def draw_single_cell(agg, df_cell):
aprops_to_display = ['category', 'maxAmp', 'sal', 'meanspect', 'q1', 'q2', 'q3',
'entropyspect', 'meantime', 'entropytime']
plist = list()
for aprop in aprops_to_display:
perfs = df_cell['perf_%s' % aprop].values
lkrats = df_cell['lkrat_%s' % aprop].values
plist.append({'perfs':perfs, 'lkrats':lkrats, 'aprop':aprop})
def _plot_perf_hist(pdata, ax):
plt.sca(ax)
plt.hist(pdata['perfs'], bins=10, color=COLOR_YELLOW_SPIKE)
plt.xlabel('R2')
plt.axis('tight')
plt.xlim(0, 0.4)
tks = [0.0, 0.1, 0.2, 0.3, 0.4]
plt.xticks(tks, ['%0.1f' % x for x in tks])
plt.title(pdata['aprop'])
def _plot_lkrat_hist(pdata, ax):
lkr = pdata['lkrats']
lkr = lkr[lkr >= 0]
plt.sca(ax)
plt.hist(lkr, bins=10, color=COLOR_YELLOW_SPIKE)
plt.xlabel('Likelihood Ratio')
plt.axis('tight')
plt.xlim(0, 60)
plt.title(pdata['aprop'])
multi_plot(plist, _plot_perf_hist, nrows=2, ncols=5)
multi_plot(plist, _plot_lkrat_hist, nrows=2, ncols=5)
def draw_neural_joint_relationships(preproc_file):
hf = h5py.File(preproc_file)
index2prop = list(hf.attrs['integer2prop'])
index2type = list(hf.attrs['integer2type'])
S = np.array(hf['S'])
X = np.array(hf['X'])
Y = np.array(hf['Y'])
hf.close()
stim_type = [index2type[k] for k in Y[:, 0]]
sub_i = [index2prop.index('maxAmp'), index2prop.index('sal'), index2prop.index('meanspect')]
S = S[:, sub_i]
ncells = X.shape[1]
nsamps = S.shape[0]
csv_file = '/auto/tdrive/mschachter/data/aggregate/big3_spike_rate.csv'
if not os.path.exists(csv_file):
# write out a csv file for analysis
cdata = {'maxAmp':list(), 'sal':list(), 'meanspect':list(), 'call_type':list(), 'cell':list(), 'spike_rate':list()}
for n in range(ncells):
for k in range(nsamps):
cdata['maxAmp'].append(S[k, 0])
cdata['sal'].append(S[k, 1])
cdata['meanspect'].append(S[k, 2])
cdata['call_type'].append(stim_type[k])
cdata['cell'].append(n)
cdata['spike_rate'].append(X[k, n])
cdf = pd.DataFrame(cdata)
cdf.to_csv(csv_file, header=True, index=False)
plist = list()
for n in range(ncells):
plist.append({'n':n, 'x':S[:, 0], 'y':S[:, 1], 'r':X[:, n], 'xlabel':'maxAmp', 'ylabel':'saliency', 'logx':True})
plist.append({'n': n, 'x': S[:, 0], 'y': S[:, 2], 'r': X[:, n], 'xlabel': 'maxAmp', 'ylabel': 'meanspect', 'logx':True})
plist.append({'n': n, 'x': S[:, 1], 'y': S[:, 2], 'r': X[:, n], 'xlabel': 'saliency', 'ylabel': 'meanspect', 'logx':False})
def _plot_scatter(_pdata, _ax):
if _pdata['logx']:
_ax.set_xscale('log')
"""
_clrs = [CALL_TYPE_COLORS[ct] for ct in stim_type]
_r = _pdata['r']
_r -= _r.min()
_r /= _r.max()
for _x,_y,_r,_c in zip(_pdata['x'], _pdata['y'], _r, _clrs):
plt.plot(_x, _y, 'o', c=_c, alpha=_r)
"""
plt.scatter(_pdata['x'], _pdata['y'], marker='o', c=_pdata['r'], cmap=plt.cm.afmhot_r, s=49, alpha=0.7)
plt.xlabel(_pdata['xlabel'])
plt.ylabel(_pdata['ylabel'])
plt.title('cell %d' % _pdata['n'])
plt.colorbar(label='Spike Rate')
multi_plot(plist, _plot_scatter, nrows=3, ncols=3, figsize=(23,13), wspace=0.25, hspace=0.25)
