import numpy as np
from ..strategy import ConnectionStrategy
[docs]class ConnectomePurkinjeDCN(ConnectionStrategy):
"""
Legacy implementation for the connection between purkinje cells and DCN cells.
Also rotates the dendritic trees of the DCN.
"""
casts = {"divergence": int}
required = ["divergence"]
[docs] def validate(self):
pass
def connect(self):
# Gather information for the legacy code block below.
from_type = self.from_cell_types[0]
to_type = self.to_cell_types[0]
purkinjes = self.from_cells[from_type.name]
dcn_cells = self.to_cells[to_type.name]
dend_tree_coeff = np.zeros((dcn_cells.shape[0], 4))
for i in range(len(dcn_cells)):
# Make the planar coefficients a, b and c.
dend_tree_coeff[i] = np.random.rand(4) * 2.0 - 1.0
# Calculate the last planar coefficient d from ax + by + cz - d = 0
# => d = - (ax + by + cz)
dend_tree_coeff[i, 3] = -np.sum(dend_tree_coeff[i, 0:2] * dcn_cells[i, 2:4])
if len(dcn_cells) == 0:
return
first_dcn = int(dcn_cells[0, 0])
divergence = self.divergence
def connectome_pc_dcn(first_dcn, purkinjes, dcn_cells, div_pc, dend_tree_coeff):
pc_dcn = np.zeros((0, 2))
# For all Purkinje cells: calculate the distance with the area around glutamatergic DCN cells soma, then choose 4-5 of them
for i in purkinjes:
distance = (
np.absolute(
(dend_tree_coeff[:, 0] * i[2])
+ (dend_tree_coeff[:, 1] * i[3])
+ (dend_tree_coeff[:, 2] * i[4])
+ dend_tree_coeff[:, 3]
)
) / (
np.sqrt(
(dend_tree_coeff[:, 0] ** 2)
+ (dend_tree_coeff[:, 1] ** 2)
+ (dend_tree_coeff[:, 2] ** 2)
)
)
dist_matrix = np.zeros((dcn_cells.shape[0], 2))
dist_matrix[:, 1] = dcn_cells[:, 0]
dist_matrix[:, 0] = distance
dcn_dist = np.random.permutation(dist_matrix)
# If the number of DCN cells are less than the divergence value, all neurons are connected to the corresponding PC
if dcn_cells.shape[0] < div_pc:
matrix = np.zeros((dcn_cells.shape[0], 2))
matrix[:, 0] = i[0]
matrix[:, 1] = dcn_cells[:, 0]
pc_dcn = np.vstack((pc_dcn, matrix))
else:
if np.random.rand() > 0.5:
connected_f = dcn_dist[0:div_pc, 1]
connected_dist = dcn_dist[0:div_pc, 0]
connected_provv = connected_f.astype(int)
connected_dcn = connected_provv
# construction of the output matrix: the first column has the PC index, while the second column has the connected DCN cell index
matrix = np.zeros((div_pc, 2))
matrix[:, 0] = i[0]
matrix[:, 1] = connected_dcn
pc_dcn = np.vstack((pc_dcn, matrix))
else:
connected_f = dcn_dist[0 : (div_pc - 1), 1]
connected_dist = dcn_dist[0 : (div_pc - 1), 0]
connected_provv = connected_f.astype(int)
connected_dcn = connected_provv
matrix = np.zeros(((div_pc - 1), 2))
matrix[:, 0] = i[0]
matrix[:, 1] = connected_dcn
pc_dcn = np.vstack((pc_dcn, matrix))
return pc_dcn
results = connectome_pc_dcn(
first_dcn, purkinjes, dcn_cells, divergence, dend_tree_coeff
)
self.scaffold.connect_cells(self, results)
