Source code for neurokit2.microstates.microstates_classify

# -*- coding: utf-8 -*-
import numpy as np

from ..misc import replace




def microstates_classify(segmentation, microstates):
    """**Reorder (sort) the microstates (experimental)**

    Reorder (sort) the microstates (experimental) based on the pattern of values in the vector of
    channels (thus, depends on how channels are ordered).

    Parameters
    ----------
    segmentation : Union[np.array, dict]
        Vector containing the segmentation.
    microstates : Union[np.array, dict]
        Array of microstates maps . Defaults to ``None``.

    Returns
    -------
    segmentation, microstates
        Tuple containing re-ordered input.

    Examples
    ------------
    .. ipython:: python

      import neurokit2 as nk

      eeg = nk.mne_data("filt-0-40_raw").filter(1, 35, verbose=False)
      eeg = nk.eeg_rereference(eeg, 'average')

      # Original order
      out = nk.microstates_segment(eeg)
      @savefig p_microstates_classify.png scale=100%
      nk.microstates_plot(out, gfp=out["GFP"][0:100])
      @suppress
      plt.close()

      # Reorder
      out = nk.microstates_classify(out["Sequence"], out["Microstates"])

    """
    # Reorder
    new_order = _microstates_sort(microstates)
    microstates = microstates[new_order]

    replacement = dict(enumerate(new_order))
    segmentation = replace(segmentation, replacement)

    return segmentation, microstates



# =============================================================================
# Methods
# =============================================================================
def _microstates_sort(microstates):

    n_states = len(microstates)
    order_original = np.arange(n_states)

    # For each state, get linear and quadratic coefficient
    coefs_quadratic = np.zeros(n_states)
    coefs_linear = np.zeros(n_states)
    for i in order_original:
        state = microstates[i, :]
        _, coefs_linear[i], coefs_quadratic[i] = np.polyfit(
            state, np.arange(len(state)), 2
        )

    # For each state, which is the biggest trend, linear or quadratic
    order_quad = order_original[np.abs(coefs_linear) <= np.abs(coefs_quadratic)]
    order_lin = order_original[np.abs(coefs_linear) > np.abs(coefs_quadratic)]

    # Reorder each
    order_quad = order_quad[np.argsort(coefs_quadratic[order_quad])]
    order_lin = order_lin[np.argsort(coefs_linear[order_lin])]

    new_order = np.concatenate([order_quad, order_lin])

    return new_order