Note
Go to the end to download the full example code.
Network sketchesΒΆ
Diagrams for illustrating network structure and connectivity.
import mesocircuit
from mesocircuit.parameterization.base_network_params import net_dict
from mesocircuit.parameterization.base_plotting_params import plot_dict
from mesocircuit.parameterization.base_plotting_params import rcParams
import os
import numpy as np
import matplotlib
import matplotlib.gridspec as gridspec
import matplotlib.pyplot as plt
from matplotlib.patches import Rectangle, Circle, RegularPolygon, FancyArrowPatch
import mpl_toolkits.mplot3d.art3d as art3d
from mpl_toolkits.mplot3d.proj3d import proj_transform
from mpl_toolkits.mplot3d.axes3d import Axes3D
matplotlib.use('Agg')
matplotlib.rcParams.update(rcParams)
def figure_network_model_sketches(output_dir):
"""
Network sketches for manuscript.
Reference model and upscaled model.
Parameters
----------
output_dir
Output directory.
"""
fig = plt.figure(figsize=(plot_dict['fig_width_2col'], 4))
gs = gridspec.GridSpec(1, 2)
gs.update(left=-0.05, right=1.05, bottom=0, top=1, hspace=0, wspace=0)
plot_network_model_sketch(gs[0], model='reference')
ax_r = plt.gca()
ax_r.text2D(0.15, 0.95, 'A',
ha='left', va='bottom',
transform=ax_r.transAxes,
weight='bold',
fontsize=matplotlib.rcParams['font.size'] * 1.2)
ax_r.text2D(0.5, 0.95, 'reference model',
ha='center', va='bottom',
transform=ax_r.transAxes,
fontsize=matplotlib.rcParams['font.size'] * 1.2)
plot_network_model_sketch(gs[1], model='upscaled')
ax_u = plt.gca()
ax_u.text2D(0.15, 0.95, 'B',
ha='left', va='bottom',
transform=ax_u.transAxes,
weight='bold',
fontsize=matplotlib.rcParams['font.size'] * 1.2)
ax_u.text2D(0.5, 0.95, 'upscaled model',
ha='center', va='bottom',
transform=ax_u.transAxes,
fontsize=matplotlib.rcParams['font.size'] * 1.2)
plt.savefig(os.path.join(output_dir, f'network_sketches.pdf'))
return
def figure_network_model_sketch(output_dir, model='upscaled'):
"""
Figure with a single network sketch of a given model name.
Parameters
----------
output_dir
Output directory.
model
Model name. Options are 'reference' or 'upscaled'.
"""
fig = plt.figure(figsize=(plot_dict['fig_width_1col'], 4))
gs = gridspec.GridSpec(1, 1)
gs.update(left=-0.07, right=1.07, bottom=0, top=1)
plot_network_model_sketch(gs[0], model=model)
plt.savefig(os.path.join(output_dir, f'network_model_sketch_{model}.pdf'))
plt.savefig(os.path.join(
output_dir, f'network_model_sketch_{model}.png'), dpi=300)
return
def figure_mesocircuit_icon(output_dir):
"""
Figure with a mesocircuit icon.
Parameters
----------
output_dir
Output directory.
"""
fig = plt.figure(figsize=(1.5, 1))
gs = gridspec.GridSpec(1, 1)
gs.update(left=-0.1, right=0.95, bottom=-0.15, top=1.25)
plot_mesocircuit_icon(gs[0])
plt.savefig(os.path.join(output_dir, 'mesocircuit_icon.pdf'))
plt.savefig(os.path.join(output_dir, 'mesocircuit_icon.png'), dpi=300)
return
def plot_mesocircuit_icon(gs, elev=12, azim=-50, scale_fs=0.7):
"""
Plots a schematic network icon to gridspec cell.
Parameters
----------
gs
A gridspec cell to plot into.
elev
Elevation angle in z-plane.
azim
Azimuth angle in x,y-plane.
scale_fs
Scaling factor for font size.
