Reversals_Debug.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np \n",
    "import pandas as pd\n",
    "import scipy\n",
    "\n",
    "\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from matplotlib.colors import ListedColormap\n",
    "import matplotlib.ticker as ticker\n",
    "import matplotlib.gridspec as gridspec\n",
    "import matplotlib.image as mpimg\n",
    "\n",
    "import seaborn as sns\n",
    "\n",
    "# to render plots with white background\n",
    "import seaborn as sns\n",
    "sns.set_style('white')\n",
    "# sns.set_context('notebook', font_scale=1.75)\n",
    "\n",
    "\n",
    "matplotlib.rcParams['pdf.fonttype'] = 42\n",
    "matplotlib.rcParams['ps.fonttype'] = 42\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.33 s, sys: 77.2 ms, total: 1.41 s\n",
      "Wall time: 1.64 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# These are the loaded dataframes that are calculated below\n",
    "\n",
    "df_corr_all = pd.read_csv(\"data/int_pc_qvals_gcamp_dms.csv\", index_col = 0) # .dropna(subset=['Q_dir_diff'])\n",
    "# should add this to my actual original df_corr_all\n",
    "# this just has unique MouseID's across recording sites. \n",
    "recordloc = np.array([0 if x == \"DMS\" else 1 for x in df_corr_all['RecordLoc']])\n",
    "df_corr_all['MouseID_unique'] = df_corr_all['MouseID'] + recordloc * 12 \n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "# update the gcamp values from strings --> arrays of floats \n",
    "df_src = df_corr_all\n",
    "\n",
    "df_src['g_np'] = df_src['g_np'].apply(\n",
    "                        lambda line: np.array([float(val) for val in line[1:-1].replace('\\n', '').split(' ') if val]))\n",
    "df_src['g_lp'] = df_src['g_lp'].apply(\n",
    "                        lambda line: np.array([float(val) for val in line[1:-1].replace('\\n', '').split(' ') if val]))\n",
    "df_src['g_choice'] = df_src['g_choice'].apply(\n",
    "                        lambda line: np.array([float(val) for val in line[1:-1].replace('\\n', '').split(' ') if val]))\n",
    "df_src['g_reward'] = df_src['g_reward'].apply(\n",
    "                        lambda line: np.array([float(val) for val in line[1:-1].replace('\\n', '').split(' ') if val]))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# some additional constants \n",
    "\n",
    "# for gcamp (dt = dt of gcamp recordings)\n",
    "dt = 15\n",
    "LEN_BEFORE = dt\n",
    "LEN_AFTER = dt * 2\n",
    "GCAMP_RANGE = np.arange(-1, 2, 1/dt)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_corr_DMS = pd.DataFrame(df_corr_all[df_corr_all['RecordLoc'] == 'DMS'])\n",
    "df_corr_DMS_CB = pd.DataFrame(df_corr_all[df_corr_all['RecordLoc'] == 'DMS_CB'])\n",
    "\n",
    "# PLOTTING CONSTANTS\n",
    "\n",
    "SZ_L = 15\n",
    "SZ = 8\n",
    "SZ_S = 6"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Actual Code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "df_corr_all is the pandas with terminals + cell-bodies data.\n",
      "df_DMS: only terminals\n",
      "df_DMS_CB: only cell-bodies\n"
     ]
    }
   ],
   "source": [
    "print(\"df_corr_all is the pandas with terminals + cell-bodies data.\")\n",
    "print(\"df_DMS: only terminals\")\n",
    "print(\"df_DMS_CB: only cell-bodies\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_corr_all['HighProb_num'] = df_corr_all['HighProb'].replace({'Ips':1, 'Con':0})\n",
    "df_corr_all['Reversal'] = df_corr_all.groupby([df_corr_all['MouseID'], df_corr_all['Session'], \n",
    "                                                   df_corr_all['RecordLoc']])['HighProb_num'].diff().fillna(0)\n",
    "# changedblock = 0 if no change;  1 or -1 denotes if it's contra -> ipsi OR ispi -> contra\n",
