tutorial_eventrelated.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:5c7927e6dfd68013edb93ea37ccd422dc2fd93900a0ddbec6b2b0c3fa0688af0"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "heading",
     "level": 1,
     "metadata": {},
     "source": [
      "Event-related Data Analysis"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "In all previous tutorial parts we have analyzed the same fMRI data. We analyzed\n",
      "it using a number of different strategies, but they all had one thing in\n",
      "common: A sample in each dataset was always a single volume from an fMRI time\n",
      "series.  Sometimes, we have limited ourselves to just a specific temporal\n",
      "windows of interest, sometimes we averaged many fMRI volumes into a single one.\n",
      "In all cases, however, a feature always corresponded to a voxel in the fMRI\n",
      "volume and appeared only once in the dataset.\n",
      "\n",
      "In this part we are going to extend the analysis beyond the spatial\n",
      "dimensions and will consider *time* as another aspect of our data.\n",
      "We will demonstrate two different approaches: 1) modeling of experimental\n",
      "conditions and proceed with an analysis of model parameter estimates, and\n",
      "2) the extraction of spatio-temporal data samples. The latter approach is\n",
      "common, for example, in ERP-analyses of EEG data.\n",
      "\n",
      "Let's start with our well-known example dataset -- this time selecting a subset\n",
      "of ventral temporal regions."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "cd data"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "/usr/share/data/pymvpa2-tutorial\n"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from mvpa2.tutorial_suite import *\n",
      "ds = get_raw_haxby2001_data(roi=(36,38,39,40))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "As we know, this dataset consists of 12 concatenated experiment sessions.\n",
      "Every session had a stimulation block spanning multiple fMRI volumes for\n",
      "each of the eight stimulus categories. Stimulation blocks were separated by\n",
      "rest periods."
     ]
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Event-related Pre-processing Is Not Event-related"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "For an event-related analysis, most of the processing is done on data\n",
      "samples that are somehow derived from a set of events. The rest of the data\n",
      "could be considered irrelevant. However, some preprocessing is only\n",
      "meaningful when performed on the full time series and not on the segmented\n",
      "event samples. An example is the detrending that typically needs to be done\n",
      "on the original, continuous time series.\n",
      "\n",
      "In its current shape our datasets consists of samples that represent contiguous\n",
      "fMRI volumes. At this stage we can easily perform linear detrending."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "poly_detrend(ds, polyord=1, chunks_attr='chunks')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 5
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "Let's make a copy of the de-trended dataset that we can use later on for\n",
      "some visualization."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "orig_ds = ds.copy()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 6
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Design Specification"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "For any event-related analysis we need some information on the experiment\n",
      "design: when was stimulated with what for how long (and maybe with what\n",
      "intensity).  In PyMVPA this is done by compiling a list of event definitions.\n",
      "In many cases, an event is defined by *onset*, *duration* and potentially a\n",
      "number of additional properties, such as stimulus condition or recording\n",
      "session number.\n",
      "\n",
      "To see how such events definitions look like, we will simply convert the\n",
      "block-design setup defined by the samples attributes of our dataset into a list\n",
      "of events.  With [find_events()](http://pymvpa.org/generated/mvpa2.datasets.eventrelated.find_events.html#mvpa2-datasets-eventrelated-find-events), PyMVPA\n",
      "provides a function to convert sequential attributes into event lists. In our\n",
      "dataset, we have the stimulus conditions of each volume sample available in the\n",
      "`targets` sample attribute."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "events = find_events(targets=ds.sa.targets, chunks=ds.sa.chunks)\n",
      "print len(events)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "204\n"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for e in events[:4]:\n",