"""
ax = plt.subplot(gs, projection='3d', computed_zorder=False)
pop_colors = plot_dict['pop_colors']
pop_labels = plot_dict['pop_labels']
zcnt = 0
for i, col in enumerate(pop_colors[::-1]):
if i == 0: # TC
patch = Circle((0.5, 0.5), 0.1, facecolor=col)
z = -5 / (len(pop_colors) - 1)
xshift = 0.4
yshift = 0.4
else:
patch = Rectangle((0, 0), 1, 1, facecolor=col)
z = 1.5*i / (len(pop_colors) - 1) + zcnt
xshift = -0.02
yshift = -0.02
if not i % 2:
zcnt += 2 / (len(pop_colors) - 1)
ax.add_patch(patch)
art3d.pathpatch_2d_to_3d(patch, z=z, zdir="z")
if i % 2:
zshift = 0.
else:
zshift = 0.07
if i == 0:
zshift = -0.05
ax.text(1. - xshift, 1 - yshift, z+zshift, pop_labels[::-1][i],
fontsize=matplotlib.rcParams['font.size'] * scale_fs,
verticalalignment='center')
ax.text(1, 1, 2.3, '4 mm', 'x',
fontsize=matplotlib.rcParams['font.size'] * scale_fs,
horizontalalignment='right')
ax.text(0, 0, 2.3, '4 mm', 'y',
fontsize=matplotlib.rcParams['font.size'] * scale_fs,
horizontalalignment='left')
# exponential profile indicating connectivity
xctr = 0.5
yctr = 0.5
X = np.arange(xctr-0.2, xctr+0.2, 0.01)
Y = np.arange(yctr - 0.2, yctr+0.2, 0.01)
X, Y = np.meshgrid(X, Y)
R = np.sqrt((X-xctr)**2 + (Y-yctr)**2)
Z = z + 0.25 * np.exp(-R/(0.2/4))
# make bottom of surface round
for i in np.arange(np.shape(Z)[0]):
for j in np.arange(np.shape(Z)[1]):
if ((X[i, j]-xctr)**2 + (Y[i, j]-yctr)**2) > 0.15**2:
Z[i, j] = np.nan
ax.plot_surface(X, Y, Z, cmap='bone')
ax.grid(False)
ax.view_init(elev=elev, azim=azim)
plt.axis('off')
return
def plot_network_model_sketch(gs, model='upscaled'):
"""
Plots a network sketch to a gridspec cell.
Shown connections have an in-degree > 300.
Parameters
----------
gs
A gridspec cell to plot into.
model
Model name. Options are 'reference' or 'upscaled'.
"""
np.random.seed(1234)
num_neurons = net_dict['num_neurons_1mm2_SvA2018']
layer_sizes = [num_neurons[i] + num_neurons[i+1]
for i in np.arange(8, step=2)]
# scaled
layer_sizes /= np.max(layer_sizes)
ax = plt.subplot(gs, projection='3d', computed_zorder=False)
pop_colors = plot_dict['pop_colors']
layer_labels = plot_dict['layer_labels']
ctr_exc = [0.3, 0.2]
ctr_inh = [0.7, 0.2]
z_ctr_offset = -0.04
conn_ctr = [0.8, 0.6]
mutation_scale = 10
# cortex ###################################################################
z = 0
z_lctrs = np.zeros(4) # layer centers
for i, ll in enumerate(layer_labels[::-1]):
z_ctr = z + layer_sizes[::-1][i] / 2 + z_ctr_offset
z_lctrs[i] = z_ctr
layer_size = layer_sizes[::-1][i]
# layer ################################################################
if model == 'reference':
layer = Circle(xy=(0.5, 0.5), radius=0.5,
facecolor='lightgrey',
edgecolor='k')
elif model == 'upscaled':
layer = Rectangle(xy=(0, 0), width=1, height=1,
facecolor='lightgrey', edgecolor='k')
ax.add_patch(layer)
art3d.pathpatch_2d_to_3d(layer, z=z, zdir="z")
# neurons ##############################################################
num_neurons_exc = int(num_neurons[::-1][1 + 2*i] / 30)
num_neurons_inh = int(num_neurons[::-1][2*i] / 30)
rnds_exc = [np.random.rand(num_neurons_exc) for x in [0, 1]]
rnds_inh = [np.random.rand(num_neurons_inh) for x in [0, 1]]
offset = 0.03
scale = 1-2*offset
