Identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fMRI
Altered cerebral blood flow (CBF), as measured by arterial spin labelling (ASL), has been observed in several psychiatric conditions, but is a generally underutilized MRI technique, especially in the study of psychosis spectrum (PS) symptoms. We aimed to determine group differences in ASL resting st...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2020-01-01
|
| Series: | NeuroImage: Clinical |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2213158220301418 |
| _version_ | 1818279428718854144 |
|---|---|
| author | Dawson J. Overton Nikhil Bhagwat Joseph D. Viviano Grace R. Jacobs Aristotle N. Voineskos |
| author_facet | Dawson J. Overton Nikhil Bhagwat Joseph D. Viviano Grace R. Jacobs Aristotle N. Voineskos |
| author_sort | Dawson J. Overton |
| collection | DOAJ |
| description | Altered cerebral blood flow (CBF), as measured by arterial spin labelling (ASL), has been observed in several psychiatric conditions, but is a generally underutilized MRI technique, especially in the study of psychosis spectrum (PS) symptoms. We aimed to determine group differences in ASL resting state functional connectivity (rsFC) between PS and non-PS youth, and the reliability of a support vector machine (SVM) classifier trained on ASL rsFC features to differentiate PS and non-PS youth, especially compared to blood oxygen level dependent (BOLD) fMRI. 1146 youth aged 8–22 with ASL and BOLD data from the Philadelphia Neurodevelopmental Cohort were analyzed. Widespread ASL hyperconnectivity was found in the left cuneus, precuneus, and dorsolateral prefrontal cortex, and hypoconnectivity was found in the left cingulate cortex and orbitofrontal area (multiple linear regression, FDR corrected). An SVM trained on ASL and BOLD features outperformed either modality alone (AUCBOTH = 0.72 versus AUCASL = 0.68 and AUCBOLD = 0.67). Classification performance and precision improved when the non-PS group had no psychiatric comorbidities. The relative success of the classifier suggests ASL rsFC changes exist in PS individuals that differ from BOLD rsFC changes, and extends previous findings of CBF dysregulation in PS. |
| first_indexed | 2024-12-12T23:33:11Z |
| format | Article |
| id | doaj.art-e780ce1f92634a2c9235e49f91f3ef43 |
| institution | Directory Open Access Journal |
| issn | 2213-1582 |
| language | English |
| last_indexed | 2024-12-12T23:33:11Z |
| publishDate | 2020-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | NeuroImage: Clinical |
| spelling | doaj.art-e780ce1f92634a2c9235e49f91f3ef432022-12-22T00:07:40ZengElsevierNeuroImage: Clinical2213-15822020-01-0127102304Identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fMRIDawson J. Overton0Nikhil Bhagwat1Joseph D. Viviano2Grace R. Jacobs3Aristotle N. Voineskos4Kimel Family Translational Imaging Genetics Research Laboratory, The Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; Campbell Family Mental Health Research Institute, The Centre for Addiction and Mental Health, University of Toronto, Toronto, CanadaKimel Family Translational Imaging Genetics Research Laboratory, The Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; Campbell Family Mental Health Research Institute, The Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; Cerebral Imaging Centre, Douglas Mental Health University Institute, Montreal, Quebec, Canada; Department of Biological and Biomedical Engineering, McGill University, Montreal, Quebec, Canada; Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Ontario, CanadaKimel Family Translational Imaging Genetics Research Laboratory, The Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; Campbell Family Mental Health Research Institute, The Centre for Addiction and Mental Health, University of Toronto, Toronto, CanadaKimel Family Translational Imaging Genetics Research Laboratory, The Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; Campbell Family Mental Health Research Institute, The Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; Institute of Medical Science, University of Toronto, Toronto, CanadaKimel Family Translational Imaging Genetics Research Laboratory, The Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; Campbell Family Mental Health Research Institute, The Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada; Department of Psychiatry, University of Toronto, Toronto, Canada; Corresponding author at: Kimel Family Translational Imaging Genetics Research Laboratory, The Centre for Addiction and Mental Health, 250 College St., Toronto, Ontario M5T 1R8, Canada.Altered cerebral blood flow (CBF), as measured by arterial spin labelling (ASL), has been observed in several psychiatric conditions, but is a generally underutilized MRI technique, especially in the study of psychosis spectrum (PS) symptoms. We aimed to determine group differences in ASL resting state functional connectivity (rsFC) between PS and non-PS youth, and the reliability of a support vector machine (SVM) classifier trained on ASL rsFC features to differentiate PS and non-PS youth, especially compared to blood oxygen level dependent (BOLD) fMRI. 1146 youth aged 8–22 with ASL and BOLD data from the Philadelphia Neurodevelopmental Cohort were analyzed. Widespread ASL hyperconnectivity was found in the left cuneus, precuneus, and dorsolateral prefrontal cortex, and hypoconnectivity was found in the left cingulate cortex and orbitofrontal area (multiple linear regression, FDR corrected). An SVM trained on ASL and BOLD features outperformed either modality alone (AUCBOTH = 0.72 versus AUCASL = 0.68 and AUCBOLD = 0.67). Classification performance and precision improved when the non-PS group had no psychiatric comorbidities. The relative success of the classifier suggests ASL rsFC changes exist in PS individuals that differ from BOLD rsFC changes, and extends previous findings of CBF dysregulation in PS.http://www.sciencedirect.com/science/article/pii/S2213158220301418 |
| spellingShingle | Dawson J. Overton Nikhil Bhagwat Joseph D. Viviano Grace R. Jacobs Aristotle N. Voineskos Identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fMRI NeuroImage: Clinical |
| title | Identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fMRI |
| title_full | Identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fMRI |
| title_fullStr | Identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fMRI |
| title_full_unstemmed | Identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fMRI |
| title_short | Identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fMRI |
| title_sort | identifying psychosis spectrum youth using support vector machines and cerebral blood perfusion as measured by arterial spin labeled fmri |
| url | http://www.sciencedirect.com/science/article/pii/S2213158220301418 |
| work_keys_str_mv | AT dawsonjoverton identifyingpsychosisspectrumyouthusingsupportvectormachinesandcerebralbloodperfusionasmeasuredbyarterialspinlabeledfmri AT nikhilbhagwat identifyingpsychosisspectrumyouthusingsupportvectormachinesandcerebralbloodperfusionasmeasuredbyarterialspinlabeledfmri AT josephdviviano identifyingpsychosisspectrumyouthusingsupportvectormachinesandcerebralbloodperfusionasmeasuredbyarterialspinlabeledfmri AT gracerjacobs identifyingpsychosisspectrumyouthusingsupportvectormachinesandcerebralbloodperfusionasmeasuredbyarterialspinlabeledfmri AT aristotlenvoineskos identifyingpsychosisspectrumyouthusingsupportvectormachinesandcerebralbloodperfusionasmeasuredbyarterialspinlabeledfmri |