def analyse(fraction_to_stim,
size,
file_to_save='rates.png',
show=True,
detailed_also=False,
average_prev_runs=False,
legend=False):
print("Will plot average of previous runs: %s" % average_prev_runs)
# transient time to discard the data (ms)
Ttrans = defaultParams.Ttrans
# simulation time before perturbation (ms)
Tblank = defaultParams.Tblank
# simulation time of perturbation (ms)
Tstim = defaultParams.Tstim
# time after perturbation
Tpost = defaultParams.Tpost
T = Ttrans + Tblank + Tstim + Tpost
defaultParams.set_total_population_size(size)
N = defaultParams.N
NE = defaultParams.NE # num exc point neurons
NE2 = 0 # num detailed cells
if detailed_also:
NE2 = 10
NE = NE - NE2
NI = defaultParams.NI # num inh neurons
NI_pert = int(NI * fraction_to_stim)
NI_nonpert = NI - NI_pert
spd = pl.loadtxt('ISN-nest-EI-0.gdf')
spt = spd[:, 0]
spi = spd[:, 1]
e_spt = []
e_spi = []
e2_spt = []
e2_spi = []
pi_spt = []
pi_spi = []
npi_spt = []
npi_spi = []
for k in range(len(spt)):
t = spt[k]
i = spi[k]
if i < NE:
e_spi.append(i)
e_spt.append(t)
elif i < NE + NE2:
e2_spi.append(i)
e2_spt.append(t)
elif i < NE + NE2 + NI_pert:
pi_spi.append(i)
pi_spt.append(t)
else:
npi_spi.append(i)
npi_spt.append(t)
if not detailed_also:
fig0, (ax1, ax2) = pl.subplots(2, 1, sharex=True)
else:
fig0, (ax1, ax1b,
ax2) = pl.subplots(3,
1,
sharex=True,
gridspec_kw={'height_ratios': [3, 1, 4]})
fig0.set_size_inches(12, 8, forward=True)
red = '#ff0000'
lblue = '#99c2ff'
lgrey = '#dddddd'
green = '#00cc00'
dgreen = '#008800'
dblue = '#0047b3'
dblue2 = '#444444'
ax1.plot(pi_spt, pi_spi, '.', color=dblue, markersize=2)
ax1.plot(npi_spt, npi_spi, '.', color=lblue, markersize=2)
if NE2 > 0:
ax1.plot(e2_spt, e2_spi, '.', color=green, markersize=7)
ax1.plot(e_spt, e_spi, '.', color=red, markersize=2)
if NE2 > 0:
ax1b.plot(e2_spt, e2_spi, '.', color=green, markersize=5)
ax1b.set_ylabel('Cell index')
ax1.set_ylabel('Cell index')
yl = ax1.get_ylim()
ax1.set_ylim([yl[1], yl[0]])
if NE2 > 0:
ylb = ax1b.get_ylim()
#ax1b.set_ylim([ylb[1]-1,ylb[0]+1])
ax1b.set_yticks([NE, NE + NE2 - 1])
bw = 1 #0
hst = np.histogram2d(spt,
spi,
range=((0, T), (0, N - 1)),
bins=(T / bw, N))
tt = hst[1][0:-1] + np.diff(hst[1])[0] / 2
rr = hst[0] / (bw / 1000)
r_exc = rr[:, 0:NE]
r_exc2 = rr[:, NE:NE + NE2]
r_inh_pert = rr[:, NE + NE2:NE + NE2 + NI_pert]
r_inh_nonpert = rr[:, N - NI_nonpert:]
sm_len = 151.0
r_exc_m = mySmooth(np.nanmean(r_exc, 1), sm_len)[0:rr.shape[0]]
r_exc2_m = mySmooth(np.nanmean(r_exc2, 1), sm_len)[0:rr.shape[0]]
r_inh_pert_m = mySmooth(np.nanmean(r_inh_pert, 1), sm_len)[0:rr.shape[0]]
r_inh_nonpert_m = mySmooth(np.nanmean(r_inh_nonpert, 1),
sm_len)[0:rr.shape[0]]
ks = open('kernelseed')
kernelseed = int(ks.read())
print('kernelseed from last simulation: %i' % kernelseed)
file_rates = 'pertinh.rate.%i.dat' % kernelseed
pertinh = open(file_rates, 'w')
print("Saving to file %s list of rates for this seed: %s" %
(pertinh, r_inh_pert_m))
for r in r_inh_pert_m:
pertinh.write('%s\n' % r)
pertinh.close()
if detailed_also:
file_rates2 = 'exc2.rate.%i.dat' % kernelseed
exc2 = open(file_rates2, 'w')
print("Saving to file %s list of rates for this seed: %s" %
