0.2_getData_rutherford.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "b7bdfc96",
   "metadata": {},
   "source": [
    "# Explaining cortical thickness development: Modelled cortical thickness data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b1849f76-b17d-4286-bf57-50ff56e81bf8",
   "metadata": {
    "id": "b1849f76-b17d-4286-bf57-50ff56e81bf8"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING (pytensor.tensor.blas): Using NumPy C-API based implementation for BLAS functions.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Working directory: /Users/llotter/projects/CTdev\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from glob import glob\n",
    "\n",
    "from os.path import join\n",
    "from matplotlib import pyplot as plt\n",
    "from matplotlib.cm import get_cmap\n",
    "from IPython.display import display\n",
    "import imageio.v2 as iio\n",
    "\n",
    "# custom functions \n",
    "from scripts.braincharts import get_centile_predictions # wrapper around PCNtoolkit/braincharts\n",
    "from scripts.templates import get_destrieux\n",
    "from scripts.plots_brain import get_parc_gifti_region, plot_surf_ax\n",
    "\n",
    "# working path\n",
    "wd = os.path.dirname(os.path.realpath(\"__file__\"))\n",
    "print(\"Working directory:\", wd)\n",
    "\n",
    "# cortical thickness model path\n",
    "ct_model_dir = join(wd, \"model_rutherford\", \"lifespan_57K_82sites\")\n",
    "plot_dir = join(wd, \"plots\", \"braincharts\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get parcellation (Destrieux, fsaverage)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "9fc1a27a",
   "metadata": {},
   "outputs": [],
   "source": [
    "parc_destrieux, destrieux_idps = get_destrieux()\n",
    "\n",
    "# save one time for reference\n",
    "np.savetxt(join(wd, \"model_rutherford\", \"idps_destrieux.txt\"), np.array(destrieux_idps), \"%s\")\n",
    "for parc, hemi in zip(parc_destrieux, [\"L\", \"R\"]):\n",
    "    parc.to_filename(join(wd, \"templates\", f\"destrieux.fsa.10k.{hemi}.label.gii\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1f2c2582",
   "metadata": {},
   "source": [
    "## Get main cortical thickness data (Rutherford)\n",
    "\n",
    "Adapted from Rutherford et al, 2022:  \n",
    "Tutorial notebook: [https://pcntoolkit.readthedocs.io/en/latest/pages/apply_normative_models.html](https://pcntoolkit.readthedocs.io/en/latest/pages/apply_normative_models.html)  \n",
    "Dataset source: [https://github.com/predictive-clinical-neuroscience/braincharts](https://github.com/predictive-clinical-neuroscience/braincharts)  \n",
    "Associated Paper: [https://elifesciences.org/articles/72904](https://elifesciences.org/articles/72904)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75210821-ccb8-4bd2-82f3-641708811b21",
   "metadata": {
    "id": "75210821-ccb8-4bd2-82f3-641708811b21"
   },
   "source": [
    "### Get \"ideal\" predicted CT data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "2d0743d8-28ca-4a14-8ef0-99bf40434b5b",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2d0743d8-28ca-4a14-8ef0-99bf40434b5b",
    "outputId": "00343be9-866d-4c9d-f243-351cbe9e4231"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6fa1078458c14b6896dd30f93b6db3b5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Iterating IDPs:   0%|          | 0/148 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "IDP 0: lh_G&S_frontomargin_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 1: lh_G&S_occipital_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 2: lh_G&S_paracentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 3: lh_G&S_subcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 4: lh_G&S_transv_frontopol_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 5: lh_G&S_cingul-Ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 6: lh_G&S_cingul-Mid-Ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 7: lh_G&S_cingul-Mid-Post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 8: lh_G_cingul-Post-dorsal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 9: lh_G_cingul-Post-ventral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 10: lh_G_cuneus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 11: lh_G_front_inf-Opercular_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 12: lh_G_front_inf-Orbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 13: lh_G_front_inf-Triangul_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 14: lh_G_front_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 15: lh_G_front_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 16: lh_G_Ins_lg&S_cent_ins_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 17: lh_G_insular_short_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 18: lh_G_occipital_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 19: lh_G_occipital_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 20: lh_G_oc-temp_lat-fusifor_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 21: lh_G_oc-temp_med-Lingual_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 22: lh_G_oc-temp_med-Parahip_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 23: lh_G_orbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 24: lh_G_pariet_inf-Angular_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 25: lh_G_pariet_inf-Supramar_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 26: lh_G_parietal_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 27: lh_G_postcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 28: lh_G_precentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 29: lh_G_precuneus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 30: lh_G_rectus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 31: lh_G_subcallosal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 32: