cortical_thickness.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "626de92c-031e-482e-8833-f38687ecc92d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from tqdm import tqdm\n",
    "from sktree import ObliqueRandomForestClassifier, PatchObliqueRandomForestClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import train_test_split\n",
    "import nibabel as nb\n",
    "from scipy import ndimage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "205e977c-230c-49f4-8de2-094770135da4",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_excel('Human.parcellated_thickness.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "72ff36da-63c2-4ec8-9110-e2793d511a69",
   "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>Unnamed: 0</th>\n",
       "      <th>sid</th>\n",
       "      <th>Markov.1</th>\n",
       "      <th>Markov.2</th>\n",
       "      <th>Markov.3</th>\n",
       "      <th>Markov.4</th>\n",
       "      <th>Markov.5</th>\n",
       "      <th>Markov.6</th>\n",
       "      <th>Markov.7</th>\n",
       "      <th>Markov.8</th>\n",
       "      <th>...</th>\n",
       "      <th>Schaefer217.191</th>\n",
       "      <th>Schaefer217.192</th>\n",
       "      <th>Schaefer217.193</th>\n",
       "      <th>Schaefer217.194</th>\n",
       "      <th>Schaefer217.195</th>\n",
       "      <th>Schaefer217.196</th>\n",
       "      <th>Schaefer217.197</th>\n",
       "      <th>Schaefer217.198</th>\n",
       "      <th>Schaefer217.199</th>\n",
       "      <th>Schaefer217.200</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>sub-OAS30876MRD4592</td>\n",
       "      <td>1.995032</td>\n",
       "      <td>2.203564</td>\n",
       "      <td>1.651978</td>\n",
       "      <td>1.969754</td>\n",
       "      <td>2.603206</td>\n",
       "      <td>2.295727</td>\n",
       "      <td>2.385144</td>\n",
       "      <td>2.719692</td>\n",
       "      <td>...</td>\n",
       "      <td>8.193966</td>\n",
       "      <td>7.736098</td>\n",
       "      <td>7.404804</td>\n",
       "      <td>7.431338</td>\n",
       "      <td>7.541022</td>\n",
       "      <td>7.433447</td>\n",
       "      <td>7.475594</td>\n",
       "      <td>7.460332</td>\n",
       "      <td>7.476401</td>\n",
       "      <td>7.609443</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>sub-HBN_CBIC_NDARXC962XNK</td>\n",
       "      <td>2.557198</td>\n",
       "      <td>2.011555</td>\n",
       "      <td>2.175673</td>\n",
       "      <td>1.863080</td>\n",
       "      <td>2.473705</td>\n",
       "      <td>2.576267</td>\n",
       "      <td>2.392282</td>\n",
       "      <td>2.242582</td>\n",
       "      <td>...</td>\n",
       "      <td>6.927265</td>\n",
       "      <td>7.487809</td>\n",
       "      <td>7.098700</td>\n",
       "      <td>6.753360</td>\n",
       "      <td>6.841211</td>\n",
       "      <td>6.707229</td>\n",
       "      <td>7.189156</td>\n",
       "      <td>6.795299</td>\n",
       "      <td>6.823550</td>\n",
       "      <td>6.533783</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>sub-AOMIC_0770</td>\n",
       "      <td>2.246607</td>\n",
       "      <td>2.295872</td>\n",
       "      <td>1.978412</td>\n",
       "      <td>2.069700</td>\n",
       "      <td>2.213602</td>\n",
       "      <td>2.449572</td>\n",
       "      <td>2.541624</td>\n",
       "      <td>2.777280</td>\n",
       "      <td>...</td>\n",
       "      <td>8.117053</td>\n",
       "      <td>7.775745</td>\n",
       "      <td>7.608771</td>\n",
       "      <td>7.579930</td>\n",
       "      <td>7.573511</td>\n",
       "      <td>7.607256</td>\n",
       "      <td>7.883723</td>\n",
       "      <td>7.630075</td>\n",
       "      <td>7.670835</td>\n",
       "      <td>7.354955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>sub-AOMIC_0344</td>\n",
