# Allen Institute Software License - This software license is the 2-clause BSD
# license plus a third clause that prohibits redistribution for commercial
# purposes without further permission.
#
# Copyright 2017. Allen Institute. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
#
# 1. Redistributions of source code must retain the above copyright notice,
# this list of conditions and the following disclaimer.
#
# 2. Redistributions in binary form must reproduce the above copyright notice,
# this list of conditions and the following disclaimer in the documentation
# and/or other materials provided with the distribution.
#
# 3. Redistributions for commercial purposes are not permitted without the
# Allen Institute's written permission.
# For purposes of this license, commercial purposes is the incorporation of the
# Allen Institute's software into anything for which you will charge fees or
# other compensation. Contact terms@alleninstitute.org for commercial licensing
# opportunities.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.
#
from .fit_parameters import get_gaussian_fit_single_channel, compute_distance, compute_overlap
from .chisquarerf import chi_square_binary, get_peak_significance, pvalue_to_NLL
from .utilities import upsample_image_to_degrees
import collections
import numpy as np
import sys
[docs]def get_gaussian_fit(rf):
fit_parameters_dict_combined = {'on':collections.defaultdict(list), 'off':collections.defaultdict(list)}
counter = {'on':0, 'off':0}
for on_off_key in ['on', 'off']:
fit_parameters_dict = fit_parameters_dict_combined[on_off_key]
for ci in range(rf[on_off_key]['fdr_mask']['attrs']['number_of_components']):
curr_component_mask = upsample_image_to_degrees(np.logical_not(rf[on_off_key]['fdr_mask']['data'][ci,:,:])) > .5
rf_response = upsample_image_to_degrees(rf[on_off_key]['rts_convolution']['data'].copy())
rf_response[curr_component_mask] = 0
if rf_response.sum() > 0:
get_gaussian_fit_single_channel(rf_response, fit_parameters_dict)
counter[on_off_key] += 1
for ii_off in range(counter['on']):
fit_parameters_dict_combined['on']['distance'].append([None]*counter['off'])
fit_parameters_dict_combined['on']['overlap'].append([None] * counter['off'])
for ii_off in range(counter['off']):
fit_parameters_dict_combined['off']['distance'].append([None] * counter['on'])
fit_parameters_dict_combined['off']['overlap'].append([None]*counter['on'])
for ii_on in range(counter['on']):
for ii_off in range(counter['off']):
center_on = fit_parameters_dict_combined['on']['center_x'][ii_on], fit_parameters_dict_combined['on']['center_y'][ii_on]
center_off = fit_parameters_dict_combined['off']['center_x'][ii_off], fit_parameters_dict_combined['off']['center_y'][ii_off]
curr_distance = compute_distance(center_on, center_off)
fit_parameters_dict_combined['on']['distance'][ii_on][ii_off] = curr_distance
fit_parameters_dict_combined['off']['distance'][ii_off][ii_on] = curr_distance
data_on = fit_parameters_dict_combined['on']['data'][ii_on]
data_off = fit_parameters_dict_combined['off']['data'][ii_off]
curr_overlap = compute_overlap(data_on, data_off)
fit_parameters_dict_combined['on']['overlap'][ii_on][ii_off] = curr_overlap
fit_parameters_dict_combined['off']['overlap'][ii_off][ii_on] = curr_overlap
return fit_parameters_dict_combined, counter
[docs]def run_postprocessing(data, rf):
stimulus = rf['attrs']['stimulus']
# Gaussian fit postprocessing:
fit_parameters_dict_combined, counter = get_gaussian_fit(rf)
for on_off_key in ['on', 'off']:
if counter[on_off_key] > 0:
rf[on_off_key]['gaussian_fit'] = {}
rf[on_off_key]['gaussian_fit']['attrs'] = {}
fit_parameters_dict = fit_parameters_dict_combined[on_off_key]
for key, val in fit_parameters_dict.items():
if key == 'data':
rf[on_off_key]['gaussian_fit']['data'] = np.array(val)
else:
rf[on_off_key]['gaussian_fit']['attrs'][key] = np.array(val)
# Chi squared test statistic postprocessing:
cell_index = rf['attrs']['cell_index']
locally_sparse_noise_template = data.get_stimulus_template(stimulus)
event_array = np.zeros((rf['event_vector']['data'].shape[0], 1), dtype=np.bool)
event_array[:,0] = rf['event_vector']['data']
chi_squared_grid = chi_square_binary(event_array, locally_sparse_noise_template)
alpha = rf['on']['fdr_mask']['attrs']['alpha']
assert rf['off']['fdr_mask']['attrs']['alpha'] == alpha
chi_square_grid_NLL = pvalue_to_NLL(chi_squared_grid)
peak_significance = get_peak_significance(chi_square_grid_NLL, locally_sparse_noise_template, alpha=alpha)
significant = peak_significance[0][0]
min_p = peak_significance[1][0]
pvalues_chi_square = peak_significance[2][0]
best_exclusion_region_mask = peak_significance[3][0]
chi_squared_grid_dict = {
'best_exclusion_region_mask':{'data':best_exclusion_region_mask},
'attrs':{'significant':significant, 'alpha': alpha, 'min_p':min_p},
'pvalues':{'data':pvalues_chi_square}
}
rf['chi_squared_analysis'] = chi_squared_grid_dict
return rf
if __name__ == "__main__":
# csid = 517472416 # triple!
csid = 517526760 # two ON
# csid = 539917553
# csid = 540988186