├── README.md ├── behavior ├── documentary_beh.mat ├── gradCPTface_beh.mat ├── gradCPTscene_beh.mat ├── hcpparticipants.csv ├── sitcomep1_beh.mat ├── sitcomep2_beh.mat └── traitscores.csv ├── code ├── fig1_3.ipynb ├── fig2.ipynb ├── fig4.ipynb ├── fig5.ipynb └── hmmfit.py └── fmri ├── MNI152_T1_3mm_brain.nii.gz ├── gradientcoeff.mat ├── hmmmodel.mat ├── hmmoutput.mat ├── state1_raw.nii.gz ├── state2_raw.nii.gz ├── state3_raw.nii.gz ├── state4_raw.nii.gz ├── ts_documentary.mat ├── ts_gradCPTface.mat ├── ts_gradCPTscene.mat ├── ts_rest1.mat ├── ts_rest2.mat ├── ts_sitcomep1.mat └── ts_sitcomep2.mat /README.md: -------------------------------------------------------------------------------- 1 | # neuraldynamics 2 | 3 | **Song, H., Shim, W. M., Rosenberg, M. D. (2022) Large-scale neural dynamics in a shared low-dimensional state space reflect cognitive and attentional dynamics.** 4 | 5 | Raw functional MRI data of the **SitcOm, Nature documentary, Gradual-onset continuous performance task (SONG)** dataset is available on OpenNeuro [https://openneuro.org/datasets/ds004592]. The associated behavioral data, processed fMRI output, and main analysis codes are published here. Processed data or analysis on datasets other than SONG are not published. For additional inquiries, please contact Hayoung Song (hyssong@uchicago.edu). 6 | 7 | **1. behavior**: behavioral experiment data 8 | - sitcomep1, sitcomep2, documentary_beh.mat 9 | - raw continuous engagement ratings from a scale of 1 to 9 10 | - engagement ratings convolved with hemodynamic response function and z-normalized across time (used to relate with the fMRI latent state dynamics) 11 | - event index: event 1 and 2 interleaved in sequence 12 | - gradCPTface, gradCPTscene_beh.mat 13 | - response time variability (deviance from the mean) 14 | - response time variability convolved with hemodynamic response function and z-normalized across time (used to relate with the fMRI latent state dynamics) 15 | - traitscores.csv 16 | - subjID, sex (male/female), handedness (right/left), age 17 | - overall engagement ratings (oral report on a scale of 1 to 9) after movie watching scans 18 | - working memory capacity (K) measured with a color working memory task (Zhang & Luck, 2008 *Nature*) using set size 6. 19 | - hcpparticipants.csv: a list of HCP participant IDs that were used in the study 20 | 21 | **2. fmri**: fMRI analysis output 22 | - ts_*.mat: 25-parcel time series of the 7 runs of the SONG dataset (subj x time x parcel), used as inputs to the HMM. 23 | - hmmmodel.mat: model parameters estimated from the HMM fit using K=4 24 | - Means: mixture Gaussian emission probability, mean activity 25 | - Covars: mixture Gaussian emission probability, covariance 26 | - transmat: transition probability (however, a transition probability matrix reported in the manuscript was separately calculated per participant’s fMRI scan) 27 | - nfeatures: number of input features 28 | - niter: number of model fitting iterations (maximum set to 1,000) 29 | - startprob: initial probability which was set as random 30 | - hmmoutput.mat: HMM-decoded latent state sequence, summarized per fMRI condition and participant. Running code/hmmfit.py outputs hmmdecode.mat. hmmdecode.mat[‘train_state’] (a decoded discrete latent state sequence) was summarized with participant and condition indices and saved as hmmoutput.mat. 31 | - state1-4_raw.nii.gz: hmmmodel.mat[‘Means’] projected onto 25 parcels of the MNI152_T1_3mm_brain.nii.gz. 32 | 33 | **3. code**: analysis code 34 | - hmmfit.py: runs HMM, i.e., infers model parameters from ts_*.mat and decodes latent state sequence 35 | - Jupyter notebooks that reproduce Figure 1 to 5 36 | -------------------------------------------------------------------------------- /behavior/documentary_beh.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/behavior/documentary_beh.mat -------------------------------------------------------------------------------- /behavior/gradCPTface_beh.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/behavior/gradCPTface_beh.mat -------------------------------------------------------------------------------- /behavior/gradCPTscene_beh.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/behavior/gradCPTscene_beh.mat -------------------------------------------------------------------------------- /behavior/hcpparticipants.csv: -------------------------------------------------------------------------------- 1 | 102311 2 | 104416 3 | 105923 4 | 108323 5 | 109123 6 | 111514 7 | 114823 8 | 115017 9 | 118225 10 | 125525 11 | 128935 12 | 130114 13 | 131217 14 | 135124 15 | 137128 16 | 146129 17 | 146432 18 | 146735 19 | 146937 20 | 148133 21 | 155938 22 | 156334 23 | 157336 24 | 158035 25 | 159239 26 | 162935 27 | 164636 28 | 165436 29 | 167036 30 | 167440 31 | 172130 32 | 173334 33 | 175237 34 | 176542 35 | 177140 36 | 177645 37 | 178243 38 | 180533 39 | 181232 40 | 182739 41 | 185442 42 | 187345 43 | 191033 44 | 191336 45 | 191841 46 | 192641 47 | 193845 48 | 195041 49 | 196144 50 | 198653 51 | 199655 52 | 200311 53 | 200614 54 | 203418 55 | 204521 56 | 209228 57 | 212419 58 | 214019 59 | 214524 60 | 239136 61 | 246133 62 | 249947 63 | 251833 64 | 263436 65 | 318637 66 | 330324 67 | 352738 68 | 360030 69 | 380036 70 | 385046 71 | 389357 72 | 393247 73 | 395756 74 | 397760 75 | 401422 76 | 406836 77 | 412528 78 | 429040 79 | 436845 80 | 463040 81 | 467351 82 | 525541 83 | 541943 84 | 573249 85 | 581450 86 | 601127 87 | 617748 88 | 627549 89 | 638049 90 | 654552 91 | 671855 92 | 680957 93 | 690152 94 | 724446 95 | 725751 96 | 732243 97 | 757764 98 | 765864 99 | 770352 100 | 782561 101 | 783462 102 | 789373 103 | 814649 104 | 818859 105 | 826353 106 | 859671 107 | 861456 108 | 872764 109 | 878877 110 | 898176 111 | 899885 112 | 901139 113 | 910241 114 | 927359 115 | 942658 116 | 943862 117 | 958976 118 | 966975 119 | 971160 -------------------------------------------------------------------------------- /behavior/sitcomep1_beh.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/behavior/sitcomep1_beh.mat -------------------------------------------------------------------------------- /behavior/sitcomep2_beh.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/behavior/sitcomep2_beh.mat -------------------------------------------------------------------------------- /behavior/traitscores.csv: -------------------------------------------------------------------------------- 1 | ID,Sex,Handedness,Overall engagement: sitcom episode 1,Overall engagement: sitcom episode 2,Overall engagement: Documentary,Working memory (Pm),Working memory (SD),Working memory (K = Pm*setsize) 2 | sub-001,F,R,7,8,3,0.34086,30.3783,2.0452 3 | sub-002,F,R,9,8,2,0.71057,17.4787,4.2634 4 | sub-003,F,R,6,3,3,0.4138,27.2877,2.4828 5 | sub-004,F,R,7,8,3,0.56275,21.0089,3.3765 6 | sub-005,M,R,7,6,4,0.48976,22.592,2.9386 7 | sub-006,M,L,7,8,4,0.60475,32.1917,3.6285 8 | sub-007,F,R,7,5,3,0.69704,23.9828,4.1822 9 | sub-008,F,R,6,5,2,0.55354,19.3968,3.3212 10 | sub-009,M,R,7,5,3,0.46813,24.86,2.8088 11 | sub-010,F,R,7,7.5,4,0.7352,18.8149,4.4112 12 | sub-011,F,R,7,7,3,0.65196,45.9873,3.9117 13 | sub-012,F,R,7,6,4,0.29255,19.5441,1.7553 14 | sub-013,M,R,5,6,4,0.50417,23.8369,3.025 15 | sub-014,F,R,7,6,4,0.52661,22.3019,3.1597 16 | sub-015,M,R,8,9,3,0.72045,34.0316,4.3227 17 | sub-016,F,L,8,9,2.5,0.41468,27.4627,2.4881 18 | sub-017,M,R,8,8,3,0.37486,23.8731,2.2491 19 | sub-018,F,R,7,6,4,0.36304,27.6886,2.1783 20 | sub-019,M,R,7,6,6,0.42932,27.7695,2.5759 21 | sub-020,M,R,8,7,3,0.59462,23.1722,3.5677 22 | sub-021,M,R,6,7.5,2,0.5757,24.0886,3.4542 23 | sub-022,M,R,4,7,6,0.69954,30.153,4.1972 24 | sub-023,F,R,6,8,4,0.3741,15.6596,2.2446 25 | sub-024,M,R,7,8,6,0.64055,23.9472,3.8433 26 | sub-025,F,R,5,7,6,0.38618,19.2188,2.3171 27 | sub-026,M,R,6,8,3,0.25477,19.2984,1.5286 28 | sub-027,F,R,7,8,2.5,0.40186,39.3792,2.4112 -------------------------------------------------------------------------------- /code/fig5.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "Song, Shim, Rosenberg (2022) Large-scale neural dynamics in a shared low-dimensional state space reflect cognitive and attentional dynamics
\n", 8 | "code created by: Hayoung Song (hyssong@uchicago.edu), March 27, 2022\n", 9 | "\n", 10 | "The code generates **Figure 5** of the paper" 11 | ] 12 | }, 13 | { 14 | "cell_type": "code", 15 | "execution_count": 1, 16 | "metadata": {}, 17 | "outputs": [], 18 | "source": [ 19 | "import numpy as np\n", 20 | "import scipy.io\n", 21 | "import matplotlib.pyplot as plt\n", 22 | "import seaborn as sns\n", 23 | "import random\n", 24 | "import pandas as pd\n", 25 | "from statsmodels.stats.multitest import fdrcorrection\n", 26 | "random.seed(1234)" 27 | ] 28 | }, 29 | { 30 | "cell_type": "code", 31 | "execution_count": 2, 32 | "metadata": {}, 33 | "outputs": [], 34 | "source": [ 35 | "nstate = 4\n", 36 | "nsubj = 27\n", 37 | "loaddir ='../'\n", 38 | "niter = 1000\n", 39 | "cmap = np.array([[219 / 255, 68 / 255, 55 / 255],\n", 40 | " [57 / 255, 0 / 255, 216 / 255],\n", 41 | " [15 / 255, 157 / 255, 88 / 255],\n", 42 | " [255 / 255, 215 / 255, 0 / 255]])\n", 43 | "statename = ['DMN','DAN','SM','base']" 44 | ] 45 | }, 46 | { 47 | "cell_type": "markdown", 48 | "metadata": {}, 49 | "source": [ 50 | "## GradCPT: Latent state dynamics & Inverted RT variability (Figure 5c)" 51 | ] 52 | }, 53 | { 54 | "cell_type": "code", 55 | "execution_count": 3, 56 | "metadata": {}, 57 | "outputs": [ 58 | { 59 | "name": "stderr", 60 | "output_type": "stream", 61 | "text": [ 62 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/802030833.py:12: RuntimeWarning: Mean of empty slice\n", 63 | " stateocc[subj,stt-1]=np.nanmean(beh[subj,np.where(seq[subj,:]==stt)[0]])\n", 64 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/802030833.py:29: RuntimeWarning: Mean of empty slice\n", 65 | " stateocc_niter[subj,stt-1]=np.nanmean(beh_null[np.where(seq[subj,:]==stt)[0]])\n", 66 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/802030833.py:53: UserWarning: FixedFormatter should only be used together with FixedLocator\n", 67 | " ax.set_xticklabels(['DMN','DAN','SM','base'])\n" 68 | ] 69 | }, 70 | { 71 | "data": { 72 | "image/png": "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\n", 73 | "text/plain": [ 74 | "