plt.show()
def draw_figure(data_dir='/auto/tdrive/mschachter/data', bird='GreBlu9508M'):
pp_file = os.path.join(data_dir, bird, 'preprocess', 'preproc_raw_coherence_band0_npc0_self_locked_all_Site4_Call1_L.h5')
print 'pp_file=%s' % pp_file
hf = h5py.File(pp_file, 'r')
X = np.array(hf['X'])
S = np.array(hf['S'])
index2electrode = list(hf.attrs['index2electrode'])
index2aprop = list(hf.attrs['integer2prop'])
freqs = list(hf.attrs['freqs'])
hf.close()
reduced_aprops = ['fund',
'sal',
'voice2percent',
'meanspect',
'skewspect',
'entropyspect',
'q2',
'meantime',
'skewtime',
'entropytime',
'maxAmp']
nelectrodes = len(index2electrode)
nfreqs = len(freqs)
electrode_order = ROSTRAL_CAUDAL_ELECTRODES_LEFT
X -= X.mean()
X /= X.std(ddof=1)
S -= S.mean()
S /= S.std(ddof=1)
cc_mats = list()
for k,aprop in enumerate(reduced_aprops):
s = S[:, k]
CC = np.zeros([nelectrodes, nfreqs])
for j,e in enumerate(index2electrode):
for m,f in enumerate(freqs):
i = j*nfreqs + m
x = X[:, i]
jj = electrode_order.index(e)
CC[jj, m] = np.corrcoef(x, s)[0, 1]
cc_mats.append({'CC':CC, 'aprop':aprop})
def _plot_cc_mat(pdata, ax):
plt.sca(ax)
plt.imshow(pdata['CC'], interpolation='nearest', aspect='auto', vmin=-0.5, vmax=0.5, cmap=plt.cm.seismic)
plt.yticks(range(nelectrodes), ['%d' % _e for _e in electrode_order])
plt.xticks(range(nfreqs), ['%d' % _f for _f in freqs])
plt.title(pdata['aprop'])
plt.colorbar()
multi_plot(cc_mats, _plot_cc_mat, nrows=4, ncols=3)
plt.show()
def draw_perf_hists(agg, df_me):
assert isinstance(agg, AggregateLFPAndSpikePSDDecoder)
freqs = agg.freqs
aprops_to_display = ['category', 'maxAmp', 'meanspect', 'stdspect', 'q1', 'q2', 'q3', 'skewspect', 'kurtosisspect',
'sal', 'entropyspect', 'meantime', 'stdtime', 'entropytime']
# make histograms of performances across sites for each acoustic property
perf_list = list()
i = (df_me.band == 0)
for aprop in aprops_to_display:
lfp_perf = df_me[i]['perf_%s_%s' % (aprop, 'lfp')].values
spike_perf = df_me[i]['perf_%s_%s' % (aprop, 'spike')].values
perf_list.append({'lfp_perf':lfp_perf, 'spike_perf':spike_perf,
'lfp_mean':lfp_perf.mean(), 'spike_mean':spike_perf.mean(),
'aprop':aprop})
# make plots
print 'len(perf_list)=%d' % len(perf_list)
def _plot_hist(pdata, ax):
plt.sca(ax)
plt.hist(pdata['lfp_perf'], bins=10, color='#00639E', alpha=0.7)
plt.hist(pdata['spike_perf'], bins=10, color='#F1DE00', alpha=0.7)
plt.legend(['LFP', 'Spike'], fontsize='x-small')
plt.title(pdata['aprop'])
if pdata['aprop'] == 'category':
plt.xlabel('PCC')
else:
plt.xlabel('R2')
plt.axis('tight')
multi_plot(perf_list, _plot_hist, nrows=3, ncols=5, hspace=0.30, wspace=0.30)
def _plot_scatter(pdata, ax):
# pmax = max(pdata['lfp_perf'].max(), pdata['spike_perf'].max())
pmax = 0.8
plt.sca(ax)
plt.plot(np.linspace(0, pmax, 20), np.linspace(0, pmax, 20), 'k-')
plt.plot(pdata['lfp_perf'], pdata['spike_perf'], 'ko', alpha=0.7, markersize=10.)