lh_G_temp_sup-G_T_transv_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 33: lh_G_temp_sup-Lateral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 34: lh_G_temp_sup-Plan_polar_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 35: lh_G_temp_sup-Plan_tempo_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 36: lh_G_temporal_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 37: lh_G_temporal_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 38: lh_Lat_Fis-ant-Horizont_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 39: lh_Lat_Fis-ant-Vertical_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 40: lh_Lat_Fis-post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 41: lh_Pole_occipital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 42: lh_Pole_temporal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 43: lh_S_calcarine_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 44: lh_S_central_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 45: lh_S_cingul-Marginalis_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 46: lh_S_circular_insula_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 47: lh_S_circular_insula_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 48: lh_S_circular_insula_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 49: lh_S_collat_transv_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 50: lh_S_collat_transv_post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 51: lh_S_front_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 52: lh_S_front_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 53: lh_S_front_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 54: lh_S_interm_prim-Jensen_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 55: lh_S_intrapariet&P_trans_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 56: lh_S_oc_middle&Lunatus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 57: lh_S_oc_sup&transversal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 58: lh_S_occipital_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 59: lh_S_oc-temp_lat_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 60: lh_S_oc-temp_med&Lingual_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 61: lh_S_orbital_lateral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 62: lh_S_orbital_med-olfact_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 63: lh_S_orbital-H_Shaped_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 64: lh_S_parieto_occipital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 65: lh_S_pericallosal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 66: lh_S_postcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 67: lh_S_precentral-inf-part_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 68: lh_S_precentral-sup-part_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 69: lh_S_suborbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 70: lh_S_subparietal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 71: lh_S_temporal_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 72: lh_S_temporal_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 73: lh_S_temporal_transverse_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 74: rh_G&S_frontomargin_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 75: rh_G&S_occipital_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 76: rh_G&S_paracentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 77: rh_G&S_subcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 78: rh_G&S_transv_frontopol_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 79: rh_G&S_cingul-Ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 80: rh_G&S_cingul-Mid-Ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 81: rh_G&S_cingul-Mid-Post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 82: rh_G_cingul-Post-dorsal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 83: rh_G_cingul-Post-ventral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 84: rh_G_cuneus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 85: rh_G_front_inf-Opercular_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 86: rh_G_front_inf-Orbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 87: rh_G_front_inf-Triangul_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 88: rh_G_front_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 89: rh_G_front_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 90: rh_G_Ins_lg&S_cent_ins_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 91: rh_G_insular_short_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 92: rh_G_occipital_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 93: rh_G_occipital_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 94: rh_G_oc-temp_lat-fusifor_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 95: rh_G_oc-temp_med-Lingual_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 96: rh_G_oc-temp_med-Parahip_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 97: rh_G_orbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 98: rh_G_pariet_inf-Angular_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 99: rh_G_pariet_inf-Supramar_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 100: rh_G_parietal_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 101: rh_G_postcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 102: rh_G_precentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 103: rh_G_precuneus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 104: rh_G_rectus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 105: rh_G_subcallosal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 106: rh_G_temp_sup-G_T_transv_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 107: rh_G_temp_sup-Lateral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 108: rh_G_temp_sup-Plan_polar_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 