-
7a1d2e6a
M2E_visualize_interareal_connectivity.py 13.43 KiB
import json
import numpy as np
import matplotlib.pyplot as pl
import os
import sys
sys.path.append('./figures/Schmidt2018')
from helpers import area_list, datapath
from matplotlib import gridspec
from matplotlib.colors import LogNorm
from matplotlib.ticker import FixedLocator
from matplotlib import rc_file
from multiarea_model import MultiAreaModel
from plotcolors import myblue
from scipy import stats
# rc_file('plotstyle.rc')
def visualize_interareal_connectivity(M):
# scale_down_to = 1
# cc_weights_factor = 1.0
# areas_simulated = ['V1', 'V2', 'VP', 'V3', 'V3A', 'MT', 'V4t', 'V4', 'VOT', 'MSTd',
# 'PIP', 'PO', 'DP', 'MIP', 'MDP', 'VIP', 'LIP', 'PITv', 'PITd',
# 'MSTl', 'CITv', 'CITd', 'FEF', 'TF', 'AITv', 'FST', '7a', 'STPp',
# 'STPa', '46', 'AITd', 'TH']
# replace_non_simulated_areas = 'het_poisson_stat'
# conn_params = {
# 'replace_non_simulated_areas': 'het_poisson_stat',
# 'g': -11.,
# 'K_stable': 'K_stable.npy',
# 'fac_nu_ext_TH': 1.2,
# 'fac_nu_ext_5E': 1.125,
# 'fac_nu_ext_6E': 1.41666667,
# 'av_indegree_V1': 3950.
# }
# input_params = {
# 'rate_ext': 10.
# }
# neuron_params = {
# 'V0_mean': -150.,
# 'V0_sd': 50.}
# network_params = {
# 'N_scaling': scale_down_to, # Scaling of population sizes, by default: 1.
# 'K_scaling': scale_down_to, # Scaling of indegrees, by default: 1.
# 'fullscale_rates': 'tests/fullscale_rates.json',
# 'input_params': input_params, # Input parameters
# 'connection_params': conn_params, # Connection parameters
# 'neuron_params': neuron_params # Neuron parameters
# }
# sim_params = {
# 'areas_simulated': areas_simulated,
# 't_sim': 2000., # Simulated time (in ms), by default: 10.0
# 'num_processes': 1, # The number of MPI processes, by default: 1
# 'local_num_threads': 1, # The number of threads per MPI process, by default: 1
# 'recording_dict': {'record_vm': False},
# 'rng_seed': 1 # global random seed
# }
# theory_params = {
# 'dt': 0.1 # The time step of the mean-field theory integration, by default: 0.01
# }
# M_full_scale = MultiAreaModel(network_params,
# simulation=True,
# sim_spec=sim_params,
# theory=True,
# theory_spec=theory_params)
"""
Figure layout
"""
# nrows = 2
nrows = 1
ncols = 2
# width = 6.8556
width = 15
panel_wh_ratio = 0.7 * (1. + np.sqrt(5)) / 2. # golden ratio
height = width / panel_wh_ratio * float(nrows) / ncols
# print(width, height)
pl.rcParams['figure.figsize'] = (width, height)
fig = pl.figure()
fig.suptitle('Your Title Here', fontsize=16)
axes = {}
# gs1 = gridspec.GridSpec(2, 2)
gs1 = gridspec.GridSpec(1, 2)
# gs1.update(left=0.06, right=0.95, top=0.95, bottom=0.1, wspace=0.1, hspace=0.3)
gs1.update(left=0.06, right=0.95, top=0.95, bottom=0.1, wspace=0.3, hspace=0.3)
# axes['A'] = pl.subplot(gs1[:1, :1])
# axes['B'] = pl.subplot(gs1[:1, 1:2])
axes['B'] = pl.subplot(gs1[:1, :1])
axes['D'] = pl.subplot(gs1[:1, 1:2])
# pos = axes['A'].get_position()
pos2 = axes['D'].get_position()
# axes['C'] = pl.axes([pos.x0 + 0.01, pos2.y0, pos.x1 - pos.x0 - 0.025, 0.23])
# print(pos.x1 - pos.x0 - 0.025)
# labels = ['A', 'B', 'C', 'D']
labels = ['B', 'D']
labels_display = ['Full-scale model', 'Down-scale model']
# for label in labels:
for i in range(labels):
label = labels[i]
label_display = labels_display[i]
if label in ['C']:
label_pos = [-0.045, 1.18]
else:
label_pos = [-0.2, 1.04]
# pl.text(label_pos[0], label_pos[1], r'\bfseries{}' + label,
# fontdict={'fontsize': 10, 'weight': 'bold',
# 'horizontalalignment': 'left', 'verticalalignment':
# 'bottom'}, transform=axes[label].transAxes)
pl.text(label_pos[0], label_pos[1], label_display,
fontdict={'fontsize': 10, 'weight': 'bold',
'horizontalalignment': 'left', 'verticalalignment':
'bottom'}, transform=axes[label].transAxes)
