deploy: braindecode/braindecode@e294fec

Author: bruAristimunha

dev/_downloads/07fcc19ba03226cd3d83d4e40ec44385/auto_examples_python.zip

@@ -15,7 +15,7 @@
       },
       "outputs": [],
       "source": [
-        "# Author: Lukas Gemein <l.gemein@gmail.com>\n#\n# License: BSD (3-clause)\n\nimport tempfile\n\nimport matplotlib.pyplot as plt\nimport mne\nimport numpy as np\nfrom numpy import multiply\n\nfrom braindecode.datasets import TUH\nfrom braindecode.preprocessing import (\n    Preprocessor,\n    create_fixed_length_windows,\n    preprocess,\n)\n\nmne.set_log_level(\"ERROR\")  # avoid messages every time a window is extracted"
+        "# Authors: Lukas Gemein <l.gemein@gmail.com>\n#          Sarthak Tayal <sarthaktayal2@gmail.com>\n#\n# License: BSD (3-clause)\n\nimport tempfile\n\nimport matplotlib.pyplot as plt\nimport mne\nimport numpy as np\nfrom numpy import multiply\n\nfrom braindecode.datasets import TUH\nfrom braindecode.preprocessing import (\n    Preprocessor,\n    create_fixed_length_windows,\n    preprocess,\n)\n\nmne.set_log_level(\"ERROR\")  # avoid messages every time a window is extracted"
       ]
     },
     {
@@ -112,7 +112,7 @@
       "cell_type": "markdown",
       "metadata": {},
       "source": [
-        "### Combining preprocessing steps\n\nNext, we use braindecode's preprocess to combine and execute several preprocessing\nsteps that are executed through 'mne':\n\n- Crop the recordings to a region of interest\n- Re-reference all recordings to 'ar' (requires load)\n- Pick channels of interest\n- Scale signals to micro volts (requires load)\n- Clip outlier values to +/- 800 micro volts (requires load)\n- Resample recordings to a common frequency (requires load)\n\n"
+        "### Combining preprocessing steps\n\nNext, we use braindecode's preprocess to combine and execute several preprocessing\nsteps that are executed through 'mne':\n\n- Crop the recordings to a region of interest\n- Pick channels of interest\n- Re-reference all recordings to average reference (CAR) (requires load)\n- Scale signals to micro volts (requires load)\n- Clip outlier values to +/- 800 micro volts (requires load)\n- Resample recordings to a common frequency (requires load)\n\n"
       ]
     },
     {
@@ -123,7 +123,7 @@
       },
       "outputs": [],
       "source": [
-        "def custom_crop(raw, tmin=0.0, tmax=None, include_tmax=True):\n    # crop recordings to tmin \u2013 tmax. can be incomplete if recording\n    # has lower duration than tmax\n    # by default mne fails if tmax is bigger than duration\n    tmax = min((raw.n_times - 1) / raw.info[\"sfreq\"], tmax)\n    raw.crop(tmin=tmin, tmax=tmax, include_tmax=include_tmax)\n\n\ntmin = 1 * 60\ntmax = 6 * 60\nsfreq = 100\nfactor = 1e6\n\npreprocessors = [\n    Preprocessor(\n        custom_crop, tmin=tmin, tmax=tmax, include_tmax=False, apply_on_array=False\n    ),\n    Preprocessor(\"set_eeg_reference\", ref_channels=\"average\", ch_type=\"eeg\"),\n    Preprocessor(\"pick_channels\", ch_names=short_ch_names, ordered=True),\n    Preprocessor(\n        lambda data: multiply(data, factor), apply_on_array=True\n    ),  # Convert from V to uV\n    Preprocessor(lambda x: np.clip(x, a_min=-800, a_max=800), apply_on_array=True),\n    Preprocessor(\"resample\", sfreq=sfreq),\n]"
+        "def custom_crop(raw, tmin=0.0, tmax=None, include_tmax=True):\n    # crop recordings to tmin \u2013 tmax. can be incomplete if recording\n    # has lower duration than tmax\n    # by default mne fails if tmax is bigger than duration\n    tmax = min((raw.n_times - 1) / raw.info[\"sfreq\"], tmax)\n    raw.crop(tmin=tmin, tmax=tmax, include_tmax=include_tmax)\n\n\ntmin = 1 * 60\ntmax = 6 * 60\nsfreq = 100\nfactor = 1e6\n\npreprocessors = [\n    Preprocessor(\n        custom_crop, tmin=tmin, tmax=tmax, include_tmax=False, apply_on_array=False\n    ),\n    # pick first so the average reference does not pull in artifacts from dropped channels\n    Preprocessor(\"pick_channels\", ch_names=short_ch_names, ordered=True),\n    Preprocessor(\"set_eeg_reference\", ref_channels=\"average\", ch_type=\"eeg\"),\n    Preprocessor(\n        lambda data: multiply(data, factor), apply_on_array=True\n    ),  # Convert from V to uV\n    Preprocessor(lambda x: np.clip(x, a_min=-800, a_max=800), apply_on_array=True),\n    Preprocessor(\"resample\", sfreq=sfreq),\n]"
       ]
     },
     {