Poster No:
558
Submission Type:
Abstract Submission
Authors:
Xiaoqian Xiao1, Ariel Rokem1, Angela Fang1
Institutions:
1University of Washington, Seattle, WA
First Author:
Co-Author(s):
Introduction:
Anxiety disorders have extensive symptom overlap and high comorbidity rates, suggesting the possibility of a common underlying neural mechanism, yet no study has tested it yet. The current research used functional connectivity to identify the common brain signatures for current active anxiety, and assess its specificity by differentiating it from neural signatures associated with asymptomatic individuals with lifetime anxiety.
Methods:
The sample included individuals with current active anxiety symptoms (CA, n=367), those with a lifetime of anxiety but no active symptoms (PA, n=874), and matched healthy controls for each group (n=337 for CA, n=819 for PA). Resting-state fMRI time series were extracted from predefined regions of interest (ROIs) based on prior researches about anxiety disorders, including the bilateral dorsolateral prefrontal cortex (dlPFC), posterior cingulate cortex (PCC), medial prefrontal cortex (mPFC) and insula (Ins). Subcortical ROIs were defined using the Melbourne Subcortex Atlas, comprising bilateral anterior hippocampus (aHIP), posterior hippocampus (pHIP), dorsal-anterior thalamus (THA-DA), ventral-anterior thalamus (THA-VA), dorsal-posterior thalamus (THA-DP), ventral-posterior thalamus (THA-VP), medial amygdala (mAMY), lateral amygdala (lAMY), nucleus accumbens core (NAc-core), and nucleus accumbens shell (NAc-shell)). Pairwise functional connectivity between ROIs was computed, Fisher's Z-transformed, and utilized as input features for classification models.
Data sets were split into training (80%) and test sets for all analyses. Model selection included a range of linear and non-linear classifiers that balance interpretability and flexibility (Logistic Regression, Ridge Classifier, Lasso, Linear Discriminant Analysis, Perceptron, SVM, and Random Forest). Recursive feature elimination (RFE) with cross-validation identified the top 10 predictive features, followed by a 10-fold cross-validation framework and two-step grid search for hyperparameter optimization. Final models were trained on the entire training dataset and evaluated on independent test datasets.
Results:
First, we trained the model to classify CA versus healthy controls. A Lasso model (L1 regularization) was trained and achieved a classification accuracy of 66.4% and an area under the curve (AUC) of 0.667 on the test dataset. The identified brain signature highlighted thalamo-cortical, thalamo-basal, and thalamo-thalamic connections, as well as the connections between medial frontal cortex and the amygdala.
We then tested the specificity of the brain signature for CA by assessing both functional and spatial similarity. To evaluate functional similarity, we applied the CA model to classify PA versus healthy controls. This analysis yielded a classification accuracy of 52.5% and an AUC of 0.518, suggesting that the CA brain signature does not generalize to PA. For spatial similarity, we used cosine similarity to compare the brain signatures of CA and PA. A separate Lasso model (L1 regularization) was trained to classify PA versus healthy controls, yielded an accuracy of 55.2% and an AUC of 0.527. While the model performance was not robust, we proceeded the comparison to preliminarily examine spatial topography similarities between CA and PA. The low spatial similarity (.148) found, indicating distinct neural mechanisms underlying current versus past anxiety symptoms.
Conclusions:
These findings support the hypothesis of a shared neural basis for current anxiety, while also highlighting the specificity. However, the study did not identify a robust model for PA, and the performance of the CA model remains moderate, suggesting the existence of neural signatures beyond the regions and features examined in this study. Future research should explore whole-brain analyses and advanced feature engineering to capture these potential signatures.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
Connectivity (eg. functional, effective, structural)
Keywords:
Anxiety
Machine Learning
Other - functional connectivity, UK Biobank, brain signatures, resting-state fMRI
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 am submitting this abstract as an original work to be reproduced. I am available to be the “source party” in an upcoming team and consent to have this work listed on the OSSIG website. I agree to be contacted by OSSIG regarding the challenge and may share data used in this abstract with another team.
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?
No
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
For human MRI, what field strength scanner do you use?
3.0T
Provide references using APA citation style.
Koskinen, M.-K., & Hovatta, I. (2023). Genetic insights into the neurobiology of anxiety. Trends in Neurosciences, 46(4), 318–331. https://doi.org/10.1016/j.tins.2023.01.007
Penninx, B. W., Pine, D. S., Holmes, E. A., & Reif, A. (2021). Anxiety disorders. The Lancet, 397(10277), 914–927. https://doi.org/10.1016/S0140-6736(21)00359-7
Springer, K. S., Levy, H. C., & Tolin, D. F. (2018). Remission in CBT for adult anxiety disorders: A meta-analysis. Clinical Psychology Review, 61, 1–8. https://doi.org/10.1016/j.cpr.2018.03.002
Vollebergh, W. A., Iedema, J., Bijl, R. V., de Graaf, R., Smit, F., & Ormel, J. (2001). The structure and stability of common mental disorders: The NEMESIS study. Archives of General Psychiatry, 58(6), 597–603. https://doi.org/10.1001/archpsyc.58.6.597
Yang, X., Fang, Y., Chen, H., Zhang, T., Yin, X., Man, J., Yang, L., & Lu, M. (2021). Global, regional and national burden of anxiety disorders from 1990 to 2019: Results from the Global Burden of Disease Study 2019. Epidemiology and Psychiatric Sciences, 30, e36. https://doi.org/10.1017/S2045796021000275
No