" 75 | ] 76 | }, 77 | "metadata": {}, 78 | "output_type": "display_data" 79 | }, 80 | { 81 | "name": "stderr", 82 | "output_type": "stream", 83 | "text": [ 84 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/802030833.py:53: UserWarning: FixedFormatter should only be used together with FixedLocator\n", 85 | " ax.set_xticklabels(['DMN','DAN','SM','base'])\n" 86 | ] 87 | }, 88 | { 89 | "data": { 90 | "image/png": "iVBORw0KGgoAAAANSUhEUgAAAU0AAAGJCAYAAAAOiWFmAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAA9hAAAPYQGoP6dpAABdBElEQVR4nO3dd1xV9f8H8Ne5C+4VuGxUjAs4cSM4UDNxoWnuX4iGqzQzF2aaDXGUK3NkuTJXGVJqliNHZo4UN1qKi5lsEC4b7ji/P+jer9d7gXuY98L7+XhQ8rnnnPu+l8ubzzqfD8OyLAtCCCFG4dV1AIQQYk4oaRJCCAeUNAkhhANKmoQQwgElTUII4YCSJiGEcEBJkxBCOKCkSQghHFDSJIQQDihpkhrl7u6OyZMn65XHxMRg1qxZaNWqFcRiMSQSCdq1a4ePP/4YiYmJ2uMmT54MhmG0XxYWFmjdujVCQ0NRVFSEuLg4ncfL+4qLi6u9F16LsrOz4ejoiAMHDmjLli5dCoZh4OzsjNzcXL1z3N3dMWzYsCo/96NHj7BgwQL4+PjA1tYW9vb26NWrFw4ePKh3bN++fcv9+aSkpAAAFAoFmjdvjo0bN1Y5vpogqOsASMNz7NgxjBs3Do6Ojpg1axa8vb3BMAz+/vtv7Nq1C8ePH8ft27e1x4vFYvzxxx8AgKysLISFhWH58uV48OAB9u3bhytXruhcf+bMmZDL5di/f79OeZMmTWr+xdWBZcuWoWnTpggMDNR7LD09HWvXrsWKFStq5LlPnz6N48ePIzg4GF27doVSqUR4eDj+7//+D8uWLcOSJUu0x27ZsgU5OTk65xcUFGDw4MHw8fFB48aNAQBCoRBLlixBSEgIgoOD4eDgUCOxVxpLyAvUajVbUFBQLdeSyWTspEmTtN/HxMSwjRo1Yr29vdns7GyDz33o0CHt95MmTWIbNWqkd9zLL7/MAmCfPn2q99grr7zCtmvXrlriN3WZmZmsWCxmt23bplMeGhrKAmAHDx7MNmrUiE1OTtZ5XCaTsUOHDq3y86enp7NqtVqvfOjQoaxEImGLiorKPX/Pnj0sAHbnzp065cXFxay9vT372WefVTnG6kbN83rsl19+QceOHWFhYQFPT09s2rRJ22x7HsMwmDVrFrZt2wYvLy9YWFhg7969AEprMd27d4e9vT1sbGzQpUsXfPvtt2BfWOdFoVBg4cKFaNy4MSQSCXr37o1r167pxbR+/Xrk5+djy5YtkEqleo8zDIPRo0dX+Np69OgBAIiPjzfqvdi6dSs6deoEKysrWFtbo02bNvjwww91jklMTMT06dPx0ksvQSQSoWnTphg7dixSU1O1x+Tk5GDBggXw8PCASCSCq6sr5s2bh/z8fL3XMWvWLHz33Xfw8vKCRCJBp06dcOzYMb3YHj9+jPHjx8PZ2RkWFhbw8vLC119/bdTr2rNnD5RKpcFaJgB8+umnUCqVWLp0qVHX48rR0VHv8wQA3bp1Q0FBAZ49e1bu+d9++y2srKz04heJRAgMDMSOHTv0Pmt1jZrn9dTJkycxevRo9OnTB+Hh4VAqlVi3bp1OAnjekSNHcPHiRSxZsgSNGzeGs7MzACAuLg5vv/023NzcAAARERGYPXs2EhMTdZpe06ZNw759+7BgwQIMHDgQ//zzD0aPHq3Xn3b69Gm4uLhok15lPXnyBADg5ORU4bEHDhzAzJkzMXv2bKxbtw48Hg9PnjzB/fv3tcckJiaia9euUCgU+PDDD9GxY0dkZmbi1KlTyMrKgouLCwoKCvDKK6/g6dOn2mPu3buHJUuW4O+//8bvv/+uk0COHz+O69evY/ny5bCyssLatWsxatQoPHz4EJ6engCA+/fvo2fPnnBzc8MXX3yBxo0b49SpU5gzZw4yMjIQGhpa7ms7fvw4vL29YWtra/BxmUyGmTNnYvPmzZg/fz5atWpV5rVUKpVRCYrH44HHK7++de7cOTg5OWk/R4Y8fvwYFy9exFtvvQUrKyu9x/v27YutW7fin3/+QYcOHSqMq9bUcU2X1JCuXbuyL730EltcXKwty83NZR0cHNgXf+wAWKlUyj579qzca6pUKlahULDLly9nHRwctM2yqKgoFgAbEhKic/z+/ftZADrNc0tLS7ZHjx5Gvw5N81yhULAKhYJNT09nN23axDIMw3bt2tXgOS82z2fNmsXa2tqW+zxTp05lhUIhe//+/TKPWbVqFcvj8djr16/rlB88eJAFwJ44cUJbBoB1cXFhc3JytGUpKSksj8djV61apS0LCAhgmzVrxsrlcp1rzpo1i7W0tKzwZyKRSNgZM2bolWua5+np6WxGRgYrlUrZMWPGaB831DyXyWQsgAq/QkNDy43pm2++YQGwmzZtKve4RYsWsQDYK1euGHz88ePHLAB269at5V6ntlHzvB7Kz8/HjRs3MHLkSIhEIm25lZUVXnvtNYPn9OvXD3Z2dnrlf/zxBwYMGACpVAo+n6/tpM/MzERaWhqA0loFAEyYMEHn3Ndffx0CQdUbM/n5+RAKhRAKhXBycsK8efMwZMgQ/Pzzz0ad361bN2RnZyMoKAi//PILMjIy9I757bff4O/vDy8vrzKvc+zYMbRv3x6dO3eGUqnUfgUEBIBhGPz55586x/v7+8Pa2lr7vYuLC5ydnbVdCkVFRTh79ixGjRoFiUSic81XX30VRUVFiIiIKDOe7OxsFBQUlFubAwAHBwcsWrQIhw4dwtWrV8s87ujRo7h+/XqFX9OnTy/zGr/99hveffddjB07FrNnzy7zOKVSib1796Jdu3Zltjo0r+v52RSmgJrn9VBWVhZYloWLi4veY4bKAMMjy9euXcOgQYPQt29ffPPNN2jWrBlEIhGOHDmCzz77DIWFhQCAzMxMANCOfmoIBAK9kU83NzfExsZyej1isRgXLlwAAFhYWEAmk8HGxsbo84ODg6FUKvHNN99gzJgxUKvV6Nq1Kz799FMMHDgQQOkoc7Nmzcq9TmpqKp48eQKhUGjw8ReTsaFRXwsLC533TalUYvPmzdi8ebNR13ye5jqWlpblxg0A8+bNw1dffYWFCxfi/PnzBo9p27at0c1zQ06dOoXRo0dj4MCB2L9/v8G+To0TJ04gJSUFixYtKvMYzevSvE5TQUmzHrKzswPDMAb7LzVz4V5k6AN+4MABCIVCHDt2TOcX88iRIzrHaZJDSkoKXF1dteVKpVKbUDUCAgKwefNmREREGN2vyePx4Ovra9SxZZkyZQqmTJmC/Px8XLhwAaGhoRg2bBgePXoEmUwGJycnPH36tNxrODo6QiwWY9euXWU+zoWdnR34fD6Cg4Px7rvvGjzGw8OjzPM173tFgy1A6R+epUuXYvr06Th+/LjBY5o3b27UwFpoaKjewNKpU6cwcuRIvPLKKzh06JBOC8eQb7/9FiKRCMHBwWUeo3ldXN/XmkZJsx5q1KgRfH19ceTIEaxbt077Ac7LyzM4elsWhmEgEAjA5/O1ZYWFhfjuu+90juvbty8AYP/+/fDx8dGW//jjj1AqlTrHhoSEYNeuXZg5cybOnTunN4LOsiyOHDmCUaNGGR0nF40aNcKQIUNQUlKCkSNH4t69e5DJZBgyZAi+++47PHz4EK1btzZ47rBhw7By5Uo4ODiUm8yMJZFI4O/vj9u3b6Njx44VJpoXiUQieHp6Ijo62qjjp06dig0bNuCDDz6AWq3We/zo0aMoLi6u8DpNmzbV+f706dMYOXIkevfujSNHjsDCwqLc81NSUnDixAmMHj263DmYMTExAEprwKaEkmY9tXz5cgwdOhQBAQGYO3cuVCoVPv/8c1hZWRlVMwGAoUOHYv369Rg/fjymT5+OzMxMrFu3Tu+XwsvLC2+88QY2btwIoVCIAQMG4J9//sG6dev0mtEeHh44cOAAAgMD0blzZ+3kdqB0JHnXrl1gWbZak+a0adMgFovRq1cvNGnSBCkpKVi1ahWkUim6du0KoPT9+u2339CnTx98+OGH6NChA7Kzs3Hy5EnMnz8fbdq0wbx583Do0CH06dMHISEh6NixI9RqNRISEnD69Gm899576N69O6fYNm3ahN69e+Pll1/GO++8A3d3d+Tm5uLJkyc4evSodlJ/Wfr27YvffvvNqOfi8/lYuXKl9r3t2LGjzuOVGaG+dOkSRo4cicaNG+PDDz9EZGSkzuNt27bV+wzs3bsXSqUSb731VrnXjoiIAJ/PR58+fTjHVaPqdhyK1KSff/6Z7dChAysSiVg3Nzd29erV7Jw5c1g7Ozud4wCw7777rsFr7Nq1i23dujVrYWHBenp6sqtWrWK//fZbFgAbGxurPa64uJh97733WGdnZ+0I+ZUrV/Qmt2tER0ezM2fOZFu0aMFaWFiwYrGYbdu2LTt//nyd65Y1ub08L46e7927l/X392ddXFxYkUjENm3alH399dfZu3fv6pz377//slOnTmUbN27MCoVC7XGpqanaY/Ly8tiPP/6Ybd26NSsSiVipVMp26NCBDQkJYVNSUip8Tw29H7GxsezUqVNZV1dXVigUsk5OTmzPnj3ZTz/9tMLXevbsWRYAe+3aNZ3y50fPX9SzZ08WQLVMbtc8T1lf586d0zunVatWrLu7u8FJ8c97+eWX2ddee63KMVY3hmVNbOYoqTEKhQKdO3eGq6srTp8+XdfhkGrSsWNH9OrVC1u3bq3rUKpNdHQ0WrZsiVOnTmkH60wFJc167M0338TAgQO1TdJt27bh/PnzOH36NAYMGFDX4ZFqcvLkSYwaNQqPHz+ucAaAuZgyZQqePn2KM2fO1HUoeqhPsx7Lzc3FggULkJ6eDqFQiC5duuDEiROUMOuZwYMH4/PPP0dsbGy9SJpKpRLNmzfH4sWL6zoUg6imSQghHNAdQYQQwgElTUII4YCSJiGEcEBJs5awLIucnByTWxuQEMINJc1akpubC6lUanC/FlOnmcry+PHjug6FE3ONm5g2SpqkQj/88AMSExMRFhZW16FwYq5xE9NG8zSJQXfv3tVubrZv3z7t/2UyGQDA29tb795lU2CucRPzQfM0a0lOTg6kUinkcjmntSDrypgxY3D48OEyHx89ejQOHTpUixEZx1zjJuaDmufEoD179uCNN94w+FhwcDD27NlTuwEZyVzjJuaDapq1xNxqmhqtW7fGo0ePdL5/8OBBHUZkHHONm5g+qmmSMsXHx2sTj0QiAQA8fPgQCQkJdRlWhcw1bmIeGlzS3LJlCzw8PGBpaQkfHx9cvHix3OOLi4vx0UcfQSaTwcLCAs2bNy9zu4P65q+//gIALFy4EGlpaXj//fcBlC48a8rMNW5iHhpU8zw8PBzBwcHYsmULevXqhe3bt2Pnzp24f/++dl/vF40YMQKpqan49NNP0aJFC6SlpUGpVKJnz56cntscm+clJSWIjY3V2f5Bs2d3WZuLmQJzjZuYhwaVNLt3744uXbroLNbq5eWFkSNHYtWqVXrHnzx5EuPGjUNMTAzs7e2r9NzmmDQJIfoaTPO8pKQEN2/exKBBg3TKBw0ahMuXLxs859dff4Wvry/Wrl0LV1dXtGrVCgsWLDC5LUUJIbWnwUxuz8jIgEql0tv328XFpcxtbWNiYnDp0iVYWlri559/RkZGBmbOnIlnz56V2a8ZFhZm8A4UhUJR9RdBCKlzDSZpary4vzfLsmVuaq9Wq8EwDPbv36/danb9+vUYO3Ysvv76a4jFYr1zgoKCEBQUpFeuaZ4TQsxbg2meOzo6gs/n69Uq09LS9GqfGk2aNIGrq6tOsvPy8gLLsnj69GmNxksIMU0NJmmKRCL4+PjobdR05syZMkfCe/XqhaSkJOTl5WnLHj16BB6PVy/2YiGEcNdgkiYAzJ8/Hzt37sSuXbsQFRWFkJAQJCQkYMaMGQCAxYsXY+LEidrjx48fDwcHB0yZMgX379/HhQsX8P7772Pq1KkGm+aEkPqvQfVpBgYGIjMzE8uXL0dycjLat2+PEydOaFfASU5O1rlrxMrKCmfOnMHs2bPh6+sLBwcHvP766/j000/r6iUQQupYg5qnWZdoniYh9UODap4TQkhVUdIkhBAOKGkSQggHlDQJIYQDSpqEEMIBJU1CCOGAkiYhhHBASZMQQjigpEkIIRxQ0iSEEA4oaRJCCAeUNAkhhANKmoQQwgElTUII4YCSJiGEcEBJkxBCOKCkSQghHFDSJBVSKBTIycmhvdsJQQPbI4hwl5aWhvj4eO3+8DKZDM7OznUdFiF1hmqapEwKhQLx8fFQKpUoKCiAUqlEfHw81ThJg0Y1TVKmwsJCZGVlIT09XVvTdHJyQmFhIYRCYV2HR0idoJomKZNAIEBGRgY0G5ayLIuMjAwIBPS3ljRclDRJmZRKJRwdHcEwDACAYRg4OjpCqVTWcWSE1B2qMpAyicVi2NnZwcrKCiUlJRCJRBAKhRCLxXUdGiF1hmqapExCoRAymQxCoRASiUTne0IaKqppknI5OzvDzs4OhYWFEIvFlDBJg0dJk1RIKBRSsiTkP9Q8J4QQDihpEkIIB5Q0CSGEA0qahBDCASVNQgjhgJImIYRw0OCS5pYtW+Dh4QFLS0v4+Pjg4sWLRp33119/QSAQoHPnzjUbICHEpDWopBkeHo558+bho48+wu3bt/Hyyy9jyJAhSEhIKPc8uVyOiRMnon///rUUKSHEVDGsZgmbBqB79+7o0qULtm7dqi3z8vLCyJEjsWrVqjLPGzduHFq2bAk+n48jR44gMjKS83Pn5ORAKpVCLpfDxsamMuETQkxAg6lplpSU4ObNmxg0aJBO+aBBg3D58uUyz9u9ezeio6MRGhpa0yESQsxAg7mNMiMjAyqVCi4uLjrlLi4uSElJMXjO48eP8cEHH+DixYtGryEZFhaGsLAwvXJa7ZyQ+qHBJE0NzdqQGpoVyV+kUqkwfvx4LFu2DK1atTL6+kFBQQgKCtIr1zTPzZFCoaAFOwj5T4NJmo6OjuDz+Xq1yrS0NL3aJwDk5ubixo0buH37NmbNmgUAUKvVYFkWAoEAp0+fRr9+/Wol9rpEG6sRoqvB9GmKRCL4+PjgzJkzOuVnzpxBz5499Y63sbHB33//jcjISO3XjBkz0Lp1a0RGRqJ79+61FXqd0Wys9vx2F7SxGmnoGkxNEwDmz5+P4OBg+Pr6ws/PDzt27EBCQgJmzJgBAFi8eDESExOxb98+8Hg8tG/fXud8Z2dnWFpa6pXXV4WFhXhxcgXLsrSxGmnQGlTSDAwMRGZmJpYvX47k5GS0b98eJ06cgEwmAwAkJydXOGezIRGLxWAYRidxMgxD212QBq1BzdOsS+Y6T5P6NAnR1aBqmoQ72u6CEF2UNEmFaLsLQv6nwYyeE0JIdaCkSQghHFDSJIQQDqqUNIuLi6srDkIIMQuckuapU6cwefJkNG/eHEKhEBKJBNbW1njllVfw2WefISkpqabiJIQQk2DUPM0jR45g0aJFkMvlePXVV9GtWze4urpCLBbj2bNn+Oeff3Dx4kVcuXIFkydPxooVK+Dk5FQb8ZsNc52nSQjRZVTS7NatGz755BMMHToUPF7ZldPExERs2rQJLi4ueO+996o1UHNHSZOQ+oHuCKollDQJqR84DwRduHABaWlpeuUKhQIXLlyolqAIIcRUcU6affv2RadOnXDlyhWd8mfPnsHf37/aAiOEEFNUqSlH48aNQ//+/bFnzx6dcmrpE0LqO85Jk2EYLF68GN9//z1mz56N+fPna5OloW0jCCGkPuGcNDUJcvTo0bhw4QIOHjyIIUOGIDs7u7pjI4QQk1OlO4K8vb1x7do1ZGdno3///tUVEyGEmCzOSXPSpEk6K3c3btwY58+fR//+/eHm5latwRHToFAokJOTQ3sDEQKap1lrzHWeJq3cXjdo22TTZdQixAkJCZxqkYmJiXB1da10UMQ0aHajVCgUKCkpgUgkQnx8POzs7OgXuQbRHyrTZlTzvGvXrpg2bRquXbtW5jFyuRzffPMN2rdvj8OHD1dbgKTuFBYWIisrC7GxsXj69CliY2ORlZWFwsLCug6t3qJtk02fUTXNqKgorFy5EoMHD4ZQKISvry+aNm0KS0tLZGVl4f79+7h37x58fX3x+eefY8iQITUdN6kFAoEAGRkZOr/AGRkZEAhol5SaQtsmmz6japr29vZYt24dkpKSsHXrVrRq1QoZGRl4/PgxAGDChAm4efMm/vrrL0qY9YhSqYSjoyPUajWKi4uhVqvh6OgIpVJZ16HVW5ptk59H2yabFk5VBktLS4wePRqjR4+uqXiICdH8Amt+iTX/pl/gmiMUCiGTyfT6NKmWaTqonUXKxTAMeDweLCwsdBIoqTm0bbJpo6RJylRYWAgbGxtIJBLt6LlAIKD+tVpA2yabLkqapEya5rlAINAO/lDzvHbQPE3TVWdJ88mTJ4iOjkafPn0gFou1/TfEdGj612JiYlBUVARLS0t4enrSL3ENo3mapo3zbZT5+flVesLMzEwMGDAArVq1wquvvork5GQAwFtvvUVbZJgolUqFwsJCqFSqug6l3qN5mqaPc9J0cXHB1KlTcenSpUo9YUhICAQCARISEiCRSLTlgYGBOHnyZKWuSWqGQqHA3bt3ERMTg6SkJMTExODu3bv0C1yDypunSUwD56QZFhYGuVyO/v37o1WrVli9ejWnrXtPnz6NNWvWoFmzZjrlLVu2RHx8PNdwSA3KyclBTEwMkpOTkZmZieTkZMTExCAnJ6euQ6u3NP3ISqUSBQUFUCqV1I9sYjgnzddeew2HDh1CUlIS3nnnHYSFhUEmk2HYsGE4fPhwhROf8/PzdWqYGhkZGbCwsOAaDqlBKpUKcrkcarUaCoUCarUacrmcmuk1SCgUwsrKCnFxcXj69Cni4uJgZWVF/cgmpNLraTo4OCAkJAR37tzB+vXr8fvvv2Ps2LFo2rQplixZgoKCAoPn9enTB/v27dN+zzAM1Go1Pv/8c9pjyMTw+XwIhUJkZ2cjJycH2dnZEAqF4PP5dR2aUcxxSTuFQoG8vDx4eHigWbNm8PDwQF5enlm9hvqu0qPnKSkp2LdvH3bv3o2EhASMHTsWb775JpKSkrB69WpERETg9OnTeud9/vnn6Nu3L27cuIGSkhIsXLgQ9+7dw7Nnz/DXX39V6cWQ6iUWi9GoUSPY2Nho52k2atTILJqK5joCrenT5PP52hYZ3XtuWjjXNA8fPozXXnsNbm5u+OGHH/Duu+8iMTER33//Pfz9/TFhwgQcOHAAf/75p8Hz27Zti7t376Jbt24YOHAg8vPzMXr0aNy+fRvNmzev6uup0JYtW+Dh4QFLS0v4+Pjg4sWLZR57+PBhDBw4EE5OTrCxsYGfnx9OnTpV4zGaCqVSCTs7O22Nk8/nw87OzuTvPTfnEWi699z0ca5pTpkyBePGjcNff/2Frl27GjzG09MTH330UZnXaNy4MZYtW8b1qassPDwc8+bNw5YtW9CrVy9s374dQ4YMwf379w2uF3rhwgUMHDgQK1euhK2tLXbv3o3XXnsNV69ehbe3d63HX5OSk5O10780iouLER8fr72VkmEYxMfH4969e3r9z02aNEGTJk1qM+QymfNKQTQ31vRxXrm9oKDA4ECOsXbv3g0rKyv83//9n075Tz/9hIKCAkyaNKnS165I9+7d0aVLF2zdulVb5uXlhZEjR2LVqlVGXaNdu3YIDAzEkiVLOD23qa/cvnTpUr0/ZCKRCC4uLpBIJODz+VCpVCgoKEBqaipKSkp0jg0NDcXSpUtrMeKyKRQKREZG6iROhmHQuXNns0g+aWlpeknTHLoWGgrONU1ra2skJyfr/RAzMzPh7Oxc4cjq6tWrsW3bNr1yZ2dnTJ8+vcaSZklJCW7evIkPPvhAp3zQoEG4fPmyUddQq9XIzc2Fvb19TYRYp95++20MHz5cp0ypVOLmzZvYtGkTAGDixIlwcXFBhw4d9NbUNJVaJmDeKwVpuhbUajWA0s8crZZvWjgnzbIqpsXFxRCJRBWeHx8fDw8PD71ymUyGhIQEruEYLSMjAyqVCi4uLjrlLi4uSElJMeoaX3zxBfLz8/H666+XeUxYWBjCwsL0yk29P81Q81qhUCAtLU37vYuLC9zc3ODt7W3yv8DmulKQZrX8lJQUKBQKCIVCNG7c2Cy6FhoKo5Pml19+CaC0mbNz505YWVlpH1OpVLhw4QLatGlT4XWcnZ1x9+5duLu765TfuXMHDg4OxoZTaS92sht7z3tYWBiWLl2KX375pdymUlBQEIKCgvTKNc1zc6LpG9S8R2q12mz6BgHzXClIIBAgPj4eGRkZKC4uhoWFBYqLi9G7d++6Do38x+ikuWHDBgClSWbbtm06c/VEIhHc3d0NNrtfNG7cOMyZMwfW1tbo06cPAOD8+fOYO3cuxo0bxzV+ozk6OoLP5+vVKtPS0vRqny8KDw/Hm2++iZ9++gkDBgyosRhNjUAgwL///qv9A5mRkQEA9Atcg4qKipCZmYnExESo1WrtAFxRUVGVxhJI9TE6acbGxgIA/P39cfjwYdjZ2VXqCT/99FPEx8ejf//+2n4xtVqNiRMnYuXKlZW6pjFEIhF8fHxw5swZjBo1Slt+5swZjBgxoszzwsLCMHXqVISFhWHo0KE1Fl9NCBr2KrISEyt/AR4PjJWNtqZ54cwZMGoVDm3bAvzX58aVnasrwo6dqHxM9VxhYSHkcjkkEok2acrlcrr33IRw7tM8d+5clZ5QJBIhPDwcK1aswJ07dyAWi9GhQwfIZLIqXdcY8+fPR3BwMHx9feHn54cdO3YgISEBM2bMAAAsXrwYiYmJ2juWwsLCMHHiRGzatAk9evTQ1lLFYrFZNLWzEhPxlWPl5/fJweA8j4cdqQXg8/no72gDCx4fPYWWkILTpAutWVVJ4g2AUqmEWCxGfn6+tjUnFotNfm6sRkNYB9SopDl//nysWLECjRo1wvz588s9dv369UY9catWrdCqVSujjq0ugYGByMzMxPLly5GcnIz27dvjxIkT2oSdnJysMxi1fft2KJVKvPvuu3j33Xe15ZMmTcKePXtqNfa60Ags7NjS2RA8Hg9qAPasCo0qmTBJxZycnNCkSRNkZmZCpVKBz+fDwcEBTk5OdR1ahcz1LiyujEqat2/f1o7+3r59u8zjjBlQUalU2LNnD86ePYu0tDTt1AqNP/74w5iQKm3mzJmYOXOmwcdeTIRl3dXUUAgAWLBqSCQS8Hg8ZPP48FSpaLn/GiSRSNC7d2/tbcYikQi+vr4m359Z1l1Y9XGqlFGf/+eb5FVtns+dOxd79uzB0KFD0b59e1qt3USkFZcgvVh3WlQJgHsSGxQUFIDH40FVZIG74IEtyMeLk8ucLIRwtqh4yhmpmJeXF2QyGbKysmBnZ2fyCRP430wLpVJZ7/eTqvVKw4EDB/Djjz/i1Vdfre2nJuU48DQNm2N1+xsbNWqks+7pL8mZAID1T5/qreA/28MVc5rrrpFKKk8ikZhFstQQi8XIyclBamqqtnnu4uJSL++ZNyppctnn/PDhw+U+LhKJ0KJFC6OvR2rHuGbO6O+kOyOiBMDNRvYo5PMAMABYiFVq+EiFBmuapGFjWRZqtVpb0+R4h7bZMCppVudI8XvvvYdNmzbhq6++oqa5CXG2EOk1r5UAsvg8POX972PSjK9EB34j6tckOjTNc5VKheLiYvD5fLO6EYILoz77u3fvrrYnvHTpEs6dO4fffvsN7dq103tDK6qpktpTBAYurAo2KhVywIMN1BD/V25FI+jkOZo7mbKysgBAu3ByfbwRotYrDLa2tjqTy4npsgSLPIbBM6Z0vmA6eLBnVbCkhFnjCgoKzGogqKioSHvLrVKphEAgAMuy9fJOJqOSZpcuXXD27FnY2dnB29u73Gb1rVu3yr1WddZaSS1gATUDKMFAABaUL2teVFQUbt++rb0jyNvbG15eXnUdVrnK6r+sj/2aRiXNESNGaBedHTlyZJWfVKlU4s8//0R0dDTGjx8Pa2trJCUlwcbGRmchEFK3isA8lyNZ7X+peV5zCgoKtPOiNTW227dvQyaTmXSNTTNKzjCMTpdbgx09Dw0NNfjvyoiPj8fgwYORkJCA4uJiDBw4ENbW1li7di2KioqMWvSD1A4BWGTz+OAB2tHybB4fAjUlzJqSlZWF3NxcyOVybZlUKkVWVpZJJ02lUgl3d3ftAtWaBazN5fZPLirdp3njxg1ERUWBYRh4eXnBx8fHqPPmzp0LX19fvaXgRo0ahbfeequy4ZAaoAQDW1aNTIanbZ47sGoo/5t+RKqftbU1cnNzdcpyc3NhbW1dRxEZRywWw87ODtbW1tol7QQCQcOtaT7v6dOnCAoKwl9//QVbW1sAQHZ2Nnr27ImwsDC89NJL5Z5/6dIl/PXXX3oLFstkMiTSYg4mxRKstoGuSZEMWBoIqmHu7u6IjY3VThJ/ce1ZU/T83kYAzGq1fK4470Y5depUKBQKREVF4dmzZ3j27BmioqLAsizefPPNCs9Xq9UGt8R4+vSpyf81bZBezI+UL2uUWCyGTCaDn58fOnToAD8/P8hkMrOpsalUKhQWFla47Y0545w0L168iK1bt6J169bastatW2Pz5s3lboerMXDgQGzcuFH7PcMwyMvLQ2hoKN1aaWKKwID9b6aEtsbJMCgC3ZRQUzQ1Nk3TViAQmEWNTaFQ4M6dO4iLi0NmZibi4uJw584dk9/mpTI4N8/d3NwMvhFKpRKurq4Vnr9+/Xr069cPbdu2RVFREcaPH4/Hjx/D0dHR4N46pO4IwCKb0f27ms3wSqceEfKc3NxcpKWl6axylJaWVi83IuScNNeuXYvZs2fj66+/ho+PDxiGwY0bNzB37lysW7euwvNdXV0RGRmJAwcO4ObNm1Cr1XjzzTcxYcIEs2mCNBRKMBCq1Ujj/6+W46pS0EBQDdIsscbj8bSj5eawxBrN03yBnZ2dzoT2/Px8dO/eXbtdhWY+2dSpU8udx6lQKNC6dWscO3YMU6ZMwZQpU6oWPalRArBQ8HiwYVUoBA9iqKHg8WjKUQ3S3MP9PHO4h9vGxgYuLi56qxzZ2NjUdWjVzqik+XwfZFUIhUIUFxfTQh1mQgkGChb4ly+ECgAffLhTTbNGicViMAyjkzgZhjH5VphQKETHjh0RExODoqIiWFpawtPT06QTfWUZlTQnTZpUbU84e/ZsrFmzBjt37tTWVImpYpHG0/RgMmD/+x4qSpg1RTMQ9OK2EeaQfJydnWFlZWVW98xXRpWyVmFhod6gUEXV8atXr+Ls2bM4ffo0OnTogEaNGuk8TqscmY4c8MCidCVN/nPzNXPAgxUqtxslqZizszPs7OzMboOy5/cISkpKath7BD0vPz8fixYtwo8//ojMzEy9xyuan2Vra4sxY8ZwfVpSByzBQgQWApaFGqXz03j/lZOaJRQKzSZZArRHULkWLlyIc+fOYcuWLZg4cSK+/vprJCYmYvv27Vi9enW55yqVSvTt2xcBAQFo3LhxpYMmtcMGLJqpFXjKE2on9DZTK2BjJkmzIWwnayrMdQCrMjgnzaNHj2Lfvn3o27cvpk6dipdffhktWrSATCbD/v37MWHChLKfTCDAO++8g6ioqCoFTWqHAEBndQnsWTXy/muSu7FKs1i13dy3kzW39TQ1A1gKhUK7YIdQKDT5AazK4Pz5f/bsGTw8PACU9l8+e/YMANC7d2+88847FZ7fvXt37VJXxPTZs2rYsCUoAgNLsGaRMM2lqZicnIzk5GS98oSEBERHR2sTfvPmzeHm5qZ3XJMmTdCkSZPaCLVCQqEQVlZWuH79OvLz89GoUSN07drVpN7v6sL5d8DT0xNxcXGQyWRo27YtfvzxR3Tr1g1Hjx7VLuBRnpkzZ+K9997D06dP4ePjozcQ1LFjR64hkRomAMxq/UxzaSpu374dy5Yt0ynj8Xho0aKFzrQ8lmXx5MkTqNW6g2+hoaFYunRpbYRaIc16FMnJyVAqlcjJyUFUVBRatGhhUu95deCcNKdMmYI7d+7glVdeweLFizF06FBs3rwZSqUS69evr/D8wMBAAMCcOXO0ZZp5aQzD1Osb/UntMJe5jm+//TaGDx+uU5aeno5bt25h7969AEqn+wmFQnTo0AFOTk46x5pKLRMobYFGR0frLEIcHR2NZ8+ewcXFpY6jq16ck2ZISIj23/7+/njw4AFu3LiB5s2bo1OnThWeHxsby/UpCeHEXOY6GmpeFxQUaJvsfD4fTk5OkEgkePnll026b1OzR9DzNHsE1TdV7qJyc3Mz2N9SFurLJLXBXOc6SiQSuLm5QSKRgGEYpKen45VXXjHphAkATk5OsLW1RXZ2trbM1tZWr3ZcHxiVNL/88ktMnz4dlpaW+PLLL8s99vlmtyH79u0r9/GJEycaExIhFTK3uY5Aad+gUChEcnIyRCKRdjBFU26qNLXhmzdvaldu9/HxMflkXxlGJc0NGzZgwoQJsLS0xIYNG8o8jmGYCpPm3Llzdb5XKBQoKCiASCSCRCKhpGmClIBZjZ6bs+cHsVQqFfh8vkkOYhni5eUFmUxmVlOlKsOo34Hn+yGr2iep2Uz+eY8fP8Y777yD999/v0rXJtXvGcNDCiPQ3k7ZmFXCnqVbKGuKWCxGTk4ObGxswDAM4uLi0KxZM5MbxCqLRCKpt8lSg1PF4fml3dq2bVttQbRs2RKrV6/GG2+8gQcPHlTbdUnVKAFtwgRK7ztPYQSwYUuoxlmGgLHDkZCeVKVrWPEttP/+Zv8eKKFG7qeLK309N6emOHXw1yrFRP6H02e/Jpd24/P5SEqq2oeNVC/dfc9L0b7n5UtIT0LypJaVPl9QzKI4tQQF26IhEAhQ0qstIOIjx54PpUUlf+/2Pq50PFw1hFtXOVcYqrq026+/6v7FY1kWycnJ+Oqrr9CrVy/O1yM1p3Q3St2VMxnQgh01SSUE2P9yo1KphIgBwJSWmzpzv3XVWJyzXlWXdntxZXeGYeDk5IR+/frhiy++4BoOZ1u2bMHnn3+O5ORktGvXDhs3bsTLL79c5vHnz5/H/Pnzce/ePTRt2hQLFy7EjBkzajxOUyBAaR/mi32a5tI0N8daD8tjkG/NA/Nf4mEZoMCGB5ZnOgt3G7r9U6lUIiYmRmeuZnR0NDw9PfUqV6Z0+2dlcP78V3VptxdvBatN4eHhmDdvHrZs2YJevXph+/btGDJkCO7fv29wrmlsbCxeffVVTJs2Dd9//z3++usvzJw5E05OTg1meTtzvPccKK31PHr0CLm5ubC2tkarVq3MptZTLGaQl5EBsViMYnsGrITzprE1ytDtnyKRCHZ2dnrHZmVloaSkRKfMlG7/rAzOvwO7d++uiThqxfr16/Hmm2/irbfeAlC6jcepU6ewdetWrFq1Su/4bdu2wc3NTbvdh5eXF27cuIF169Y1mKQJmN+95wqFApcuXdJZ9CItLQ2vvfaaWdQ4LQpZWDo6lt76+YxFAaNGsQklTkO3fyqVSjx48AALFy6EQCDAihUrIBaLy6xpmrNarziMHTsWvr6++OCDD3TKP//8c1y7dg0//fRTjTxvSUkJbt68qfe8gwYNwuXLlw2ec+XKFQwaNEinLCAgAN9++63JTzZuyDIzMxEdHQ2lUgm1Wg0ej4fo6GhkZmaa1Dqu6uxCsPJCnTJGDYizWBT/N9iqTs+HOIuHQms12BfyJiMVg2db+1ORympeCwQCSKVSMAwDhmHQqVMneHl51Xp8Na1SSfPgwYP48ccfkZCQoFf1vnXrVrnnnj9/HqGhoXrlgwcPNmoL4MrKyMiASqXSWzzAxcUFKSkpBs9JSUkxeLxSqURGRobBD05YWJjB/ds124IEBgbWWrL9OykFPWMKKz6wFgktxXq1FEOuXrmJokJFhccZIpaI4dn8JfAFAu1AlkqpxBeff4XCgsq/H5ZiIbr7+ZR7TObjp1C9/8So6xUXFkFRWKxTJhKJUPxcM7f44J3S/xto5grFFrAQW1b4PJkWlka957duXoFSUbn3h2EY2NjYISdPAT6fj1UrlwJgkSN/VqVtfAVCMbr4+FX6fK5eHKg2hHPS/PLLL/HRRx9h0qRJ+OWXXzBlyhRER0fj+vXrePfddys8Py8vDyKRSK9cKBQiJyeHazicvThdStN843K8oXKNoKAgBAUF6ZXn5ORAKpUiPDy8Xm5rWt06tx4MyaNNlTu5KA9Wsp/AZwqgyZoqpQSez/4PUFlVOqYC17lG/VIZy9gBFYZhamVAZfjg1vh146NKnZuTL0LkoxK8HlL6Gr5ZpoSlBdBGVgKbRiUVX6CsmOa5Vut7Xh04J80tW7Zgx44dCAoKwt69e7Fw4UJ4enpiyZIl2gWJy9O+fXuEh4djyZIlOuUHDhyo1gnzL3J0dASfz9erVaalpZW5dFXjxo0NHi8QCODg4FBjsZKqYfgMlM/cAelTMHwlWJUASnkzMHwGrAmtPFhW0nN3d8fjx4+1dwa1bNmyVgaxpHYyDJ9X+fNVagmAZAgEAqzZ4wQ+jwFYaZVjMjWck2ZCQgJ69uwJoPSWr9zcXABAcHAwevToga+++qrc8z/55BOMGTMG0dHR6NevHwDg7NmzCAsLq7H+TKC02ePj44MzZ85g1KhR2vIzZ85gxIgRBs/x8/PD0aNHdcpOnz4NX19f6s80YazCAqq8xlAX2oEnLIRaIQarsgCrsKj4ZBOQmZmJ2NhYFBcXIzMzE/b29rWSNL8LO12l82/duoWgoCDw+XyMCZwNPz+/etmnyXlIrnHjxtpdKGUyGSIiIgCUTs8xpu9i+PDhOHLkCJ48eaKzivvvv/+uN4ezus2fPx87d+7Erl27EBUVhZCQECQkJGjnXS5evFhnwZAZM2YgPj4e8+fPR1RUFHbt2oVvv/0WCxYsqNE4SRWxAqhyG4NVWUBdLAWrsoAqtzHAmv6EqYKCAty+fRsAYGFRmuRv376NgoKCugyrQgqFAklJSdpBIJZl8fTpU70tvusDzp+ifv364ejRo+jSpQvefPNNhISE4ODBg7hx4wZGjx5t1DWGDh2KoUOHcg62qgIDA5GZmYnly5cjOTkZ7du3x4kTJ7RrfCYnJyMhIUF7vIeHB06cOIGQkBB8/fXXaNq0Kb788ssGNd3IXKkL7aAuFoMnzoG60AZQVzxgYgqysrL05jKr1WpkZWWZ9EIYubm5SE9PB/C/fv+0tDTk5ubC3t6+LkOrdgzLcWhLrVZDrVZrO6V//PFHXLp0CS1atMCMGTMMDvI87/r161Cr1ejevbtO+dWrV8Hn8+Hr68vxJZgHzUCQXC6ngSAjVGkgCABPnAW+TSJ4ghKolSKoclyhLtSffM1FQau5iHx4skrXqPA5Cgpw5MgRncTJ4/EwcuRIk0mahgawcnJy8Oeff2q7s/r37w8bGxv07t1b7/Pe4O4I4vF44PH+16p//fXX8frrrxt9/rvvvouFCxfqJc3ExESsWbMGV69e5RoSIboYJQQOTyCQPAMYFnyWASMsREmit8k30SUSCby9vXH79m3tHFNvb2+TSZhA2RvCtWnTBpaWluDxeDh79iyKiooQGhpq0hvCVQbnT5CHhwfeeOMNTJgwAW3atOH8hPfv30eXLl30yr29vXH//n3O1yPkRYwoX5swSwtYCCTPoBDlgy2u2mhubTD1xXwN3RGUn5+P27dvIykpSbt4ctOmTeHt7a23PoU51zKBSq5yFBYWhs8++wze3t4IDg5GYGCg0W+EhYUFUlNT4enpqVOenJxcqVWTCNHDQrvAyHNFMKM7QSEUCmFtbW2SszTK2hDu33//ha2trfZuOYFAUC+3vOA8ej5//nxcv34dDx48wLBhw7B161a4ublh0KBBFe7/AwADBw7E4sWLIZfLtWXZ2dn48MMPMXDgQK7hEKKHVTSCusAeLFiArwALtvR7RaOKTzYBaWlpiIyMxIMHDxAZGYm0tLS6DqlCSqUSjv/dLw+UTsh3dHSEUqms48iqX6VXAWjVqhWWLVuGhw8f4uLFi0hPT8eUKVMqPO+LL77Av//+C5lMBn9/f/j7+8PDwwMpKSm1sjQcaQBYAVT5Tv9bmJJl/vve9FsyCoVCuyYlUDoSHR8fb/JTd17ca15zp525bNPBRZU+RdeuXcMPP/yA8PBwyOVyjB07tsJzXF1dcffuXezfvx937tyBWCzGlClTEBQUZJJNEWKGGCUYYSFUOY3BiArBlojBCAsBRmnyifP5jdU0zGVjNYZhwOfzwePxtPM16yPOn6BHjx5h//79+OGHHxAXFwd/f3+sXr0ao0ePhrW1tVHXaNSoEaZPn845WEKMwQiLwbPIBU+cBYZhwVrmQF1oB0ZYDLbEtJOmpsamUChQUlICkUgEoVBo8jW2wsJC2NjYQCKRaOMWCARmkey54vwJatOmDXx9ffHuu+9i3LhxlVpqKzo6Ghs3bkRUVBQYhoGXlxfmzp2L5s2bc74WIS9ilXzwxNlgGM3iKix44mywSn4dR1YxoVAIKysrvSlHpp54NMn++T5Nap7/58GDB2jVqlWln/DUqVMYPnw4OnfujF69eoFlWVy+fBnt2rXD0aNHaTCIVBkjUEFdaAu+OBPgKwCVEOpCWzACFdjKL7hTKxQKBfLy8uDh4YHi4mJYWFggLy/P5Ndv1ST7mzdvauP28fEx6Zgri3PSrErCBIAPPvgAISEhWL16tV75okWLKGmSKmMVFmBZBuCV9m2yLFP6ZQYLdmj6NPl8vnaqjjn0aSoUCiQmJuoMBCUmJqJFixYmHXdl1Poa+lFRUXjzzTf1yqdOnUqT20m14TdKh8DuXwisMiCw+xf8Rul1HZJRNM3c55lDMzcnJwepqalgGAYWFhZgGAapqam1skZubav1pOnk5ITIyEi98sjISLPZ+IqYNkaUA74kC2yxFdQljcAWW4EvyQIjMv1fYKFQCJlMptM3KJPJTL62VtZIeX0cQa/1ocRp06Zh+vTpiImJQc+ePcEwDC5duoQ1a9bgvffeq+1wSD3ECIvBaO4JUpcO/jBgS0fPi8s/1xQ4OzvDzs7OrLYftra2hrOzM9LS0rRzNJ2dnY2eUWNOjE6anp6euH79epVXLP/kk09gbW2NL774AosXLwYANG3aFEuXLsWcOXOqdG1CAEBdYAdVsTX4Frml95+zDFTF1lAXVG2Vo9okFArNIllqCIVCdOrUCTExMSgqKoKlpSU8PT3N6jUYy+ikGRcXB5Wq6nsFMAyDkJAQhISEaFd9r49/jeoThUJhVrUeqC1RktwJQud7pbVLhQUUae3MZk1Nc+Xs7AwrKyuTXWikutTpTF9KlqYvLS1Ne1ufpn/NHPqe1UVSKOXNwBcVQFUigbrI9Fc3MnfPf1aSkpLM5rPCFaekef/+/TK3u9Xo2LFjlQIipqOs+6Dt7OxMu8bJKMG3TgHDCqEuloIBwLdOgbrI2uRvozRXms9KUVER8vLyYGVlZR6flUrg9Anq37+/wX2ANDfqMwxTLU14YhrM4T7oEqShBLrTiXjCIlgw6QBU4AlUUCv5AMNHsbAA6hLdJroIThCh/tWGalthYSHi4+O1e4UxDAMPDw+0bt3aZD4r1YVT0rx69SqcnJxqKhZiYl5cuQYwvTmDqQjHU3ytU8YoGLgKXSGRSMDn86FSqVBQUIBERWLpcnHPaYZ38RJm12bI9VZcXJxOqyQuLq5uA6ohnJKmm5tbtfZRaEbZiGnSzBl8sU/TlGoOLgiEHfq9UKqClfAuihodB8MwsGH9UKJwhB3aAdC9/1wEqgRUh9zcXFhbW+usk2ttbY3c3Nx6tydWrXfwqNVqfPbZZ9i2bRtSU1Px6NEjeHp64pNPPoG7u7vBu4VI3TH1OYMiOOs1rxmRHCJhAp4VFIDP50OqcoNYaAmIZGax3YU5srOzg7W1NSwsLFBQUACJRAKRSAQ7O/OZ5mUso+8IeuWVV8rdaTI5ORmzZs2q8Dqffvop9uzZg7Vr1+pcr0OHDti5c6ex4ZBaJBQKYWNjY3IJs0zs/7a6KO1j55nddhfmRiKRwN3dHRkZGcjLy0NGRgbc3d3r5bQjo5PmuXPnkJSUhK+//ho7duxAdnY2ACAjIwMhISHw9PTEH3/8UeF19u3bhx07dmDChAng8//XVOrYsSMePHjA/RUQ8gJW0QiqAtv/9cWyjFltdwGUjkbn5OSY/IrtGppVmHx9fdGyZUv4+vpCKBSaTfxcGN08P3bsGMaMGaN9E9auXYtvvvkGr7/+Otq3b4+ffvoJw4YNq/A6mpVPXqRWq+vlG0zqACtASaYn5Ao5BAIBFMomYHM8zWa6kTnOjS0sLERWVhbS09PBsiwKCgrg5ORkUjMtqovRNc3PPvsMM2bMQE5ODtatW4eYmBjMmDEDhw4dwrlz54xKmADQrl07XLx4Ua/8p59+gre3t/GRE1IOdaEt0tLSkJ6ejpK0llAXmkffmrnuESQQCJCRkQGlUoni4mIolUpkZGTUyx1mjX5FUVFR2Lt3L6ysrDBnzhwsXLgQGzduRJ8+fTg9YWhoKIKDg5GYmAi1Wo3Dhw/j4cOH2LdvH44dO8b5BZD6yVVmh0TMrfT5KrUS7BMWJSUlKGixGHxe1X95XWU1n3jNYW6sIUqlEiKRSLumpmaeZn3cjdLoT1JOTg5sbW1LTxIIIBaLK7Ug8WuvvYbw8HCsXLkSDMNgyZIl6NKlC63aTnQcPx1WpfPz8/NhZWUFAIiI/AWNGplHf6ZmbqymxmZhYaH9fTNlAoEAJSUlcHZ2hlKp1H7foGuagO5tlCzL4uHDh8jPz9c5xpjbKAMCAhAQEMDlqQnhjGEYCIVCs6rtaLaNuHbtmnbqTrdu3Uy6lgn8b9/z9PR08Pn8er3veZVuo9T0Y3K5jbKsJeays7PRpUsXxMTEcAmJEIPS09Ph5OQEhmFw9+5dtG7d2uQHU4DSPs2oqCikpKRAqVQiJycHUVFRJr9thFgs1s7VNKcacmUYnTRjY2Or5QnLWmKuuLgYiYmJ1fIcpGFTKBT4999/tauGm81CIwAyMzMRHR2trSUDpbu3ZmZmVmrn19ry/N1jmpqmqd09Vl2MTpoymaxKT/Trr79q/33q1ClIpf+7M0OlUuHs2bNwd3ev0nMQApjvYApQWnkwFHtxsekvOW/qd49Vl1rrpR05ciSA0qb8pEmTdB4TCoVwd3fHF198UVvhkHoiOTkZycnJOmVKpVJnsYjY2FhYWJTuRPniwESTJk3QpEmTGo/TWE5OTrCzs0NWVpa2zM7OzmwWyjG3Fecrg2ENrfVWgzw8PHD9+nU4OjrW5tPWuZycHEilUsjl8nq3gEFdWrp0KZYtW6ZXLhaLYW1tre1vz83NRWFhod5xoaGhWLp0aS1EaryoqCjcuHEDJSUlEIlE8PX1hZeXV12HRf5T60mzrmRlZWHOnDnaboLhw4dj8+bN2mlUL1IoFPj4449x4sQJxMTEQCqVYsCAAVi9ejWaNm3K+fkpadYMQzVNjRen7RhiajVNjYKCgnq/bYS5qpOkmZ+fj/PnzyMhIQElJSU6j9XU5mpDhgzB06dPsWPHDgDA9OnT4e7ujqNHjxo8Xi6XY+zYsZg2bRo6deqErKwszJs3D0qlEjdu3OD8/JQ0Cakfaj1p3r59G6+++ioKCgqQn58Pe3t7ZGRkQCKRwNnZuUamHEVFRaFt27aIiIhA9+7dAQARERHw8/PDgwcP0Lp1a6Ouc/36dXTr1g3x8fFwc3PjFAMlTULqB6MGgry9vY3e9P3WrVvlPh4SEoLXXnsNW7duha2tLSIiIiAUCvHGG29g7tzK3zZXnitXrkAqlWoTJgD06NEDUqkUly9fNjppyuVyMAxTZpOeEFL/GZU0NSPfQOlq61u2bEHbtm3h5+cHoLTWdu/ePcycObPCa0VGRmL79u3g8/ng8/koLi6Gp6cn1q5di0mTJmH06NGVeyXlSElJMTix2dnZucKN4jSKiorwwQcfYPz48eXWFMPCwhAWpn8LoKkvuFAfmd3Ww8QsGJU0Q0NDtf9+6623MGfOHKxYsULvmH///bfCawmFQm2t1cXFBQkJCfDy8oJUKkVCQgKX2MscOX3e9evXAcBgTVlzF1NFFAoFxo0bB7VajS1btpR7bFBQEIKCgvTKNc1zUjvMcXk1Yh44z9P86aefDA6EvPHGG/D19cWuXbvKPd/b2xs3btxAq1at4O/vjyVLliAjIwPfffcdOnTowCmWWbNmYdy4ceUe4+7ujrt37yI1NVXvsfT0dLi4uJR7vkKhwOuvv47Y2Fj88ccf1B9pBsx262FiFjgnTbFYjEuXLqFly5Y65ZcuXTJqk7SVK1ciNzcXALBixQpMmjQJ77zzDlq0aFFhwn2Ro6OjUfM9/fz8IJfLce3aNXTr1g1A6c6acrkcPXv2LPM8TcJ8/Pgxzp07p3e/PDFN5nxHEDF9nJPmvHnz8M477+DmzZvo0aMHgNI+zV27dmHJkiUVnu/r66v9t5OTE06cOME1BM68vLwwePBgTJs2Ddu3bwdQOuVo2LBhOoNAbdq0wapVqzBq1CgolUqMHTsWt27dwrFjx6BSqbT9n/b29uXul0TqljlsPUzMGFsJ4eHhbM+ePVk7OzvWzs6O7dmzJxseHm7Uuf7+/mxWVpZeuVwuZ/39/SsTjlEyMzPZCRMmsNbW1qy1tTU7YcIEvTgAsLt372ZZlmVjY2NZlG7Fpfd17tw5zs8vl8tZAKxcLq/6iyEVSk1NZa9du8ZevXqVvXbtGpuamlrXIZF6otbnafJ4PIOj2WlpaXB1da23o8w0T7P20V01pCZUasGO7OxsHDx4EDExMViwYAHs7e1x69YtuLi4wNXV1eA5d+/e1f77+cWMgdJVjk6ePFnmuYRwlZaWhsePHyMnJwc2NjZo2bIljZ6TasE5ad69excDBgyAVCpFXFwc3nrrLdjb2+Pnn39GfHw89u3bZ/C8zp07g2EYMAyDfv366T0uFouxefNm7q+AkBcoFApcvHgRMTEx2ilHKSkpGD58OA0EkSrjnDTnz5+PyZMnY+3atbC2ttaWDxkyBOPHjy/zvNjYWLAsC09PT1y7dk1nqSuRSARnZ2edfdAJqaysrCxtwgRKR85jYmKQlZVFtU1SZZyT5vXr17Uj0M9zdXUt9+4azSLGarWa61MSwklBQYHBKUcFBQV1FBGpT4ze91zD0tISOTk5euUPHz40aqHUvXv34vjx49rvFy5cCFtbW/Ts2