plt.title(pdata['aprop'])
pstr = 'R2'
if pdata['aprop'] == 'category':
pstr = 'PCC'
plt.xlabel('LFP %s' % pstr)
plt.ylabel('Spike %s' % pstr)
plt.axis('tight')
plt.xlim(0, pmax)
plt.ylim(0, pmax)
multi_plot(perf_list, _plot_scatter, nrows=3, ncols=5, hspace=0.30, wspace=0.30)
def draw_all_encoder_perfs_and_decoder_weights(agg, aprops=('sal', 'q2', 'maxAmp', 'meantime', 'entropytime')):
freqs,lags = get_freqs_and_lags()
font = {'family':'normal', 'weight':'bold', 'size':10}
plt.matplotlib.rc('font', **font)
plist = list()
for (bird,block,segment,hemi),gdf in agg.df.groupby(['bird', 'block', 'segment', 'hemi']):
bstr = '%s_%s_%s_%s' % (bird,hemi,block,segment)
ii = (gdf.decomp == 'self_locked')
assert ii.sum() == 1
wkey = gdf[ii]['wkey'].values[0]
lfp_eperf = agg.encoder_perfs[wkey]
plist.append({'type':'encoder', 'X':lfp_eperf, 'title':bstr})
lfp_decoder_weights = agg.decoder_weights[wkey]
lfp_decoder_perfs = agg.decoder_perfs[wkey]
for k,aprop in enumerate(aprops):
ai = USED_ACOUSTIC_PROPS.index(aprop)
lfp_weights = lfp_decoder_weights[:, :, ai]
dstr = '%0.2f (%s)' % (lfp_decoder_perfs[ai], aprop)
plist.append({'type':'decoder', 'X':lfp_weights, 'title':dstr})
def _plot_X(_pdata, _ax):
plt.sca(_ax)
_X = _pdata['X']
if _pdata['type'] == 'decoder':
_absmax = np.abs(_X).max()
_vmin = -_absmax
_vmax = _absmax
_cmap = plt.cm.seismic
else:
_vmin = 0.
_vmax = 0.35
_cmap = magma
plt.imshow(_X, interpolation='nearest', aspect='auto', cmap=_cmap, vmin=_vmin, vmax=_vmax)
plt.title(_pdata['title'])
plt.xticks([])
plt.yticks([])
multi_plot(plist, _plot_X, nrows=5, ncols=6, figsize=(23, 13))
plt.show()
def draw_lags_vs_perf(data_dir='/auto/tdrive/mschachter/data'):
pfile = os.path.join(data_dir, 'GreBlu9508M', 'transforms', 'PairwiseCF_GreBlu9508M_Site4_Call1_L_raw.h5')
hf = h5py.File(pfile, 'r')
full_lags_ms = hf.attrs['lags']
hf.close()
max_lag = full_lags_ms.max()
agg_file = os.path.join(data_dir, 'aggregate', 'pard.h5')
agg = AcousticEncoderDecoderAggregator.load(agg_file)
print 'keys=',agg.df.keys()
lag_bnds = [ (-1., 1.), (-6., 6.), (-13., 13.), (-18., 18), (-23., 23), (-28., 28.), (-34., 34), (None,None)]
perfs_by_bound = list()
for lb,ub in lag_bnds:
if lb is None:
decomp = 'self+cross_locked'
else:
decomp = 'self+cross_locked_lim_%d_%d' % (lb, ub)
i = agg.df.decomp == decomp
print 'decomp=%s, i.sum()=%d' % (decomp, i.sum())
wkeys = agg.df.wkey[i].values
dperfs = np.array([agg.decoder_perfs[wkey] for wkey in wkeys])
perfs_by_bound.append(dperfs)
perfs_by_bound = np.array(perfs_by_bound)
pbb_mean = perfs_by_bound.mean(axis=1)
pbb_std = perfs_by_bound.std(axis=1)
plist = list()
for k,aprop in enumerate(REDUCED_ACOUSTIC_PROPS):
plist.append({'aprop':aprop, 'mean':pbb_mean[:, k], 'std':pbb_std[:, k]})
def _plot_pbb(pdata, ax):
_lagw = [2*x[1] for x in lag_bnds if x[1] is not None]