# """
# Load data
# """
# M = MultiAreaModel({})
# with open(os.path.join(datapath, 'viscortex_processed_data.json'), 'r') as f:
# proc = json.load(f)
# with open(os.path.join(datapath, 'viscortex_raw_data.json'), 'r') as f:
# raw = json.load(f)
# FLN_Data_FV91 = proc['FLN_Data_FV91']
# cocomac_data = raw['cocomac_data']
# median_distance_data = raw['median_distance_data']
# cocomac = np.zeros((32, 32))
# conn_matrix = np.zeros((32, 32))
# for i, area1 in enumerate(area_list[::-1]):
# for j, area2 in enumerate(area_list):
# if M.K_areas[area1][area2] > 0. and area2 in cocomac_data[area1]:
# cocomac[i][j] = 1.
# if area2 in FLN_Data_FV91[area1]:
# conn_matrix[i][j] = FLN_Data_FV91[area1][area2]
# """
# Panel A: CoCoMac Data
# """
# ax = axes['A']
# ax.yaxis.set_ticks_position("left")
# ax.xaxis.set_ticks_position("bottom")
# ax.set_aspect(1. / ax.get_data_ratio())
# ax.yaxis.set_ticks_position("none")
# ax.xaxis.set_ticks_position("none")
# masked_matrix = np.ma.masked_values(cocomac, 0.0)
# cmap = pl.cm.binary
# cmap.set_bad('w', 1.0)
# x = np.arange(0, len(area_list) + 1)
# y = np.arange(0, len(area_list[::-1]) + 1)
# X, Y = np.meshgrid(x, y)
# ax.set_xticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_xticklabels(area_list, rotation=90, size=6.)
# ax.set_yticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_yticklabels(area_list[::-1], size=6.)
# ax.set_ylabel('Target area')
# ax.set_xlabel('Source area')
# im = ax.pcolormesh(masked_matrix, cmap=cmap,
# edgecolors='None', vmin=0., vmax=1.)
# t = FixedLocator([])
# cbar = pl.colorbar(im, ticks=t, fraction=0.046, ax=ax)
# cbar.set_alpha(0.)
# cbar.remove()
# """
# Panel B: Data from Markov et al. (2014) "A weighted and directed
# interareal connectivity matrix for macaque cerebral cortex."
# Cerebral Cortex, 24(1), 17–36.
# """
# ax = axes['B']
# ax.set_aspect(1. / ax.get_data_ratio())
# ax.yaxis.set_ticks_position("none")
# ax.xaxis.set_ticks_position("none")
# masked_matrix = np.ma.masked_values(conn_matrix, 0.0)
# cmap = pl.get_cmap('inferno')
# cmap.set_bad('w', 1.0)
# x = np.arange(0, len(area_list) + 1)
# y = np.arange(0, len(area_list[::-1]) + 1)
# X, Y = np.meshgrid(x, y)
# ax.set_xticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_xticklabels(area_list, rotation=90, size=6.)
# ax.set_yticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_yticklabels(area_list[::-1], size=6.)
# im = ax.pcolormesh(masked_matrix, cmap=cmap,
# edgecolors='None', norm=LogNorm(vmin=1e-6, vmax=1.))
# t = FixedLocator([1e-6, 1e-4, 1e-2, 1])
# cbar = pl.colorbar(im, ticks=t, fraction=0.046, ax=ax)
# cbar.set_alpha(0.)
"""
Panel B: Interareal connectivity of full-scaling multi-area model
"""
conn_matrix_full_scale = np.zeros((32, 32))
for i, area1 in enumerate(area_list[::-1]):
for j, area2 in enumerate(area_list):
conn_matrix_full_scale[i][j] = M.K_areas[area1][
area2] / np.sum(list(M.K_areas[area1].values()))
ax = axes['B']
ax.yaxis.set_ticks_position("none")
ax.xaxis.set_ticks_position("none")
ax.set_aspect(1. / ax.get_data_ratio())
masked_matrix_full_scale = np.ma.masked_values(conn_matrix_full_scale, 0.0)
cmap = pl.get_cmap('inferno')
cmap.set_bad('w', 1.0)
x = np.arange(0, len(area_list) + 1)
y = np.arange(0, len(area_list[::-1]) + 1)
X, Y = np.meshgrid(x, y)
ax.set_xticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_xticklabels(area_list, rotation=90, size=6.)
ax.set_yticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_yticklabels(area_list[::-1], size=6.)
ax.set_ylabel('Target area')
ax.set_xlabel('Source area')
im = ax.pcolormesh(masked_matrix_full_scale, cmap=cmap,
edgecolors='None', norm=LogNorm(vmin=1e-6, vmax=1.))
t = FixedLocator([1e-6, 1e-4, 1e-2, 1])
cbar = pl.colorbar(im, ticks=t, fraction=0.046, ax=ax)
cbar.set_alpha(0.)