       "      <td>2.219745</td>\n",
       "      <td>2.366237</td>\n",
       "      <td>2.036068</td>\n",
       "      <td>2.173696</td>\n",
       "      <td>2.508847</td>\n",
       "      <td>2.408997</td>\n",
       "      <td>2.430510</td>\n",
       "      <td>2.882698</td>\n",
       "      <td>...</td>\n",
       "      <td>7.908504</td>\n",
       "      <td>7.856069</td>\n",
       "      <td>7.756918</td>\n",
       "      <td>7.526684</td>\n",
       "      <td>7.410575</td>\n",
       "      <td>7.654072</td>\n",
       "      <td>7.952062</td>\n",
       "      <td>7.682724</td>\n",
       "      <td>7.444550</td>\n",
       "      <td>7.697996</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>sub-Narratives_150</td>\n",
       "      <td>2.131236</td>\n",
       "      <td>2.432549</td>\n",
       "      <td>2.066158</td>\n",
       "      <td>2.352589</td>\n",
       "      <td>2.331565</td>\n",
       "      <td>2.799966</td>\n",
       "      <td>2.490590</td>\n",
       "      <td>2.818574</td>\n",
       "      <td>...</td>\n",
       "      <td>8.229232</td>\n",
       "      <td>7.847060</td>\n",
       "      <td>7.821367</td>\n",
       "      <td>7.739102</td>\n",
       "      <td>8.001608</td>\n",
       "      <td>7.803380</td>\n",
       "      <td>7.750091</td>\n",
       "      <td>7.730715</td>\n",
       "      <td>7.858832</td>\n",
       "      <td>7.892026</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 384 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0                        sid  Markov.1  Markov.2  Markov.3  \\\n",
       "0           0        sub-OAS30876MRD4592  1.995032  2.203564  1.651978   \n",
       "1           1  sub-HBN_CBIC_NDARXC962XNK  2.557198  2.011555  2.175673   \n",
       "2           2             sub-AOMIC_0770  2.246607  2.295872  1.978412   \n",
       "3           3             sub-AOMIC_0344  2.219745  2.366237  2.036068   \n",
       "4           4         sub-Narratives_150  2.131236  2.432549  2.066158   \n",
       "\n",
       "   Markov.4  Markov.5  Markov.6  Markov.7  Markov.8  ...  Schaefer217.191  \\\n",
       "0  1.969754  2.603206  2.295727  2.385144  2.719692  ...         8.193966   \n",
       "1  1.863080  2.473705  2.576267  2.392282  2.242582  ...         6.927265   \n",
       "2  2.069700  2.213602  2.449572  2.541624  2.777280  ...         8.117053   \n",
       "3  2.173696  2.508847  2.408997  2.430510  2.882698  ...         7.908504   \n",
       "4  2.352589  2.331565  2.799966  2.490590  2.818574  ...         8.229232   \n",
       "\n",
       "   Schaefer217.192  Schaefer217.193  Schaefer217.194  Schaefer217.195  \\\n",
       "0         7.736098         7.404804         7.431338         7.541022   \n",
       "1         7.487809         7.098700         6.753360         6.841211   \n",
       "2         7.775745         7.608771         7.579930         7.573511   \n",
       "3         7.856069         7.756918         7.526684         7.410575   \n",
       "4         7.847060         7.821367         7.739102         8.001608   \n",
       "\n",
       "   Schaefer217.196  Schaefer217.197  Schaefer217.198  Schaefer217.199  \\\n",
       "0         7.433447         7.475594         7.460332         7.476401   \n",
       "1         6.707229         7.189156         6.795299         6.823550   \n",
       "2         7.607256         7.883723         7.630075         7.670835   \n",
       "3         7.654072         7.952062         7.682724         7.444550   \n",
       "4         7.803380         7.750091         7.730715         7.858832   \n",
       "\n",
       "   Schaefer217.200  \n",
       "0         7.609443  \n",
       "1         6.533783  \n",
       "2         7.354955  \n",
       "3         7.697996  \n",
       "4         7.892026  \n",
       "\n",