_lagw.append(max_lag*2)
plt.plot(_lagw, pdata['mean'], 'k-', linewidth=3.0, alpha=0.7)
plt.xlabel('Lag Width (ms)')
plt.ylabel('R2')
plt.axis('tight')
plt.title(pdata['aprop'])
multi_plot(plist, _plot_pbb, nrows=3, ncols=4)
plt.show()
def draw_figures(data_dir='/auto/tdrive/mschachter/data'):
bird = 'GreBlu9508M'
block = 'Site4'
seg = 'Call1'
hemi = 'L'
file_ext = '_'.join([bird, block, seg, hemi])
pcf_file = os.path.join(data_dir, bird, 'transforms', 'PairwiseCF_%s_%s_new.h5' % (file_ext, 'raw'))
pcf = PairwiseCFTransform.load(pcf_file)
g = pcf.df.groupby(['stim_id', 'order', 'stim_type'])
plist = list()
i = (pcf.df.decomp == 'full') & (pcf.df.electrode1 == pcf.df.electrode2)
assert i.sum() > 0
xfull_indices = list(pcf.df['index'][i].values)
print 'len(pcf.df)=%d' % len(pcf.df)
print 'pcf.df[index].max()=%d' % pcf.df['index'].max()
print 'pcf.psds.shape=',pcf.psds.shape
print 'xfull_indices.max()=%d' % max(xfull_indices)
print 'len(xfull_indices)=%d' % len(xfull_indices)
Xfull = pcf.psds[xfull_indices, :]
Xfull /= Xfull.max()
pcf.log_transform(Xfull)
Xfull -= Xfull.mean(axis=0)
Xfull /= Xfull.std(axis=0, ddof=1)
# take log transform of power spectra
i = (pcf.df.decomp == 'onewin') & (pcf.df.electrode1 == pcf.df.electrode2)
assert i.sum() > 0
xone_indices = list(pcf.df['index'][i].values)
Xonewin = pcf.psds[xone_indices, :]
Xonewin /= Xonewin.max()
pcf.log_transform(Xonewin)
Xonewin -= Xonewin.mean(axis=0)
Xonewin /= Xonewin.std(axis=0, ddof=1)
for (stim_id,order,stim_type),gdf in g:
electrodes = gdf.electrode1.unique()
stim_dur = gdf.stim_duration.values[0]
if stim_dur < 0.050 or stim_dur > 0.400:
continue
for e in electrodes:
i = (gdf.decomp == 'full') & (gdf.electrode1 == e) & (gdf.electrode1 == gdf.electrode2)
assert i.sum() == 1
xi = gdf['index'][i].values[0]
ii = xfull_indices.index(xi)
full_psd = Xfull[ii, :]
i = (gdf.decomp == 'onewin') & (gdf.electrode1 == e) & (gdf.electrode1 == gdf.electrode2)
assert i.sum() == 1
xi = gdf['index'][i].values[0]
ii = xone_indices.index(xi)
onewin_psd = Xonewin[ii, :]
plist.append({'full_psd':full_psd, 'onewin_psd':onewin_psd, 'stim_id':stim_id, 'stim_order':order,
'stim_type':stim_type, 'electrode':e, 'stim_dur':stim_dur})
np.random.shuffle(plist)
plist.sort(key=operator.itemgetter('stim_dur'))
short_plist = [x for k,x in enumerate(plist) if k % 20 == 0]
print 'len(short_plist)=%d' % len(short_plist)
def _plot_psds(_pdata, _ax):
absmax = max(np.abs(_pdata['full_psd']).max(), np.abs(_pdata['onewin_psd']).max())
plt.axhline(0, c='k')
plt.plot(pcf.freqs, _pdata['full_psd'], 'k-', linewidth=3.0, alpha=0.7)
plt.plot(pcf.freqs, _pdata['onewin_psd'], 'g-', linewidth=3.0, alpha=0.7)
plt.title('e%d: %d_%d (%s) %0.3fs' % (_pdata['electrode'], _pdata['stim_id'], _pdata['stim_order'], _pdata['stim_type'], _pdata['stim_dur']))