      "   print e"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "{'chunks': 0, 'duration': 6, 'onset': 0, 'targets': 'rest'}\n",
        "{'chunks': 0, 'duration': 9, 'onset': 6, 'targets': 'scissors'}\n",
        "{'chunks': 0, 'duration': 6, 'onset': 15, 'targets': 'rest'}\n",
        "{'chunks': 0, 'duration': 9, 'onset': 21, 'targets': 'face'}\n"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "We are feeding not only the `targets` to the function, but also the\n",
      "`chunks` attribute, since we do not want to have events spanning multiple\n",
      "recording sessions. [find_events()](http://pymvpa.org/generated/mvpa2.datasets.eventrelated.find_events.html#mvpa2-datasets-eventrelated-find-events)\n",
      "sequentially parses all provided attributes and records an event whenever the\n",
      "value in *any* of the attributes changes. The generated event definition is a\n",
      "dictionary that contains:\n",
      "\n",
      "1. Onset of the event as an index in the sequence (in this example this is a\n",
      "volume id)\n",
      "\n",
      "2. Duration of the event in \"number of sequence elements\" (i.e. number of\n",
      "volumes). The duration is determined by counting the number of identical\n",
      "attribute combinations following an event onset.\n",
      "\n",
      "3. Attribute combination of this event, i.e. the actual values of all given\n",
      "attributes at the particular position.\n",
      "\n",
      "Let's limit ourselves to `face` and `house` stimulation blocks for now.\n",
      "We can easily filter out all other events."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "events = [ev for ev in events if ev['targets'] in ['house', 'face']]\n",
      "print len(events)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "24\n"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for e in events[:4]:\n",
      "   print e"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "{'chunks': 0, 'duration': 9, 'onset': 21, 'targets': 'face'}\n",
        "{'chunks': 0, 'duration': 9, 'onset': 63, 'targets': 'house'}\n",
        "{'chunks': 1, 'duration': 9, 'onset': 127, 'targets': 'face'}\n",
        "{'chunks': 1, 'duration': 9, 'onset': 213, 'targets': 'house'}\n"
       ]
      }
     ],
     "prompt_number": 10
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Response Modeling"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "Whenever we have to deal with data were multiple concurrent signals\n",
      "are overlapping in time, such as in fast event-related fMRI studies,\n",
      "it often makes sense to fit an appropriate model to the data and\n",
      "proceed with an analysis of model parameter estimates, instead of\n",
      "the raw data.\n",
      "\n",
      "PyMVPA can make use of NiPy's GLM modeling capabilities. It expects\n",
      "information on stimulation events to be given as actual time stamps\n",
      "and not data sample indices, hence we have to convert our event list."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "TR = np.median(np.diff(ds.sa.time_coords))\n",
      "for ev in events:\n",
      "    ev['onset'] = (ev['onset'] * TR)\n",
      "    ev['duration'] = ev['duration'] * TR"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 11
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "Now we can fit a model of the hemodynamic response to all relevant\n",
      "stimulus conditions. The function\n",
      "[eventrelated_dataset()](http://pymvpa.org/generated/mvpa2.datasets.eventrelated.eventrelated_dataset.html#mvpa2-datasets-eventrelated-eventrelated-dataset) does everything\n",
      "for us. For a given input dataset we need to provide a list of events,\n",
      "the name of an attribute with a time stamp for each sample, and information\n",
      "on what conditions we would like to have modeled. The latter is specified\n",
      "to the `condition_attr` argument. This can be a single attribute name\n",
      "in which case all unique values will be used as conditions. It can also\n",
      "be a sequence of multiple attribute names, and all combinations of unique\n",
      "values of the attributes will be used as conditions. In the following example\n",
      "`('targets', 'chunks')` indicates that we want a separate model for each\n",
      "stimulation condition (`targets`) for each run of our example dataset\n",
      "(`chunks`)."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "evds = fit_event_hrf_model(ds,\n",
      "                           events,\n",
      "                           time_attr='time_coords',\n",
      "                           condition_attr=('targets', 'chunks'))\n",
      "print len(evds)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "24\n"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "This yields one parameter estimate sample for each target value for each\n",
      "chunks."