RPx8fFcwyFEj2YF8efZ2dlRLZNUC85Jc8SIEVi+fLl2lJthGCQkJOCDDz7AmDFjKjx/5cqV2vlyV65cwVdffYW1a9fC0dERISEhXMMhRI9EIkHv3r3h4uICa2truLi4oHfv3jSCTqoF5ylHOTk5ePXVV3Hv3j3k5uaiadOmSElJgZ+fH06cOFHh/tISiQQPHjyAm5sbFi1ahOTkZOzbtw/37t1D3759kZ6eXqUXZKpoylHtotFzUlM492na2Njg0qVL+OOPP3Dr1i2o1Wp06dIFAwYMMOp8KysrZGZmws3NDadPn9bWLi0tLQ1uR0AIV5p7z4VCofbWV7r3nFQXzklz3759CAwMRL9+/XSmDpWUlODAgQOYOHFiuecPHDgQb731Fry9vfHo0SMMHToUAHDv3j3ajZJUC7r3nNQkzn2aU6ZMgVwu1yvPzc3FlClTKjz/66+/hp+fH9LT03Ho0CFtTeDmzZsGl1QjhCvNvefPo3vPSXXh3KfJ4/GQmpqqN1J+584d+Pv749mzZ9UaYH1BfZq1y9zX06QFlE2X0c1zzZYXDMOgf//+Orv7qVQqxMbGYvDgwTUSJCFcaaYdmWPiMfeEX98ZnTQ1W15ERkYiICAAVlZW2sdEIhHc3d2NmnJESG0RCoVmlSwBWkDZHBidNENDQ6FSqSCTyRAQEGCSe0UTYu5oEMv0cRoI4vP5mDFjBoqKimoqHkIaNBrEMn2cR887dOhQpb3J+/Xrh+zsbL3ynJwcg6sfEdKQCIVCyGQybeLU9GlSLdN0cB49P336NBYtWoQVK1bAx8dH7w6gikaGaRFiGj0nFaPRc9PFeXK7ZoR8+PDhOs0IzUifSqUyeB4tQkyI8cxxEKuh4Jw0z507V6knokWICSH1Qa3tEaSZRtFQFyGm5jkh9UOl9gi6ePEitm/fjpiYGPz0009wdXXFd999Bw8PD/Tu3dvgObQIMSGkPuCcNA8dOoTg4GBMmDABt27dQnFxMYDSe89Xrlxp9D7m9+/fR0JCAkpKSnTKhw8fzjUkQgipNZyb597e3ggJCcHEiRNhbW2NO3fuwNPTE5GRkRg8eHC5W14AQExMDEaNGoW///4bDMNoJ/JqBpXKGkgyd9Q8r300Al37GsJ7znme5sOHD9GnTx+9chsbG4PzL180d+5ceHh4IDU1FRKJBPfu3cOFCxfg6+uLP//8k2s4hBiUlpaGyMhIPHjwAJGRkUhLS6vrkOq9hvKec06aTZo0wZMnT/TKL126BE9PzwrPv3LlCpYvXw4nJyfweDzweDz07t0bq1atwpw5c7iGQ4iesu7frq9zgE1BQ3rPOSfNt99+G3PnzsXVq1fBMAySkpKwf/9+LFiwADNnzqzwfJVKpV3sw9HREUlJSQBKB4oePnzINRxC9JR3/zapGQ3pPec8ELRw4ULI5XL4+/ujqKgIffr0gYWFBRYsWIBZs2ZVeH779u1x9+5deHp6onv37li7di1EIhF27NhhVE2VkIpo7t9WKpUoLi6GhYUFBAKBWd2/bW59g5r3/PnEWV/vma/0PM2CggLcv38farUabdu21VkqrjynTp1Cfn4+Ro8ejZiYGAwbNgwPHjyAg4MDwsPD6+395zQQVLuioqJw+/ZtqNVq8Hg8eHt7w8vLq67DMoq5rqdprnFzxTlp7t27F2PHjq1w10kunj17Bjs7O73VXeoTSpq1R6FQIDIyEgqFAiUlJRCJRBAKhejcubPJ19o0sb9YYzOH2AHzqyFXBuc+zQULFsDZ2Rnjxo3DsWPHoFQqqxyEvb19vU6YpHZp+tcEAgEkEgkEAoHZ9K+Ze9+gUCiEjY1NvU2YQCWSZnJyMsLDw8Hn8zFu3Dg0adIEM2fOxOXLl2siPkI4e75Ps6CgAEql0mz612g9TdPHOWkKBAIMGzYM+/fvR1paGjZu3Ij4+Hj4+/ujefPmNREjIZwIhUJYWVkhLi4OT58+RVxcHKysrMyi9kPraZq+St17riGRSBAQEICsrCzEx8cjKiqquuIipNIUCgXy8vLg4eGhHT3Py8uDQqEwi+RjzpvCNQSca5pA6cj5/v378eqrr6Jp06bYsGEDRo4ciX/++ae646s2WVlZCA4OhlQqhVQqRXBwsFF3MGm8/fbbYBgGGzdurLEYSfXQ9Avy+XxIJBLw+Xyz6hcEGkbfoLniXNMMCgrC0aNHIZFI8H//93/4888/0bNnz5qIrVqNHz8eT58+xcmTJwEA06dPR3BwMI4ePVrhuUeOHMHVq1fRtGnTmg6TVIOGNGeQ1D7OSZNhGISHhyMgIEBn73NTFhUVhZMnTyIiIgLdu3cHAHzzzTfw8/PDw4cP0bp16zLPTUxMxKxZs3Dq1CkMHTq0tkImVaDpF3xxziDV2kh14Jz1fvjhh5qIo0ZduXIFUqlUmzABoEePHpBKpbh8+XKZSVOtViM4OBjvv/8+2rVrV1vhkmpA/YKkplSqqnj27FmcPXsWaWlpeosK79q1q1oCq06GNnIDSn+xylvKbs2aNRAIBJwWEgkLC0NYWJheeX1cuMDU0T47pCZwTprLli3D8uXL4evriyZNmtTppPSlS5di2bJl5R5z/fp1ADAYp6bpZsjNmzexadMm3Lp1i9NrDAoKQlBQkF655o4gQoh545w0t23bhj179iA4OLgm4uFk1qxZGDduXLnHuLu74+7du0hNTdV7LD09HS4uLgbPu3jxItLS0uDm5qYtU6lUeO+997Bx40bExcVVKXZCiHninDRLSkpMZrTc0dERjo6OFR7n5+cHuVyOa9euoVu3bgCAq1evQi6Xl/lagoODMWDAAJ2ygIAABAcHY8qUKVUPnhBiljjP03zrrbfMbjDIy8sLgwcPxrRp0xAREYGIiAhMmzYNw4YN0xkEatOmDX7++WcAgIODA9q3b6/zJRQK0bhx43JH2wkh9RvnmmZRURF27NiB33//HR07dtTraF+/fn21BVed9u/fjzlz5mDQoEEASjdw++qrr3SOefjwIeRyeV2ERwgxE5yXhvP39y/7YgyDP/74o8pB1Ue0NBwh9UOlFyEm3FDSJKR+qNS954QQ0lAZ3ac5evRoo447fPhwpYMhpDo1hFXESe0zOmnSxGxiThrKfjWk9lGfZi2hPs3aY+777BDTRn2apN4x9312iGmjpEnqHdpnh9QkSpqk3qF9dkhNMo9VhAnhyNnZGVZWVsjKyoKdnR0kEkldh0TqCUqapF56fvQ8KSmJRs9JtaHmOal3FAqFNmECpYNA8fHxtBA0qRaUNEm9Q6PnpCZR0iT1Do2ek5pESZPUOzR6TmoSDQSReol2oyQ1hZImqbdoN0pSE6h5TgghHFDSJIQQDihpEkIIB5Q0CSGEA0qahBDCASVNQgjhgJImIYRwQEmTEEI4oKRJCCEcUNIkhBAOKGkSQggHlDQJIYQDSpqEEMIBJU1CCOGAkiYhhHBASZMQQjigpEkIIRw0mKSZlZWF4OBgSKVSSKVSBAcHIzs7u8LzoqKiMHz4cEilUlhbW6NHjx5ISEio+YAJISapwSTN8ePHIzIyEidPnsTJkycRGRmJ4ODgcs+Jjo5G79690aZNG/z555+4c+cOPvnkE1haWtZS1IQQU8OwL24QXQ9FRUWhbdu2iIiIQPfu3QEAERER8PPzw4MHD9C6dWuD540bNw5CoRDfffddlWPIycmBVCqFXC6HjY1Nla9HCKkbDaKmeeXKFUilUm3CBIAePXpAKpXi8uXLBs9Rq9U4fvw4WrVqhYCAADg7O6N79+44cuRILUVNCDFFDWI3ypSUFDg7O+uVOzs7IyUlxeA5aWlpyMvLw+rVq/Hpp59izZo1OHnyJEaPHo1z587hlVdeMXheWFgYwsLC9MoVCkXVXgQhxCSYddJcunQpli1bVu4x169fBwAwDKP3GMuyBsuB0pomAIwYMQIhISEAgM6dO+Py5cvYtm1bmUkzKCgIQUFBeuWa5jkhxLyZddKcNWsWxo0bV+4x7u7uuHv3LlJTU/UeS09Ph4uLi8HzHB0dIRAI0LZtW51yLy8vXLp0qfJBE0LMmlknTUdHRzg6OlZ4nJ+fH+RyOa5du4Zu3boBAK5evQq5XI6ePXsaPEckEqFr1654+PChTvmjR48gk8mqHjwhxCw1iIEgLy8vDB48GNOmTUNERAQiIiIwbdo0DBs2TGfkvE2bNvj555+137///vsIDw/HN998gydPnuCrr77C0aNHMXPmzLp4GYQQE9AgkiYA7N+/Hx06dMCgQYMwaNAgdOzYUW8q0cOHDyGXy7Xfjxo1Ctu2bcPatWvRoUMH7Ny5E4cOHULv3r1rO3xCiIloEPM0TQHN0ySkfmgwNU1CCKkOlDQJIYQDSpqEEMIBJU1CCOGAkiYhhHBASZMQQjigpEkIIRxQ0iSEEA4oaRJCCAeUNAkhhANKmoQQwgElTUII4YCSJiGEcEBJkxBCOKCkSQghHFDSJIQQDihpEkIIB5Q0CSGEA0qahBDCASVNQgjhgJImIYRwQEmTEEI4oKRJCCEcUNIkhBAOKGkSQggHgroOoKFgWRYAkJOTU8eREELKY21tDYZhynyckmYtyc3NBQC89NJLdRwJIaQ8crkcNjY2ZT7OsJoqEKlRarUaSUlJFf4VM1WBgYEIDw+v6zA4M9e4AfON3Vzj1qCapong8Xho1qxZXYdRaUKhsNy/vqbKXOMGzDd2c43bWDQQRAghHFDSJIQQDihpEkIIB5Q0CSGEA0qaxChBQUF1HUKlmGvcgPnGbq5xG4umHBFCCAdU0ySEEA4oaRJCCAeUNAkhhANKmoQQPX379sW8efPqOgyTREmzAZk8eTIYhgHDMBAKhXBxccHAgQOxa9cuqNVq7XHu7u5gGAYHDhzQu0a7du3AMAz27Nmjd3xERITOsfPmzUPfvn1rNXaNQYMGgc/n68X0/LVWr16tU37kyJFaWxcgLS0Nb7/9Ntzc3GBhYYHGjRsjICAAV65cAcD9Z0BqDyXNBmbw4MFITk5GXFwcfvvtN/j7+2Pu3LkYNmwYlEql9riXXnoJu3fv1jk3IiICKSkpaNSokd51LS0tsWjRIpOIPSEhAVeuXMGsWbPw7bffGryWpaUl1qxZg6ysrBqNuSxjxozBnTt3sHfvXjx69Ai//vor+vbti2fPnmmP4fozILWDkmYDo6nVuLq6okuXLvjwww/xyy+/4LffftOpuUyYMAHnz5/Hv//+qy3btWsXJkyYAIFAf52Xt99+GxEREThx4kSdx757924MGzYM77zzDsLDw5Gfn693rQEDBqBx48ZYtWpVjcVbluzsbFy6dAlr1qyBv78/ZDIZunXrhsWLF2Po0KHa47j+DKqbUqnErFmzYGtrCwcHB3z88cfadWG///57+Pr6wtraGo0bN8b48eORlpamPTcrKwsTJkyAk5MTxGIxWrZsqfMHIDExEYGBgbCzs4ODgwNGjBiBuLi4Gn9N1YGSJkG/fv3QqVMnHD58WFvm4uKCgIAA7N27FwBQUFCA8PBwTJ061eA13N3dMWPGDCxevNhgc7mmvBg7y7LYvXs33njjDbRp0watWrXCjz/+qHcen8/HypUrsXnzZjx9+rTW4gUAKysrWFlZ4ciRIyguLi7zOK4/g+q2d+9eCAQCXL16FV9++SU2bNiAnTt3AgBKSkqwYsUK3LlzB0eOHEFsbCwmT56sPfeTTz7B/fv38dtvvyEqKgpbt26Fo6Oj9nX4+/vDysoKFy5cwKVLl2BlZYXBgwejpKSkVl5bVVDSJACANm3a6P2lnzp1Kvbs2QOWZXHw4EE0b94cnTt3LvMaH3/8MWJjY7F///6aDfYFz8f++++/o6CgAAEBAQCAN954o8wm+qhRo9C5c2eEhobWVqgAAIFAgD179mDv3r2wtbVFr1698OGHH+Lu3bt6x3L9GVSnl156CRs2bEDr1q0xYcIEzJ49Gxs2bNDGNWTIEHh6eqJHjx748ssv8dtvvyEvLw9AaReJt7c3fH194e7ujgEDBuC1114DABw4cAA8Hg87d+5Ehw4d4OXlhd27dyMhIQF//vlnrby2qqCkSQCU1tBeHAQZOnQo8vLycOHCBezatavCGo6TkxMWLFiAJUuW1GqN4fnYv/32WwQGBmqbr0FBQbh69SoePnxo8Nw1a9Zg7969uH//fq3FC5T2aSYlJeHXX39FQEAA/vzzT3Tp0kVvcIfrz6A69ejRQ+cz4efnh8ePH0OlUuH27dsYMWIEZDIZrK2ttQN+CQkJAIB33nkHBw4cQOfOnbFw4UJcvnxZe52bN2/iyZMnsLa21ta67e3tUVRUhOjo6Fp7fZVFSZMAAKKiouDh4aFTJhAIEBwcjNDQUFy9ehUTJkyo8Drz589HYWEhtmzZUlOh6tHE/uzZMxw5cgRbtmyBQCCAQCCAq6srlEoldu3aZfDcPn36ICAgAB9++GGtxathaWmJgQMHYsmSJbh8+TImT56sV+utzM+gphUVFWHQoEGwsrLC999/j+vXr+Pnn38GAO0fyyFDhiA+Ph7z5s1DUlIS+vfvjwULFgAo3cXAx8cHkZGROl+PHj3C+PHj6+x1GYuSJsEff/yBv//+G2PGjNF7bOrUqTh//jxGjBgBOzu7Cq9lZWWFTz75BJ999lmtbCL3fOz79+9Hs2bNcOfOHZ1fxo0bN2Lv3r06I+zPW716NY4ePapTG6oLbdu2NThoxfVnUF1enK4VERGBli1b4sGDB8jIyMDq1avx8ssvo02bNjqDQBpOTk6YPHkyvv/+e2zcuBE7duwAAHTp0gWPHz+Gs7MzWrRoofMllUpr5bVVBW130cAUFxcjJSUFKpUKqampOHnyJFatWoVhw4Zh4sSJesd7eXkhIyMDEonE6OeYPn06NmzYgLCwMHTv3r3WYvfx8cHYsWPRvn17nfNkMhkWLVqE48ePY8SIEXrX7dChAyZMmIDNmzdXW6zlyczMxP/93/9h6tSp6NixI6ytrXHjxg2sXbvWYHyV+RlUh3///Rfz58/H22+/jVu3bmHz5s344osv4ObmBpFIhM2bN2PGjBn4559/sGLFCp1zlyxZAh8fH7Rr1w7FxcU4duwYvLy8AJTOCvj8888xYsQILF++HM2aNUNCQgIOHz6M999/3+S3haGk2cCcPHkSTZo0gUAggJ2dHTp16oQvv/wSkyZNAo9nuOHh4ODA6TmEQiFWrFhR7U2t8mK/ffs27ty5g2+++UbvPGtrawwaNAjffvutwaQEACtWrDA4yl4TrKys0L17d2zYsAHR0dFQKBR46aWXMG3atDK7Cbj+DKrDxIkTUVhYiG7duoHP52P27NmYPn26dmL9hx9+iC+//BJdunTBunXrMHz4cO25IpEIixcvRlxcHMRiMV5++WXtRH2JRIILFy5g0aJFGD16NHJzc+Hq6or+/fubxd5CtDQcIYRwQH2ahBDCASVNQgjhgJImIYRwQEmTEEI4oKRJCCEcUNIkhBAOKGkSQggHlDQJIYQDSpqEEMIBJU1CDJg8eTJGjhzJ+bylS5fW2nqXpG5Q0iSEEA4oaZIG7eDBg+jQoQPEYjEcHBwwYMAAvP/++9i7dy9++eUX7Q6YmhXFFy1ahFatWkEikcDT0xOffPIJFAoFAGDPnj1YtmwZ7ty5oz1Ps6iwXC7H9OnT4ezsDBsbG/Tr1w937typo1dNqoJWOSINVnJyMoKCgrB27VqMGjUKubm5uHjxIiZOnIiEhATk5ORoNwOzt7cHULpi0p49e9C0aVP8/fffmDZtGqytrbFw4UIEBgbin3/+wcmTJ/H7778DAKRSKViWxdChQ2Fvb48TJ05AKpVi+/bt6N+/Px49eqS9NjETLCEN1M2bN1kAbFxcnN5jkyZNYkeMGFHhNdauXcv6+Phovw8NDWU7deqkc8zZs2dZGxsbtqioSKe8efPm7Pbt2ysVO6k7VNMkDVanTp3Qv39/dOjQAQEBARg0aBDGjh1b7uroBw8exMaNG/HkyRPk5eVBqVRWuAbkzZs3kZeXp7cmZmFhoVnsiUN0UdIkDRafz8eZM2dw+fJlnD59Gps3b8ZHH32Eq1evGjw+IiIC48aNw7JlyxAQEACpVIoDBw7giy++KPd51Go1mjRpYnCnRVtb22p4JaQ2UdIkDRrDMOjVqxd69eqFJUuWQCaT4eeff4ZIJIJKpdI59q+//oJMJsNHH32kLYuPj9c5xtB5Xbp0QUpKCgQCAdzd3WvstZDaQUmTNFhXr17F2bNnMWjQIDg7O+Pq1atIT0+Hl5cXioqKcOrUKTx8+BAODg6QSqVo0aIFEhIScODAAXTt2hXHjx/X7sKo4e7ujtjYWERGRqJZs2awtrbGgAED4Ofnh5EjR2LNmjVo3bo1kpKScOLECYwcORK+vr519A6QSqnrTlVC6sr9+/fZgIAA1snJibWwsGBbtWrFbt68mWVZlk1LS2MHDhzIWllZsQDYc+fOsSzLsu+//z7r4ODAWllZsYGBgeyGDRtYqVSqvWZRURE7ZswY1tbWlgXA7t69m2VZls3JyWFnz57NNm3alBUKhexLL73ETpgwgU1ISKjlV02qivYIIoQQDmhyOyGEcEBJkxBCOKCkSQghHFDSJIQQDihpEkIIB5Q0CSGEA0qahBDCASVNQgjhgJImIYRwQEmTEEI4oKRJCCEc/D+5KDf6o7OefQAAAABJRU5ErkJggg==\n", 91 | "text/plain": [ 92 | "
" 93 | ] 94 | }, 95 | "metadata": {}, 96 | "output_type": "display_data" 97 | } 98 | ], 99 | "source": [ 100 | "condition=['gradCPTface','gradCPTscene']\n", 101 | "for i, cdt in enumerate(condition):\n", 102 | " # Load HMM-derived latent state sequence\n", 103 | " seq = scipy.io.loadmat(loaddir+'fmri/hmmoutput.mat')[cdt]\n", 104 | " # Load time-resolved attention measures: Inverse of RT variability time course (z-normalized)\n", 105 | " beh = scipy.io.loadmat(loaddir+'behavior/'+cdt+'_beh.mat')['rtvariability_conv'].T*(-1)\n", 106 | " \n", 107 | " # Categorize attention measure of each TR to latent state identification\n", 108 | " stateocc = np.zeros((nsubj,nstate))\n", 109 | " for subj in range(nsubj):\n", 110 | " for stt in range(1,nstate+1):\n", 111 | " stateocc[subj,stt-1]=np.nanmean(beh[subj,np.where(seq[subj,:]==stt)[0]])\n", 112 | " stateocc_real=np.nanmean(stateocc,0)\n", 113 | " \n", 114 | " # to visualize data points\n", 115 | " valus = np.concatenate((stateocc[:,0], stateocc[:,1], stateocc[:,2],stateocc[:,3]),0)\n", 116 | " categ = np.concatenate((np.repeat('state1',nsubj), np.repeat('state2',nsubj), np.repeat('state3',nsubj), np.repeat('state4',nsubj)),0)\n", 117 | " categ = np.delete(categ,np.where(np.isnan(valus))[0])\n", 118 | " valus = np.delete(valus,np.where(np.isnan(valus))[0])\n", 119 | " df = pd.DataFrame(data={'state':categ,'values':valus})\n", 120 | " \n", 121 | " # Chance distribution: circular-shift attention time course 1,000 times and relate them with the latent state sequence\n", 122 | " stateocc_null = []\n", 123 | " for null in range(niter):\n", 124 | " stateocc_niter = np.zeros((nsubj,nstate))\n", 125 | " for subj in range(nsubj):\n", 126 | " beh_null = np.roll(beh[subj,:],np.random.randint(beh[subj,:].shape[0]-20)+10)\n", 127 | " for stt in range(1,nstate+1):\n", 128 | " stateocc_niter[subj,stt-1]=np.nanmean(beh_null[np.where(seq[subj,:]==stt)[0]])\n", 129 | " stateocc_null.append(np.nanmean(stateocc_niter,0))\n", 130 | " stateocc_null=np.array(stateocc_null)\n", 131 | " \n", 132 | " # two-tailed test\n", 133 | " pvals = []\n", 134 | " for stt in range(nstate):\n", 135 | " rl = stateocc_real[stt] - np.mean(stateocc_null[:,stt])\n", 136 | " nl = stateocc_null[:,stt] - np.mean(stateocc_null[:,stt])\n", 137 | " pvals.append((1+np.where(np.abs(rl)<=np.abs(nl))[0].shape[0])/(1+niter))\n", 138 | " qvals=fdrcorrection(np.array(pvals))[1]\n", 139 | " \n", 140 | " # Bar plot\n", 141 | " fig, ax = plt.subplots(1,1, figsize=(3,4))\n", 142 | " ax.bar(np.arange(nstate),np.nanmean(stateocc,0), color=cmap, edgecolor='black',linewidth=0.5)\n", 