109: rh_G_temp_sup-Plan_tempo_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 110: rh_G_temporal_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 111: rh_G_temporal_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 112: rh_Lat_Fis-ant-Horizont_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 113: rh_Lat_Fis-ant-Vertical_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 114: rh_Lat_Fis-post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 115: rh_Pole_occipital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 116: rh_Pole_temporal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 117: rh_S_calcarine_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 118: rh_S_central_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 119: rh_S_cingul-Marginalis_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 120: rh_S_circular_insula_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 121: rh_S_circular_insula_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 122: rh_S_circular_insula_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 123: rh_S_collat_transv_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 124: rh_S_collat_transv_post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 125: rh_S_front_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 126: rh_S_front_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 127: rh_S_front_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 128: rh_S_interm_prim-Jensen_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 129: rh_S_intrapariet&P_trans_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 130: rh_S_oc_middle&Lunatus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 131: rh_S_oc_sup&transversal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 132: rh_S_occipital_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 133: rh_S_oc-temp_lat_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 134: rh_S_oc-temp_med&Lingual_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 135: rh_S_orbital_lateral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 136: rh_S_orbital_med-olfact_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 137: rh_S_orbital-H_Shaped_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 138: rh_S_parieto_occipital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 139: rh_S_pericallosal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 140: rh_S_postcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 141: rh_S_precentral-inf-part_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 142: rh_S_precentral-sup-part_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 143: rh_S_suborbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 144: rh_S_subparietal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 145: rh_S_temporal_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 146: rh_S_temporal_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 147: rh_S_temporal_transverse_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5071c9c62a634322bdd4bd2df935f105",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Iterating IDPs:   0%|          | 0/148 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "IDP 0: lh_G&S_frontomargin_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 1: lh_G&S_occipital_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 2: lh_G&S_paracentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 3: lh_G&S_subcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 4: lh_G&S_transv_frontopol_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 5: lh_G&S_cingul-Ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 6: lh_G&S_cingul-Mid-Ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 7: lh_G&S_cingul-Mid-Post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 8: lh_G_cingul-Post-dorsal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 9: lh_G_cingul-Post-ventral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 10: lh_G_cuneus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 11: lh_G_front_inf-Opercular_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 12: lh_G_front_inf-Orbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 13: lh_G_front_inf-Triangul_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 14: lh_G_front_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 15: lh_G_front_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 16: lh_G_Ins_lg&S_cent_ins_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 17: lh_G_insular_short_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 18: lh_G_occipital_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 19: lh_G_occipital_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 20: lh_G_oc-temp_lat-fusifor_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 21: lh_G_oc-temp_med-Lingual_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 22: lh_G_oc-temp_med-Parahip_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 23: lh_G_orbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 24: lh_G_pariet_inf-Angular_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 25: lh_G_pariet_inf-Supramar_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 26: lh_G_parietal_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 27: lh_G_postcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 28: lh_G_precentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 29: lh_G_precuneus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 30: lh_G_rectus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 31: lh_G_subcallosal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 32: lh_G_temp_sup-G_T_transv_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 33: lh_G_temp_sup-Lateral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 34: lh_G_temp_sup-Plan_polar_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 