Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
[Preprint]. 2025 Jul 1:rs.3.rs-6866544.
doi: 10.21203/rs.3.rs-6866544/v1.

Brain Age Prediction in Generalized Anxiety Disorder using a Convolutional Neural Network

Corey Richier  1   2 André Zugman  2 Anita Harrewijn  2   3 Elise M Cardinale  2   4 Parmis Khosravi  2 Moji Aghajani  5 Willem B Bruin  5 Kevin Hilbert  6 Narcis Cardoner  7   8   9 Daniel Porta-Casteràs  7 Marta Cano  7   8   9 Savannah Gosnell  10 Ramiro Salas  10 Andrea P Jackowski  11 Pedro M Pan  11 Giovanni A Salum  12 Karina S Blair  13 James R Blair  14 Mohammed R Milad  15 Katie L Burkhouse  16 K Luan Phan  17 Heidi K Schroeder  18 Jeffrey R Strawn  18 Katja Beesdo-Baum  19 Neda Jahanshad  20 Sophia I Thomopoulos  20 Jared A Nielsen  21   22   23 Jordan W Smoller  22 Jair C Soares  24 Benson Mwangi  24 Mon-Ju Wu  24 Giovana B Zunta-Soares  24 Michal Assaf  25   26 Gretchen J Diefenbach  26   27 Paolo Brambilla  28   29 Eleonora Maggioni  28 David Hofmann  30 Thomas Straube  30 Carmen Andreescu  31 Rebecca B Price  31   32 Gisele G Manfro  33 Federica Agosta  34   35 Elisa Canu  34 Camilla Cividini  34   35 Massimo Filippi  34   35   36   37   38 Milutin Kostić  39   40 Ana Munjiza Jovanovic  39 Brenda Benson  2 Gabrielle F Freitag  2 Ellen Leibenluft  2 Grace V Ringlein  2   41 Kathryn Werwath  42 Hannah Zwiebel  2 Hans J Grabe  43   44 Sandra Van der Auwera  43   44 Katharina Wittfeld  43 Henry Völzke  45 Robin Bülow  46 Nicholas L Balderston  47 Monique Ernst  48 Lilianne R Mujica-Parodi  49 Helena van Nieuwenhuizen  50 Hugo D Critchley  51 Elena Makovac  52 Matteo Mancini  53 Frances Meeten  54 Cristina Ottaviani  55   56 Gregory A Fonzo  57 Martin P Paulus  58 Murray B Stein  59 Raquel E Gur  60 Ruben C Gur  60 Antonia N Kaczkurkin  61 Bart Larsen  60   62 Theodore D Satterthwaite  60 Jennifer Harper  63 Michael T Perino  63 Chad M Sylvester  63 Qiongru Yu  63 Patrick McClure  64 Francisco Pereira  64 Ulrike Lueken  65 Dick J Veltman  66 Paul M Thompson  20 Nynke A Groenewold  67 Janna Marie Bas-Hoogendam  68   69   70 Dan J Stein  71   67 Nic J A Van der Wee  68   70 Anderson M Winkler  2   72 Daniel S Pine  2 Chelsea K Sawyers  2   73
Affiliations

Brain Age Prediction in Generalized Anxiety Disorder using a Convolutional Neural Network

Corey Richier et al. Res Sq. .

Abstract

Higher predicted brain age difference has been associated with several psychiatric disorders. Generalized anxiety disorder (GAD) is associated with markers of accelerated aging. In this study, we determined brain predicted age difference (PAD) in individuals with GAD and healthy controls (HC) as well as group differences in PAD variability using voxel-wise structural MRI. The training dataset included 3,511 controls, and the testing dataset included 1,595 individuals with GAD and 4,552 HC from the ENIGMA-Anxiety GAD Working Group. A convolutional neural network model using four input modalities per subject and a model ensemble approach was used to predict brain age. The PAD was then calculated by subtracting chronological age. Model performance was consistent with other image-based brain age prediction models with similar accuracy across the training set (mean absolute error (MAE) = 2.95 years) and HC in the testing set (MAE = 2.94). We found no evidence of accelerated brain aging in individuals with GAD, though we did find evidence for greater variation in PAD for individuals with GAD (Levene's test: W = 442.98, p < .001) and evidence for greater variability in PAD of those with GAD over 25 years of age. No relationships between PAD and clinical or demographic measures were found. To conclude, using large training and testing samples, the study found no significant association between GAD and PAD, although individuals with GAD had greater heterogeneity in brain-predicted age.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Overview of the CNN Model Architecture. Note. The CNN architecture was designed with four input channels for gray and white matter segmentations, as well as the T1-weighted and Jacobian images. The model consists of five repeated convolutional blocks, each containing 3D convolutional layers, batch normalization, ReLU activations and max pooling. The number of filters used in each block are denoted above. After the last convolutional block, the model includes a flattened layer.
Figure 2
Figure 2
Model Ensemble Procedure Note. Procedure for generating average predictions for the model ensemble. Eight splits of the training data were created, which used a different 90/10% split of training to validation data. Each of these splits was given to a unique instantiation of the model with identical architecture but instantiated with random weights.
Figure 3
Figure 3
Scatter Plot of True Age vs. Predicted Age Sorted By Diagnostic Status
Figure 4
Figure 4
Violin Plot of Variance in PAD across Age and Diagnostic Groups

References

    1. Szuhany K.L. and Simon N.M., Anxiety Disorders: A Review. JAMA, 2022. 328(24): p. 2431–2445. - PubMed
    1. Wittchen H.-U., Generalized anxiety disorder: prevalence, burden, and cost to society. Depression and Anxiety, 2002. 16(4): p. 162–171. - PubMed
    1. Kessler R.C., et al. , Prevalence, Severity, and Comorbidity of 12-Month DSM-IV Disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 2005. 62(6): p. 617. - PMC - PubMed
    1. Kessler R.C., et al. , Lifetime Prevalence and Age-of-Onset Distributions of DSM-IV Disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry, 2005. 62(6): p. 593. - PubMed
    1. Cole J.H., et al. , Predicting brain age with deep learning from raw imaging data results in a reliable and heritable biomarker. NeuroImage, 2017. 163: p. 115–124. - PubMed

Publication types

LinkOut - more resources