143 | " ax.errorbar(np.arange(nstate),np.nanmean(stateocc,0),\n", 144 | " np.nanstd(stateocc,0)/np.sqrt(stateocc.shape[0]-np.sum(np.isnan(stateocc[:,0]))),\n", 145 | " capsize=6,ecolor='black',ls='none')\n", 146 | " sns.stripplot(x=\"state\", y=\"values\", data=df, color=[0.3,0.3,0.3],jitter=False,alpha=0.3, size=4)\n", 147 | " plt.hlines(0, -1, 4, colors='black', linewidth=0.5)\n", 148 | " ax.set_ylim([-0.75, 0.65])\n", 149 | " ax.set_xlim([-1,4])\n", 150 | " ax.tick_params(direction='out', length=6, width=0.6)\n", 151 | " ax.set_ylabel('Inverted RT variability (z)\\nat state occurrence')\n", 152 | " ax.set_xticklabels(['DMN','DAN','SM','base'])\n", 153 | " ax.tick_params(length=5, width=0.5)\n", 154 | " ax.spines['right'].set_visible(False)\n", 155 | " ax.spines['top'].set_visible(False)\n", 156 | " ax.set_title(cdt+' (N='+str(stateocc.shape[0]-np.sum(np.isnan(stateocc[:,0])))+')')\n", 157 | " for stt in range(nstate):\n", 158 | " if qvals[stt]<0.01:\n", 159 | " plt.scatter(stt,0.63,30,'k','*')\n", 160 | " plt.show()" 161 | ] 162 | }, 163 | { 164 | "cell_type": "markdown", 165 | "metadata": {}, 166 | "source": [ 167 | "## Movie watching: Latent state dynamics & Self-reported narrative engagement (Figure 5e, 5g)" 168 | ] 169 | }, 170 | { 171 | "cell_type": "code", 172 | "execution_count": 4, 173 | "metadata": {}, 174 | "outputs": [ 175 | { 176 | "name": "stderr", 177 | "output_type": "stream", 178 | "text": [ 179 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/459608050.py:12: RuntimeWarning: Mean of empty slice\n", 180 | " stateocc[subj,stt-1]=np.nanmean(beh[subj,np.where(seq[subj,:]==stt)[0]])\n", 181 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/459608050.py:29: RuntimeWarning: Mean of empty slice\n", 182 | " stateocc_niter[subj,stt-1]=np.nanmean(beh_null[np.where(seq[subj,:]==stt)[0]])\n", 183 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/459608050.py:53: UserWarning: FixedFormatter should only be used together with FixedLocator\n", 184 | " ax.set_xticklabels(['DMN','DAN','SM','base'])\n" 185 | ] 186 | }, 187 | { 188 | "data": { 189 | "image/png": "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\n", 190 | "text/plain": [ 191 | "
" 192 | ] 193 | }, 194 | "metadata": {}, 195 | "output_type": "display_data" 196 | }, 197 | { 198 | "name": "stderr", 199 | "output_type": "stream", 200 | "text": [ 201 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/459608050.py:12: RuntimeWarning: Mean of empty slice\n", 202 | " stateocc[subj,stt-1]=np.nanmean(beh[subj,np.where(seq[subj,:]==stt)[0]])\n", 203 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/459608050.py:29: RuntimeWarning: Mean of empty slice\n", 204 | " stateocc_niter[subj,stt-1]=np.nanmean(beh_null[np.where(seq[subj,:]==stt)[0]])\n", 205 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/459608050.py:53: UserWarning: FixedFormatter should only be used together with FixedLocator\n", 206 | " ax.set_xticklabels(['DMN','DAN','SM','base'])\n" 207 | ] 208 | }, 209 | { 210 | "data": { 211 | "image/png": "iVBORw0KGgoAAAANSUhEUgAAAU0AAAGJCAYAAAAOiWFmAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAA9hAAAPYQGoP6dpAABZIklEQVR4nO3dd1xT1/sH8M8lCRAIhA2KGNygYlFwoNaJoLXFOqqiRanW1bqrVWvde1VtrXWv1qKtba39uutGAQeirbiKjIpgQDFsCMn9/WGTnzEBciMjgef9evHSnNxz8yTAw7nnnHsOw7IsC0IIIXoxq+4ACCHElFDSJIQQDihpEkIIB5Q0CSGEA0qahBDCASVNQgjhgJImIYRwQEmTEEI4oKRJCCEcUNIknIWHh8PT01OjbPny5Th8+HC1xFOdsrOzsWzZMnTr1g1ubm4QiUTw8fHBqlWrUFhYqPd5EhISYGFhgaioKHVZeHg4GIZBixYtoFAotOowDIOJEydWyPv48ssv8e6778Ld3R0MwyA8PLzUY3/55Rd06tQJDg4OsLOzQ7t27fD9999rHJOVlQU7O7sa+TNBSZNwNm/ePPz2228aZbU1aaakpGDDhg1o06YNtm3bhiNHjmDQoEFYuHAh3n33Xeh7l/KMGTPQq1cvBAQEaD0XHx+PPXv2VHDkmtavX49nz54hJCQE5ubmpR63a9cuDBo0CHXq1MH+/ftx4MABNGrUCCNGjMD69evVx9nb22PatGmYOXMmiouLKzX2KscSUgGsra3ZkSNHVncYVS43N5fNzc3VKl+zZg0LgL106VK554iPj2cBsCdOnNAoHzlyJGttbc2+/fbbrLu7O5ufn6/xPAD2008/fbM38B+FQqH+f1nfy06dOrESiUTjeKVSyXp5ebGtWrXSODY9PZ3l8/ns/v37KyRGY0EtTaIhIyMDY8eOhYeHBywsLODs7IxOnTrhzz//VB/z+uU5wzDIy8vD3r17wTAMGIZBt27d1M+npqaqz2lubo66deti0KBBePr0qfqYlJQUfPjhh3BxcYGFhQW8vb2xbt06KJVK9TFJSUlgGAZr1qzBqlWr4OnpCaFQiG7duuHBgweQy+WYPXs26tatC7FYjP79+0MqlWq9x4MHDyIgIADW1tYQiUQIDg7GzZs3NY4JDw+HSCTCnTt30LNnT1hbW8PZ2RkTJ05Efn6++jhra2tYW1trvUa7du0AAP/++2+5n/l3330HNzc39OrVS+fzq1atQmpqKjZu3FjuuQxlZqZfKhAIBBCJRBrHMwwDW1tbWFpaahzr6uqKXr16YcuWLRUaa3WjpEk0hIWF4fDhw5g/fz5OnTqFHTt2IDAwEM+ePSu1TlRUFIRCId555x1ERUUhKioKmzdvBvAyYbZt2xa//fYbpk+fjuPHj2PDhg0Qi8XIysoC8DJRd+zYEadOncKSJUtw5MgRBAYGYsaMGTr77L799ltcvnwZ3377LXbs2IF79+7hvffew+jRo5GRkYFdu3Zh9erV+PPPP/Hxxx9r1F2+fDlCQ0PRvHlz/PTTT/j++++Rk5ODt99+G/Hx8RrHyuVyvPPOO+jZsycOHz6MiRMnYuvWrRgyZEi5n+PZs2cBAC1atCj32KNHj6JLly6lJq6AgAD0798fq1atwvPnz8s8V0lJiV5frIGLm02aNAl3797FsmXLkJGRgczMTKxduxY3btzAjBkztI7v1q0bLl++jBcvXhj0ekapupu6xLiIRCJ26tSpZR4zcuRIViKRaJSVdkk3atQoViAQsPHx8aWeb/bs2SwANiYmRqN8woQJLMMw7P3791mWZdnExEQWAPvWW29pXB5u2LCBBcCGhIRo1J86dSoLgJXJZCzLsmxKSgrL5/PZSZMmaRyXk5PDurm5sYMHD9Z4jwDYjRs3ahy7bNkyFgAbGRlZ6vu5desWKxQK2f79+5d6jMrTp09ZAOzKlSu1nlNdnrMsy967d4/l8XjsZ599pn4eOi7PAej1tXv37lJjKq+r5fDhw6xYLFafSygUsj/88IPOY0+fPs0CYI8fP17Gp2Ba+FWaoYnRa9euHfbs2QNHR0cEBgbCz88PAoHA4PMdP34c3bt3h7e3d6nHnD17Fs2bN1df0qqEh4fju+++w9mzZ9G0aVN1+TvvvKPRKlOdu2/fvhr1VeUpKSlo2bIlTp48iZKSEowYMQIlJSXq4ywtLdG1a1ecO3dOK7bhw4drPB42bBjmzp2Lc+fOoVOnTlrHJyUl4d1334WHhwd27NhR6ntWefLkCQDAxcWlzOOaNWuG0aNHY9OmTZg8eTLq16+v87hr166V+5oA0KBBA72Oe92JEyfw4Ycf4oMPPsDgwYPB5/Nx5MgRhIeHo7i4GB999JHG8ar3lZqaatDrGSNKmkTDwYMHsXTpUuzYsQPz5s2DSCRC//79sXr1ari5uXE+X0ZGBurVq1fmMc+ePdOawgQAdevWVT//KgcHB43HqtHe0spVU39Ufaht27bVGcfrl8d8Ph+Ojo4aZarPQFd3RXJyMrp37w4+n48zZ85oxaNLQUEBAGj1B+qycOFC/PDDD5g3bx727t2r8xhfX99yzwMAPB5Pr+NexbIsRo0ahS5dumDXrl3q8sDAQMhkMkyaNAmDBw/W6ONVvS/V+6wJqE+TaHBycsKGDRuQlJSE5ORkrFixAr/++muZ8/bK4uzsjMePH5d5jKOjI9LS0rTKVa0wJycng177darzHDp0CNeuXdP6iomJ0Ti+pKREKzmmp6erY35VcnIyunXrBpZlce7cuXL/ULweU3l9lQBQp04dTJ06FT/88ANu376t8xiBQKDXV2lJtyxPnz5FWlqa1hUB8PIPUV5eHpKSkjTKVe+ror6HxoBamqRU9evXx8SJE3HmzBlcvny5zGMtLCx0tib69OmD77//Hvfv30ezZs101u3ZsydWrFiB2NhYtGnTRl2+b98+MAyD7t27v9kb+U9wcDD4fD4SEhIwcOBAvers378fkydPVj/+8ccfAUBjdkBKSgq6desGhUKB8+fPQyKR6B2TRCKBUChEQkKCXsfPmjUL27Ztw+zZs3U+X5mX5/b29rC0tER0dLTWc1FRUTAzM0OdOnU0yh89egQAaN68OefXM1aUNImaTCZD9+7dMWzYMHh5ecHGxgbXrl3DiRMnMGDAgDLr+vj44Pz58/jjjz9Qp04d2NjYoFmzZli8eDGOHz+OLl264IsvvoCPjw9evHiBEydOYPr06fDy8sK0adOwb98+9O3bF4sXL4ZEIsHRo0exefNmTJgwQaM/8014enpi8eLFmDt3Lh49eoTevXvD3t4eT58+xdWrV2FtbY1Fixapjzc3N8e6deuQm5uLtm3b4sqVK1i6dCn69OmDzp07AwCkUim6d++OtLQ07Ny5E1KpVGOaU7169cpsdZqbmyMgIEBnItLF1tYWc+fOxbRp03Q+7+/vr9d5XnfhwgVkZGQAABQKBZKTk3Ho0CEAQNeuXeHs7AwLCwt88skn+OqrrzBixAgMGTIEPB4Phw8fxo8//ojRo0drdUlER0fD0dERPj4+BsVllKp7JIoYj8LCQnb8+PFsq1atWFtbW1YoFLLNmjVjFyxYwObl5amP0zV6HhcXx3bq1Im1srJiAbBdu3ZVP/fvv/+yo0aNYt3c3FiBQMDWrVuXHTx4MPv06VP1McnJyeywYcNYR0dHViAQsM2aNWPXrFmjMUquGj1fs2aNxmufO3eOBcD+/PPPGuW7d+9mAbDXrl3TKD98+DDbvXt31tbWlrWwsGAlEgk7aNAg9s8//9R4j9bW1uzt27fZbt26sUKhkHVwcGAnTJigMZld9dqlfS1YsKDcz33nzp0sj8djnzx5olH+6uj5q4qKitgGDRpU6OT2rl27lvoezp07pz5OoVCw27dvZ/39/Vk7OzvW1taWbd26Nbtp0ya2uLhY45xKpZKVSCRasxVMHcOytBslIa8LDw/HoUOHkJubW+mvVVhYiPr16+Ozzz7DrFmzKv31qsqZM2cQFBSEO3fuwMvLq7rDqTA0EERINbO0tMSiRYvw1VdfIS8vr7rDqTBLly7FqFGjalTCBKhPkxCjMHbsWLx48QKPHj2qEf1/WVlZ6Nq1Kz755JPqDqXC0eU5IYRwQJfnhBDCASVNQgjhgJImIYRwQEmzirAsi+zsbIOX5CKEGAdKmlUkJycHYrEYOTk51R0KZw8fPkS9evXw8OHD6g6l1jDVz9xU4+aCkiYp148//ojU1FRERERUdyi1hql+5qYaNxc0T5PodPv2bfUWEPv27VP/q1qMonXr1mjVqlW1xVcTmepnbqpxG6w67+GsTWQymcYq4sZuwIABZd5TPWDAgOoOscYx1c/cVOM2FCXNKmJqSTM7O5v98MMPdf4ShIWFsdnZ2dUdYo1jqp+5qcZtKLojqIpkZ2dDLBZDJpPB1ta2usPRW7NmzfDgwQONx/fu3avGiGo+U/3MTTVurmggiJQqOTlZ/UtgZWUFALh//z5SUlKqM6wazVQ/c1ON2xC1Lmlu3rwZDRo0gKWlJfz8/HDp0qUyjy8qKsLcuXMhkUhgYWGBRo0aaeyPUpOpVmv//PPPIZVKMXPmTABAZGRkdYZVo5nqZ26qcRukuvsHqtKBAwdYgUDAbt++nY2Pj2enTJnCWltbs8nJyaXWCQkJYdu3b8+ePn2aTUxMZGNiYtjLly9zfm1T69Nk2ZeL3d67d0+j7N69e1qLzZKKY6qfuanGbYha1afZvn17tGnTBt999526zNvbG++//z5WrFihdfyJEycwdOhQPHr0SK+dBctiqn2ahBBNtebyvLi4GDdu3EBQUJBGeVBQEK5cuaKzzpEjR+Dv74/Vq1fD3d0dTZs2xYwZM2rUdqSEEG5qzeT2zMxMKBQKuLq6apS7urqqt2V93aNHjxAZGQlLS0v89ttvyMzMxCeffILnz5+X2q8ZERGh824IuVz+5m+CEFLtak3SVGEYRuMxy7JaZSpKpRIMw2D//v0Qi8UAgK+++gqDBg3Ct99+C6FQqFUnNDQUoaGhWuWqy3NCiGmrNZfnTk5O4PF4Wq1KqVSq1fpUqVOnDtzd3TWSnbe3N1iWxePHjys1XkKIcao1SdPc3Bx+fn44ffq0Rvnp06fRsWNHnXU6deqEJ0+eaOxI+ODBA5iZmZW5lzUhpOaqNUkTAKZPn44dO3Zg165duHv3LqZNm4aUlBSMHz8eADBnzhyMGDFCffywYcPg6OiIjz76CPHx8bh48SJmzpyJUaNG6bw0J4TUfLWqT3PIkCF49uwZFi9ejLS0NLRs2RLHjh1Tr8aSlpamcQeDSCTC6dOnMWnSJPj7+8PR0RGDBw/G0qVLq+stEEKqWa2ap1mdaJ4mITVDrbo8J4SQN0VJkxBCOKCkSQghHFDSJIQQDihpEkIIB5Q0CSGEA0qahBDCASVNQgjhgJImIYRwQEmTEEI4oKRJCCEcUNIkhBAOKGkSQggHlDQJIYQDSpqEEMIBJU1CCOGAkiYhhHBASZMQQjigpEkIIRxQ0iSEEA4oaRJCCAeUNAkhhANKmoQQwgElTUII4YCSJiGEcEBJkxBCOKCkSQghHFDSJIQQDihpEkIIB5Q0CSGEA0qahBDCASVNQgjhgJImIYRwUOuS5ubNm9GgQQNYWlrCz88Ply5d0qve5cuXwefz4evrW7kBEkKMWq1KmgcPHsTUqVMxd+5c3Lx5E2+//Tb69OmDlJSUMuvJZDKMGDECPXv2rKJICSHGimFZlq3uIKpK+/bt0aZNG3z33XfqMm9vb7z//vtYsWJFqfWGDh2KJk2agMfj4fDhw4iLi+P82tnZ2RCLxZDJZLC1tTUkfEKIEag1Lc3i4mLcuHEDQUFBGuVBQUG4cuVKqfV2796NhIQELFiwoLJDJISYAD7XCklJSbh06RKSkpKQn58PZ2dntG7dGgEBAbC0tKyMGCtEZmYmFAoFXF1dNcpdXV2Rnp6us87Dhw8xe/ZsXLp0CXy+fh9VREQEIiIitMrlcjn3oAkhRkfvpPnjjz/i66+/xtWrV+Hi4gJ3d3cIhUI8f/4cCQkJsLS0xPDhwzFr1ixIJJLKjPmNMAyj8ZhlWa0yAFAoFBg2bBgWLVqEpk2b6n3+0NBQhIaGapWrLs8JIaZNr6TZpk0bmJmZITw8HD/99BPq16+v8XxRURGioqJw4MAB+Pv7Y/Pmzfjggw8qJWBDOTk5gcfjabUqpVKpVusTAHJycnD9+nXcvHkTEydOBAAolUqwLAs+n49Tp06hR48eVRI7IcR46DUQdPToUfTt21evE2ZmZiIxMRFt27Z94+AqWvv27eHn54fNmzery5o3b45+/fppDQQplUrEx8drlG3evBlnz57FoUOH0KBBA1hbW+v92jQQREjNoFdLU9+ECbxs0Tk5ORkcUGWaPn06wsLC4O/vj4CAAGzbtg0pKSkYP348AGDOnDlITU3Fvn37YGZmhpYtW2rUd3FxgaWlpVY5IaT24Dx63qNHDyxatEirPCsry+gvV4cMGYINGzZg8eLF8PX1xcWLF3Hs2DF1H2xaWlq5czYJIbUb53maZmZmcHR0RKdOnbB//371JerTp09Rt25dKBSKSgnU1NHlOSE1g0HzNP/880+kp6ejQ4cOSEpKquCQCCHEeBmUNOvUqYMLFy6gVatWaNu2Lc6fP1/BYRFCiHHinDRVcxotLCywf/9+TJkyBb1799YYkSaEkJqK8x1Br3eBfvnll/D29sbIkSMrLChCCDFWnJNmYmIinJ2dNcoGDhwILy8vXL9+vcICI4QQY1SrVjmqTjR6TmoDuVyOgoICCIVCCASC6g6nUujVpzl+/Hj8+++/ep3w4MGD2L9//xsFRQgxPVKpFHFxcbh37x7i4uIglUqrO6RKodflubOzM1q2bImOHTsiJCQE/v7+qFu3LiwtLZGVlYX4+HhERkbiwIEDcHd3x7Zt2yo7bkKIEZHL5UhOTlaPebAsi+TkZNjb29e4Fqfel+dSqRQ7d+7EgQMH8Pfff2s8Z2Njg8DAQIwdO1ZrvUryEl2ek5osOzsb9+7d0yr38vKqcT/vBvVpvnjxAsnJySgoKICTkxMaNWqkc3k18v8oaZKaTC6XIy4uTmN2DcMw8PX1rXEtTc6j5wBgZ2cHOzu7Cg6FEGKqBAIBJBKJ+hKdYRhIJJIalzABA5MmIYS8zsXFBfb29jV+9JySJiGkwggEghqbLFVqzcZqhBBSEShpEkIIBwYtQvzixQut8uzsbKNfhJgQQt4U56R5/vx5FBcXa5UXFhbi0qVLFRIUIYQYK70Hgm7fvq3+f3x8vMaujgqFAidOnIC7u3vFRkcIIUZG76Tp6+sLhmHAMIzOy3ChUIhvvvmmQoMjhBBjo3fSTExMBMuyaNiwIa5evaqxPJy5uTlcXFzA4/EqJUhCCDEWeidN1Y6NSqWy0oIhhJi22rA0nEGT2x88eIDz589DKpVqJdH58+dXSGCEENMilUq1bqN0cXGp7rAqHOekuX37dkyYMAFOTk5wc3PTWKiDYRhKmoTUQrVpaTjOSXPp0qVYtmwZZs2aVRnxEEJMUEFBgdb+YSzLoqCgoMYlTc7zNLOysvDBBx9URiyEEBMlFAq1lodkGAZCobCaIqo8nJPmBx98gFOnTlVGLIQQE6VaGk6VOGlpuFc0btwY8+bNQ3R0NHx8fLQ+lMmTJ+t1nn/++QcJCQno0qULhEKhuvOYEGKaasvScJxXbm/QoEHpJ2MYPHr0qMz6z549w5AhQ3D27FkwDIOHDx+iYcOGGD16NOzs7LBu3Tou4ZgMWrmd1AY05UiHxMTEN3rBadOmgc/nIyUlBd7e3uryIUOGYNq0aTU2aRJS00mlUjx69AiFhYWwtLREw4YNacrRq4qLi5GYmIhGjRqBz9f/NKdOncLJkydRr149jfImTZogOTnZ0HAIIdVILpfj1q1bkEql6q62nJwcdOvWrca1ODkPBOXn52P06NGwsrJCixYtkJKSAuBlX+bKlSvLrZ+XlwcrKyut8szMTFhYWHANhxBiBHJyctQJE3g53UgqlSInJ6eaI6t4nJPmnDlzcOvWLZw/fx6Wlpbq8sDAQBw8eLDc+l26dMG+ffvUjxmGgVKpxJo1a9C9e3eu4RBCjEBpQyMGbHZr9Dhfnh8+fBgHDx5Ehw4dNEa7mzdvjoSEhHLrr1mzBt26dcP169dRXFyMzz//HHfu3MHz589x+fJlruEQQoyAra0tXF1d8fTpU/Xluaura40c9OTc0szIyNDZuZuXl6fXlKHmzZvj9u3baNeuHXr16oW8vDwMGDAAN2/eRKNGjbiGw9nmzZvRoEEDWFpaws/Pr8yFk3/99Vf06tULzs7OsLW1RUBAAE6ePFnpMRJiagQCAVq1aoVGjRqhXr16aNSoEVq1alXj+jMBACxHXbp0Yb/++muWZVlWJBKxjx49YlmWZT/99FM2ODiY6+mq1IEDB1iBQMBu376djY+PZ6dMmcJaW1uzycnJOo+fMmUKu2rVKvbq1avsgwcP2Dlz5rACgYCNjY3l/NoymYwFwMpksjd9G4QYreLiYlYmk7HFxcXVHUql4Zw0L1++zNrY2LDjx49nLS0t2SlTprCBgYGstbU1e/369XLr79q1i/3pp5+0yn/66Sd2z549XMPhpF27duz48eM1yry8vNjZs2frfY7mzZuzixYt4vzalDQJqRk4X5537NgRly9fRn5+Pho1aoRTp07B1dUVUVFR8PPzK7f+ypUr4eTkpFXu4uKC5cuXcw1Hb8XFxbhx4waCgoI0yoOCgnDlyhW9zqFUKpGTkwMHB4fKCJEQYgIMmqfp4+ODvXv3GvSCycnJOu8qkkgk6ulLlSEzMxMKhQKurq4a5a6urhr7HZVl3bp1yMvLw+DBg0s9JiIiAhEREVrlcrmcW8DkjdWGu1NI1TN4crtUKtW5CHGrVq3KrOfi4oLbt2/D09NTo/zWrVtwdHQ0NBy9vT5Yxep5z3tERAQWLlyI33//vcy7HEJDQxEaGqpVrrqNklSN2rIgLql6nJPmjRs3MHLkSNy9e1drDhbDMFAoFGXWHzp0KCZPngwbGxt06dIFAHDhwgVMmTIFQ4cO5RqO3pycnMDj8bRalVKpVKv1+bqDBw9i9OjR+PnnnxEYGFhpMZKKUZsWxCVVj3PS/Oijj9C0aVPs3LkTrq6unFcmWrp0KZKTk9GzZ0/17ZdKpRIjRoyo1D5Nc3Nz+Pn54fTp0+jfv7+6/PTp0+jXr1+p9SIiIjBq1ChERESgb9++lRYfqTi1aUFcY1MbukQ4r3JkY2ODmzdvonHjxm/0wg8ePMCtW7cgFArh4+Oj3ritMh08eBBhYWHYsmULAgICsG3bNmzfvh137tyBRCLBnDlzkJqaqr5jKSIiAiNGjMDGjRsxYMAA9XmEQiHnS21a5ajqyOVyxMXFaSROhmHg6+tbY3+RjUFt6RLh3NLs2bMnbt269cZJs2nTpmjatOkbnYOrIUOG4NmzZ1i8eDHS0tLQsmVLHDt2TJ2w09LSNAajtm7dipKSEnz66af49NNP1eUjR47Enj17qjR2oj/Vgriv/wJTwqw8qi6RwsJC5ObmQiQS1dguEc4tzczMTIwcORLt2rVDy5YttT6QkJCQMusrFArs2bMHZ86c0TmQdPbsWS7hmAxqaVa92nCpaCyys7Nx6tQpJCYmqv9QNWjQAEFBQTXu551zS/PKlSuIjIzE8ePHtZ7TZyBoypQp2LNnD/r27YuWLVvSau2k0ggEAkqWVSgpKUlj8C0pKal6A6oknJPm5MmTERYWhnnz5pU76qzLgQMH8NNPP+Gdd97hXJcQYpxycnJgY2MDmUymLrOxsUFOTg61NJ89e4Zp06YZlDCBl6PYb9ofSggxLvb29rCxsYGFhYW6S8Tc3Bz29vbVHVqF43wb5YABA3Du3DmDX/Czzz7Dxo0ba+Q6e4TUVlZWVvD09ERmZiZycnKQmZkJT09PnQuOmzrOLc2mTZtizpw5iIyMNGg3ysjISJw7dw7Hjx9HixYttOr/+uuvXEMihFQzuVwOgUAAf39/PH/+HA4ODhAIBOrymqTKd6P86KOPynx+9+7dXMIxGTR6Tmqy7OxsREdHIz09XZ0o3dzc0KFDhxr38845aRLDUNKsejTlqHKkpaUhLS1No6yoqAhnz55Fbm6uukwkEqFHjx5ae3/VqVMHderUqZJYK4PBSdPQ3SgBoKSkBOfPn0dCQgKGDRsGGxsbPHnyBLa2thCJRIaEY/QoaVat2nJ3SnVYuHAhFi1apFFmYWGBxo0bw9zcHAzDgGVZFBcX459//kFRUZHGsQsWLMDChQurMOKKxTlp5ufnY9KkSeql4R48eICGDRti8uTJqFu3LmbPnl1m/eTkZPTu3RspKSkoKipS1586dSoKCwuxZcsWw9+NEaOkWXXoNsrKpaulqbo837NnD3g8HoYPHw5zc3P4+/tr/bybekuT80DQq7tR9u7dW10eGBiIBQsWlJs0p0yZAn9/f62l4Pr374+PP/6YaziEaKkJC3bk5+cjKysL9vb2RjcCrSvpyeVy5OXlgWVZKBQK1KlTBx4eHujUqZPJfOb6qvLdKCMjI3H58mWYm5trlEskEqSmpnINhxAtQqFQfYmowjAMhEJhNUalv7t37+LGjRsoKiqChYUF/Pz84O3tXd1hlUkgEKBFixbIzs6Gubk5JBKJztusawLOSfNNd6NUKpU6b7V8/PgxbGxsuIZDiBZTXrAjPz8fly5dwosXLzTKJBKJ0bU4y1KTx5c5T25v27Ytjh49qn6sSpTbt29HQEBAufV79eqFDRs2aNTPzc3FggUL6NZKUmFcXFzg6+sLLy8v+Pr6mswgUEZGhkbCBIAXL14gIyOjegLSk1wux99//w1bW1sIhUKkpKTg1q1bNXKbF84tzRUrVqB3796Ij49HSUkJNm7ciDt37iAqKgoXLlwot/5XX32FHj16oHnz5igsLMSwYcPw8OFDODk56dxbh1Q/U526Y4oLdlhaWursWrC0tKz01w4LDYIsK9nA2jwoGLG6EXXop23gMQy+WfcCQNmL+JRFbC/B9xGnDK5fGTgnTdVulGvXrlXvRtmmTRtERUXBx8en3Pru7u6Ii4vDgQMHcOPGDSiVSowePRrDhw83mT6n2sSUp+6YYrJ3cHBAo0aNkJCQoP7MGzVqVCU7oMqyknFkwwOD6j6TmePIJU+s2vEyaQ7qmQVzAYveHZLgKC42OKaQqQZXrTRVuhulXC5Hs2bN8L///Q8fffRRuXcHkeplynvtmGqyFwgE6Ny5M1xdXZGdnQ1bW1s0adLE6D9vW2s5XOzzwbJW/7U2Wbg65MPWmi7PkZ2drbOcYRhYWFhojYq/SiAQoKioiNbQNBGmOnXHlJM98LI/1t7e3qRayQI+ixYNnkEmk4PP56NhnRJ4N8yGgF/zBoQ4DwTZ2dnB3t5e68vOzg5CoRASiQQLFizQWpFdZdKkSVi1ahVKSkreOHhSuVRTd15lClN3ykr2pkIgEMDW1tYkEqaKs10BpFIpMjIy4NtEChd70/m8ueDc0tyzZw/mzp2L8PBwtGvXDizL4tq1a9i7dy++/PJLZGRkYO3atbCwsMAXX3yhVT8mJgZnzpzBqVOn4OPjA2tra43naZUj42GqU3dMfZ6msUuTAmk6BvMLCqG+ffLvh4CwlLGrOs5AHePvKSkV56S5d+9erFu3DoMHD1aXhYSEwMfHB1u3bsWZM2dQv359LFu2TGfStLOzw8CBA98salJlXFxcIBKJjPbuFF1MNdmbiq0/AYu+LfuYzh+W/tyCT4GFEys2pqrEOWlGRUXpvD+8devWiIqKAgB07txZY1dHlZKSEnTr1g3BwcFwc3MzIFxS1V4dUHny5InJDKiYYr+gqRg3GAjprvu5EgWDIrkAFgI5+Dzd/Zl1nCsxuCrAOWnWq1cPO3fuxMqVKzXKd+7cCQ8PDwAvt8TQtcw9n8/HhAkTcPfuXQPDJVXJ1AdUTHGepimo46L78lqaJURyug0s+AwYhkU9t5wa2a/JOWmuXbsWH3zwAY4fP462bduCYRhcu3YN9+7dw6FDhwAA165dw5AhQ3TWb9++PW7evKnea5wYL1MdPSdVT17CIDndBiz7cuCQZV8+trcprHEj6JyTZkhICB48eIAtW7bg/v37YFkWffr0weHDh+Hp6QkAmDBhQqn1P/nkE3z22Wd4/Pgx/Pz8tAaCWrVqxTUkUklMfUDFFCe3m6qCIgFYloFCwaBIzoOFQAEe72W5gG/45HZjZNDkdolEghUrVhj0gqoW6Kt7Cal+MfXZN51UHdWAysOHD01qojVgupPbTZXQQg5ZngDS51ZgwYABCxeHfAgtaHK7Wn5+PlJSUlBcrPlXpLyWYmJioqEvSarBs2fPkJiYiKKiIjx79gwODg5Gn3xMvS/WVL1+y0pNvYXFoKXhPvroIxw/flzn8+W1FKkv03Tk5+fj5s2bAKDe50XVH23MU49qQl+sqXUtFBQJYGsth5WlDMVyPswFJeDT5flLU6dORVZWFqKjo9G9e3f89ttvePr0KZYuXYp169aVW3/fvn1lPj9ixAiuIZFKkpWVpXVnl1KpRFZWllEnTVVfrFwuR3FxMczNzSEQCEymL9YUuxaEFnIwDAs+jwGf9/JuP4Zh6fIcAM6ePYvff/8dbdu2hZmZGSQSCXr16gVbW1usWLECffv2LbP+lClTNB7L5XLk5+fD3NwcVlZWlDSNiL29PczMzDQSp5mZmc7pZMZEIBBAJBLh5s2bUCqVMDMzQ+vWrU2ixWaqXQsCPguJW456BJ1hXj6uaSPngAH3nufl5an/6jk4OKgXR/Xx8UFsbGy59bOysjS+cnNzcf/+fXTu3JnW0zQyVlZWaN26NQCodxRs3bq1UbcygZeJJzc3Fx4eHnB0dISHhwdyc3NNYkFcU75v3sW+AL5NMuAlyYJvk4waOUcTMKCl2axZM9y/fx+enp7w9fXF1q1b4enpiS1bthi8w1yTJk2wcuVKfPjhh7h3755B5yCVw9HREQ0aNFCPnr+6GZ6xKigoQFZWFtLT0yGXyyEQCODm5mYSfZqmPs1LwGdrXB/m6wzq01Rt37lgwQIEBwdj//79MDc3x549ewwOhMfj4cmTJwbXJxVPdakoEAjUydIULhX5fD6Sk5ORlZWlLisqKkLnzp2rMSr90H3zxo9z0hw+fLj6/61bt0ZSUhLu3buH+vXrw8nJqdz6R44c0XjMsizS0tKwadMmdOrUiWs4pBKZwih0aXtwS6VSZGdnq/s05XI5rl69ahJ7cNN988bN4HmaKlZWVmjTpo3ex7///vsajxmGgbOzM3r06KHX6Pub2rx5M9asWYO0tDS0aNECGzZswNtvv13q8RcuXMD06dNx584d1K1bF59//jnGjx9f6XEaA1O4VNy6dSsWLVqkUWZubo7GjRvD0tJSPZBVWFiIxYsXa80rXrBgARYuXFiFEeuH7ps3XpyT5vTp03WWqzZ/aty4Mfr161fqnialLU5cFQ4ePIipU6di8+bN6NSpE7Zu3Yo+ffogPj4e9evX1zo+MTER77zzDsaMGYMffvgBly9fxieffAJnZ+dasbyd6lLx0aNHKCwshKWlJRo2bGhUv8zjxo1DSEiIRllRURFOnTqF33//HSzLokePHrCzs0OPHj3U801VjK2VSYwf56R58+ZNxMbGQqFQoFmzZmBZFg8fPgSPx4OXlxc2b96Mzz77DJGRkWjevHllxGywr776CqNHj8bHH38MANiwYQNOnjyJ7777TudtoVu2bEH9+vXVWw57e3vj+vXrWLt2ba1ImqZA1+V1dnY2EhMTwTAM+Hw+bGxs0KxZM7Ro0ULr8pwQrjhPOerXrx8CAwPx5MkT3LhxA7GxsUhNTUWvXr0QGhqK1NRUdOnSBdOmTdNZf9CgQVrLygHAmjVr8MEHH3B/B3oqLi7GjRs3EBQUpFEeFBSEK1eu6KwTFRWldXxwcDCuX79uEtNX3pRqIMjMzAxWVlYwMzNDcnKy0b93Pp+PtLQ0WFhYQCAQQCaT4fHjx+Dz37g3ihDuLc01a9bg9OnTGn+xbW1tsXDhQgQFBWHKlCmYP3++VrJRuXDhAhYsWKBV3rt3b6xdu5ZrOHrLzMyEQqGAq6urRrmrqyvS09N11klPT9d5fElJCTIzM3Ve2kVEROicb6pKNEOGDKmyy9sbMTGQFxo+V47PF0Ak1m6Z5cqyUVJiWOIUWArh1759ucfFRN1AYYFhr2FuYY6mzRpAIBCAYRjExsYhJvoqli1Zi+Iiw6fDWAoFaB/gV+Yx0devoqCo0ODXUGEYBnweDyUKhdZgHFdCC0t08G9X7nGxfz2HWxfrco+rSnzBc63ul8r0+kC1LpyTpkwmg1Qq1br0zsjIUO9UaWdnp9XhrpKbm6tzx0qBQFDqTpcV6fWNwlTTOrgcr6tcJTQ0FKGhoVrl2dnZEIvFOHjwYJVdIvZu/RY2ORk+aFMC4IGZOV79lWUANFUWGzyCODGzQK8fzL5BoUhNzir3OF34AsBKyEN+4YuXA1cW1hAILSGp0xIG5noAgLvEHkeOlH0DhndXf6SNbGL4iwCwyFdC9EIBsxJAyQdy7XgosuJ8UajmuPehXp850Q/nn/1+/fph1KhRWLdunXoR4qtXr2LGjBnqkfGrV6+iadOmOuu3bNkSBw8exPz58zXKDxw4UKl9oE5OTuDxeFqtSqlUqtWaVHFzc9N5PJ/PN4lJ3m+KD8CNLUEqw0cxGJiDhTtb8uZTLvRw9JThd4elp6djy5YtOHXqFADAt01z2NraYsyYMUa/zQqjZGGXoYBFnhIvdw8H+HJA6sGANaup6waZFs4//1u3bsW0adMwdOhQ9Ta8fD4fI0eOxPr16wEAXl5e2LFjh8768+bNw8CBA5GQkIAePXoAAM6cOYOIiAj8/PPPhr6Pcpmbm8PPzw+nT59G//791eWnT59Gv379dNYJCAjAH3/8oVF26tQp+Pv7G9UIclUwpTuIHR0d4eXlhSNHjoDH48HOzg7e3t4m8YeOX8SqEybwsmVvkacEv4iFXEhJ0xhwTpoikQjbt2/H+vXr8ejRI7Asi0aNGkEkEqmP8fX1LbV+SEgIDh8+jOXLl+PQoUMQCoVo1aoV/vzzT3Tt2tWgN6Gv6dOnIywsDP7+/ggICMC2bduQkpKinnc5Z84cpKamqldiGj9+PDZt2oTp06djzJgxiIqKws6dO2vNPfIlANIZPswACP9Lm+kMH7as4ZfnFU3X5HbgZT97WloazM3N4ebmBjs7O/z1119axxnj5HZG1cR89TExGgb/7ItEIoO3pujbt2+5qyFVhiFDhuDZs2dYvHgx0tLS0LJlSxw7dky9xmdaWprGLpoNGjTAsWPHMG3aNHz77beoW7cuvv7661oz3agQDFgACgByMBCABe+/cpGRtD11TW5XYRgGAoEAY8aMKXUwxdgmt5dYMCi0MoNlvvJl4mSAQiszlFhQ5jQWDMtxaC4vLw8rV67EmTNnIJVKtSarP3r0qMz6165dg1KpRPvXRlBjYmLA4/Hg7+/PJRyToRoIkslkJjUQdMPMAs/NeOoyB6UCfsqiNxoIOnHzlsExva60lqa+KrqlWVEDQdYyJXhyFgoBgzyx2RsNBNXZ+xB3L1x/o5jI/+P8s//xxx/jwoULCAsLQ506dcocedbl008/xeeff66VNFNTU7Fq1SrExMRwDYlUJgZQ4v9bmsa2h4ExXl6/qSIrMxRbMuDJAYUANABkZDgnzePHj+Po0aMGL64RHx+v81711q1bIz4+3qBzksrx8vKcUSdNHliwYIzq8rymYs0YlFiUfxypepzb/Pb29qXeV64PCwsLPH36VKs8LS2N7tgwMnywSGN4yGJ4KGAYZDE8pDE88ClhklLI5XJkZ2cb/V1jb4Jz0lyyZAnmz5+P/Px8g16wV69emDNnDmQymbrsxYsX+OKLL9CrVy+DzkkqR9F/A0GvYv8rJ+R1UqlUfWv1jRs3IJVKqzukSsG5abdu3TokJCTA1dUVnp6eWvMVy9vyYt26dejSpQskEol6K4W4uDi4urri+++/5xoOqSDSomJkFGm2DvIYM+QK+VAygBIMzMAilwXuF+TDmtUcAHS2EMDFQvtOr+pkajs6mjK5XI5bt25BKpWq77LLyclBt27datxnzzlpvr4eJlfu7u64ffs29u/fj1u3bkEoFOKjjz5CaGhojftwTcmBx1J8k5iqUcYwDNzd3WFlZaVeVzM/Px+pqalaU3gmNXDH5Eb1qjLkMpnijo6mLCcnR50wgZe3G0ulUuTk5LxRd54x4pw0dS22wZW1tTXGjh37xuchFWdoPRf0dNbeZTKTYTA36eWto1M93eCmsIJNXTut45wtjOcPnqnu6GjKSpu5+KaLjRgjg0ZeXrx4gUOHDiEhIQEzZ86Eg4MDYmNj4erqCnd393LrJyQkYMOGDbh79y4YhoG3tzemTJmCRo0aGRIOqQAuFuY6L6/zFQpIpVIIBAL0aFoHtjwLAMY9rKvapqOkpES97zmfzzeqbTpqGltbW7i6uuLp06fq1r2rq2uNXL+Uc9K8ffs2AgMDIRaLkZSUhDFjxsDBwQG//fYbkpOT1bcglubkyZMICQmBr68vOnXqBJZlceXKFbRo0QJ//PEHDQYZIZZlUVxsPLdOlkcoFCI7OxtpaWnqpFmnTh2j2qajphEIBGjVqpVRr/JfUQza7iI8PByrV6+GjY2NurxPnz4YNmxYufVnz56NadOmaS1EPHv2bMyaNYuSJqkQMpkMT548QUlJCfh8vtHv1V4TuLi4QCQSISsrC/b29jX2M+c85ejatWsYN26cVrm7u3upi/m+6u7duxg9erRW+ahRo2hyO6kQOTk5ePbsmUaf5rNnz5CTk1PNkdVsUqkUd+7cQWpqKu7cuVNjpxxxTpqWlpY6Fwu+f/8+nJ2dy63v7OyMuLg4rfK4uDga3TRSDMPA3NwcJdUdiJ7kcjlkMhnMzMwgEAhgZmYGmUxWoydcV7fSBt9q4mdu0CLEixcvxk8//QTg5S9USkoKZs+erdfqP2PGjMHYsWPx6NEjdOzYEQzDIDIyEqtWrcJnn33G/R2QUtm7u2Niamr5B5ZBATM4OzuDYRjMzmNgpigGFIo3iqmy8fl8iMVivHjxQl0mFovpjrNKpBp8exXLsjVy8I3zT9HatWvxzjvvwMXFBQUFBejatSvS09MREBCAZcuWlVt/3rx5sLGxwbp16zBnzhwAQN26dbFw4UJMnjyZ+zsgpYr437E3qi+XyxEdHY3BgweDx+NhzspVEIlE8PX1NepfBFtbWzRs2FBrIKgmjuQaC6FQqJ7Lq8IwTI0cfOOcNG1tbREZGYmzZ88iNjYWSqUSbdq0QWBgoF71GYbBtGnTMG3aNHUf06sDSsR4FBQU4MWLF7C1tQWPx8OjR4/g4eFh9K0H1UiuSCSq8SO5xkIgEEAikWjdUFATP3ODr1d69Oih3q7CUJQsjRufz8e///6rXpVftX5q586dqzmy8pn6SC6jZE1uaTgXFxfY29vX+FtXqZOHlKqwsBAsy6pbDqr/FxYWGn0SevU2yidPnpjUbZQW+UpYZ///yu15tm+2CHFVEggENTZZqpjGd4JUC5ZlYWVlhfz8fBQWFsLFxQXW1tZGf2ucKY/kMkr2/xMmALCAdbYSjNK4P/PahJImKZWtrS1cXFzAsiwUCgV4PJ5J3BpX1kiusePJob31J/tfOTEK1Xp5ruqkJ9WvtL12GIaBTCZTT9cxMzMz+l0dTXkkVyEA8NpulGD+KzcBtWE5Pr2Spq7J7KUprxWiVCqxbNkybNmyBU+fPsWDBw/QsGFDzJs3D56enjrvFiKVT59dHUePHm0Suzqaykiu8kUBWJl26ze7GBAV/P9FYK5QiZIi7fqMWAgzO+P5Q1BbluPTK2na2dnpvYGaopyJz0uXLsXevXuxevVqjBkzRl3u4+OD9evXU9KsJuPGjUNISIjB9Y2llaliCiO5xRf+QdGRO1rluQCe//eHSi6Xl/qHyiKkBSz7+VRylPqpTcvx6ZU0z507p/5/UlISZs+ejfDwcAQEBAAAoqKisHfvXqxYsaLcc+3btw/btm1Dz549MX78eHV5q1atcO/ePa7xkwpiTJfXFcXYR3LNuzaGwFf7Dim2WIG8lWdQXFwM69k9wZjzdNR+2dI0FnRH0Gu6du2q/v/ixYvx1VdfITQ0VF0WEhICHx8fbNu2DSNHjizzXKmpqWjcuLFWuVKpNInRTUIqipmdENBxec0W/f9d/rz69mAsjH9moCn3I3PFefQ8KioK/v7+WuX+/v64evVqufVbtGiBS5cuaZX//PPP6j2DCCGmRdWPrOrGM9Z+5IrA+U+Yh4cHtmzZgnXr1mmUb926FR4eHuXWX7BgAcLCwpCamgqlUolff/0V9+/fx759+/C///2PaziE1EiqwTdTmp9pCv3IFYFz0ly/fj0GDhyIkydPokOHDgCA6OhoJCQk4Jdffim3/nvvvYeDBw9i+fLlYBgG8+fPR5s2bWjVdkL+Y1HAwvK/laWEmUrkOyrpjiAjwjlpvvPOO3jw4AG+++473Lt3DyzLol+/fhg/frxeLU0ACA4ORnBwMOdgCanpGCUL6xwlClSXuf/dEVRsyZjMPeg1nUE9zB4eHli+fLlBL9iwYUNcu3YNjo6OGuUvXrxAmzZt8OjRI4POS8jrTHGiNU/+MlFq+O+OoBLj3s+u1jCozX/p0iV8+OGH6NixI1L/W+T2+++/R2RkZLl1k5KSdM7lLCoqUp+LGBe5XI7s7GyTmt0glUoRFxeHe/fuIS4uzmS2XlAIAPb1BqUJ3RFUG3Buaf7yyy8ICwvD8OHDERsbi6Kil7cq5OTkYPny5Th2TPfCt0eOHFH//+TJkxCLxerHCoUCZ86cgaenJ9dwSCUzxbs8THmiNWvGIM/GDIxqZSkGyLc1o0tzI8I5aS5duhRbtmzBiBEjcODAAXV5x44dsXjx4lLrvf/++wBejgq+PpdTIBDA09NTa0SeVC9TTT7VOdG6vnNdYO/DNzqHUqHAw4wMCAQCiE8nwpKne3I7p5hIheGcNO/fv48uXbpoldva2mrsyfI6pVIJAGjQoAGuXbsGJycnri9Nqpip3uVRnROtTx46Uv5B5cjLy4NIJEJxcTFiT0bC2tq6AiIjFYVzn2adOnXwzz//aJVHRkaiYcOG5dZPTEysloSZlZWFsLAwiMViiMVihIWFlZnk5XI5Zs2aBR8fH1hbW6Nu3boYMWIEnjx5UnVBVzNV8nmVKdzlURMmWqt3AC0xlT1Aaw/OLc1x48ZhypQp2LVrFxiGwZMnTxAVFYUZM2Zg/vz5ep0jLy8PFy5cQEpKCoqLizWeq6zN1YYNG4bHjx/jxIkTAICxY8ciLCwMf/zxh87j8/PzERsbi3nz5uGtt95CVlYWpk6dipCQEFy/fr1SYjQ2quTz6NEjk9trx5QnWmdkZKh3AL19+zaaNWtm9P3ItQprgC+++IIVCoUswzAswzCspaUl++WXX+pVNzY2lnVzc2NtbW1ZHo/HOjs7swzDsNbW1myDBg0MCadc8fHxLAA2OjpaXRYVFcUCYO/du6f3ea5evcoCYJOTkznHIJPJWACsTCbjXLc6PX36lI2KimLPnTvHRkVFsU+fPq3ukGq04uJi9uLFi6yrqyvr6urKXrhwgb169SpbXFxc3aGR/xg05WjZsmXIzMzE1atXER0djYyMDCxZskSvutOmTcN7772H58+fQygUIjo6GsnJyfDz88PatWsNCadcUVFREIvFaN++vbqsQ4cOEIvFuHLlit7nkclkYBgGdnZ2lRCl8THlbSNMlSmvOl9bcL48HzVqFDZu3AgbGxuNhTvy8vIwadIk7Nq1q8z6cXFx2Lp1K3g8Hng8HoqKitCwYUOsXr0aI0eOxIABA7i/i3Kkp6frvLxxcXFBenq6XucoLCzE7NmzMWzYsDIXWo6IiEBERIRWuSkmmoKCAmRlZSEjI0M95cjZ2dnoB4JM2av9yHw+HwqFwiT6kWsTzi3NvXv36vyrV1BQgH379pVbXyAQqH8oXF1dkZKSAgAQi8Xq/+tr4cKFYBimzC9V/6OuRZRViaA8crkcQ4cOhVKpxObNm8s8NjQ0FEeOHNH6OnjwIKf3Zgz4fD4yMzM1WpqZmZnqrS9IxRMIBLC2toZYLIZIJEJycjJEIhH9kTIiev/0Z2dnq7dwzcnJ0djbR6FQ4NixY3p1Vrdu3RrXr19H06ZN0b17d8yfPx+ZmZn4/vvv4ePDbRXqiRMnYujQoWUe4+npidu3b+Pp06daz2VkZMDV1bXM+nK5HIMHD0ZiYiLOnj1r9JuKVaSSkhI4OTlptDSdnJxoRLcSyeVy5OXlITs7GzweDxKJBLm5uZDL5ZQ4jYTeSVO15QXDMGjatKnW8wzDlLrHzKuWL1+OnJwcAMCSJUswcuRITJgwAY0bNy730v51Tk5Oek1fCggIgEwmw9WrV9GuXTsAQExMDGQyGTp27FhqPVXCfPjwIc6dO6d1v3xNJxQKYW9vr54zaG5uDoFAQJeKFUTXZnZ5eXka6y8kJSXB0tIShYWFWvM1a+Jq+6aAYV/vdS7FhQsXwLIsevTogV9++QUODg7q58zNzSGRSFC3rvHeedCnTx88efIEW7duBfByypFEItGYcuTl5YUVK1agf//+KCkpwcCBAxEbG4v//e9/Gi1SBwcHmJubc3r97OxsiMViyGQyk2qtmuJtlKZi4cKFWg0NhmHg7u4OKysr9QT9/Px8pKamag0QGdNmdrWJ3kkTeHm59vHHH2PJkiV6LwP3uh49euDXX3/VGoHOzs7G+++/j7Nnzxp03vI8f/4ckydPVt8DHxISgk2bNmnEwTAMdu/ejfDwcCQlJaFBgwY6z3Xu3Dl069aN0+ubatIETHO1IFOgq6VZUlKCv/76C1lZWeoye3t7+Pj4aPUlU0uzenBKmgBgY2ODv/76y+DFNczMzHSOZkulUri7u5vkKLM+TDlpkqqTnZ2Ne/fuoaSkRN0lwufz4eXlRT83RoLzMGjPnj1x/vx5hIeHc6p3+/Zt9f/j4+M1pvooFAqcOHEC7u7aO/MRUpuophzx+Xx1y5KmHBkXzkmzT58+mDNnDv7++2/4+flpdU6Xtne2r6+veiCpR48eWs8LhUJ88803XMMhpEZR3br6ej8ydYsYD86X52ZmpU/tZBhG5wLDANQ/BA0bNsTVq1fh7Oysfs7c3BwuLi7gveESWMaMLs8JF9SPbLw4tzRVS7xxJZFI3qg+IYQYgyrf4m7v3r04evSo+vHnn38OOzs7dOzYEcnJyVUdDiFGx1S36qgtDEqaFy5cwHvvvYfGjRujSZMmCAkJwaVLl/Squ3z5cnWndlRUFDZt2oTVq1fDyckJ06ZNMyQcQmoMWiTF+HFOmj/88AMCAwNhZWWFyZMnY+LEiRAKhejZsyd+/PHHcuv/+++/aNy4MQDg8OHDGDRoEMaOHYsVK1bonXgJqalUqxwpFArk5+dDoVDQKkdGhnOf5rJly7B69WqNVuGUKVPw1VdfYcmSJRg2bFiZ9UUiEZ49e4b69evj1KlT6vNYWlrSDwap9YRCIWQyGdLT09XzNN3c3GjKkRHhnDQfPXqE9957T6s8JCQEX3zxRbn1e/XqhY8//hitW7fGgwcP0LdvXwDAnTt3aDdKQvByZ9f09HT1lCPaI8i4cL489/DwwJkzZ7TKz5w5o9etld9++y0CAgKQkZGBX375Rb0Ixo0bNxAaGso1HEJqlOzsbBQUFMDFxQWOjo5wcXFBQUEBsrOzqzs08h/OLc3PPvsMkydPRlxcHDp27AiGYRAZGYk9e/Zg48aN5da3s7PDpk2btMr1WSGJkJqutPVd9Vn3lVQNzklzwoQJcHNzw7p16/DTTz8BALy9vXHw4EH069evwgMkpDaxsbGBUChEQkKC+vK8UaNGsLGxqe7QyH8MWoK7f//+6N+/f0XHQgjBy8Tp6uqqXniYEqZxMXjfguvXr+Pu3btgGAbe3t7w8/OryLgIqZUKCgrUW10UFRXBwsICPB6P9mUyIpyT5uPHjxEaGorLly+r16J88eIFOnbsiIiICIPX2SSE/P8qRzweD1ZWVgBolSNjw3n0fNSoUZDL5bh79y6eP3+O58+f4+7du2BZFqNHjy63fo8ePfDixQut8uzsbJ2rHxFSm6hWOVIN/NAqR8aH8ypHQqEQV65cQevWrTXKY2Nj0alTp3InqNMixLTKESkfrXJkvDhfntevX19nYispKSlzEWFahJgQ/QkEAkqWRopz0ly9ejUmTZqEb7/9Fn5+fuq9xadMmYK1a9eWWo8WISaE1AScL8/t7e2Rn5+PkpIS9XL8qv+/frvX8+fP1f+nRYhN9/KcLhUJ+X+cW5obNmww6IVoEWLTRFv4EqKJc0uzosTHxyMlJQXFxcUa5aXtMWTqTLGlKZfLERcXp7HfNsMw8PX1pRYnqbUMmtyekJCA3bt3IyEhARs3boSLiwtOnDgBDw8PtGjRosy6jx49Qv/+/fHXX3+BYRj1L6RqikVpewyRqqda2/FVqrUdKWmS2orzPM0LFy7Ax8cHMTEx+PXXX5Gbmwvg5ej4ggULyq0/ZcoUNGjQAE+fPoWVlRXu3LmDixcvwt/fH+fPn+f8BkjlUU20fhVNtCa1HeekOXv2bCxduhSnT5+Gubm5urx79+6Iiooqt35UVBQWL14MZ2dnmJmZwczMDJ07d8aKFSswefJkruGQSkQTrauPXC5HdnZ2jZ23bMo4X57/9ddfOre1cHZ2xrNnz8qtr1AoIBKJAABOTk548uQJmjVrBolEgvv373MNh1QyFxcX2Nvb0+h5FaLBN+PGOWna2dkhLS0NDRo00Ci/efOmXpPTW7Zsidu3b6Nhw4Zo3749Vq9eDXNzc2zbtg0NGzbkGg6pAjTRuuqUtrGavb09fQ+MBOfL82HDhmHWrFlIT08HwzBQKpW4fPkyZsyYgREjRpRb/8svv1RPO1q6dCmSk5Px9ttv49ixY/j666+5vwNCapCyBt+IceA85UgulyM8PBwHDhwAy7Lg8/lQKBQYNmwY9uzZY9AE9efPn8Pe3r5Gr05tilOOSNWjaV7Gj1PSZFkWKSkpcHZ2Rnp6OmJjY6FUKtG6dWs0adKkMuM0eZQ0ib6oT9O4cUqaSqUSlpaWuHPnDiVJjihpEi7o1lXjxalP08zMDE2aNNFrlJwQYjiBQABbW1tKmEaI80DQ6tWrMXPmTPz999+VEQ8hhBg1zknzww8/xNWrV/HWW29BKBTCwcFB48tYZWVlISwsDGKxGGKxGGFhYTpXkC/NuHHjwDCMwQuWEEJqhipb5ai6DRs2DI8fP8aJEycAAGPHjkVYWBj++OOPcusePnwYMTExqFu3bmWHSQgxcpyT5siRIysjjkp19+5dnDhxAtHR0Wjfvj0AYPv27QgICMD9+/fRrFmzUuumpqZi4sSJOHnyJPr27VtVIRNCjBTny/NX9e3bF2lpaRUVS6WJioqCWCxWJ0wA6NChA8RiMa5cuVJqPaVSibCwMMycObPc1ZsIIbWDwfueA8DFixdN4k4FXRu5AS/vq351r6LXrVq1Cnw+n9NCIhEREYiIiNAqp4UXCKkZ3ihpVreFCxdi0aJFZR5z7do1ANB5t5Fq8rAuN27cwMaNGxEbG8vpTqXQ0FCEhoZqlavmaRJCTJteSfPrr7/G2LFjYWlpiZSUFHh4eBjFMmETJ07E0KFDyzzG09MTt2/fxtOnT7Wey8jIgKurq856ly5dglQqRf369dVlCoUCn332GTZs2ICkpKQ3ip0QYpr0uiOIz+fjyZMn6s3P0tLSTOq2rrt376J58+aIiYlBu3btAAAxMTHo0KED7t27p3Mg6NmzZ1r9tcHBwQgLC8NHH31U5uCRLnRHECE1g14tzbp16+KXX37BO++8A5Zl8fjxYxQWFuo89tWWmbHw9vZG7969MWbMGGzduhXAyylH7777rkby8/LywooVK9C/f384OjrC0dFR4zwCgQBubm6cEyYhpObQK2l++eWXmDRpEiZOnAiGYdC2bVutY1T9g8a6x8/+/fsxefJkBAUFAXi5gdumTZs0jrl//z5kMll1hEcIMRF6L9iRk5OD5ORktGrVCn/++adWK0zlrbfeqtAAawq6PCekZtB79NzGxgYtW7bE7t270alTJ1hYWFRmXIQQYpQ4T24fOXIkCgoKsGPHDsyZMwfPnz8HAMTGxiI1NbXCAySkNqKN1YwX53mat2/fRmBgIMRiMZKSkjBmzBg4ODjgt99+Q3JyMvbt21cZcRJSa9AixMaNc0tz2rRpCA8Px8OHD2Fpaaku79OnDy5evFihwRFS25S2sRq1OI0H56R5/fp1jBs3Tqvc3d29zFsSCSHlo43VjB/npGlpaYns7Gyt8vv378PZ2blCgiKkthIKhVq37TIMA6FQWE0RkddxTpr9+vXD4sWL1ZcLDMMgJSUFs2fPxsCBAys8QFL9aFCi6ggEAkgkEnXiNIbblYkmzlv4Zmdn45133sGdO3eQk5ODunXrIj09HQEBATh27Bisra0rK1aTZqrzNGlQonrQxmrGi3PSVDl79qx6C982bdogMDCwomOrUUwxadIe3IRoM3hpuB49eqBHjx4VGQsxMmUNSlDSJLWV3n2aMTExOH78uEbZvn370KBBA7i4uGDs2LEoKiqq8ABJ9aFBCUK06Z00Fy5ciNu3b6sf//XXXxg9ejQCAwMxe/Zs/PHHH1ixYkWlBEmqBw1KEKJN7z7NOnXq4I8//oC/vz8AYO7cubhw4QIiIyMBAD///DMWLFiA+Pj4yovWhJlin6YKDUoQ8v/07tPMysrSWOX8woUL6N27t/px27Zt8e+//1ZsdMQoCAQCSpaE/Efvy3NXV1ckJiYCAIqLixEbG4uAgAD18zk5OfSLRQip8fROmr1798bs2bNx6dIlzJkzB1ZWVnj77bfVz9++fRuNGjWqlCAJIcRY6H15vnTpUgwYMABdu3aFSCTC3r17YW5urn5+165d6lXRCSGkpuI8uV0mk0EkEoHH42mUP3/+HCKRSCORkv9nygNBhJD/x3lye2l7dzs4OLxxMIQQYuw4L9hBCCG1GSVNQgjhgJImIYRwQEmTEEI4oKRJCCEcUNIkhBAOKGkSQggHlDQJIYQDSpqEEMIBJU1CCOGAkiYhhHBASZMQQjigpEkIIRxQ0iSEEA5qTdLMyspCWFgYxGIxxGIxwsLC8OLFi3Lr3b17FyEhIRCLxbCxsUGHDh2QkpJS+QETQoxSrUmaw4YNQ1xcHE6cOIETJ04gLi4OYWFhZdZJSEhA586d4eXlhfPnz+PWrVuYN28eLC0tqyhqQoix4bxyuym6e/cumjdvjujoaLRv3x4AEB0djYCAANy7dw/NmjXTWW/o0KEQCAT4/vvv3zgGWrmdkJqhVrQ0o6KiIBaL1QkTADp06ACxWIwrV67orKNUKnH06FE0bdoUwcHBcHFxQfv27XH48OEqipoQYow4b3dhitLT0+Hi4qJV7uLigvT0dJ11pFIpcnNzsXLlSixduhSrVq3CiRMnMGDAAJw7dw5du3bVWS8iIgIRERFa5XK5/M3eBCHEKJh00ly4cCEWLVpU5jHXrl0DADAMo/Ucy7I6y4GXLU0A6NevH6ZNmwYA8PX1xZUrV7Bly5ZSk2ZoaChCQ0O1ylWX54QQ02bSSXPixIkYOnRomcd4enri9u3bePr0qdZzGRkZcHV11VnPyckJfD4fzZs31yj39vZGZGSk4UETQkyaSSdNJycnODk5lXtcQEAAZDIZrl69inbt2gEAYmJiIJPJ0LFjR511zM3N0bZtW9y/f1+j/MGDB5BIJG8ePCHEJNWKgSBvb2/07t0bY8aMQXR0NKKjozFmzBi8++67GiPnXl5e+O2339SPZ86ciYMHD2L79u34559/sGnTJvzxxx/45JNPquNtEEKMQK1ImgCwf/9++Pj4ICgoCEFBQWjVqpXWVKL79+9DJpOpH/fv3x9btmzB6tWr4ePjgx07duCXX35B586dqzp8UsvI5XJkZ2fTAKIRqhXzNI0BzdMk+pJKpUhOTlYPVEokEp2zP0j1qDUtTUJMgVwuVydM4OUMj+TkZGpxGhFKmoQYkYKCArx+8ceyLAoKCqopIvI6SpqEGBGhUKg1d5hhGAiFwmqKiLyOkiYhRkQgEEAikagTp6pPUyAQVHNkRMWk52kSUhO5uLjA3t4eBQUFEAqFlDCNDCVNQoyQQCCgZGmk6PKcEEI4oKRJCCEcUNIkhBAOKGkSQggHlDQJIYQDSpqEEMIBJU1CCOGAkiYhhHBASZMQQjigpEkIIRxQ0iSEEA4oaRJCCAeUNAkhhANKmoQQwgElTUII4YCSJiGEcEBJkxBCOKCkSQghHFDSJIQQDmiPoCqi2ss6Ozu7miMhhJTFxsZGaxvlV1HSrCI5OTkAAA8Pj2qOhBBSFplMBltb21KfZ1hVE4hUKqVSiSdPnpT7V8xYDRkyBAcPHqzuMDgz1bgB043dVONWoZamkTAzM0O9evWqOwyDCQSCMv/6GitTjRsw3dhNNW590UAQIYRwQEmTEEI4oKRJCCEcUNIkhBAOKGkSvYSGhlZ3CAYx1bgB043dVOPWF005IoQQDqilSQghHFDSJIQQDihpEkIIB5Q0CSFaunXrhqlTp1Z3GEaJkmYtEh4eDoZhwDAMBAIBXF1d0atXL+zatQtKpVJ9nKenJxiGwYEDB7TO0aJFCzAMgz179mgdHx0drXHs1KlT0a1btyqNXSUoKAg8Hk8rplfPtXLlSo3yw4cPV9m6AFKpFOPGjUP9+vVhYWEBNzc3BAcHIyoqCgD37wGpOpQ0a5nevXsjLS0NSUlJOH78OLp3744pU6bg3XffRUlJifo4Dw8P7N69W6NudHQ00tPTYW1trXVeS0tLzJo1yyhiT0lJQVRUFCZOnIidO3fqPJelpSVWrVqFrKysSo25NAMHDsStW7ewd+9ePHjwAEeOHEG3bt3w/Plz9TFcvwekalDSrGVUrRp3d3e0adMGX3zxBX7//XccP35co+UyfPhwXLhwAf/++6+6bNeuXRg+fDj4fO11XsaNG4fo6GgcO3as2mPfvXs33n33XUyYMAEHDx5EXl6e1rkCAwPh5uaGFStWVFq8pXnx4gUiIyOxatUqdO/eHRKJBO3atcOcOXPQt29f9XFcvwcVraSkBBMnToSdnR0cHR3x5ZdfqteF/eGHH+Dv7w8bGxu4ublh2LBhkEql6rpZWVkYPnw4nJ2dIRQK0aRJE40/AKmpqRgyZAjs7e3h6OiIfv36ISkpqdLfU0WgpEnQo0cPvPXWW/j111/VZa6urggODsbevXsBAPn5+Th48CBGjRql8xyenp4YP3485syZo/NyubK8HjvLsti9ezc+/PBDeHl5oWnTpvjpp5+06vF4PCxfvhzffPMNHj9+XGXxAoBIJIJIJMLhw4dRVFRU6nFcvwcVbe/eveDz+YiJicHXX3+N9evXY8eOHQCA4uJiLFmyBLdu3cLhw4eRmJiI8PBwdd158+YhPj4ex48fx927d/Hdd9/ByclJ/T66d+8OkUiEixcvIjIyEiKRCL1790ZxcXGVvLc3QUmTAAC8vLy0/tKPGjUKe/bsAcuyOHToEBo1agRfX99Sz/Hll18iMTER+/fvr9xgX/Nq7H/++Sfy8/MRHBwMAPjwww9LvUTv378/fH19sWDBgqoKFQDA5/OxZ88e7N27F3Z2dujUqRO++OIL3L59W+tYrt+DiuTh4YH169ejWbNmGD58OCZNmoT169er4+rTpw8aNmyIDh064Ouvv8bx48eRm5sL4GUXSevWreHv7w9PT08EBgbivffeAwAcOHAAZmZm2LFjB3x8fODt7Y3du3cjJSUF58+fr5L39iYoaRIAL1torw+C9O3bF7m5ubh48SJ27dpVbgvH2dkZM2bMwPz586u0xfBq7Dt37sSQIUPUl6+hoaGIiYnB/fv3ddZdtWoV9u7di/j4+CqLF3jZp/nkyRMcOXIEwcHBOH/+PNq0aaM1uMP1e1CROnTooPEzERAQgIcPH0KhUODmzZvo168fJBIJbGxs1AN+KSkpAIAJEybgwIED8PX1xeeff44rV66oz3Pjxg38888/sLGxUbe6HRwcUFhYiISEhCp7f4aipEkAAHfv3kWDBg00yvh8PsLCwrBgwQLExMRg+PDh5Z5n+vTpKCgowObNmysrVC2q2J8/f47Dhw9j8+bN4PP54PP5cHd3R0lJCXbt2qWzbpcuXRAcHIwvvviiyuJVsbS0RK9evTB//nxcuXIF4eHhWq1eQ74Hla2wsBBBQUEQiUT44YcfcO3aNfz2228AoP5j2adPHyQnJ2Pq1Kl48uQJevbsiRkzZgB4uYuBn58f4uLiNL4ePHiAYcOGVdv70hclTYKzZ8/ir7/+wsCBA7WeGzVqFC5cuIB+/frB3t6+3HOJRCLMmzcPy5Ytq5JN5F6Nff/+/ahXrx5u3bql8cu4YcMG7N27V2OE/VUrV67EH3/8odEaqg7NmzfXOWjF9XtQUV6frhUdHY0mTZrg3r17yMzMxMqVK/H222/Dy8tLYxBIxdnZGeHh4fjhhx+wYcMGbNu2DQDQpk0bPHz4EC4uLmjcuLHGl1gsrpL39iZou4tapqioCOnp6VAoFHj69ClOnDiBFStW4N1338WIESO0jvf29kZmZiasrKz0fo2xY8di/fr1iIiIQPv27assdj8/PwwaNAgtW7bUqCeRSDBr1iwcPXoU/fr10zqvj48Phg8fjm+++abCYi3Ls2fP8MEHH2DUqFFo1aoVbGxscP36daxevVpnfIZ8DyrCv//+i+nTp2PcuHGIjY3FN998g3Xr1qF+/fowNzfHN998g/Hjx+Pvv//GkiVLNOrOnz8ffn5+aNGiBYqKivC///0P3t7eAF7OClizZg369euHxYsXo169ekhJScGvv/6KmTNnGv22MJQ0a5kTJ06gTp064PP5sLe3x1tvvYWvv/4aI0eOhJmZ7gsPR0dHTq8hEAiwZMmSCr/UKiv2mzdv4tatW9i+fbtWPRsbGwQFBWHnzp06kxIALFmyROcoe2UQiURo37491q9fj4SEBMjlcnh4eGDMmDGldhNw/R5UhBEjRqCgoADt2rUDj8fDpEmTMHbsWPXE+i+++AJff/012rRpg7Vr1yIkJERd19zcHHPmzEFSUhKEQiHefvtt9UR9KysrXLx4EbNmzcKAAQOQk5MDd3d39OzZ0yT2FqKl4QghhAPq0ySEEA4oaRJCCAeUNAkhhANKmoQQwgElTUII4YCSJiGEcEBJkxBCOKCkSQghHFDSJIQQDihpEqJDeHg43n//fc71Fi5cWGXrXZLqQUmTEEI4oKRJarVDhw7Bx8cHQqEQjo6OCAwMxMyZM7F37178/vvv6h0wVSuKz5o1C02bNoWVlRUaNmyIefPmQS6XAwD27NmDRYsW4datW+p6qkWFZTIZxo4dCxcXF9ja2qJHjx64detWNb1r8iZolSNSa6WlpSE0NBSrV69G//79kZOTg0uXLmHEiBFISUlBdna2ejMwBwcHAC9XTNqzZw/q1q2Lv/76C2PGjIGNjQ0+//xzDBkyBH///TdOnDiBP//8EwAgFovBsiz69u0LBwcHHDt2DGKxGFu3bkXPnj3x4MED9bmJiWAJqaVu3LjBAmCTkpK0nhs5ciTbr1+/cs+xevVq1s/PT/14wYIF7FtvvaVxzJkzZ1hbW1u2sLBQo7xRo0bs1q1bDYqdVB9qaZJa66233kLPnj3h4+OD4OBgBAUFYdCgQWWujn7o0CFs2LAB//zzD3Jzc1FSUlLuGpA3btxAbm6u1pqYBQUFJrEnDtFESZPUWjweD6dPn8aVK1dw6tQpfPPNN5g7dy5iYmJ0Hh8dHY2hQ4di0aJFCA4OhlgsxoEDB7Bu3boyX0epVKJOnTo6d1q0s7OrgHdCqhIlTVKrMQyDTp06oVOnTpg/fz4kEgl+++03mJubQ6FQaBx7+fJlSCQSzJ07V12WnJyscYyuem3atEF6ejr4fD48PT0r7b2QqkFJk9RaMTExOHPmDIKCguDi4oKYmBhkZGTA29sbhYWFOHnyJO7fvw9HR0eIxWI0btwYKSkpOHDgANq2bYujR4+qd2FU8fT0RGJiIuLi4lCvXj3Y2NggMDAQAQEBeP/997Fq1So0a9YMT548wbFjx/D+++/D39+/mj4BYpDq7lQlpLrEx8ezwcHBrLOzM2thYcE2bdqU/eabb1iWZVmpVMr26tWLFYlELAD23LlzLMuy7MyZM1lHR0dWJBKxQ4YMYdevX8+KxWL1OQsLC9mBAweydnZ2LAB29+7dLMuybHZ2Njtp0iS2bt26rEAgYD08PNjhw4ezKSkpVfyuyZuiPYIIIYQDmtxOCCEcUNIkhBAOKGkSQggHlDQJIYQDSpqEEMIBJU1CCOGAkiYhhHBASZMQQjigpEkIIRxQ0iSEEA4oaRJCCAf/B+XAA1qJ+QuvAAAAAElFTkSuQmCC\n", 212 | "text/plain": [ 213 | "
