Bin Wan, Seok-Jun Hong, Richard AI Bethlehem, Dorothea L Floris, Boris C Bernhardt, Sofie L Valk.
Molecular Psychiatry (2023).
https://www.nature.com/articles/s41380-023-02220-x
Preprint version
https://www.biorxiv.org/content/10.1101/2023.04.05.535683v1

How to run the codes

Activate the enviornment

• $ conda activate autism
• $ cd [working directory]/autism/

Clean the data according to the requirements

• Here, I deleted the subjects with bad quality of MRI
• And remove IQ<70 or without IQ information
• Keep all the boys
• Age from 5-40 years
• Head motion with mean_FD<0.3mm
• Finally there are 5 datasites survived including 283 subjects.

Prepare the phenotype dataframe

• ID, site, group, age, FIQ, ADOS_social, ADOS_communication, ADOS_rrb, mean_FD
  $ python scripts/data_sort.py
  The output here is: 'abide_demo_sort.csv'

Process the fMRI data from time series to FC to gradients

• FC
  Input: '../data/data_autism/1_fc/'
  $ python scripts/data_process_fc.py
  Output: 'results/fc/'
• Gradients
  Input: 'results/fc/'
  $ python scripts/data_process_grad_HCP_template.py
  Output: 'results/grad/'

Statics

$ jupyter-lab

1. Demographics
   click the 'scripts/vis_basic_stas.ipython' (Table S1)
2. Comparisons between ASD and controls and age effcts click the 'scripts/vis_main.ipython' (Figures 1, 2, S1, S2, S8, S9 and Tables S2, S3, S4, S5, S6)
3. Enrichment analyses
   click the 'scripts/vis_enrichment.ipynb' (Figures 3, S6, S7, S8, and Table S7)
4. Machine learning prediction
   $ python scripts/prediction.py
   Output: 'results/prediction/'
   then in ipython notebook, click the 'scripts/vis_EN_ML.ipynb' (Figures 4, S9, and S10)
5. Global signal regression
   Input: '../data/data_autism/1_fc/'
   $ python scripts/data_process_GSR.py
   Output: 'results/GSR/'
6. FIQ, head motion removal
   click the 'scripts/vis_main_fIQ_HeadMotion.ipynb'

PS: Bulit-in functions are shown in 'scripts/func_utils.py'

Main dependencies based on Python 3.9

• BrainSpace
• BrainStat
• Scikit-learn
• SciPy
• neuroCombat

Acknowdgements

• Autism Brain Imaging Data Exchange
• Human Connectome Project
• Funding sources from Boris C. Bernhardt & Sofie L. Valk