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "**Exercise**\n",
      "\n",
      "\n",
      "\n",
      "Explore the `evds` dataset. It contains the generated HRF model.\n",
      "Find and plot (some of) them. Take a look at the parameter estimate\n",
      "samples themselves -- can you spot a pattern?"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# you can use this cell for this exercise\n",
      "evds.shape"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 15,
       "text": [
        "(24, 941)"
       ]
      }
     ],
     "prompt_number": 15
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print evds"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "<Dataset: 24x941@float64, <sa: chunks,regressors,targets>, <fa: voxel_indices>, <a: add_regs,imgaffine,imghdr,imgtype,mapper,voxel_dim,voxel_eldim>>\n"
       ]
      }
     ],
     "prompt_number": 22
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%matplotlib inline"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 20
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "pl.plot(np.mean(evds, axis=0))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 21,
       "text": [
        "[<matplotlib.lines.Line2D at 0x7f43c3512950>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEACAYAAAC6d6FnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQkERARAEQSgoIgCCIAgF5f8DFVDxRBk2dPIAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f43c6c3f4d0>"
       ]
      }
     ],
     "prompt_number": 21
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "pl.plot(evds.sa.regressors[0][:100])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 24,
       "text": [
        "[<matplotlib.lines.Line2D at 0x7f43c3852490>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEACAYAAAC6d6FnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nADAz61IOADOzLuUAMDPrUv8f+HqcDwxFhGoAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f43c6c3f510>"
       ]
      }
     ],
     "prompt_number": 24
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print evds.a.add_regs.sa"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "<SampleAttributesCollection: regressor_names,regressors>\n"
       ]
      }
     ],
     "prompt_number": 25
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "pl.plot(evds.a.add_regs.sa.regressors[-2])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 26,
       "text": [
        "[<matplotlib.lines.Line2D at 0x7f43c36ad150>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYkAAAEACAYAAABGYoqtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f43c34856d0>"
       ]
      }
     ],
     "prompt_number": 26
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "pl.matshow(evds.samples)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 33,
       "text": [
        "<matplotlib.image.AxesImage at 0x7f43c603ff10>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAA5oAAAA0CAYAAADrC0eIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f43c60db490>"
       ]
      }
     ],
     "prompt_number": 33
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "evds.targets"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 32,
       "text": [
        "array(['face', 'face', 'face', 'face', 'face', 'face', 'face', 'face',\n",
        "       'face', 'face', 'face', 'face', 'house', 'house', 'house', 'house',\n",
        "       'house', 'house', 'house', 'house', 'house', 'house', 'house',\n",
        "       'house'], \n",
        "      dtype='|S5')"
       ]
      }
     ],
     "prompt_number": 32
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "\n",
      "\n",
      "Before we can run a classification analysis we still need to normalize each\n",
      "feature (GLM parameters estimates for each voxel at this point)."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "zscore(evds, chunks_attr=None)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 27
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "The rest is straight-forward: we set up a cross-validation analysis with\n",
      "a chosen classifier and run it:"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf = kNN(k=1, dfx=one_minus_correlation, voting='majority')\n",
      "cv = CrossValidation(clf, NFoldPartitioner(attr='chunks'))\n",
      "cv_glm = cv(evds)\n",
      "print '%.2f' % np.mean(cv_glm)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "0.04\n"
       ]
      }
     ],
     "prompt_number": 28
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "Not bad! Let's compare that to a simpler approach that is also suitable for\n",
      "block-design experiments like this one."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "zscore(ds, param_est=('targets', ['rest']))\n",
      "avgds = ds.get_mapped(mean_group_sample(['targets', 'chunks']))\n",
      "avgds = avgds[np.array([t in ['face', 'house'] for t in avgds.sa.targets])]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 29
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "We normalize all voxels with respect to the `rest` condition. This yields\n",
      "some crude kind of \"activation\" score for all stimulation conditions.\n",
      "Subsequently, we average all sample of a condition in each run. This yield\n",
      "a dataset of the same size as from the GLM modeling. We can re-use the\n",
      "cross-validation setup."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "cv_avg = cv(avgds)\n",
      "print '%.2f' % np.mean(cv_avg)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "0.04\n"
       ]
      }
     ],
     "prompt_number": 30
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "Not bad either. However, it is worth repeating that this simple average-sample\n",
      "approach is limited to block-designs with a clear temporal separation of\n",
      "all signals of interest, whereas the HRF modeling is more suitable for\n",
      "experiments with fast stimulation alternation."