pos_x_exc = offset + scale*rnds_exc[0]
pos_y_exc = offset + scale*rnds_exc[1]
pos_x_inh = offset + scale*rnds_inh[0]
pos_y_inh = offset + scale*rnds_inh[1]
if model == 'reference':
pxexc = []
pyexc = []
for p in np.arange(len(pos_x_exc)):
if (pos_x_exc[p]-0.5)**2 + (pos_y_exc[p]-0.5)**2 <= (0.5 - offset)**2:
pxexc.append(pos_x_exc[p])
pyexc.append(pos_y_exc[p])
# reduce density from 16 mm2 to 1 mm2
pos_x_exc = np.array(pxexc)[::16]
pos_y_exc = np.array(pyexc)[::16]
pxinh = []
pyinh = []
for p in np.arange(len(pos_x_inh)):
if (pos_x_inh[p]-0.5)**2 + (pos_y_inh[p]-0.5)**2 <= (0.5 - offset)**2:
pxinh.append(pos_x_inh[p])
pyinh.append(pos_y_inh[p])
pos_x_inh = np.array(pxinh)[::16]
pos_y_inh = np.array(pyinh)[::16]
ax.scatter(xs=pos_x_exc, ys=pos_y_exc, zs=z,
marker=',',
s=matplotlib.rcParams['lines.markersize'] * 0.01,
color=pop_colors[::-1][2 + 2*i], alpha=1)
ax.scatter(xs=pos_x_inh, ys=pos_y_inh, zs=z,
marker=',',
s=matplotlib.rcParams['lines.markersize'] * 0.01,
color=pop_colors[::-1][1 + 2*i], alpha=1)
if i == 3:
conns = RegularPolygon(xy=conn_ctr, radius=0.1, numVertices=3, orientation=12.05,
facecolor=pop_colors[::-1][2 + 2*i],
edgecolor='k')
ax.add_patch(conns)
art3d.pathpatch_2d_to_3d(conns, z=z, zdir="z")
# exponential profile indicating connectivity ##########################
if i == 2 and model == 'upscaled':
xctr = conn_ctr[0]
yctr = conn_ctr[1]
X = np.arange(xctr-0.2, xctr+0.2, 0.01)
Y = np.arange(yctr - 0.2, yctr+0.2, 0.01)
X, Y = np.meshgrid(X, Y)
R = np.sqrt((X-xctr)**2 + (Y-yctr)**2)
Z = z + layer_size * np.exp(-R/(0.06))
# make bottom of surface round
for k in np.arange(np.shape(Z)[0]):
for l in np.arange(np.shape(Z)[1]):
if ((X[k, l]-xctr)**2 + (Y[k, l]-yctr)**2) > 0.16**2:
Z[k, l] = np.nan
ax.plot_surface(X, Y, Z, rstride=1, cstride=1, shade=True,
antialiased=False, color=pop_colors[::-1][2 + 2*i])
if i == 2 and model == 'reference':
for j in np.arange(20):
ax.plot(xs=[pos_x_exc[j], conn_ctr[0]],
ys=[pos_y_exc[j], conn_ctr[1]],
zs=[z, z+layer_size],
color=pop_colors[::-1][2+2*i],
zorder=1)
# front ################################################################
front = Rectangle(xy=(0, z), width=1, height=layer_sizes[::-1][i],
facecolor='white', edgecolor='k')
ax.add_patch(front)
art3d.pathpatch_2d_to_3d(front, z=0, zdir="y")
# excitatory neurons
exc = RegularPolygon(xy=(ctr_exc[0], z_ctr-0.025), radius=0.1,
numVertices=3, orientation=0,
edgecolor=pop_colors[::-1][2 + 2*i],
facecolor='white',
linewidth=2)
ax.add_patch(exc)
art3d.pathpatch_2d_to_3d(exc, z=0, zdir="y")
# inhibitory neurons
inh = Circle(xy=(ctr_inh[0], z_ctr), radius=0.08,
edgecolor=pop_colors[::-1][1 + 2*i],
facecolor='white',
linewidth=2)
ax.add_patch(inh)
art3d.pathpatch_2d_to_3d(inh, z=0, zdir="y")
ax.text(x=ctr_exc[0]-0.015, y=0, z=z_lctrs[i]-0.03, s='E', zdir='x',
fontsize=matplotlib.rcParams['font.size'], color='k',
horizontalalignment='center', verticalalignment='center')
ax.text(x=ctr_inh[0]-0.01, y=0, z=z_lctrs[i], s='I', zdir='x',
fontsize=matplotlib.rcParams['font.size'], color='k',
horizontalalignment='center', verticalalignment='center')
# excitatory connections ###############################################
# same layer, E -> E
e_e_list = ['L2/3', 'L4']