    "# nan originally meant -> new block/session. I've set it to 0 (since this is REVERSALS we're talkinga bout. )\n",
    "\n",
    "\n",
    "\n",
    "Trials_with_Switch = np.array(np.where(abs(df_corr_all['Reversal']) == 1))[0]\n",
    "Switch_type = df_corr_all[abs(df_corr_all['Reversal']) == 1]['Reversal'].values\n",
    "\n",
    "df_corr_all['TrialsAfterSwitch'] = np.nan\n",
    "df_corr_all['TrialsBeforeSwitch'] = np.nan\n",
    "\n",
    "df_corr_all.loc[Trials_with_Switch]['TrialsAfterSwitch'] = 0\n",
    "df_corr_all.loc[Trials_with_Switch]['TrialsBeforeSwitch'] = 0\n",
    "\n",
    "# at this point, you should have \n",
    "# ChangedBlock: 0 for no new block in this trial; 1 if there is a new block (in other words, indicator variable for changed block)\n",
    "# TrialsAfter/BeforeSwitch: 0...10 for trials after/before the switch\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# this is setting trials before/after switch properly \n",
    "for i in np.flip(np.arange(11), 0):\n",
    "    forward = np.array([min(trial,len(df_corr_all) - 1) for trial in Trials_with_Switch + i])\n",
    "    backward = np.array([max(trial,0) for trial in Trials_with_Switch - i])\n",
    "    df_corr_all.loc[forward, 'TrialsAfterSwitch'] = np.round(i)\n",
    "    df_corr_all.loc[backward,'TrialsBeforeSwitch'] = np.round(i)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ACTION: (old code)\n"
     ]
    }
   ],
   "source": [
    "print(\"ACTION: (old code)\")\n",
    "prob_prev = np.zeros((2,21))\n",
    "prob_prev_std = np.zeros((2, 21))\n",
    "\n",
    "num_recording_sites = len(np.unique(df_corr_all['MouseID_unique']))\n",
    "\n",
    "\n",
    "# for trials -10, +10 from block switch \n",
    "for t_i in np.arange(-10, 11):\n",
    "    if t_i < 0:\n",
    "        trials_i = df_corr_all[df_corr_all['TrialsBeforeSwitch'] == -t_i]\n",
    "    else:\n",
    "        trials_i = df_corr_all[df_corr_all['TrialsAfterSwitch'] == t_i]\n",
    "\n",
    "    for i, [switch_type_i, switch_prev, switch_after] in enumerate(zip([-1, 1], ['Ips', 'Con'], ['Con', 'Ips'])):\n",
    "        if (len(trials_i) == len(Switch_type) - 1):\n",
    "            actions = trials_i[Switch_type[:-1] == switch_type_i].groupby('Action')['MouseID_unique'].value_counts()\n",
    "        else:\n",
    "            actions = trials_i[Switch_type == switch_type_i].groupby('Action')['MouseID_unique'].value_counts()\n",
    "        \n",
    "        prev_count = actions[switch_prev].sort_index().reindex(np.arange(num_recording_sites), fill_value=0)\n",
    "        after_count = actions[switch_after].sort_index().reindex(np.arange(num_recording_sites), fill_value=0)\n",
    "\n",
    "        prob = prev_count / (prev_count + after_count)\n",
    "        \n",
    "        prob_prev[i, t_i + 10] = prob.mean()\n",
    "        prob_prev_std[i, t_i + 10] = prob.std()\n",
    "\n",
    "prob_prev_sem = prob_prev_std / np.sqrt(len(np.unique(df_corr_all['MouseID_unique'])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "REWARD (new code)\n"
     ]
    }
   ],
   "source": [
    "print(\"REWARD (new code)\")\n",
    "prob_reward = np.zeros((2,21))\n",
    "\n",
    "\n",
    "num_recording_sites = len(np.unique(df_corr_all['MouseID_unique']))\n",
    "\n",
    "\n",
    "# for trials -10, +10 from block switch \n",
    "for t_i in np.arange(-10, 11):\n",
    "    if t_i < 0:\n",
    "        trials_i = df_corr_all[df_corr_all['TrialsBeforeSwitch'] == -t_i]\n",
    "    else:\n",
    "        trials_i = df_corr_all[df_corr_all['TrialsAfterSwitch'] == t_i]\n",
    "\n",
    "    for i, [switch_type_i, switch_prev, switch_after] in enumerate(zip([-1, 1], ['Ips', 'Con'], ['Con', 'Ips'])):\n",
    "        if (len(trials_i) == len(Switch_type) - 1):\n",