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "**Exercise**\n",
      "\n",
      "\n",
      "\n",
      "Think about what need to be done to perform odd/even run GLM modeling."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "ds.sa['oddeven'] = [['even', 'odd'][c%2] for c in ds.sa.chunks]\n",
      "print ds.sa.oddeven"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "['even' 'even' 'even' ..., 'odd' 'odd' 'odd']\n"
       ]
      }
     ],
     "prompt_number": 36
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for e in events:\n",
      "    e['oddeven'] = ['even', 'old'][e['chunks'] % 2]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 37
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# you can use this cell for this exercise\n",
      "# print evds.summary()\n",
      "evds2 = fit_event_hrf_model(ds,\n",
      "                           events,\n",
      "                           time_attr='time_coords',\n",
      "                           condition_attr=('targets', 'oddeven'))\n",
      "# print len(evds)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 41
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print len(evds2)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "4\n"
       ]
      }
     ],
     "prompt_number": 42
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "pl.plot(evds2.sa.regressors[0])\n",
      "pl.plot(evds2.sa.regressors[1])\n",
      "# clf2 = kNN(k=1, dfx=one_minus_correlation, voting='majority')\n",
      "# cv2 = CrossValidation(clf2, NFoldPartitioner(attr='oddeveb'))\n",
      "# cv2_glm = cv(evds2)\n",
      "# print '%.2f' % np.mean(cv2_glm)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 52,
       "text": [
        "[<matplotlib.lines.Line2D at 0x7f43c4127990>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAEACAYAAABRQBpkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nnEEQBEFAOIMgCIIA+P/xOjqSneSIOgAAAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f43c408f3d0>"
       ]
      }
     ],
     "prompt_number": 52
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n"
     ]
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "From Timeseries To Spatio-temporal Samples"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "Now we want to try something different. Instead of compressing all temporal\n",
      "information into a single model parameter estimate, we can also consider the\n",
      "entire spatio-temporal signal across our region of interest and the full\n",
      "duration of the stimulation blocks. In other words, we can perform a\n",
      "sensitivity analysis (see *chap_tutorial_sensitivity*) revealing the\n",
      "spatio-temporal distribution of classification-relevant information.\n",
      "\n",
      "Before we start with our event-extraction, we want to normalize each feature\n",
      "(i.e. a voxel at this point). In this case we are, again, going to Z-score\n",
      "them, using the mean and standard deviation from the experiment's rest\n",
      "condition, and the resulting values might be interpreted as \"activation\n",
      "scores\"."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "zscore(ds, chunks_attr='chunks', param_est=('targets', 'rest'))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 60
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "For this analysis we do not have to convert event onset information into\n",
      "time-stamp, but can operate on sample indices, hence we start with the\n",
      "original event list again."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "events = find_events(targets=ds.sa.targets, chunks=ds.sa.chunks)\n",
      "events = [ev for ev in events if ev['targets'] in ['house', 'face']]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 61
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "All of our events are of the same length, 9 consecutive fMRI volumes. Later\n",
      "on we would like to view the temporal sensitivity profile from *before* until\n",
      "*after* the stimulation block, hence we should extend the duration of the\n",
      "events a bit."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "event_duration = 13\n",
      "for ev in events:\n",
      "    ev['onset'] -= 2\n",
      "    ev['duration'] = event_duration"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 62
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "The next and most important step is to actually segment the original\n",
      "time series dataset into event-related samples. PyMVPA offers\n",
      "[eventrelated_dataset()](http://pymvpa.org/generated/mvpa2.datasets.eventrelated.eventrelated_dataset.html#mvpa2-datasets-eventrelated-eventrelated-dataset) as a function to\n",
      "perform this conversion. Let's just do it, it only needs the original\n",
      "dataset and our list of events."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "evds = eventrelated_dataset(ds, events=events)\n",
      "len(evds) == len(events)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 63,
       "text": [
        "True"
       ]
      }
     ],
     "prompt_number": 63
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "evds.nfeatures == ds.nfeatures * event_duration"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 64,
       "text": [
        "True"
       ]
      }
     ],
     "prompt_number": 64
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print evds.a.mapper[-2:].reverse1(evds[0]).shape"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "(1, 12233)\n"
       ]
      }
     ],
     "prompt_number": 74
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "**Exercise**\n",
      "\n",
      "\n",
      "\n",
      "Inspect the `evds` dataset. It has a fairly large number of attributes\n",