if model == 'upscaled':
e_e_list += ['L6']
if ll in e_e_list:
draw_edge_arrow_xzplane(ax=ax, x=ctr_exc[0]-0.05, y=0, z=z_ctr,
xshift1=-0.08, zshift1=0.16,
xshift2=0.13, zshift2=-0.065,
color=pop_colors[::-1][2 + 2*i], sign='exc')
# same layer, E -> I
if ll in ['L2/3', 'L4', 'L6']:
ax.arrow3D(x=ctr_exc[0]+0.02, y=0, z=z_ctr+0.025,
dx=0.32, dy=0, dz=0,
mutation_scale=mutation_scale,
arrowstyle='-|>',
color=pop_colors[::-1][2+2*i])
# inhibitory connections ###############################################
# same layer, I -> I
if ll in ['L2/3', 'L4']:
draw_edge_arrow_xzplane(ax=ax, x=ctr_inh[0]+0.09, y=0, z=z_ctr,
xshift1=+0.04, zshift1=0.16,
xshift2=-0.13, zshift2=-0.06,
color=pop_colors[::-1][1+2*i], sign='inh')
inhibitory_arrowhead_front(ax=ax, x=ctr_inh[0], z=z_ctr+0.11,
color=pop_colors[::-1][1+2*i])
# same layer, I -> E
i_e_list = ['L2/3', 'L4']
if model == 'reference':
i_e_list += ['L6']
if ll in i_e_list:
ax.arrow3D(x=ctr_inh[0]-0.07, y=0, z=z_ctr-0.025,
dx=-0.26, dy=0, dz=0,
mutation_scale=mutation_scale,
arrowstyle='-',
color=pop_colors[::-1][1+2*i])
inhibitory_arrowhead_front(ax=ax, x=ctr_inh[0]-0.3, z=z_ctr-0.025,
color=pop_colors[::-1][1+2*i])
# layer annotations ####################################################
ax.text(x=1, y=0, z=z + layer_sizes[::-1][i]+0.03, zdir='x',
s=ll, fontsize=matplotlib.rcParams['font.size']*1.8,
horizontalalignment='right', verticalalignment='top')
z += layer_sizes[::-1][i]
# layer crossing connections ###############################################
# reverse layer center such that top layer (2/3) has index 0
z_lctrs = z_lctrs[::-1]
# L2/3E -> L5E
draw_edge_arrow_xzplane(ax=ax, x=ctr_exc[0]-0.08, y=0, z=z_lctrs[0]-0.05,
xshift1=-0.15, zshift1=-1.49,
xshift2=0.15, zshift2=0,
color=pop_colors[0], sign='exc')
# L4E -> L2/3E
ax.arrow3D(x=ctr_exc[0], y=0, z=z_lctrs[1] + 0.16,
dx=0, dy=0, dz=0.62,
mutation_scale=mutation_scale,
arrowstyle='-|>',
color=pop_colors[2])
# L6E -> L4E
draw_edge_arrow_xzplane(ax=ax, x=ctr_exc[0]-0.06, y=0, z=z_lctrs[3]-0.02,
xshift1=-0.11, zshift1=0.88,
xshift2=0.085, zshift2=0,
color=pop_colors[6], sign='exc')
# L4E -> L4I
ax.arrow3D(x=ctr_exc[0]+0.02, y=0, z=z_lctrs[1] + 0.03,
dx=0.335, dy=0, dz=0.77,
mutation_scale=mutation_scale,
arrowstyle='-|>',
color=pop_colors[2])
# L4E -> L5E
ax.arrow3D(x=ctr_exc[0], y=0, z=z_lctrs[1] - 0.07,
dx=0, dy=0, dz=-0.51,
mutation_scale=mutation_scale,
arrowstyle='-|>',
color=pop_colors[2])
if model == 'upscaled':
# L4E -> L6E
draw_edge_arrow_xzplane(ax=ax, x=ctr_exc[0]-0.06, y=0, z=z_lctrs[1]-0.02,
xshift1=-0.14, zshift1=-0.96,
xshift2=0.12, zshift2=0,
color=pop_colors[2], sign='exc')
# L2/3E -> L4I
ax.arrow3D(x=ctr_exc[0]+0.05, y=0, z=z_lctrs[0] - 0.07,
dx=0.3, dy=0, dz=-0.73,
mutation_scale=mutation_scale,
arrowstyle='-|>',
color=pop_colors[0])
# L2/3E -> L5I
ax.arrow3D(x=ctr_exc[0]+0.04, y=0, z=z_lctrs[0] - 0.07,
dx=0.32, dy=0, dz=-1.36,
mutation_scale=mutation_scale,
arrowstyle='-|>',
color=pop_colors[0])
# L2/3E -> L5I
ax.arrow3D(x=ctr_exc[0]+0.03, y=0, z=z_lctrs[0] - 0.07,
dx=0.31, dy=0, dz=-1.66,
mutation_scale=mutation_scale,
arrowstyle='-|>',
color=pop_colors[0])