35: lh_G_temp_sup-Plan_tempo_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 36: lh_G_temporal_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 37: lh_G_temporal_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 38: lh_Lat_Fis-ant-Horizont_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 39: lh_Lat_Fis-ant-Vertical_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 40: lh_Lat_Fis-post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 41: lh_Pole_occipital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 42: lh_Pole_temporal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 43: lh_S_calcarine_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 44: lh_S_central_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 45: lh_S_cingul-Marginalis_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 46: lh_S_circular_insula_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 47: lh_S_circular_insula_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 48: lh_S_circular_insula_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 49: lh_S_collat_transv_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 50: lh_S_collat_transv_post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 51: lh_S_front_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 52: lh_S_front_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 53: lh_S_front_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 54: lh_S_interm_prim-Jensen_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 55: lh_S_intrapariet&P_trans_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 56: lh_S_oc_middle&Lunatus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 57: lh_S_oc_sup&transversal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 58: lh_S_occipital_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 59: lh_S_oc-temp_lat_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 60: lh_S_oc-temp_med&Lingual_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 61: lh_S_orbital_lateral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 62: lh_S_orbital_med-olfact_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 63: lh_S_orbital-H_Shaped_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 64: lh_S_parieto_occipital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 65: lh_S_pericallosal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 66: lh_S_postcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 67: lh_S_precentral-inf-part_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 68: lh_S_precentral-sup-part_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 69: lh_S_suborbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 70: lh_S_subparietal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 71: lh_S_temporal_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 72: lh_S_temporal_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 73: lh_S_temporal_transverse_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 74: rh_G&S_frontomargin_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 75: rh_G&S_occipital_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 76: rh_G&S_paracentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 77: rh_G&S_subcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 78: rh_G&S_transv_frontopol_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 79: rh_G&S_cingul-Ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 80: rh_G&S_cingul-Mid-Ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 81: rh_G&S_cingul-Mid-Post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 82: rh_G_cingul-Post-dorsal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 83: rh_G_cingul-Post-ventral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 84: rh_G_cuneus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 85: rh_G_front_inf-Opercular_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 86: rh_G_front_inf-Orbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 87: rh_G_front_inf-Triangul_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 88: rh_G_front_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 89: rh_G_front_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 90: rh_G_Ins_lg&S_cent_ins_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 91: rh_G_insular_short_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 92: rh_G_occipital_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 93: rh_G_occipital_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 94: rh_G_oc-temp_lat-fusifor_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 95: rh_G_oc-temp_med-Lingual_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 96: rh_G_oc-temp_med-Parahip_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 97: rh_G_orbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 98: rh_G_pariet_inf-Angular_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 99: rh_G_pariet_inf-Supramar_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 100: rh_G_parietal_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 101: rh_G_postcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 102: rh_G_precentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 103: rh_G_precuneus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 104: rh_G_rectus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 105: rh_G_subcallosal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 106: rh_G_temp_sup-G_T_transv_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 107: rh_G_temp_sup-Lateral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 108: rh_G_temp_sup-Plan_polar_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 109: rh_G_temp_sup-Plan_tempo_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 110: rh_G_temporal_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 111: rh_G_temporal_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 112: rh_Lat_Fis-ant-Horizont_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 113: rh_Lat_Fis-ant-Vertical_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 114: rh_Lat_Fis-post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 115: rh_Pole_occipital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 116: rh_Pole_temporal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 117: rh_S_calcarine_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 118: rh_S_central_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 119: rh_S_cingul-Marginalis_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 120: rh_S_circular_insula_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 121: rh_S_circular_insula_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 122: rh_S_circular_insula_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 123: rh_S_collat_transv_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 124: rh_S_collat_transv_post_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 125: rh_S_front_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 126: rh_S_front_middle_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 127: rh_S_front_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 128: rh_S_interm_prim-Jensen_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 129: rh_S_intrapariet&P_trans_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 130: rh_S_oc_middle&Lunatus_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 131: rh_S_oc_sup&transversal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 132: rh_S_occipital_ant_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 133: rh_S_oc-temp_lat_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 134: rh_S_oc-temp_med&Lingual_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 135: rh_S_orbital_lateral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 136: rh_S_orbital_med-olfact_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 137: rh_S_orbital-H_Shaped_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 138: rh_S_parieto_occipital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 139: rh_S_pericallosal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 140: rh_S_postcentral_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 141: rh_S_precentral-inf-part_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 142: rh_S_precentral-sup-part_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 143: rh_S_suborbital_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 144: rh_S_subparietal_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 145: rh_S_temporal_inf_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 146: rh_S_temporal_sup_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n",
      "IDP 147: rh_S_temporal_transverse_thickness\n",
      "Loading data ...\n",
      "Prediction by model  1 of 1\n",
      "Writing outputs ...\n"
     ]
    }
   ],
   "source": [
    "# centiles to extract (as used by rutherford et al.)\n",
    "centiles = [0.01, 0.05, 0.25, 0.5, 0.75, 0.95, 0.99]\n",
    "step = 2\n",
    "\n",
    "# iterate sexes\n",
    "temp = dict()\n",
    "for sex in [\"f\", \"m\"]:\n",
    "    \n",
    "    # ideal predictions by sex\n",
    "    # covariate lists 'age' & 'sex' will be overwritten in iteration but they are the same anyways\n",
    "    temp[sex], age, sex = get_centile_predictions(\n",
    "        model_dir=ct_model_dir,\n",
    "        site_ids_tr=list(np.loadtxt(join(wd, \"model_rutherford\", \"site_ids_82sites.txt\"), dtype=str)),\n",
    "        idps=destrieux_idps,\n",
    "        sex=sex, \n",
    "        age_min=-5, \n",
    "        age_max=110, \n",
    "        age_step=step, \n",
    "        centiles=centiles,\n",
    "    )\n",
    "    \n",
    "# store in independent arrays for ease of use  \n",
    "predicted_ct_f, predicted_ct_m = temp[\"f\"], temp[\"m\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1a457bd",
   "metadata": {},
   "source": [
    "### Check and save data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9c3d3265",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape ct_predicted_f: parcels x tps x centiles (148, 230, 7)\n",
      "Shape ct_predicted_m: parcels x tps x centiles (148, 230, 7)\n"
     ]
    }
   ],
   "source": [
    "# save 3d data array and index vectors\n",
    "np.save(join(wd, \"data_rutherford\", \"ct_predicted_f.npy\"), predicted_ct_f)\n",
    "np.save(join(wd, \"data_rutherford\", \"ct_predicted_m.npy\"), predicted_ct_m)\n",
    "np.savetxt(join(wd, \"data_rutherford\", \"ct_predicted_parcels.txt\"), np.array(destrieux_idps), '%s')\n",
    "np.savetxt(join(wd, \"data_rutherford\", \"ct_predicted_centiles.txt\"), np.array(centiles), '%.3f')\n",
    "np.savetxt(join(wd, \"data_rutherford\", \"ct_predicted_tps.txt\"), np.array(age), '%.2f')\n",
    "\n",
    "print(f\"Shape ct_predicted_f: parcels x tps x centiles {predicted_ct_f.shape}\")\n",
    "print(f\"Shape ct_predicted_m: parcels x tps x centiles {predicted_ct_m.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b21a0ce",
   "metadata": {},
   "source": [
    "## Plot cortex development"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot percentiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# If this is rerun and single-subject data is not available (i.e., if you downloaded this file from \n",
    "# GitHub), you will have to set 'plot_scatter' to False.\n",
    "plot_lines = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0458cb5bcba2456882dbf8298515677e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/74 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Context leak detected, msgtracer returned -1\n",
      "Context leak detected, msgtracer returned -1\n",
      "Context leak detected, msgtracer returned -1\n"
     ]
    }