    "            actions = trials_i[Switch_type[:-1] == switch_type_i].groupby(['Action', 'Reward'])['MouseID_unique'].value_counts()\n",
    "        else:\n",
    "            actions = trials_i[Switch_type == switch_type_i].groupby(['Action', 'Reward'])['MouseID_unique'].value_counts()\n",
    "        \n",
    "        prev_reward_count = actions[switch_prev, 1].sort_index().reindex(np.arange(num_recording_sites), fill_value=0)\n",
    "        \n",
    "        # hacky way of counting per recording site the reward/no-reward, so forth. \n",
    "        prev_nr_count = actions[switch_prev, 0].sort_index().reindex(np.arange(num_recording_sites), fill_value=0)\n",
    "\n",
    "        prob = prev_reward_count / (prev_reward_count + prev_nr_count )\n",
    "        \n",
    "        prob_reward[i, t_i + 10] = prob.mean()\n",
    "#         prob_prev_std[i, t_i + 10] = prob.std()\n",
    "\n",
    "# prob_prev_sem = prob_prev_std / np.sqrt(len(np.unique(df_corr_all['MouseID_unique'])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ips to Con\n"
     ]
    }
   ],
   "source": [
    "print(switch_name[switch_type])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.patches.Polygon at 0x1a1ff92898>"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "switch_type = 0\n",
    "switch_name = {0:'Con to Ips', 1:'Ips to Con'}\n",
    "nathan_reward = np.array([0.623543936092956,    0.751324601324601,    0.694633723054439,    0.714706780147957 ,   0.810514786211381 ,   0.699269151900731 ,   0.715130787828156  ,  0.682767373673507  ,  0.830285160542208  ,  0.739236678309108   , 0.855465131111016   , 0.158399787347156   , 0.101568051019805    ,0.185068471833178 ,   0.0377087872185911   , 0.191078658183921 ,   0.134263448969331  ,  0.132192460317460    ,0.110317460317460  ,  0.122817460317460   , 0.113795518207283])\n",
    "plt.plot(np.arange(-10, 11), prob_reward[switch_type,:], label = \"Reward\")\n",
    "plt.plot(np.arange(-10, 11), prob_prev[switch_type,:], label = \"Prob Contra\")\n",
    "# plt.plot(np.arange(-10, 11), nathan_ips_con_reward, label = \"Nathan_reward\")\n",
    "\n",
    "plt.title(\"Prob of \" + switch_name[switch_type])\n",
    "plt.legend()\n",
    "plt.axvspan(0, 11, color = 'grey', alpha = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18    2\n",
       "19    1\n",
       "20    1\n",
       "21    1\n",
       "Name: num_trials, dtype: int64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actions.loc[(actions['Action'] == 'Con') & (actions['Reward'] == 1), 'num_trials']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.71423077, 0.6821811 , 0.69196488, 0.6974153 , 0.75342837,\n",
       "        0.73844324, 0.72846452, 0.72019887, 0.81578427, 0.84170252,\n",
       "        0.77633789, 0.70007508, 0.55174716, 0.40413118, 0.41229605,\n",
       "        0.32045538, 0.30569232, 0.26484475, 0.26132589, 0.24790267,\n",
       "        0.23315097],\n",
       "       [0.72257571, 0.773229  , 0.82604352, 0.81256813, 0.79627855,\n",
       "        0.79035768, 0.82769603, 0.84874418, 0.8464366 , 0.90049508,\n",
       "        0.83574191, 0.76202189, 0.58856034, 0.54291925, 0.46352244,\n",
       "        0.41635696, 0.39022177, 0.33693168, 0.39441955, 0.3708359 ,\n",
       "        0.30802617]])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prob_prev\n",
    "print(\"First row is ips->con switches\")\n",
    "print(\"second row is con->ips switches\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x1a1f974f98>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "nathan_ips_con = np.array([0.677772529708960,    0.698081548225467 ,   0.666434660030906 ,   0.694352022188414 ,   0.692972844738602 ,   0.716832546856411  ,  0.734943889192130  ,  0.702393421183573 ,   0.717172857039767   , 0.823730421467626   , 0.865700146168827 ,   0.806793437344511    ,0.701525240327009    ,0.563628605889851 ,   0.430423366776213    ,0.419238232558338 ,   0.310126504118761 ,   0.274841517330203    ,0.249851720948033 ,   0.252367115005606  ,  0.243308501986484])\n",