# """
# Panel C: Exponential decay of FLN with distance
# """
# FLN_values_FV91 = np.array([])
# distances_FV91 = np.array([])
# for target_area in FLN_Data_FV91:
# for source_area in FLN_Data_FV91[target_area]:
# if target_area in median_distance_data and source_area in median_distance_data:
# if FLN_Data_FV91[target_area][source_area]:
# FLN_values_FV91 = np.append(FLN_values_FV91, FLN_Data_FV91[
# target_area][source_area])
# distances_FV91 = np.append(distances_FV91, median_distance_data[
# target_area][source_area])
# # Linear fit of distances vs. log FLN
# print("\n \n Linear fit to logarithmic values")
# gradient, intercept, r_value, p_value, std_err = stats.linregress(
# distances_FV91, np.log(FLN_values_FV91))
# print("Raw parameters: ", gradient, intercept)
# print("Transformed parameters: ", -gradient, np.exp(intercept))
# print('r_value**2', r_value ** 2)
# print('p_value', p_value)
# print('std_err', std_err)
# ax = axes['C']
# ax.yaxis.set_ticks_position("left")
# ax.xaxis.set_ticks_position("bottom")
# ax.yaxis.set_ticks_position("left")
# ax.xaxis.set_ticks_position("bottom")
# ax.spines['right'].set_color('none')
# ax.spines['top'].set_color('none')
# ax.yaxis.set_ticks_position("left")
# ax.xaxis.set_ticks_position("bottom")
# ax.plot(distances_FV91, np.log10(FLN_values_FV91), '.', color=myblue)
# x = np.arange(np.min(distances_FV91), np.max(distances_FV91), 1)
# ax.plot(x, (intercept + gradient * x) / np.log(10), linewidth=2.0,
# color='Black', label='Linear regression fit')
# ax.set_xlabel('Distance (mm)', labelpad=7)
# ax.set_ylabel(r'$\log(FLN)$')
# ax.set_yticks([-6, -4, -2, 0])
# print("log fit")
# print(np.corrcoef(gradient * distances_FV91 + intercept, np.log(FLN_values_FV91))[0][1])
# """
# Panel D: Resulting connectivity matrix
# """
# conn_matrix = np.zeros((32, 32))
# for i, area1 in enumerate(area_list[::-1]):
# for j, area2 in enumerate(area_list):
# conn_matrix[i][j] = M.K_areas[area1][
# area2] / np.sum(list(M.K_areas[area1].values()))
# ax = axes['D']
# ax.yaxis.set_ticks_position("none")
# ax.xaxis.set_ticks_position("none")
# ax.set_aspect(1. / ax.get_data_ratio())
# masked_matrix = np.ma.masked_values(conn_matrix, 0.0)
# cmap = pl.get_cmap('inferno')
# cmap.set_bad('w', 1.0)
# x = np.arange(0, len(area_list) + 1)
# y = np.arange(0, len(area_list[::-1]) + 1)
# X, Y = np.meshgrid(x, y)
# ax.set_xticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_xticklabels(area_list, rotation=90, size=6.)
# ax.set_yticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_yticklabels(area_list[::-1], size=6.)
# ax.set_ylabel('Target area')
# ax.set_xlabel('Source area')
# im = ax.pcolormesh(masked_matrix, cmap=cmap,
# edgecolors='None', norm=LogNorm(vmin=1e-6, vmax=1.))
# t = FixedLocator([1e-6, 1e-4, 1e-2, 1])
# cbar = pl.colorbar(im, ticks=t, fraction=0.046, ax=ax)
# cbar.set_alpha(0.)
"""
Panel D: Interareal connectivity of down-scaling multi-area model
"""
conn_matrix_down_scale = np.zeros((32, 32))
for i, area1 in enumerate(area_list[::-1]):
for j, area2 in enumerate(area_list):
conn_matrix_down_scale[i][j] = M.K_areas[area1][
area2] / np.sum(list(M.K_areas[area1].values()))
ax = axes['D']
ax.yaxis.set_ticks_position("none")
ax.xaxis.set_ticks_position("none")
ax.set_aspect(1. / ax.get_data_ratio())
masked_matrix_down_scale = np.ma.masked_values(conn_matrix_down_scale, 0.0)
cmap = pl.get_cmap('inferno')
cmap.set_bad('w', 1.0)
x = np.arange(0, len(area_list) + 1)
y = np.arange(0, len(area_list[::-1]) + 1)
X, Y = np.meshgrid(x, y)
ax.set_xticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_xticklabels(area_list, rotation=90, size=6.)
ax.set_yticks([i + 0.5 for i in np.arange(0, len(area_list) + 1, 1)])
# ax.set_yticklabels(area_list[::-1], size=6.)
ax.set_ylabel('Target area')
ax.set_xlabel('Source area')
im = ax.pcolormesh(masked_matrix_down_scale, cmap=cmap,
edgecolors='None', norm=LogNorm(vmin=1e-6, vmax=1.))
t = FixedLocator([1e-6, 1e-4, 1e-2, 1])
cbar = pl.colorbar(im, ticks=t, fraction=0.046, ax=ax)
cbar.set_alpha(0.)
# """
# Save figure
# """
# pl.savefig('Fig4_connectivity.eps')