       "[5 rows x 384 columns]"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "79f09473-1d7c-4be1-93fb-87a260116c68",
   "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>ID</th>\n",
       "      <th>Age</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Dataset</th>\n",
       "      <th>Dataset-ID</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>sub-ABIDE1050339</td>\n",
       "      <td>18.0000</td>\n",
       "      <td>MALE</td>\n",
       "      <td>ABIDE</td>\n",
       "      <td>50339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>sub-ABIDE1050701</td>\n",
       "      <td>18.0000</td>\n",
       "      <td>MALE</td>\n",
       "      <td>ABIDE</td>\n",
       "      <td>50701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>sub-ABIDE1050445</td>\n",
       "      <td>18.1383</td>\n",
       "      <td>MALE</td>\n",
       "      <td>ABIDE</td>\n",
       "      <td>50445</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>sub-ABIDE1050459</td>\n",
       "      <td>18.1547</td>\n",
       "      <td>MALE</td>\n",
       "      <td>ABIDE</td>\n",
       "      <td>50459</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>sub-ABIDE1050341</td>\n",
       "      <td>18.2000</td>\n",
       "      <td>FEMALE</td>\n",
       "      <td>ABIDE</td>\n",
       "      <td>50341</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 ID      Age     Sex Dataset Dataset-ID\n",
       "0  sub-ABIDE1050339  18.0000    MALE   ABIDE      50339\n",
       "1  sub-ABIDE1050701  18.0000    MALE   ABIDE      50701\n",
       "2  sub-ABIDE1050445  18.1383    MALE   ABIDE      50445\n",
       "3  sub-ABIDE1050459  18.1547    MALE   ABIDE      50459\n",
       "4  sub-ABIDE1050341  18.2000  FEMALE   ABIDE      50341"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sex = pd.read_excel('~/data_MRI/subjects_age_sex_data_MRI.xlsx')\n",
    "df_sex.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "8a5c96b7-c722-4776-9fd8-746edd3a1e2a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████| 14465/14465 [00:32<00:00, 441.35it/s]\n"
     ]
    }
   ],
   "source": [
    "X1 = []\n",
    "X2 = []\n",
    "y = []\n",
    "IDs = set(df['sid'])\n",
    "ref_IDs = set(df_sex['ID'])\n",
    "\n",
    "for subject in tqdm(IDs):\n",
    "    if subject in ref_IDs:\n",
    "        features = np.array(df[df['sid']==subject]).reshape(-1)[2:]\n",
    "        gender = list(df_sex[df_sex['ID']==subject]['Sex'])\n",
    "        sex = int(gender[0]=='FEMALE')\n",
    "             \n",
    "        X1.append(list(features[:182]))\n",
    "        X2.append(list(features[182:]))\n",
    "        y.append(sex)\n",
    "\n",
    "X1 = np.array(X1)\n",
    "X2 = np.array(X2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "70fb902e-dd1f-469d-aa77-fc529d2dcb17",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10648, 182) (10648, 200)\n"
     ]
    }
   ],
   "source": [
    "print(X1.shape, X2.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "848eb6fd-3718-4b92-98ba-be171ab0e406",
   "metadata": {},
   "source": [
    "### Try random forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "464c2761-da21-441c-a57b-cfde674d1452",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:20<00:00,  4.19s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy is  0.7057276995305164\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "reps = 5\n",
    "accuracy = 0.0\n",
    "\n",
    "for ii in tqdm(range(reps)):\n",
    "    x_train, x_test, y_train, y_test = train_test_split(\n",
    "                    X1, y, train_size=0.8, random_state=ii, stratify=y)\n",
    "    clf = RandomForestClassifier(n_estimators=1000, n_jobs=-1)\n",
    "    clf.fit(x_train,y_train)\n",
    "    accuracy += np.mean(clf.predict(x_test)==y_test)\n",
    "\n",