(exc2, r_exc2_m))
for r in r_exc2_m:
exc2.write('%s\n' % r)
exc2.close()
all_ri = []
all_re2 = []
if average_prev_runs:
# Load all rates files in & average
for f in os.listdir('.'):
if f.startswith('pertinh.rate'):
r = pl.loadtxt(f)
all_ri.append(r)
print("Loaded rates from %s: %s" % (f, r))
if f.startswith('exc2.rate'):
r = pl.loadtxt(f)
all_re2.append(r)
print("Loaded rates from %s: %s" % (f, r))
#ax2.plot(tt, r, lgrey, lw=0.3)
avg_ri = []
avg_re2 = []
if len(all_ri) > 0:
for i_t in range(len(all_ri[0])):
r = 0
for i_r in range(len(all_ri)):
r += all_ri[i_r][i_t]
r /= len(all_ri)
avg_ri.append(r)
if len(all_re2) > 0:
for i_t in range(len(all_re2[0])):
r = 0
for i_r in range(len(all_re2)):
r += all_re2[i_r][i_t]
r /= len(all_re2)
avg_re2.append(r)
if len(all_ri) > 0:
ax2.plot(tt, avg_ri, dblue2, lw=4, linestyle=':')
if NE2 > 0:
ax2.plot(tt, avg_re2, dgreen, lw=4, linestyle=':')
ax2.plot(tt, r_exc_m, red, lw=2)
if NE2 > 0:
ax2.plot(tt, r_exc2_m, green, lw=2)
ax2.plot(tt, r_inh_pert_m, dblue, lw=2)
ax2.plot(tt, r_inh_nonpert_m, lblue, lw=2)
pl.ylabel('Firing rate (Hz)')
pl.xlabel('Time (ms)')
height = ax2.get_ylim()[1]
if detailed_also: height = 12
pl.xlim([0, T])
pl.ylim([0, height])
ax2.add_patch(
patches.Rectangle(
(Ttrans + Tblank, 0), # (x,y)
Tstim, # width
height, # height
facecolor='#f8f5dd'))
t1 = (tt > Ttrans) * (tt < (Ttrans + Tblank))
t2 = (tt > (Ttrans + Tblank)) * (tt < (Ttrans + Tblank + Tstim))
if legend:
j = 3
matplotlib.pyplot.text(100, j + 17, 'Exc.', color=red)
i = 0
if detailed_also:
matplotlib.pyplot.text(100, j + 15, 'Detailed Exc.', color=green)
matplotlib.pyplot.text(100,
j + 13,
'Detailed Exc. avg %s' % len(all_re2),
color=dgreen)
i = -4
matplotlib.pyplot.text(100,
j + 15 + i,
'Inh. (non pert.)',
color=lblue)
matplotlib.pyplot.text(100, j + 13 + i, 'Inh. (pert.)', color=dblue)
matplotlib.pyplot.text(100,
j + 11 + i,
'Inh. (pert.) avg %s' % len(all_ri),
color=dblue2)
print('before vs after perturbation')
print('exc: ', np.nanmean(r_exc[t1]), np.nanmean(r_exc[t2]))
print('inh (pert): ', np.nanmean(r_inh_pert_m[t1]),
np.nanmean(r_inh_pert_m[t2]))
print('inh (non-pert): ', np.nanmean(r_inh_nonpert_m[t1]),
np.nanmean(r_inh_nonpert_m[t2]))
all_ax = [ax1, ax2]
if detailed_also: all_ax.append(ax1b)
for ax in all_ax:
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.yaxis.set_ticks_position('left')
ax.xaxis.set_ticks_position('bottom')
fig0.savefig(file_to_save, dpi=150, bbox_inches='tight')
if show:
pl.show()
"Voltage traces: Be=%s; Bi=%s; N=%s; p=%s" %
(Be, Bi, N, nn_stim),
xaxis="Time (s)",
yaxis="Membrane potential (V)",
colors=colors,
grid=False,
show_plot_already=False)
if __name__ == '__main__':
Be = 0.1
Bi = -0.2
if '-small' in sys.argv:
defaultParams.set_total_population_size(100)
Be = .1
Bi = -.1
nn_stim_rng = (np.array([0.1, .25, .5, .75, 1]) *
defaultParams.NI).astype('int')
nn_stim_rng = (np.array([.75]) * defaultParams.NI).astype('int')
if '-small' in sys.argv:
nn_stim_rng = (np.array([0.75]) * defaultParams.NI).astype('int')
if '-nogui' in sys.argv:
show_gui = False
else:
import matplotlib.pyplot as plt
show_gui = True