# L6E -> L4I
ax.arrow3D(x=ctr_exc[0]+0.017, y=0, z=z_lctrs[3]-0.01,
dx=0.4, dy=0, dz=0.87,
mutation_scale=mutation_scale,
arrowstyle='-|>',
color=pop_colors[6])
# thalamus #################################################################
y_tc = -0.15
tc = RegularPolygon(xy=(0.5, y_tc), radius=0.1, numVertices=3, orientation=0,
facecolor=pop_colors[-1],
edgecolor='k')
ax.add_patch(tc)
art3d.pathpatch_2d_to_3d(tc, z=0, zdir="y")
# TC to L4 and L6
draw_edge_arrow_xzplane(ax=ax, x=0.5-0.07, y=0, z=y_tc,
xshift1=-0.55,
zshift1=-y_tc +
np.sum(layer_sizes[2:]) +
layer_sizes[1]/2. + z_ctr_offset,
xshift2=0.12, zshift2=0,
color=pop_colors[-1], sign='exc')
draw_edge_arrow_xzplane(ax=ax, x=0.5-0.07, y=0, z=y_tc,
xshift1=-0.55,
zshift1=-y_tc +
layer_sizes[3]/2. + z_ctr_offset,
xshift2=0.12, zshift2=0,
color=pop_colors[-1], sign='exc')
# thalamus label
ax.text(x=1, y=0, z=0+0.03, zdir='x',
s='TC', fontsize=matplotlib.rcParams['font.size']*1.8,
horizontalalignment='right', verticalalignment='top')
# size label ###############################################################
if model == 'reference':
model_label = r'$1 \mathrm{mm}^2$' # r'$1\times 1 \mathrm{mm}^2$'
ax.text(x=0.97, y=0.6, z=0, s=model_label, zdir='y',
fontsize=matplotlib.rcParams['font.size'] * 1.8,
horizontalalignment='center', verticalalignment='top')
elif model == 'upscaled':
model_label = r'$4\times 4 \mathrm{mm}^2$'
ax.text(x=0.9, y=0.8, z=0, s=model_label, zdir='y',
fontsize=matplotlib.rcParams['font.size'] * 1.8,
horizontalalignment='center', verticalalignment='top')
# ax.grid(False)
ax.set_zlim(bottom=-0.2, top=np.sum(layer_sizes))
ax.set_box_aspect(aspect=(1, 1, 2.2))
ax.view_init(elev=20, azim=-50)
# ax.view_init(elev=20, azim=-90) # front view
plt.axis('off')
return
def choose_connections_to_draw(output_dir, threshold=300):
"""
Helps to decide where to put the in-degree threshold for network sketch.
Data from reference model.
Parameters
----------
threshold
Maximum in-degree to be shown.
"""
matrix = net_dict['indegrees_1mm2_SvA2018']
matrix_int = np.round(net_dict['indegrees_1mm2_SvA2018']).astype(int)
for i, src in enumerate(plot_dict['pop_labels'][:-1]):
for j, tgt in enumerate(plot_dict['pop_labels'][:-1]):
if src[:2] == tgt[:2]:
same_layer = True
else:
same_layer = False
if matrix_int[j, i] > threshold:
draw = True
else:
draw = False
if same_layer != draw and same_layer:
emph = '|->'
elif same_layer != draw and not same_layer:
emph = ' ->'
else:
emph = ' '
print(emph, src, tgt, same_layer, draw, matrix_int[j, i])
# set bad
matrix[np.where(matrix < threshold)] = np.nan
# image
matrix = np.ma.masked_invalid(matrix)
cm = matplotlib.cm.get_cmap('viridis').copy()
cm.set_bad('white')
fig = plt.figure()
plt.imshow(matrix, cmap=cm) # , vmin=vmin, vmax=vmax)
ax = plt.gca()
ax.set_xticks(np.arange(8))
ax.set_yticks(np.arange(8))
ax.set_xticklabels(plot_dict['pop_labels'][:-1])
ax.set_yticklabels(plot_dict['pop_labels'][:-1])
ax.set_title(f'threshold={threshold}')
plt.savefig(os.path.join(
output_dir, f'indegrees_threshold{threshold}.pdf'))
helper functions #############################################################
class Arrow3D(FancyArrowPatch):
"""
Arrow class for 3D sketches.