    "print('Accuracy is ',accuracy/reps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "4e5553c9-45a5-45fd-a1dd-aea8c29efae2",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:19<00:00,  3.93s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy is  0.7063849765258217\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "reps = 5\n",
    "accuracy = 0.0\n",
    "\n",
    "for ii in tqdm(range(reps)):\n",
    "    x_train, x_test, y_train, y_test = train_test_split(\n",
    "                    X1, y, train_size=0.8, random_state=ii, stratify=y)\n",
    "    clf = RandomForestClassifier(n_estimators=1000, n_jobs=-1)\n",
    "    clf.fit(x_train,y_train)\n",
    "    accuracy += np.mean(clf.predict(x_test)==y_test)\n",
    "\n",
    "print('Accuracy is ',accuracy/reps)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "982c208e-4be2-452d-a8da-412706947044",
   "metadata": {},
   "source": [
    "## Try SPORF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "3fbb8067-9731-42e0-a032-ab75ad0c6c19",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:23<00:00,  4.79s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy is  0.7112676056338028\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "reps = 5\n",
    "accuracy = 0.0\n",
    "\n",
    "for ii in tqdm(range(reps)):\n",
    "    x_train, x_test, y_train, y_test = train_test_split(\n",
    "                    X1, y, train_size=0.8, random_state=ii, stratify=y)\n",
    "    clf = ObliqueRandomForestClassifier(n_estimators=1000, n_jobs=-1, feature_combinations=2.3)\n",
    "    clf.fit(x_train,y_train)\n",
    "    accuracy += np.mean(clf.predict(x_test)==y_test)\n",
    "\n",
    "print('Accuracy is ',accuracy/reps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "25e423e6-9a6b-402d-b941-5adb8cdc50ad",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:23<00:00,  4.73s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy is  0.7070422535211267\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "reps = 5\n",
    "accuracy = 0.0\n",
    "\n",
    "for ii in tqdm(range(reps)):\n",
    "    x_train, x_test, y_train, y_test = train_test_split(\n",
    "                    X2, y, train_size=0.8, random_state=ii, stratify=y)\n",
    "    clf = ObliqueRandomForestClassifier(n_estimators=1000, n_jobs=-1, feature_combinations=1.8)\n",
    "    clf.fit(x_train,y_train)\n",
    "    accuracy += np.mean(clf.predict(x_test)==y_test)\n",
    "\n",
    "print('Accuracy is ',accuracy/reps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "fff96dfb-7484-4135-8479-275cb83f2ade",
   "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>Unnamed: 0</th>\n",
       "      <th>participant_id</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>Markov.1</th>\n",
       "      <th>Markov.2</th>\n",
       "      <th>Markov.3</th>\n",
       "      <th>Markov.4</th>\n",
       "      <th>Markov.5</th>\n",
       "      <th>Markov.6</th>\n",
       "      <th>...</th>\n",
       "      <th>Schaefer217.191</th>\n",
       "      <th>Schaefer217.192</th>\n",
       "      <th>Schaefer217.193</th>\n",
       "      <th>Schaefer217.194</th>\n",
       "      <th>Schaefer217.195</th>\n",
       "      <th>Schaefer217.196</th>\n",
       "      <th>Schaefer217.197</th>\n",
       "      <th>Schaefer217.198</th>\n",
       "      <th>Schaefer217.199</th>\n",
       "      <th>Schaefer217.200</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>sub-1001</td>\n",
       "      <td>1.756164</td>\n",
       "      <td>M</td>\n",
       "      <td>3.048436</td>\n",
       "      <td>3.908286</td>\n",
       "      <td>3.221595</td>\n",
       "      <td>3.615675</td>\n",
       "      <td>4.662432</td>\n",
       "      <td>3.707754</td>\n",
       "      <td>...</td>\n",
       "      <td>4.231826</td>\n",
       "      <td>4.908868</td>\n",
       "      <td>4.522730</td>\n",
       "      <td>2.294943</td>\n",