    "plt.plot(np.arange(-10, 10), nathan_ips_con[1:], label = \"Nathan\")\n",
    "plt.plot(np.arange(-10, 10), prob_prev[0,:-1], label = \"Rachel\")\n",
    "plt.title(\"Prob of Ips->Con switches\")\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.003195181655815893"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(prob_prev[0,:-1] - nathan_ips_con[1:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x1a1fa79240>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "nathan_con_ips = np.array([0.678742327503937,    0.725201657647478,    0.756837929871986,    0.823564464431338,    0.814517158108489 ,   0.810130608985098 ,   0.765555397474902,    0.792203974216358  ,  0.839537466317652  ,  0.829236319236319 ,   0.871756929001511   , 0.805791016688850   , 0.757488624206890  ,  0.589139701028246    ,0.541719680388411    ,0.457260520697053   , 0.434965134624577 ,   0.402521723202838    ,0.375491611931240   , 0.423961731423032  ,  0.380972082612949])\n",
    "plt.plot(np.arange(-10,10), nathan_con_ips[1:], label = \"Nathan\")\n",
    "plt.plot(np.arange(-10,10), prob_prev[1,:-1], label = \"Rachel\")\n",
    "plt.title(\"Prob of Con->Ips switches\")\n",
    "plt.legend()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.002405123188183436"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(prob_prev[1,:-1] - nathan_con_ips[1:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RESPONSE FIGURE: IPS TO CON REVERSALS\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"RESPONSE FIGURE: IPS TO CON REVERSALS\")\n",
    "figure = plt.figure(figsize=(5, 3))\n",
    "ax = plt.gca()\n",
    "ax.errorbar(np.arange(-10, 11), prob_prev[0,:], yerr = prob_prev_sem[0, :], color = 'k')\n",
    "ax.set_xlabel('Trials from block switch', size = SZ_L)\n",
    "ax.set_ylabel('Probability of\\nIpsi Lever Press', size = SZ_L )\n",
    "ax.axvspan(0, 11, color = 'grey', alpha = 0.5)\n",
    "ax.set_xlim([-10.5, 10.5])\n",
    "ax.set_ylim([-0.01, 1.1])\n",
    "ax.set_xticks([-10, -5, 0, 5, 10])\n",
    "ax.set_yticks([0, 0.5, 1.0])\n",
    "ax.text(0.15, 0.05, 'Contra Lever\\nHigh Prob', size = SZ_L)\n",
    "ax.text(-10, 0.05, 'Ipsi Lever\\nHigh Prob', size = SZ_L)\n",
    "ax.tick_params(labelsize = SZ_L, length = 2)\n",
    "\n",
    "plt.tight_layout()\n",
    "# plt.savefig('figs/FigR9a.pdf', transparent=True)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RESPONSE FIGURE: CON TO IPS REVERSALS\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print(\"RESPONSE FIGURE: CON TO IPS REVERSALS\")\n",
    "\n",
    "\n",
    "figure = plt.figure(figsize=(5, 3))\n",
    "ax = plt.gca()\n",
    "ax.errorbar(np.arange(-10, 11), prob_prev[1,:], yerr = prob_prev_sem[1, :], color = 'k')\n",
    "ax.set_xlabel('Trials from block switch', size = SZ_L)\n",
    "ax.set_ylabel('Probability of\\nContra Lever Press', size = SZ_L )\n",
    "ax.axvspan(0, 11, color = 'grey', alpha = 0.5)\n",
    "ax.set_xlim([-10.5, 10.5])\n",
    "ax.set_ylim([-0.01, 1.1])\n",
    "ax.set_xticks([-10, -5, 0, 5, 10])\n",
    "ax.set_yticks([0, 0.5, 1.0])\n",
    "ax.text(0.15, 0.05, 'Ipsi Lever\\nHigh Prob', size = SZ_L)\n",
    "ax.text(-10, 0.05, 'Contra Lever\\nHigh Prob', size = SZ_L)\n",
    "ax.tick_params(labelsize = SZ_L, length = 2)\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.savefig('figs/FigR9b.pdf', transparent=True)\n",
    "\n",
    "\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}