      "-- both for samples and for features. Look at each of them and think\n",
      "about what it could be useful for."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# you can use this cell for this exercise\n",
      "print evds.fa.event_offsetidx"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[ 0  0  0 ..., 12 12 12]\n"
       ]
      }
     ],
     "prompt_number": 58
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "\n",
      "\n",
      "At this point it is worth looking at the dataset's mapper -- in particular at\n",
      "the last two items in the chain mapper that have been added during the\n",
      "conversion into events."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print evds.a.mapper[-2:]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "**Exercise**\n",
      "\n",
      "\n",
      "\n",
      "Reverse-map a single sample through the last two items in the chain\n",
      "mapper. Inspect the result and make sure it doesn't surprise. Now,\n",
      "reverse-map multiple samples at once and compare the result. Is this what\n",
      "you would expect?"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# you can use this cell for this exercise"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "\n",
      "\n",
      "The rest of our analysis is business as usual and is quickly done.  We want to\n",
      "perform a cross-validation analysis of an SVM classifier. We are not\n",
      "primarily interested in its performance, but in the weights it assigns to\n",
      "the features. Remember, each feature is now voxel-at-time-point, so we get a\n",
      "chance of looking at the spatio-temporal profile of classification-relevant\n",
      "information in the data. We will nevertheless enable computing of a confusion\n",
      "matrix, so we can assure ourselves that the classifier is performing\n",
      "reasonably well, because only a generalizing model is worth\n",
      "inspecting, as otherwise it overfits and the assigned weights\n",
      "could be meaningless."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "sclf = SplitClassifier(LinearCSVMC(),\n",
      "                       enable_ca=['stats'])\n",
      "sensana = sclf.get_sensitivity_analyzer()\n",
      "sens = sensana(evds)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 75
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "sens.shape"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 76,
       "text": [
        "(12, 12233)"
       ]
      }
     ],
     "prompt_number": 76
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "**Exercise**\n",
      "\n",
      "\n",
      "\n",
      "Check that the classifier achieves an acceptable accuracy. Is it\n",
      "enough above chance level to allow for an interpretation of the\n",
      "sensitivities?"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# you can use this cell for this exercise"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 77
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "**Exercise**\n",
      "\n",
      "\n",
      "\n",
      "Using what you have learned in the last tutorial part: Combine the\n",
      "sensitivity maps for all splits into a single map. Project this map into\n",
      "the original dataspace. What is the shape of that space? Store the\n",
      "projected map into a NIfTI file and inspect it using an MRI viewer.\n",
      "Viewer needs to be capable of visualizing time series (hint: for FSLView\n",
      "the time series image has to be opened first)!"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print sens"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "<Dataset: 12x12233@float64, <sa: biases,splits,targets>>\n"
       ]
      }
     ],
     "prompt_number": 81
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# you can use this cell for this exercise\n",
      "evds.a.mapper[-2:].reverse1(sens).shape"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 82,
       "text": [
        "(156, 941)"
       ]
      }
     ],
     "prompt_number": 82
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n"
     ]
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "A Plotting Example"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "We have inspected the spatio-temporal profile of the sensitivities using\n",
      "some MRI viewer application, but we can also assemble an informative figure\n",
      "right here. Let's compose a figure that shows the original peri-stimulus\n",
      "time series, the effect of normalization, as well as the corresponding\n",
      "sensitivity profile of the trained SVM classifier. We are going to do that\n",
      "for two example voxels, whose coordinates we might have derived from\n",
      "inspecting the full map."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "example_voxels = [(28,25,25), (28,23,25)]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 78
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "The plotting will be done by the popular [matplotlib](http://matplotlib.sourceforge.net/) package.\n",
      "\n",
      "First, we plot the original signal after initial detrending. To do this, we\n",
      "apply the same time series segmentation to the original detrended dataset\n",
      "and plot the mean signal for all face and house events for both of our\n",
      "example voxels. The code below will create the plot using matplotlib's\n",
      "`pylab` interface (imported as `pl`). If you are familiar with Matlab's\n",
      "plotting facilities, this shouldn't be hard to read."
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "*Note*\n",
      "\n",
      "\n",
      "\n",
      "`_ =` is used in the examples below simply to absorb output of plotting\n",