       "      <td>2.853976</td>\n",
       "      <td>3.406234</td>\n",
       "      <td>4.261370</td>\n",
       "      <td>4.131977</td>\n",
       "      <td>3.387978</td>\n",
       "      <td>3.451267</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>sub-1002</td>\n",
       "      <td>1.783562</td>\n",
       "      <td>F</td>\n",
       "      <td>3.053520</td>\n",
       "      <td>3.748308</td>\n",
       "      <td>3.043567</td>\n",
       "      <td>3.764927</td>\n",
       "      <td>4.708283</td>\n",
       "      <td>4.060617</td>\n",
       "      <td>...</td>\n",
       "      <td>4.384853</td>\n",
       "      <td>4.849508</td>\n",
       "      <td>4.589500</td>\n",
       "      <td>2.443734</td>\n",
       "      <td>2.855187</td>\n",
       "      <td>3.344378</td>\n",
       "      <td>3.926697</td>\n",
       "      <td>3.477919</td>\n",
       "      <td>2.962553</td>\n",
       "      <td>3.474969</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>sub-1003</td>\n",
       "      <td>1.756164</td>\n",
       "      <td>M</td>\n",
       "      <td>3.211265</td>\n",
       "      <td>4.122524</td>\n",
       "      <td>3.374628</td>\n",
       "      <td>4.022762</td>\n",
       "      <td>4.759439</td>\n",
       "      <td>4.182558</td>\n",
       "      <td>...</td>\n",
       "      <td>4.570739</td>\n",
       "      <td>4.921833</td>\n",
       "      <td>4.770724</td>\n",
       "      <td>3.106145</td>\n",
       "      <td>3.094785</td>\n",
       "      <td>3.350355</td>\n",
       "      <td>4.562199</td>\n",
       "      <td>4.212585</td>\n",
       "      <td>3.582792</td>\n",
       "      <td>3.827813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>sub-1004</td>\n",
       "      <td>1.756164</td>\n",
       "      <td>M</td>\n",
       "      <td>3.004275</td>\n",
       "      <td>3.681716</td>\n",
       "      <td>3.227427</td>\n",
       "      <td>3.762712</td>\n",
       "      <td>4.555942</td>\n",
       "      <td>3.984013</td>\n",
       "      <td>...</td>\n",
       "      <td>4.264869</td>\n",
       "      <td>4.935628</td>\n",
       "      <td>4.505048</td>\n",
       "      <td>3.337418</td>\n",
       "      <td>2.892611</td>\n",
       "      <td>3.690076</td>\n",
       "      <td>4.095378</td>\n",
       "      <td>4.328465</td>\n",
       "      <td>3.763171</td>\n",
       "      <td>3.758017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>sub-1005</td>\n",
       "      <td>1.742466</td>\n",
       "      <td>M</td>\n",
       "      <td>2.868796</td>\n",
       "      <td>3.837011</td>\n",
       "      <td>2.997172</td>\n",
       "      <td>3.724171</td>\n",
       "      <td>4.537298</td>\n",
       "      <td>3.816082</td>\n",
       "      <td>...</td>\n",
       "      <td>4.154663</td>\n",
       "      <td>4.817727</td>\n",
       "      <td>4.695378</td>\n",
       "      <td>3.965287</td>\n",
       "      <td>3.219764</td>\n",
       "      <td>3.268439</td>\n",
       "      <td>4.115168</td>\n",
       "      <td>3.889531</td>\n",
       "      <td>3.271547</td>\n",
       "      <td>4.040183</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 386 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0 participant_id       age sex  Markov.1  Markov.2  Markov.3  \\\n",
       "0           0       sub-1001  1.756164   M  3.048436  3.908286  3.221595   \n",
       "1           1       sub-1002  1.783562   F  3.053520  3.748308  3.043567   \n",
       "2           2       sub-1003  1.756164   M  3.211265  4.122524  3.374628   \n",
       "3           3       sub-1004  1.756164   M  3.004275  3.681716  3.227427   \n",
       "4           4       sub-1005  1.742466   M  2.868796  3.837011  2.997172   \n",
       "\n",
       "   Markov.4  Markov.5  Markov.6  ...  Schaefer217.191  Schaefer217.192  \\\n",
