# Code Tracker & Paper Overview
Find all tables, figures and results in the order in which they appear in-text. Supplementary materials are listed in the end. Name of the files for code, which can be downloaded, can be found for each element.
[Introduction]
[Results Section]
Table 1. Brain Age Prediction Model Performance
grid_search.py & prediction.py
Corrected brain age is calculated in
compare_correlations.R
Table 2. Features ranked by gain by age prediction model
read out from slurm files when applying
grid_search.py & prediction.py
Figure 1. Differences between correlations of chronological and corrected predicted age across models with 95% confidence interval
model_split_performance.R
Figure 2. Correlations of corrected BAG and age across models
compare_correlations.R
Figure 3. Correlation Matrix for Fornix Diffusion Metrics and Age
correlation_plots.R
Figure 4. Correlations between Diffusion Metrics and Age
correlation_plots.R
Testing linear models vs GAM
linear_vs_gam_no_ouliers.R
Figure 5. Fornix diffusion features across age
correlation_plots.R
Figure 6. Raw and Predicted Mean Diffusion Metrics by Age
Age curves
correlation_plots.R
Equivalence tests
prediction_vs_raw.R
[Discussion Section]
[Methods Section]
Figure 7. Density Plots for the Sample’s Age by Sex and Scanner Site
- simple ggplot application
Figure 8. Model Performance for Different Train-Test Splits
model_split_performance.R
Supplement
SF1: Correlations of uncorrected BAG and age across models
compare_correlations.R
SF2: Comparison of predicted and raw Fornix Z-scored diffusion metrics’ density
density_predicted_vs_raw.R
SF3: Comparison of predicted and raw Fornix Z-scored diffusion metrics’ density including outliers
compare_correlations.R
SF4: Correlations between Fornix diffusion metrics and chronological age for data including outliers
correlation_plots.R
SF5: Density plots for the sample’s age by sex and scanner site for data including outliers
age_sensitivity.R
SF6: Model performance for different train-test splits for data including outliers
model_split_performance.R
SF7: Correlations between diffusion metrics and chronological age for data including outliers
correlation_plots.R
ST1: Brain age predictions from different train-test splits
find these results in slurm and text files from
grid_search.py & prediction.py
ST2: Differences between correlations of chronological and corrected predicted age across models with 95% confidence interval
compare_correlations.R
ST3: Differences between correlations of uncorrected predicted and chronological age across models with 95% confidence interval
age_predictions
(Made by hand based on output from compare_correlations.R!!!)
ST4: Metrics' age sensitivity: comparing models with and without age as predictors
ST5: Model summaries for all 28 Fornix models
model_summaries.R
ST6: Brain age prediction model performance on data including outliers
grid_search.py & prediction.py
ST7: Top five features ranked by gain by age prediction model
grid_search.py & prediction.py
(slurm and output file readout)
ST8: Brain age predictions from different train-test splits on data including outliers
grid_search.py & prediction.py
ST9: Comparisons of linear and generalized additive models predicting Fornix features
age_sensitivity.R
