2024_Ruthe_SND/code/correlation_between_bahavior_and_DTI_anovaSpecific.py

import os
import pandas as pd
from scipy.stats import spearmanr, shapiro, pearsonr
import numpy as np
from statsmodels.stats.multitest import multipletests

# Get the directory where the code file is located
code_dir = os.path.dirname(os.path.abspath(__file__))

# Get the parent directory of the code directory
parent_dir = os.path.dirname(code_dir)

# Step 4: Save the resulting dataframe to a CSV file
input_file_path = os.path.join(parent_dir, 'output', "Final_Quantitative_output", 'Quantitative_results_from_dwi_processing_merged_with_behavior_data.csv')

input_anova_file = os.path.join(parent_dir, 'output', "Final_Quantitative_output","mixed_model_analysis",'mixed_model_results.csv')
df_annova = pd.read_csv(input_anova_file)
df_annova2 = df_annova[df_annova["p_value"]<=0.05]






df_temp = pd.read_csv(input_file_path)
# Filter df based on condition
df_f1 = df_temp[df_temp["dialation_amount"] == 2]

# Merge df_f1 and df_annova2 on common columns "Qtype" and "mask_name"
merged_df = pd.merge(df_f1, df_annova2[['Qtype', 'mask_name', 'p_value']], on=['Qtype', 'mask_name'], how='left')

# Display the merged DataFrame
print(merged_df)


#%%
# Empty list to store the results
results = []

df = merged_df[(merged_df["p_value"]<=0.05) & (merged_df["merged_timepoint"]<40)]
df.dropna(subset=['p_value'], inplace=True)


# Iterate over unique values of 'Group', 'merged_timepoint', 'Qtype', and 'mask_name'
for ss in df["Group"].unique():
    for time_point in df["merged_timepoint"].unique():
        for qq in df["Qtype"].unique():
            for mm in df["mask_name"].unique():
                # Filter the DataFrame for the current combination of 'Group', 'merged_timepoint', 'Qtype', and 'mask_name'
                df_f2 = df[(df["Group"] == ss) & (df["merged_timepoint"] == time_point) & (df["Qtype"] == qq) & (df["mask_name"] == mm)]

                # Remove rows with NaN values in 'Value' or 'averageScore' columns
                df_f2 = df_f2.dropna(subset=['Value', 'averageScore'])

                if not df_f2.empty:
                    
                    shapiro_statValue, shapiro_pvalueQValue = shapiro(df_f2["Value"])
                    shapiro_statScore, shapiro_pvalueBehavior = shapiro(df_f2["averageScore"])
                    
                    if shapiro_pvalueQValue < 0.05 or shapiro_pvalueBehavior < 0.05:
                        correlation_coefficient, p_value = pearsonr(df_f2["Value"], df_f2["averageScore"])
                    else:
                        correlation_coefficient, p_value = pearsonr(df_f2["Value"], df_f2["averageScore"])
                   

                    # Store the results in a dictionary
                    result = {'Group': ss, 'merged_timepoint': time_point, 'Qtype': qq, 'mask_name': mm, 
                              'Pval': p_value, 'R': correlation_coefficient,
                              'shapiro-wilk_pvalue_qtype': shapiro_pvalueQValue,'shapiro-wilk_pvalue_behavior': shapiro_pvalueBehavior}

                    # Append the dictionary to the results list
                    results.append(result)
                else:
                    print(
                        f"No valid data found for Group: {ss}, merged_timepoint: {time_point}, Qtype: {qq}, and mask_name: {mm}. Skipping.")

# Create a DataFrame from the results list
correlation_results_df = pd.DataFrame(results)
unique_groups = df["Group"].unique()
unique_masks = df["mask_name"].unique()
unique_tp = df["merged_timepoint"].unique()

for tt in unique_tp:
    for mask in unique_masks:
        for group in unique_groups:
            time_mask = correlation_results_df['merged_timepoint'] == tt
            mask_mask = correlation_results_df['mask_name'] == mask
            group_mask = correlation_results_df['Group'] == group
            combined_mask = time_mask & mask_mask & group_mask
            
            p_values = correlation_results_df[combined_mask]['Pval']
            rejected, p_values_corrected, _, _ = multipletests(p_values, method='fdr_bh')

            # Assign the corrected p-values to the DataFrame
            correlation_results_df.loc[combined_mask, 'Pval_corrected'] = p_values_corrected
# Define the output file path
output_file_path = os.path.join(parent_dir, 'output', "Correlation_with_behavior", 'correlation_dti_with_behavior.csv')

# Save the correlation results DataFrame to a CSV file
correlation_results_df.to_csv(output_file_path, index=False)

print("Correlation results with corrected p-values saved successfully to 'correlation_results.csv' in the output folder.")