       "0  3.615675  4.662432  3.707754  ...         4.231826         4.908868   \n",
       "1  3.764927  4.708283  4.060617  ...         4.384853         4.849508   \n",
       "2  4.022762  4.759439  4.182558  ...         4.570739         4.921833   \n",
       "3  3.762712  4.555942  3.984013  ...         4.264869         4.935628   \n",
       "4  3.724171  4.537298  3.816082  ...         4.154663         4.817727   \n",
       "\n",
       "   Schaefer217.193  Schaefer217.194  Schaefer217.195  Schaefer217.196  \\\n",
       "0         4.522730         2.294943         2.853976         3.406234   \n",
       "1         4.589500         2.443734         2.855187         3.344378   \n",
       "2         4.770724         3.106145         3.094785         3.350355   \n",
       "3         4.505048         3.337418         2.892611         3.690076   \n",
       "4         4.695378         3.965287         3.219764         3.268439   \n",
       "\n",
       "   Schaefer217.197  Schaefer217.198  Schaefer217.199  Schaefer217.200  \n",
       "0         4.261370         4.131977         3.387978         3.451267  \n",
       "1         3.926697         3.477919         2.962553         3.474969  \n",
       "2         4.562199         4.212585         3.582792         3.827813  \n",
       "3         4.095378         4.328465         3.763171         3.758017  \n",
       "4         4.115168         3.889531         3.271547         4.040183  \n",
       "\n",
       "[5 rows x 386 columns]"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_excel('Macaque.parcellated_thickness.xlsx')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "939c47d8-09b4-476d-827d-967f87dd65d1",
   "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>participant_id</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>sub-1001</td>\n",
       "      <td>1.756164</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>sub-1002</td>\n",
       "      <td>1.783562</td>\n",
       "      <td>F</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>sub-1003</td>\n",
       "      <td>1.756164</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>sub-1004</td>\n",
       "      <td>1.756164</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>sub-1005</td>\n",
       "      <td>1.742466</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  participant_id       age sex\n",
       "0       sub-1001  1.756164   M\n",
       "1       sub-1002  1.783562   F\n",
       "2       sub-1003  1.756164   M\n",
       "3       sub-1004  1.756164   M\n",
       "4       sub-1005  1.742466   M"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sex = pd.read_csv('~/spmmouse_segment/uwmadison.csv')\n",
    "df_sex.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "217af12b-5540-4fb9-af5b-99546e535008",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 592/592 [00:00<00:00, 1329.79it/s]\n"
     ]
    }
   ],
   "source": [
    "X1 = []\n",
    "X2 = []\n",
    "y = []\n",
    "IDs = set(df['participant_id'])\n",
    "ref_IDs = set(df_sex['participant_id'])\n",
    "\n",
    "for subject in tqdm(IDs):\n",
    "    if subject in ref_IDs:\n",
    "        features = np.array(df[df['participant_id']==subject]).reshape(-1)[4:]\n",
    "        gender = list(df_sex[df_sex['participant_id']==subject]['sex'])\n",
    "        sex = int(gender[0]=='F')\n",
    "             \n",
    "        X1.append(list(features[:182]))\n",
    "        X2.append(list(features[182:]))\n",
    "        y.append(sex)\n",
    "\n",
    "X1 = np.array(X1)\n",
    "X2 = np.array(X2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "d26412ad-dbe8-49c8-8edf-ec7bf1e14155",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(592, 182) (592, 200)\n"
     ]
    }
   ],
   "source": [
    "print(X1.shape, X2.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "762d5800-2d5c-4429-bb1f-5df61ffcc51a",
