series of subcortical regions and a network-based cortical template (King et al., 2018). Following multimodal feature extraction, all data was standardly scaled yielding 934 neural features for each individual.
Multimodal MR features were then reduced to 40 features using gradient-boosted recursive feature elimination on the entire dataset. These 40 features then underwent a 70/30 train-test split, and classifier training was conducted with a 10-fold cross-validated grid-search random forest classifier. The classifier was tested on 1000 training-naïve, bootstrapped testing datasets to gain confidence intervals and mean performance for model accuracy, sensitivity, specificity, and ROC-AUC. Following testing, feature importance in the classification model was explained using random-forest specific Shapely Additive Explanations (SHAP).
Results:
Our modeling approach resulted in a classification accuracy of 89.1 ± 5.1%, sensitivity of 94.2 ± 5.6%, specificity of 84.5 ± 8.3%, and ROC-AUC of 0.91 ± 0.05. Moreover, SHAP values revealed a cohesive picture of sMRI, dMRI, and fMRI cingulate-related neural features driving the classification model. Specifically, our results show decreased sMRI cingulate volumes, decreased cingulate – limbic fMRI connectivity, increased dMRI corpus callosum radial diffusivity, and decreased dMRI internal capsule fractional anisotropy as driving characteristics of our ASD classification model.
Conclusions:
Using a series of machine learning ensemble methods, our model effectively classified ASD and TD individuals with high confidence. The use of MRI to diagnose ASD may help those individuals who do not have access to early historical records or informants. Further, explainable Shapley estimations revealed multimodal structural and functional cingulate differences in the ASD brain. The cingulate, which is involved in behavioral adaptation, emotional regulation, and social cognition, may be an appropriate neural biomarker for altered cognitive functions in ASD.
Disorders of the Nervous System:
Neurodevelopmental/ Early Life (eg. ADHD, autism) 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
Connectivity (eg. functional, effective, structural)
Diffusion MRI Modeling and Analysis
Novel Imaging Acquisition Methods:
Multi-Modal Imaging
Keywords:
Autism
Data analysis
FUNCTIONAL MRI
Machine Learning
STRUCTURAL MRI
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
1|2Indicates the priority used for review
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
I accept
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information.
Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
I do not want to participate in the reproducibility challenge.
Please indicate below if your study was a "resting state" or "task-activation” study.
Resting state
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Patients
Was this research conducted in the United States?
Yes
Are you Internal Review Board (IRB) certified?
Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Yes, I have IRB or AUCC approval
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel?
NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Yes
Were any animal research approved by the relevant IACUC or other animal research panel?
NOTE: Any animal studies without IACUC approval will be automatically rejected.
Not applicable
Please indicate which methods were used in your research:
Functional MRI
Structural MRI
Diffusion MRI
For human MRI, what field strength scanner do you use?
3.0T
Which processing packages did you use for your study?
AFNI
FSL
Free Surfer
Other, Please list
-
MRtrix3, ANTs, DIPY, DMIPY
Provide references using APA citation style.
DiPiero, M. (2023). Tract- and gray matter- based spatial statistics show white matter and
gray matter microstructural differences in autistic males. Frontiers in Neuroscience, 17. https://doi.org/10.3389/fnins.2023.1231719
Horlin, C. (2014). The Cost of Autism Spectrum Disorders. PLOS ONE, 9(9), e106552.
https://doi.org/10.1371/journal.pone.0106552
King, J. B. (2018). Evaluation of Differences in Temporal Synchrony Between Brain
Regions in Individuals With Autism and Typical Development. JAMA Network Open, 1(7), e184777. https://doi.org/10.1001/jamanetworkopen.2018.4777
Knussmann, G. N. (2022). Test-retest reliability of FreeSurfer-derived volume, area and
cortical thickness from MPRAGE and MP2RAGE brain MRI images. Neuroimage. Reports, 2(2), 100086. https://doi.org/10.1016/j.ynirp.2022.100086
Kuhlthau, K. (2010). Health-Related Quality of Life in Children with Autism Spectrum
Disorders: Results from the Autism Treatment Network. Journal of Autism and Developmental Disorders, 40(6), 721–729. https://doi.org/10.1007/s10803-009-0921-2
Lewis, L. F. (2017). A Mixed Methods Study of Barriers to Formal Diagnosis of Autism
Spectrum Disorder in Adults. Journal of Autism and Developmental Disorders, 47(8), 2410–2424. https://doi.org/10.1007/s10803-017-3168-3
Lindly, O. J. (2022). Caregiver strain among North American parents of children from the
Autism Treatment Network Registry Call-Back Study. Autism, 26(6), 1460–1476. https://doi.org/10.1177/13623613211052108
Rogge, N. (2019). The Economic Costs of Autism Spectrum Disorder: A
Literature Review. Journal of Autism and Developmental Disorders, 49(7), 2873–2900. https://doi.org/10.1007/s10803-019-04014-z
Schielen, S. J. C. (2024). The diagnosis of ASD with MRI: A systematic review and
meta-analysis. Translational Psychiatry, 14(1), 1–11. https://doi.org/10.1038/s41398-024-03024-5
Zeidan, J. (2022). Global prevalence of autism: A systematic review update. Autism
Research: Official Journal of the International Society for Autism Research, 15(5), 778–790. https://doi.org/10.1002/aur.2696
No