      "functions.  You do not have to swallow output in your interactive sessions.\n",
      "- - -\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "vx_lty = ['-', '--']\n",
      "t_col = ['b', 'r']\n",
      "for i, v in enumerate(example_voxels):\n",
      "    # get a slicing array matching just to current example voxel\n",
      "    slicer = np.array([tuple(idx) == v for idx in ds.fa.voxel_indices])\n",
      "    # perform the timeseries segmentation just for this voxel\n",
      "    evds_detrend = eventrelated_dataset(orig_ds[:, slicer], events=events)\n",
      "    # now plot the mean timeseries and standard error\n",
      "    for j, t in enumerate(evds.uniquetargets):\n",
      "        l = plot_err_line(evds_detrend[evds_detrend.sa.targets == t].samples,\n",
      "                          fmt=t_col[j], linestyle=vx_lty[i])\n",
      "        # label this plot for automatic legend generation\n",
      "        l[0][0].set_label('Voxel %i: %s' % (i, t))\n",
      "_ = pl.ylabel('Detrended signal')\n",
      "_ = pl.axhline(linestyle='--', color='0.6')\n",
      "# pl.plot(sens_mapped[:,792]*10000, '--')\n",
      "# _ = pl.legend()\n",
      "_ = pl.xlim((0,12))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEACAYAAACpoOGTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUQWhKIqiGPJ/WI0BdQ73nCYAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f43c3c56710>"
       ]
      }
     ],
     "prompt_number": 85
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "In the next figure we do exactly the same again, but this time for the\n",
      "normalized data."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for i, v in enumerate(example_voxels):\n",
      "    slicer = np.array([tuple(idx) == v for idx in ds.fa.voxel_indices])\n",
      "    evds_norm = eventrelated_dataset(ds[:, slicer], events=events)\n",
      "    for j, t in enumerate(evds.uniquetargets):\n",
      "        l = plot_err_line(evds_norm[evds_norm.sa.targets == t].samples,\n",
      "                          fmt=t_col[j], linestyle=vx_lty[i])\n",
      "        l[0][0].set_label('Voxel %i: %s' % (i, t))\n",
      "_ = pl.ylabel('Normalized signal')\n",
      "_ = pl.axhline(linestyle='--', color='0.6')\n",
      "_ = pl.xlim((0,12))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "Finally, we plot the associated SVM weight profile for each peri-stimulus\n",
      "time-point of both voxels. For easier selection we do a little trick and\n",
      "reverse-map the sensitivity profile through the last mapper in the\n",
      "dataset's chain mapper (look at `evds.a.mapper` for the whole chain).\n",
      "This will reshape the sensitivities into `cross-validation fold x volume x\n",
      "voxel features`."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "normed = sens.get_mapped(FxMapper(axis='features', fx=l1_normed))\n",
      "smaps = evds.a.mapper[-1].reverse(normed)\n",
      "for i, v in enumerate(example_voxels):\n",
      "    slicer = np.array([tuple(idx) == v for idx in ds.fa.voxel_indices])\n",
      "    smap = smaps.samples[:,:,slicer].squeeze()\n",
      "    l = plot_err_line(smap, fmt='ko', linestyle=vx_lty[i], errtype='std')\n",
      "_ = pl.xlim((0,12))\n",
      "_ = pl.ylabel('Sensitivity')\n",
      "_ = pl.axhline(linestyle='--', color='0.6')\n",
      "_ = pl.xlabel('Peristimulus volumes')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "\n",
      "\n",
      "That was it. Perhaps you are scared by the amount of code. Please note that\n",
      "it could have been done shorter, but this way allows for plotting any other voxel\n",
      "coordinate combination as well. matplotlib also allows for saving this figure in\n",
      "[SVG](http://en.wikipedia.org/wiki/Scalable_Vector_Graphics) format, allowing for convenient post-processing in [Inkscape](http://www.inkscape.org/) -- a\n",
      "publication quality figure is only minutes away.\n",
      "\n",
      "\\[Visit [http://pymvpa.org/examples/tutorial_eventrelated.html](http://pymvpa.org/examples/tutorial_eventrelated.html) to view this figure\\]"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "**Exercise**\n",
      "\n",
      "\n",
      "\n",
      "What can we say about the properties of the example voxel's signal from\n",
      "the peri-stimulus plot?"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# you can use this cell for this exercise"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- - -\n",
      "\n",
      "\n",
      "This demo showed an event-related data analysis. Although we have performed\n",
      "it on fMRI data, an analogous analysis can be done for any time series based\n",
      "data in an almost identical fashion. Moreover, if a dataset has information\n",
      "about acquisition time (e.g. like the ones created by\n",
      "[fmri_dataset()](http://pymvpa.org/generated/mvpa2.datasets.mri.fmri_dataset.html#mvpa2-datasets-mri-fmri-dataset))\n",
      "[eventrelated_dataset()](http://pymvpa.org/generated/mvpa2.datasets.eventrelated.eventrelated_dataset.html#mvpa2-datasets-eventrelated-eventrelated-dataset) can also convert\n",
      "event-definition in real time, making it relatively easy to \"convert\"\n",
      "experiment design logfiles into event lists. In this case there would be no\n",
      "need to run a function like\n",
      "[find_events()](http://pymvpa.org/generated/mvpa2.datasets.eventrelated.find_events.html#mvpa2-datasets-eventrelated-find-events), but instead they could be\n",
      "directly specified and passed to\n",
      "[eventrelated_dataset()](http://pymvpa.org/generated/mvpa2.datasets.eventrelated.eventrelated_dataset.html#mvpa2-datasets-eventrelated-eventrelated-dataset)."
     ]
    }
   ],
   "metadata": {}
  }
 ]
}