   ],
   "source": [
    "from matplotlib.collections import LineCollection\n",
    "from scripts.utils import na\n",
    "from tqdm.auto import tqdm\n",
    "\n",
    "if plot_lines:\n",
    "    # single subject data\n",
    "    data_cross = pd.read_feather(join(wd, \"data_ABCD-IMGN\", \"data_cross.ftr\")) \\\n",
    "        .set_index([\"ds\", \"id\", \"tp\"]).sort_index()\n",
    "    imgn = pd.read_feather(join(wd, \"data_ABCD-IMGN\", \"imgn.ftr\")) \\\n",
    "        .set_index([\"id\", \"tp\"], drop=False).sort_index()\n",
    "    abcd = pd.read_feather(join(wd, \"data_ABCD-IMGN\", \"abcd.ftr\")) \\\n",
    "        .set_index([\"id\", \"tp\"], drop=False).sort_index()\n",
    "\n",
    "# x axis (years)\n",
    "xx_plot = np.arange(0, 90.1, 1/step)\n",
    "# percentiles\n",
    "percentiles_plot = [0.01, 0.05, 0.25, 0.5, 0.75, 0.95, 0.99]\n",
    "# position of median\n",
    "med_idx = int(len(percentiles_plot) / 2)\n",
    "# IDPs\n",
    "destrieux_idps_bilat = [l.replace(\"lh_\",\"\") for l in destrieux_idps[:int(len(destrieux_idps)/2)]]\n",
    "# ylims\n",
    "ylims = list()\n",
    "for a in [predicted_ct_f, predicted_ct_m]:\n",
    "    ylims.append(a[:,[age.index(i) for i in xx_plot],:].min())\n",
    "    ylims.append(a[:,[age.index(i) for i in xx_plot],:].max())\n",
    "ylims = (min(ylims), max(ylims))\n",
    "\n",
    "for idp in tqdm(destrieux_idps_bilat):\n",
    "    # get idps\n",
    "    idp_destrieux_lh = \"lh_\" + idp\n",
    "    idp_destrieux_rh = \"rh_\" + idp\n",
    "    \n",
    "    # this idp data\n",
    "    idp_data = dict(\n",
    "        destrieux_lh_f = pd.DataFrame(predicted_ct_f[destrieux_idps.index(idp_destrieux_lh),:,:], \n",
    "                                      index=age, columns=centiles).loc[xx_plot, percentiles_plot],\n",
    "        destrieux_lh_m = pd.DataFrame(predicted_ct_m[destrieux_idps.index(idp_destrieux_lh),:,:], \n",
    "                                      index=age, columns=centiles).loc[xx_plot, percentiles_plot],\n",
    "        destrieux_rh_f = pd.DataFrame(predicted_ct_f[destrieux_idps.index(idp_destrieux_rh),:,:], \n",
    "                                      index=age, columns=centiles).loc[xx_plot, percentiles_plot],\n",
    "        destrieux_rh_m = pd.DataFrame(predicted_ct_m[destrieux_idps.index(idp_destrieux_rh),:,:], \n",
    "                                      index=age, columns=centiles).loc[xx_plot, percentiles_plot],\n",
    "    )\n",
    "\n",
    "    # figure\n",
    "    fig = plt.figure(figsize=(10.5,11))\n",
    "    gs = fig.add_gridspec(\n",
    "        3,2, \n",
    "        height_ratios=(0.55,1,1), \n",
    "        wspace=0.1,\n",
    "        hspace=0.15)\n",
    "    \n",
    "    # plot brain\n",
    "    ax_brain = fig.add_subplot(gs[0,:])\n",
    "    lh, rh = get_parc_gifti_region(\n",
    "        region_lh=destrieux_idps.index(idp_destrieux_lh)+1, \n",
    "        region_rh=destrieux_idps.index(idp_destrieux_rh)+1,\n",
    "        parc_gifti=parc_destrieux\n",
    "    )\n",
    "    plot_surf_ax(ax=ax_brain, fig=fig, lh=lh, rh=rh, plot=\"region\", c_fill=\"Reds_r\", c_outline=\"Greys\")\n",
    "    ax_brain.set_title(idp, weight=\"semibold\", size=13)\n",
    "    \n",
    "    # plot trajectories\n",
    "    for i_hemi, hemi in zip([0, 1],\n",
    "                            [\"lh\", \"rh\"]):\n",
    "        for i_sex, sex, color in zip([1, 2],\n",
    "                                     [\"f\", \"m\"], \n",
    "                                     [\"tab:blue\", \"tab:green\"]):\n",
    "            # get data\n",
    "            data = idp_data[f\"destrieux_{hemi}_{sex}\"]\n",
    "            # get ylims\n",
    "            if hemi==\"rh\":\n",
    "                ylims = ax.get_ylim()\n",
    "            # subplot\n",
    "            ax = fig.add_subplot(gs[i_sex, i_hemi])\n",
    "            # fill space\n",
    "            for i in range(med_idx):\n",
    "                ax.fill_between(xx_plot, data.iloc[:, med_idx-i-1], data.iloc[:, med_idx+i+1], \n",
    "                                alpha=0.1, color=color)   \n",
    "            # plot actual centile lines\n",
    "            for p in range(len(percentiles_plot)):\n",
    "                ax.plot(xx_plot, data.iloc[:,p], color=color, alpha=0.7, linewidth=1)\n",
    "            # highlight median\n",
    "            ax.plot(xx_plot, data.iloc[:,med_idx], color)\n",
    "            \n",
    "            # plot single subject data\n",
    "            if plot_lines:\n",
    "                \n",
    "                for df_single, ds, c in zip(\n",
    "                    [abcd, imgn],\n",
    "                    [\"abcd\", \"imgn\"],\n",
    "                    [\"tab:purple\", \"tab:orange\"]):\n",
    "                    \n",
    "                    # Sex-specific single subject data\n",
    "                    subs = data_cross.loc[(ds, na(), na()), \"EstimatedTotalIntraCranialVol-rescaled\"].dropna().index.get_level_values(\"id\").unique().tolist()\n",
    "                    subs = [sub for sub in subs if sub in df_single.loc[df_single.sex_str==sex.upper(), \"id\"]]\n",
    "\n",
    "                    # data\n",
    "                    temp = data_cross.loc[(ds, subs, na()), [\"age\", f\"{hemi}_{idp}-rescaled\"]].droplevel(\"ds\")\n",
    "                   \n",
    "                    # plot\n",
    "                    ax.add_collection(\n",
    "                        LineCollection(\n",
    "                            [temp.loc[sub] for sub in temp.index.get_level_values(\"id\")],\n",
    "                            alpha=0.05 if ds==\"imgn\" else 0.03,\n",
    "                            color=c,\n",
    "                            lw=0.1,\n",
    "                            zorder=100,\n",