   "metadata": {},
   "source": [
    "### Try random forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "05b5d0e3-3bb4-4d44-a3c4-a01334e96899",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:09<00:00,  1.85s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy is  0.6453781512605041\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "reps = 5\n",
    "accuracy = 0.0\n",
    "\n",
    "for ii in tqdm(range(reps)):\n",
    "    x_train, x_test, y_train, y_test = train_test_split(\n",
    "                    X1, y, train_size=0.8, random_state=ii, stratify=y)\n",
    "    clf = RandomForestClassifier(n_estimators=1000, n_jobs=-1)\n",
    "    clf.fit(x_train,y_train)\n",
    "    accuracy += np.mean(clf.predict(x_test)==y_test)\n",
    "\n",
    "print('Accuracy is ',accuracy/reps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "733c24a0-6055-49ae-bfd6-7a55964c0116",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:09<00:00,  1.94s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy is  0.6504201680672269\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "reps = 5\n",
    "accuracy = 0.0\n",
    "\n",
    "for ii in tqdm(range(reps)):\n",
    "    x_train, x_test, y_train, y_test = train_test_split(\n",
    "                    X1, y, train_size=0.8, random_state=ii, stratify=y)\n",
    "    clf = RandomForestClassifier(n_estimators=1000, n_jobs=-1)\n",
    "    clf.fit(x_train,y_train)\n",
    "    accuracy += np.mean(clf.predict(x_test)==y_test)\n",
    "\n",
    "print('Accuracy is ',accuracy/reps)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "65b1a923-42f0-4898-aa6b-4fb91105e88e",
   "metadata": {},
   "source": [
    "### Try SPORF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "f57bbb90-569f-4a18-ae33-ba90861585cf",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:10<00:00,  2.02s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy is  0.6420168067226891\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "reps = 5\n",
    "accuracy = 0.0\n",
    "\n",
    "for ii in tqdm(range(reps)):\n",
    "    x_train, x_test, y_train, y_test = train_test_split(\n",
    "                    X1, y, train_size=0.8, random_state=ii, stratify=y)\n",
    "    clf = ObliqueRandomForestClassifier(n_estimators=1000, n_jobs=-1, feature_combinations=1.8)\n",
    "    clf.fit(x_train,y_train)\n",
    "    accuracy += np.mean(clf.predict(x_test)==y_test)\n",
    "\n",
    "print('Accuracy is ',accuracy/reps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "id": "672f0d12-5d3e-4265-aec2-3e246c2c6b77",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5/5 [00:10<00:00,  2.00s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy is  0.6084388185654009\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "reps = 5\n",
    "accuracy = 0.0\n",
    "\n",
    "for ii in tqdm(range(reps)):\n",
    "    x_train, x_test, y_train, y_test = train_test_split(\n",
    "                    X2, y, train_size=0.6, random_state=ii, stratify=y)\n",
    "    clf = ObliqueRandomForestClassifier(n_estimators=1000, n_jobs=-1, feature_combinations=1.8)\n",
    "    clf.fit(x_train,y_train)\n",
    "    accuracy += np.mean(clf.predict(x_test)==y_test)\n",
    "\n",
    "print('Accuracy is ',accuracy/reps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "5bb8c3c1-a1ca-4758-9971-8096ef238fc5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "cfb69260-0489-4a6a-a652-465654e5d1a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('feature_imp_white_pval_NHP.pickle', 'rb') as f:\n",
    "    p = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "a812dac7-c29b-475b-9c89-9ec7dc04a3b3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x723ae5dbe760>"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "idx = np.where(p>=1e-4)[0]\n",
    "mask = np.ones(p.shape)\n",