" 214 | ] 215 | }, 216 | "metadata": {}, 217 | "output_type": "display_data" 218 | }, 219 | { 220 | "name": "stderr", 221 | "output_type": "stream", 222 | "text": [ 223 | "/var/folders/qv/ycsblvjs0bnd5ssgv07h6s6m0000gn/T/ipykernel_5124/459608050.py:53: UserWarning: FixedFormatter should only be used together with FixedLocator\n", 224 | " ax.set_xticklabels(['DMN','DAN','SM','base'])\n" 225 | ] 226 | }, 227 | { 228 | "data": { 229 | "image/png": "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\n", 230 | "text/plain": [ 231 | "
" 232 | ] 233 | }, 234 | "metadata": {}, 235 | "output_type": "display_data" 236 | } 237 | ], 238 | "source": [ 239 | "condition=['sitcomep1','sitcomep2','documentary']\n", 240 | "for i, cdt in enumerate(condition):\n", 241 | " # Load HMM-derived latent state sequence\n", 242 | " seq = scipy.io.loadmat(loaddir+'fmri/hmmoutput.mat')[cdt]\n", 243 | " # Load time-resolved attention measures: Self-reported narrative engagement (z-normalized)\n", 244 | " beh = scipy.io.loadmat(loaddir+'behavior/'+cdt+'_beh.mat')['engagement_conv'].T\n", 245 | " \n", 246 | " # Categorize attention measure of each TR to latent state identification\n", 247 | " stateocc = np.zeros((nsubj,nstate))\n", 248 | " for subj in range(nsubj):\n", 249 | " for stt in range(1,nstate+1):\n", 250 | " stateocc[subj,stt-1]=np.nanmean(beh[subj,np.where(seq[subj,:]==stt)[0]])\n", 251 | " stateocc_real=np.nanmean(stateocc,0)\n", 252 | " \n", 253 | " # to visualize data points\n", 254 | " valus = np.concatenate((stateocc[:,0], stateocc[:,1], stateocc[:,2],stateocc[:,3]),0)\n", 255 | " categ = np.concatenate((np.repeat('state1',nsubj), np.repeat('state2',nsubj), np.repeat('state3',nsubj), np.repeat('state4',nsubj)),0)\n", 256 | " categ = np.delete(categ,np.where(np.isnan(valus))[0])\n", 257 | " valus = np.delete(valus,np.where(np.isnan(valus))[0])\n", 258 | " df = pd.DataFrame(data={'state':categ,'values':valus})\n", 259 | " \n", 260 | " # Chance distribution: circular-shift attention time course 1,000 times and relate them with the latent state sequence\n", 261 | " stateocc_null = []\n", 262 | " for null in range(niter):\n", 263 | " stateocc_niter = np.zeros((nsubj,nstate))\n", 264 | " for subj in range(nsubj):\n", 265 | " beh_null = np.roll(beh[subj,:],np.random.randint(beh[subj,:].shape[0]-20)+10)\n", 266 | " for stt in range(1,nstate+1):\n", 267 | " stateocc_niter[subj,stt-1]=np.nanmean(beh_null[np.where(seq[subj,:]==stt)[0]])\n", 268 | " stateocc_null.append(np.nanmean(stateocc_niter,0))\n", 269 | " stateocc_null=np.array(stateocc_null)\n", 270 | " \n", 271 | " # two-tailed test\n", 272 | " pvals = []\n", 273 | " for stt in range(nstate):\n", 274 | " rl = stateocc_real[stt] - np.mean(stateocc_null[:,stt])\n", 275 | " nl = stateocc_null[:,stt] - np.mean(stateocc_null[:,stt])\n", 276 | " pvals.append((1+np.where(np.abs(rl)<=np.abs(nl))[0].shape[0])/(1+niter))\n", 277 | " qvals=fdrcorrection(np.array(pvals))[1]\n", 278 | "\n", 279 | " # Bar plot\n", 280 | " fig, ax = plt.subplots(1,1, figsize=(3,4))\n", 281 | " ax.bar(np.arange(nstate),np.nanmean(stateocc,0), color=cmap, edgecolor='black',linewidth=0.5)\n", 282 | " ax.errorbar(np.arange(nstate),np.nanmean(stateocc,0),\n", 283 | " np.nanstd(stateocc,0)/np.sqrt(stateocc.shape[0]-np.sum(np.isnan(stateocc[:,0]))),\n", 284 | " capsize=6,ecolor='black',ls='none')\n", 285 | " sns.stripplot(x=\"state\", y=\"values\", data=df, color=[0.3,0.3,0.3],jitter=False,alpha=0.3, size=4)\n", 286 | " plt.hlines(0, -1, 4, colors='black', linewidth=0.5)\n", 287 | " ax.set_ylim([-0.75, 0.65])\n", 288 | " ax.set_xlim([-1,4])\n", 289 | " ax.tick_params(direction='out', length=6, width=0.6)\n", 290 | " ax.set_ylabel('Sef-reported engagement (z)\\nat state occurrence')\n", 291 | " ax.set_xticklabels(['DMN','DAN','SM','base'])\n", 292 | " ax.tick_params(length=5, width=0.5)\n", 293 | " ax.spines['right'].set_visible(False)\n", 294 | " ax.spines['top'].set_visible(False)\n", 295 | " ax.set_title(cdt+' (N='+str(stateocc.shape[0]-np.sum(np.isnan(stateocc[:,0])))+')')\n", 296 | " for stt in range(nstate):\n", 297 | " if qvals[stt]<0.01:\n", 298 | " plt.scatter(stt,0.63,30,'k','*')\n", 299 | " plt.show()" 300 | ] 301 | } 302 | ], 303 | "metadata": { 304 | "kernelspec": { 305 | "display_name": "Python 3 (ipykernel)", 306 | "language": "python", 307 | "name": "python3" 308 | }, 309 | "language_info": { 310 | "codemirror_mode": { 311 | "name": "ipython", 312 | "version": 3 313 | }, 314 | "file_extension": ".py", 315 | "mimetype": "text/x-python", 316 | "name": "python", 317 | "nbconvert_exporter": "python", 318 | "pygments_lexer": "ipython3", 319 | "version": "3.9.13" 320 | } 321 | }, 322 | "nbformat": 4, 323 | "nbformat_minor": 2 324 | } 325 | -------------------------------------------------------------------------------- /code/hmmfit.py: -------------------------------------------------------------------------------- 1 | # # # import libraries 2 | import numpy as np 3 | import scipy.io 4 | from sklearn.cluster import KMeans 5 | from hmmlearn import hmm 6 | import os 7 | import random 8 | import timeit 9 | import pickle 10 | from datetime import datetime 11 | from scipy import stats 12 | 13 | # # # parameters 14 | nstate=4 15 | nsubj=27 16 | condition=['rest1','rest2','gradCPTface','gradCPTscene','sitcomep1','sitcomep2','documentary'] 17 | print('numState=' + str(nstate)) 18 | random.seed(datetime.now()) 19 | loaddir = '../fmri' 20 | savedir = '../fmri' 21 | 22 | # # # Load fMRI time series and concatenate 23 | # Load time series extracted from preprocessed EPI file 24 | # 25 ROI: 17 cortical networks (Yeo et al., 2011) and 8 subcortical regions (Tian et al., 2020) 25 | data={'rest1':scipy.io.loadmat(loaddir+'/ts_rest1.mat')['ts'], 26 | 'rest2':scipy.io.loadmat(loaddir+'/ts_rest2.mat')['ts'], 27 | 'gradCPTface':scipy.io.loadmat(loaddir+'/ts_gradCPTface.mat')['ts'], 28 | 'gradCPTscene':scipy.io.loadmat(loaddir+'/ts_gradCPTscene.mat')['ts'], 29 | 'sitcomep1':scipy.io.loadmat(loaddir+'/ts_sitcomep1.mat')['ts'], 30 | 'sitcomep2':scipy.io.loadmat(loaddir+'/ts_sitcomep2.mat')['ts'], 31 | 'documentary':scipy.io.loadmat(loaddir+'/ts_documentary.mat')['ts']} 32 | 33 | # Concatenate z-normalized time series of all participants' all fMRI scan runs 34 | concatts, subjid, epiid = [], [], [] 35 | for subj in range(nsubj): 36 | for ep, cdt in enumerate(condition): 37 | if subj==nsubj-2 and cdt=='sitcomep1': # one participant with missing fMRI scan run 38 | pass 39 | else: 40 | # z-normalize ROI time series 41 | if len(concatts)==0: 42 | concatts = stats.zscore(data[cdt][subj,:,:], axis=0, ddof=1) 43 | subjid = np.repeat(subj,stats.zscore(data[cdt][subj,:,:], axis=0, ddof=1).shape[0], 0) 44 | epiid = np.repeat(ep, stats.zscore(data[cdt][subj, :, :], axis=0, ddof=1).shape[0], 0) 45 | else: 46 | concatts = np.concatenate((concatts, stats.zscore(data[cdt][subj, :, :], axis=0, ddof=1)), 0) 47 | subjid = np.concatenate((subjid, np.repeat(subj, stats.zscore(data[cdt][subj, :, :], axis=0, ddof=1).shape[0], 0)), 0) 48 | epiid = np.concatenate((epiid, np.repeat(ep, stats.zscore(data[cdt][subj, :, :], axis=0, ddof=1).shape[0], 0)), 0) 49 | 50 | print('nTime: '+str(concatts.shape[0])) 51 | print('nRegion: '+str(concatts.shape[1])) 52 | print('subj: '+str(np.unique(subjid))) 53 | print('condition: '+str(np.unique(epiid))) 54 | 55 | # # # Initialize with k-means clustering 56 | start = timeit.default_timer() 57 | print('Doing k-means clustering') 58 | kmeans_train = KMeans(n_clusters=nstate, init='k-means++', n_init=50, max_iter=200, tol=0.001).fit(concatts) 59 | stop = timeit.default_timer() 60 | print('Time: ', stop - start) 61 | 62 | # # # HMM fit 63 | start = timeit.default_timer() 64 | print('Fitting HMM') 65 | hmmmodel = hmm.GaussianHMM(n_components=nstate, covariance_type='full', means_prior=kmeans_train.cluster_centers_, n_iter=1000, tol=0, init_params='m') 66 | hmmmodel.fit(concatts) 67 | print('HMM fitting finished at ' + str(hmmmodel.monitor_.iter) + ' iterations') 68 | stop = timeit.default_timer() 69 | print('Time: ', stop - start) 70 | 71 | # # # Save HMM fit 72 | print('Saving Data') 73 | HMMMODEL = {'niter': hmmmodel.monitor_.iter, 74 | 'nfeatures': hmmmodel.n_features, 75 | 'transmat': hmmmodel.transmat_, 76 | 'startprob': hmmmodel.startprob_, 77 | 'Means': hmmmodel.means_, 78 | 'Covars': hmmmodel.covars_} 79 | scipy.io.savemat(savedir + '/hmmmodel.mat', HMMMODEL) 80 | pickle.dump(hmmmodel, open(savedir + '/hmmfit.pkl', 'wb')) 81 | 82 | # # # Save decoded latent state sequence 83 | HMMOUTPUT = {'train_state': hmmmodel.decode(concatts)[1], 84 | 'train_logprob': hmmmodel.decode(concatts)[0], 85 | 'train_posterior': hmmmodel.predict_proba(concatts), 86 | 'subjid':subjid, 87 | 'epiid':epiid} 88 | scipy.io.savemat(savedir + '/hmmdecode.mat', HMMOUTPUT) 89 | 90 | 91 | -------------------------------------------------------------------------------- /fmri/MNI152_T1_3mm_brain.nii.gz: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/MNI152_T1_3mm_brain.nii.gz -------------------------------------------------------------------------------- /fmri/gradientcoeff.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/gradientcoeff.mat -------------------------------------------------------------------------------- /fmri/hmmmodel.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/hmmmodel.mat -------------------------------------------------------------------------------- /fmri/hmmoutput.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/hmmoutput.mat -------------------------------------------------------------------------------- /fmri/state1_raw.nii.gz: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/state1_raw.nii.gz -------------------------------------------------------------------------------- /fmri/state2_raw.nii.gz: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/state2_raw.nii.gz -------------------------------------------------------------------------------- /fmri/state3_raw.nii.gz: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/state3_raw.nii.gz -------------------------------------------------------------------------------- /fmri/state4_raw.nii.gz: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/state4_raw.nii.gz -------------------------------------------------------------------------------- /fmri/ts_documentary.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/ts_documentary.mat -------------------------------------------------------------------------------- /fmri/ts_gradCPTface.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/ts_gradCPTface.mat -------------------------------------------------------------------------------- /fmri/ts_gradCPTscene.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/ts_gradCPTscene.mat -------------------------------------------------------------------------------- /fmri/ts_rest1.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/ts_rest1.mat -------------------------------------------------------------------------------- /fmri/ts_rest2.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/ts_rest2.mat -------------------------------------------------------------------------------- /fmri/ts_sitcomep1.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/ts_sitcomep1.mat -------------------------------------------------------------------------------- /fmri/ts_sitcomep2.mat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/hyssong/neuraldynamics/800fffae55a1a157b73f107e627c3cb4fc7517b2/fmri/ts_sitcomep2.mat --------------------------------------------------------------------------------