    "                            label=f\"{'ABCD' if ds=='abcd' else 'IMAGEN'}\\n$(n={len(subs)})$\"\n",
    "                        )\n",
    "                    )\n",
    "                    \n",
    "            # # details\n",
    "            ax.set_title(f\"{'Left' if hemi=='lh' else 'Right'} hemisphere: {'female' if sex=='f' else 'male'}\")   \n",
    "            ax.set_xlim((0,90))\n",
    "            ax.set_ylim(ylims)\n",
    "            if sex==\"m\":\n",
    "                ax.set_xlabel(\"Age\")\n",
    "            else:\n",
    "                 ax.set_xticklabels([])\n",
    "            ax.grid()\n",
    "            if hemi==\"rh\":\n",
    "                ax.set_yticklabels([])\n",
    "                if plot_lines:\n",
    "                    leg = ax.legend(loc=\"center left\", bbox_to_anchor=(1,0.5), handletextpad=0.2)\n",
    "                    for handle in leg.legend_handles:\n",
    "                        handle._linewidth = 2\n",
    "                        handle._alpha = 0.5\n",
    "            else:\n",
    "                ax.set_ylabel(\"Cortical thickness\")\n",
    "            \n",
    "    plt.savefig(\n",
    "        join(plot_dir, f\"braincharts_{destrieux_idps_bilat.index(idp):02d}_{idp}.png\"),  \n",
    "        bbox_inches=\"tight\", dpi=150\n",
    "    )\n",
    "    #plt.show()\n",
    "    plt.close()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create gif"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "imgs = glob(join(plot_dir, \"braincharts_*.png\"))\n",
    "imgs.sort()\n",
    "frames = []\n",
    "[frames.append(iio.imread(i)) for i in imgs]\n",
    "iio.mimwrite(join(plot_dir, \"braincharts_animation.gif\"), frames, fps=1.5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Age distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>n_sites</th>\n",
       "      <th>n_subs</th>\n",
       "      <th>n_female</th>\n",
       "      <th>n_male</th>\n",
       "      <th>age_mean</th>\n",
       "      <th>age_sd</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>82</td>\n",
       "      <td>58836</td>\n",
       "      <td>30006</td>\n",
       "      <td>28830</td>\n",
       "      <td>46.9</td>\n",
       "      <td>24.4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   n_sites  n_subs  n_female  n_male  age_mean  age_sd\n",
       "0       82   58836     30006   28830      46.9    24.4"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# data from publication (Rutherford et al., 2022, eLife)\n",
    "\n",
    "n_subs = 58836\n",
    "age_pub = pd.DataFrame([dict(\n",
    "        n_sites=82,\n",
    "        n_subs=n_subs,\n",
    "        n_female=round(n_subs * ((0.511+0.509)/2)),\n",
    "        n_male=round(n_subs * ((0.489+0.491)/2)),\n",
    "        age_mean=46.9,\n",
    "        age_sd=24.4\n",
    "    )])\n",
    "age_pub.to_csv(join(wd, \"data_rutherford\", \"age_pub.csv\"), index=None)\n",
    "display(age_pub)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "All subjects (test+train)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>58834.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>46.967462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>24.393133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2.091667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>19.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>57.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>67.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>100.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                age\n",
       "count  58834.000000\n",
       "mean      46.967462\n",
       "std       24.393133\n",
       "min        2.091667\n",
       "25%       19.000000\n",
       "50%       57.000000\n",
       "75%       67.000000\n",
       "max      100.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "By group\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"8\" halign=\"left\">age</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_label</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>controls test</th>\n",
       "      <td>29416.0</td>\n",
       "      <td>46.967618</td>\n",
       "      <td>24.390227</td>\n",
       "      <td>2.234167</td>\n",
       "      <td>19.0</td>\n",
       "      <td>57.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>controls train</th>\n",
       "      <td>29418.0</td>\n",
       "      <td>46.967305</td>\n",
       "      <td>24.396454</td>\n",
       "      <td>2.091667</td>\n",
       "      <td>19.0</td>\n",
       "      <td>57.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    age                                                    \\\n",
       "                  count       mean        std       min   25%   50%   75%   \n",
       "test_label                                                                  \n",
       "controls test   29416.0  46.967618  24.390227  2.234167  19.0  57.0  67.0   \n",
       "controls train  29418.0  46.967305  24.396454  2.091667  19.0  57.0  67.0   \n",
       "\n",
       "                       \n",
       "                  max  \n",
       "test_label             \n",
       "controls test   100.0  \n",
       "controls train  100.0  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# data from repo\n",
    "\n",
    "age_rutherford = pd.read_csv(join(wd, \"model_rutherford\", \"all_age_eLife.csv\"))\n",
    "age_main = age_rutherford.query(\"test_label in ['controls test', 'controls train']\")\n",
    "\n",
    "# summary stats\n",
    "print(\"All subjects (test+train)\")\n",
    "display(age_main.describe())\n",
    "print(\"By group\")\n",
    "display(age_main.groupby(\"test_label\").describe())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Function to add developmental stages (aligned to CT change period mean, e.g. delta(5,10) = 7.5)\n",
    "def add_stages(\n",
