{
  "nbformat_minor": 0, 
  "nbformat": 4, 
  "cells": [
    {
      "execution_count": null, 
      "cell_type": "code", 
      "source": [
        "%matplotlib inline"
      ], 
      "outputs": [], 
      "metadata": {
        "collapsed": false
      }
    }, 
    {
      "source": [
        "\n# Decoding real-time data\n\n\nSupervised machine learning applied to MEG data in sensor space.\nHere the classifier is updated every 5 trials and the decoding\naccuracy is plotted\n\n"
      ], 
      "cell_type": "markdown", 
      "metadata": {}
    }, 
    {
      "execution_count": null, 
      "cell_type": "code", 
      "source": [
        "# Authors: Mainak Jas <mainak@neuro.hut.fi>\n#\n# License: BSD (3-clause)\n\nimport numpy as np\nimport matplotlib.pyplot as plt\n\nimport mne\nfrom mne.realtime import MockRtClient, RtEpochs\nfrom mne.datasets import sample\n\nprint(__doc__)\n\n# Fiff file to simulate the realtime client\ndata_path = sample.data_path()\nraw_fname = data_path + '/MEG/sample/sample_audvis_filt-0-40_raw.fif'\nraw = mne.io.read_raw_fif(raw_fname, preload=True)\n\ntmin, tmax = -0.2, 0.5\nevent_id = dict(aud_l=1, vis_l=3)\n\ntr_percent = 60  # Training percentage\nmin_trials = 10  # minimum trials after which decoding should start\n\n# select gradiometers\npicks = mne.pick_types(raw.info, meg='grad', eeg=False, eog=True,\n                       stim=True, exclude=raw.info['bads'])\n\n# create the mock-client object\nrt_client = MockRtClient(raw)\n\n# create the real-time epochs object\nrt_epochs = RtEpochs(rt_client, event_id, tmin, tmax, picks=picks, decim=1,\n                     reject=dict(grad=4000e-13, eog=150e-6), baseline=None,\n                     isi_max=4.)\n\n# start the acquisition\nrt_epochs.start()\n\n# send raw buffers\nrt_client.send_data(rt_epochs, picks, tmin=0, tmax=90, buffer_size=1000)\n\n# Decoding in sensor space using a linear SVM\nn_times = len(rt_epochs.times)\n\nfrom sklearn import preprocessing  # noqa\nfrom sklearn.svm import SVC  # noqa\nfrom sklearn.pipeline import Pipeline  # noqa\nfrom sklearn.cross_validation import cross_val_score, ShuffleSplit  # noqa\nfrom mne.decoding import Vectorizer, FilterEstimator  # noqa\n\n\nscores_x, scores, std_scores = [], [], []\n\n# don't highpass filter because it's epoched data and the signal length\n# is small\nfilt = FilterEstimator(rt_epochs.info, None, 40)\nscaler = preprocessing.StandardScaler()\nvectorizer = Vectorizer()\nclf = SVC(C=1, kernel='linear')\n\nconcat_classifier = Pipeline([('filter', filt), ('vector', vectorizer),\n                              ('scaler', scaler), ('svm', clf)])\n\ndata_picks = mne.pick_types(rt_epochs.info, meg='grad', eeg=False, eog=True,\n                            stim=False, exclude=raw.info['bads'])\nax = plt.subplot(111)\nax.set_xlabel('Trials')\nax.set_ylabel('Classification score (% correct)')\nax.set_title('Real-time decoding')\nax.set_xlim([min_trials, 50])\nax.set_ylim([30, 105])\nplt.axhline(50, color='k', linestyle='--', label=\"Chance level\")\nplt.show(block=False)\n\nfor ev_num, ev in enumerate(rt_epochs.iter_evoked()):\n\n    print(\"Just got epoch %d\" % (ev_num + 1))\n\n    if ev_num == 0:\n        X = ev.data[None, data_picks, :]\n        y = int(ev.comment)  # the comment attribute contains the event_id\n    else:\n        X = np.concatenate((X, ev.data[None, data_picks, :]), axis=0)\n        y = np.append(y, int(ev.comment))\n\n    if ev_num >= min_trials:\n\n        cv = ShuffleSplit(len(y), 5, test_size=0.2, random_state=42)\n        scores_t = cross_val_score(concat_classifier, X, y, cv=cv,\n                                   n_jobs=1) * 100\n\n        std_scores.append(scores_t.std())\n        scores.append(scores_t.mean())\n        scores_x.append(ev_num)\n\n        # Plot accuracy\n\n        plt.plot(scores_x[-2:], scores[-2:], '-x', color='b',\n                 label=\"Classif. score\")\n        ax.hold(True)\n        ax.plot(scores_x[-1], scores[-1])\n\n        hyp_limits = (np.asarray(scores) - np.asarray(std_scores),\n                      np.asarray(scores) + np.asarray(std_scores))\n        fill = plt.fill_between(scores_x, hyp_limits[0], y2=hyp_limits[1],\n                                color='b', alpha=0.5)\n        plt.pause(0.01)\n        plt.draw()\n        ax.collections.remove(fill)  # Remove old fill area\n\nplt.fill_between(scores_x, hyp_limits[0], y2=hyp_limits[1], color='b',\n                 alpha=0.5)\nplt.draw()  # Final figure"
      ], 
      "outputs": [], 
      "metadata": {
        "collapsed": false
      }
    }
  ], 
  "metadata": {
    "kernelspec": {
      "display_name": "Python 2", 
      "name": "python2", 
      "language": "python"
    }, 
    "language_info": {
      "mimetype": "text/x-python", 
      "nbconvert_exporter": "python", 
      "name": "python", 
      "file_extension": ".py", 
      "version": "2.7.13", 
      "pygments_lexer": "ipython2", 
      "codemirror_mode": {
        "version": 2, 
        "name": "ipython"
      }
    }
  }
}