From https://gist.github.com/WetHat/1d6cd0f7309535311a539b42cccca89c
Parameters
----------
FancyArrowPatch
"""
def __init__(self, x, y, z, dx, dy, dz, *args, **kwargs):
super().__init__((0, 0), (0, 0), *args, **kwargs)
self._xyz = (x, y, z)
self._dxdydz = (dx, dy, dz)
def draw(self, renderer):
x1, y1, z1 = self._xyz
dx, dy, dz = self._dxdydz
x2, y2, z2 = (x1 + dx, y1 + dy, z1 + dz)
xs, ys, zs = proj_transform((x1, x2), (y1, y2), (z1, z2), self.axes.M)
self.set_positions((xs[0], ys[0]), (xs[1], ys[1]))
super().draw(renderer)
def do_3d_projection(self, renderer=None):
x1, y1, z1 = self._xyz
dx, dy, dz = self._dxdydz
x2, y2, z2 = (x1 + dx, y1 + dy, z1 + dz)
xs, ys, zs = proj_transform((x1, x2), (y1, y2), (z1, z2), self.axes.M)
self.set_positions((xs[0], ys[0]), (xs[1], ys[1]))
return np.min(zs)
def _arrow3D(ax, x, y, z, dx, dy, dz, *args, **kwargs):
"""
Adds a 3d arrow to an `Axes3D` instance.
"""
arrow = Arrow3D(x, y, z, dx, dy, dz, *args, **kwargs)
ax.add_artist(arrow)
return
setattr(Axes3D, 'arrow3D', _arrow3D)
def draw_edge_arrow_xzplane(
ax, x, y, z, xshift1, zshift1, xshift2, zshift2,
color='k', mutation_scale=10, sign='exc'):
"""
Draws an arrow in the x-z plane.
"""
if sign == 'exc':
head = '-|>'
elif sign == 'inh':
head = '-'
# xshift
ax.arrow3D(x=x, y=y, z=z,
dx=xshift1, dy=0, dz=0,
mutation_scale=mutation_scale,
arrowstyle='-',
shrinkA=0, shrinkB=0,
color=color)
# zshift
ax.arrow3D(x=x+xshift1, y=y, z=z,
dx=0, dy=0, dz=zshift1,
mutation_scale=mutation_scale,
arrowstyle='-',
shrinkA=0, shrinkB=0,
color=color)
if zshift2 == 0:
arrowstyle = head
else:
arrowstyle = '-'
# xshift back
ax.arrow3D(x=x+xshift1, y=y, z=z+zshift1,
dx=xshift2, dy=0, dz=0,
mutation_scale=mutation_scale,
arrowstyle=arrowstyle,
shrinkA=0, shrinkB=0,
color=color)
if zshift2 == 0:
return
# zshift back
ax.arrow3D(x=x+xshift1+xshift2, y=y, z=z+zshift1,
dx=0, dy=0, dz=zshift2,
mutation_scale=mutation_scale,
arrowstyle=head,
shrinkA=0, shrinkB=0,
color=color)
def inhibitory_arrowhead_front(ax, x, z, color):
"""
Circular arrow head for inhibitory connections.
"""
head = Circle(xy=(x, z), radius=0.02,
facecolor=color,
edgecolor=None)
ax.add_patch(head)
art3d.pathpatch_2d_to_3d(head, z=0, zdir='y')
if __name__ == '__main__':
output_dir = 'network_sketches'
if not os.path.isdir(output_dir):
os.makedirs(output_dir)
figure_mesocircuit_icon(output_dir)
figure_network_model_sketch(output_dir, model='reference')
figure_network_model_sketch(output_dir, model='upscaled')
# choose_connections_to_draw(output_dir)
# manuscript figure of network sketches for both reference and upscaled model
output_dir = 'ms_figures'
if not os.path.isdir(output_dir):
os.makedirs(output_dir)
figure_network_model_sketches(output_dir)