    "    ax, height=0.09,\n",
    "    alignment=-2.5,\n",
    "    stages=[(6,12), (12,20), (20,40), (40,60), (60,100)],\n",
    "    stage_names=[\"Late childhood\", \"Adolescence\", \"Young adulthood\", \"Middle adulthood\", \"Late adulthood\"]\n",
    "    ):\n",
    "    \n",
    "    ax_stages = ax.inset_axes((0, 1, 1, height))\n",
    "    ax_stages.set_xlim(ax.get_xlim())\n",
    "    ax_stages.set_ylim(0,1)\n",
    "    ax_stages.set_axis_off()\n",
    "    \n",
    "    for i, x_range in enumerate(stages):\n",
    "        x0 = x_range[0] - 5 + alignment\n",
    "        x1 = x_range[1] - 5 + alignment\n",
    "        if x_range[0] > 10:\n",
    "            ax_stages.axvline(\n",
    "                x0, \n",
    "                c=\"0.3\",\n",
    "                lw=1,\n",
    "                zorder=0\n",
    "            )\n",
    "        if x0 < 2:\n",
    "            x0 = -5\n",
    "        if x1 > 90:\n",
    "            x1 = 80\n",
    "        ax_stages.text(\n",
    "            x=(x0 + x1) / 2,\n",
    "            y=0.5,\n",
    "            s=f\"{stage_names[i]}\\n${x_range[0]} - {x_range[1]}\\ y.$\",\n",
    "            ha=\"center\",\n",
    "            va=\"center\"\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0–1        0\n",
       "1–2        0\n",
       "2–3       58\n",
       "3–4       74\n",
       "4–5       15\n",
       "5–6       54\n",
       "6–7       64\n",
       "7–8      108\n",
       "8–9      315\n",
       "9–10    5683\n",
       "dtype: int64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# distribution\n",
    "hist_train = np.histogram(age_main[age_main.test_label==\"controls train\"].age, range=(0,110), bins=110)\n",
    "hist_test = np.histogram(age_main[age_main.test_label==\"controls test\"].age, range=(0,110), bins=110)\n",
    "age_stats = age_main.describe().drop([\"std\", \"count\"])\n",
    "display(pd.Series(hist_train[0] + hist_test[0], index=[f\"{i}–{i+1}\" for i in np.arange(0,110,1)]).head(10))\n",
    "\n",
    "# plot\n",
    "fig, axes = plt.subplots(2,1, figsize=(10,8))\n",
    "\n",
    "# iterate all and zoomed-in axes\n",
    "for ax in axes:\n",
    "    \n",
    "    # hist\n",
    "    for hist, lab in zip([hist_train, hist_test], [\"Training\", \"Testing\"]):\n",
    "        ax.bar(\n",
    "            range(0,110),\n",
    "            height=hist[0],\n",
    "            bottom=hist_train[0] if lab==\"Testing\" else None,\n",
    "            label=lab\n",
    "        )\n",
    "        \n",
    "    # stats\n",
    "    for stat in age_stats.index:\n",
    "        ax.axvline(\n",
    "            age_stats.loc[stat, \"age\"], \n",
    "            c=\"k\", \n",
    "            lw=2\n",
    "        )\n",
    "        \n",
    "    # design\n",
    "    ax.set_ylabel(\"Number of subjects\")\n",
    "    ax.set_xticks(np.arange(0,101,10))\n",
    "    ax.set_xticklabels([f\"{i}–{i+1}\" for i in np.arange(0,101,10)])\n",
    "    ax.set_xticks(np.arange(0,101,1), minor=True)\n",
    "    ax.set_xlim(-1,101)\n",
    "    ax.legend()\n",
    "    \n",
    "    \n",
    "# stats annotation\n",
    "for stat in age_stats.index:\n",
    "    axes[0].annotate(\n",
    "        f\"{stat}\\n({age_stats.loc[stat,'age']:.02f})\", \n",
    "        xy=(age_stats.loc[stat,\"age\"]/102+0.008, 1.03), \n",
    "        xycoords=\"axes fraction\", \n",
    "        ha=\"center\"\n",
    "    )\n",
    "    \n",
    "# design\n",
    "axes[1].set_ylim(0,500)\n",
    "\n",
    "# developmental stages\n",
    "stages=[(6,12), (12,20), (20,40), (40,60), (60,100)]\n",
    "stage_names=[\"Late childhood\", \"Adolescence\", \"Young adulthood\", \"Middle adulthood\", \"Late adulthood\"]\n",
    "    \n",
    "ax_stages = axes[1].inset_axes((0, -0.25, 1, 0.125))\n",
    "ax_stages.set_xlim(ax.get_xlim())\n",
    "ax_stages.set_ylim(0,1)\n",
    "ax_stages.set_axis_off()\n",
    "\n",
    "for i, x_range in enumerate(stages):\n",
    "    # full lines\n",
    "    [ax_stages.axvline(x - 0.5, c=\"0.3\", lw=1) for x in [1, 12, 20, 100]]\n",
    "    # half lines\n",
    "    [ax_stages.axvline(x - 0.5, c=\"0.3\", lw=1, ymin=0.6) for x in [6, 40, 60]]\n",
    "    # headings\n",
    "    for s, x in zip([\"Childhood\", \"Adulthood\"], [np.mean([1,12]), np.mean([20,100])]):\n",
    "        ax_stages.text(x=x - 0.5, y=0.25, s=s, ha=\"center\", va=\"center\")\n",
    "    # subheadings\n",
    "    for s, x in zip([\"Early\", \"Late\", \"Adoles.\", \"Young\", \"Middle\", \"Late\"], \n",
    "                    [np.mean([1,6]), np.mean([6,12]), np.mean([12,20]), np.mean([20,40]), np.mean([40,60]), np.mean([60,100])]):\n",
    "        ax_stages.text(x=x - 0.5, y=0.5 if \"Adol\" in s else 0.75, s=s, ha=\"center\", va=\"center\")\n",
    "    # label\n",
    "    ax_stages.text(\n",
    "        s=\"Age bins [years]\",\n",
    "        x=50,\n",
    "        y=-0.55,\n",
    "        size=12,\n",
    "        ha=\"center\",\n",
    "        va=\"center\"\n",
    "    )\n",
    "\n",
    "# finalize\n",
    "plt.tight_layout()\n",
    "fig.savefig(join(plot_dir, \"braincharts_age.pdf\"), bbox_inches=\"tight\")\n",
    "\n",
    "# source\n",
    "pd.DataFrame(\n",
    "    {\n",
    "        \"age_bins\": [f\"{i:.0f}-{i+1:.0f}\" for i in hist_train[1][:-10]],\n",
    "        \"hist_train\": hist_train[0][:-9],\n",
    "        \"hist_test\": hist_test[0][:-9]\n",
    "    }\n",
    ") \\\n",
    "    .set_index(\"age_bins\") \\\n",
    "    .to_csv(join(plot_dir, \"braincharts_age_source.csv\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "55a5a13bab2c2636ad1654b177bb15c7b38b3aeaddbba167703e2619b356d1fa"
  },
  "kernelspec": {
   "display_name": "Python 3.9.11 ('ntdev')",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.11"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}