    "mask[idx] = 0\n",
    "slice = 55\n",
    "\n",
    "#img_gray = mask.reshape(113,137,113)\n",
    "img_gray = -np.log(p).reshape(113,137,113)\n",
    "\n",
    "sliced_img_graymatter = img_gray[:,:,slice]\n",
    "\n",
    "fig, ax = plt.subplots(1, 2, figsize=(16,8))\n",
    "ax[0].imshow(sliced_img_graymatter)\n",
    "ax[1].imshow(-np.log(p).reshape(113,137,113)[:,:,slice])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "20dc9fea-d330-4d5f-9bf3-64d5f20fb4b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "current_depth = 275\n",
    "current_width = 347\n",
    "current_height = 245\n",
    "\n",
    "desired_depth = 113\n",
    "desired_width = 137\n",
    "desired_height = 113\n",
    "\n",
    "depth = current_depth / desired_depth\n",
    "width = current_width / desired_width\n",
    "height = current_height / desired_height\n",
    "\n",
    "depth_factor = 1 / depth\n",
    "width_factor = 1 / width\n",
    "height_factor = 1 / height\n",
    "\n",
    "img_gray = ndimage.zoom(img_gray, (1.0/depth_factor, 1.0/width_factor, 1.0/height_factor), order=1)\n",
    "\n",
    "demo_filepath_whitematter = '/cis/home/jdey4/spmmouse_segment/sub-1001/mwc2T1w_brain.nii.gz'\n",
    "img_demo = nb.load(demo_filepath_whitematter)\n",
    "img_gray_to_save = nb.Nifti1Image(img_gray, header = img_demo.header, affine = img_demo.affine)\n",
    "\n",
    "nb.save(img_gray_to_save, 'feature_imp_white_rank_pval_NHP.nii')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "4dec11fc-a1b6-4e90-95f0-9e66568fc6e3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([   2653.,    8853.,   27601.,    6496.,   69209.,   64131.,\n",
       "           9664.,   60652.,   25374., 1474720.]),\n",
       " array([6.99930007e-04, 1.00629937e-01, 2.00559944e-01, 3.00489951e-01,\n",
       "        4.00419958e-01, 5.00349965e-01, 6.00279972e-01, 7.00209979e-01,\n",
       "        8.00139986e-01, 9.00069993e-01, 1.00000000e+00]),\n",
       " <BarContainer object of 10 artists>)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(p)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "534c9c86-0beb-4054-b901-405b39ed0b7e",
   "metadata": {},
   "source": [
    "# To-dos\n",
    "\n",
    "1. make a clear bug report on MORF\n",
    "2. get feature map parcellation data for humans and monkeys\n",
    "3. give white matter feature map, monkey map\n",
    "4. try naive transfer, silly-F\n",
    "5. do grid search\n",
    "6. add age as a feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "dbbf80d8-15e1-455c-9d7f-dfd94c96bc68",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([6.99930007e-04, 8.99910009e-04, 1.09989001e-03, ...,\n",
       "       1.00000000e+00, 1.00000000e+00, 1.00000000e+00])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sort(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "fc9e5ea3-e8ba-4424-8b01-8f81a3492af5",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('feature_imp_gray_NHP_random.pickle', 'rb') as f:\n",
    "    imp = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "eb290bfe-7abe-4603-8921-222c410c70f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "img = np.mean(imp[:500],axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "bb032842-26ab-4024-85e3-9945cad66177",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x723ae75a7d00>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, 1, figsize=(8,8))\n",
    "\n",
    "slice=59\n",
    "ax.imshow(img.reshape(113,137,113)[:,:,slice])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "08603491-e54f-4121-a9f1-6726803439da",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1., 1., 1., ..., 1., 1., 1.])"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "p"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d493eb96-9978-4e1f-9bee-8d9fb4d08acd",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}