plt.axis('tight')
plt.ylim(-absmax, absmax)
multi_plot(short_plist, _plot_psds, nrows=5, ncols=9)
plt.show()
def draw_freq_lkrats(agg, df_me):
# aprops_to_display = ['category', 'maxAmp', 'meanspect', 'stdspect', 'q1', 'q2', 'q3', 'skewspect', 'kurtosisspect',
# 'sal', 'entropyspect', 'meantime', 'stdtime', 'entropytime']
#aprops_to_display = ['category', 'maxAmp', 'sal',
# 'entropytime', 'meanspect', 'entropyspect',
# 'q1', 'q2', 'q3']
aprops_to_display = ['category', 'maxAmp', 'sal',
'q1', 'q2', 'q3']
assert isinstance(agg, AggregateLFPAndSpikePSDDecoder)
freqs = agg.freqs
nbands = len(freqs)
# compute the significance threshold for each site, use it to normalize the likelihood ratio
i = agg.df.bird != 'BlaBro09xxF'
g = agg.df[i].groupby(['bird', 'block', 'segment', 'hemi'])
num_sites = len(g)
normed_lkrat_lfp = dict()
normed_lkrat_spike = dict()
for aprop in aprops_to_display:
normed_lkrat_lfp[aprop] = np.zeros([num_sites, nbands])
normed_lkrat_spike[aprop] = np.zeros([num_sites, nbands])
for k,((bird,block,segment,hemi),gdf) in enumerate(g):
i = (gdf.e1 != -1) & (gdf.e1 == gdf.e2) & (gdf.cell_index != -1) & (gdf.decomp == 'spike_psd') & \
(gdf.exel == False) & (gdf.exfreq == False) & (gdf.aprop == 'q2')
ncells = i.sum()
# print '%s,%s,%s,%s # of cells: %d' % (bird, block, segment, hemi, ncells)
# get the likelihood ratios and normalize them by threshold
for aprop in aprops_to_display:
for b in range(1, nbands+1):
i = (df_me.bird == bird) & (df_me.block == block) & (df_me.segment == segment) & (df_me.hemi == hemi) & \
(df_me.band == b)
if i.sum() != 1:
print 'i.sum()=%d, b=%d, (%s,%s,%s,%s)' % (i.sum(), b, bird, block, segment, hemi)
assert i.sum() == 1
lkrats_lfp = df_me[i]['lkrat_%s_%s' % (aprop, 'lfp')].values[0]
lkrats_spike = df_me[i]['lkrat_%s_%s' % (aprop, 'spike')].values[0]
normed_lkrat_lfp[aprop][k, b-1] = lkrats_lfp
normed_lkrat_spike[aprop][k, b-1] = lkrats_spike
# make a list of data for multi plot
plist = list()
for aprop in aprops_to_display:
plist.append({'lfp':normed_lkrat_lfp[aprop], 'spike':normed_lkrat_spike[aprop], 'freqs':freqs, 'aprop':aprop})
def _plot_freqs(pdata, ax):
plt.sca(ax)
nsamps_lfp = len(pdata['lfp'])
lkrat_lfp_mean = pdata['lfp'].mean(axis=0)
lkrat_lfp_std = pdata['lfp'].std(axis=0, ddof=1) / np.sqrt(nsamps_lfp)
nsamps_spike = len(pdata['spike'])
lkrat_spike_mean = pdata['spike'].mean(axis=0)
lkrat_spike_std = pdata['spike'].std(axis=0, ddof=1) / np.sqrt(nsamps_spike)
if pdata['aprop'] != 'category':
plt.axhline(1.0, c='k', linestyle='dashed', alpha=0.7, linewidth=2.0)
plt.errorbar(pdata['freqs'], lkrat_lfp_mean, yerr=lkrat_lfp_std, c=COLOR_BLUE_LFP, linewidth=7.0, alpha=0.9, ecolor='k', elinewidth=2.0)
plt.errorbar(pdata['freqs']+2., lkrat_spike_mean, yerr=lkrat_spike_std, c=COLOR_YELLOW_SPIKE, linewidth=7.0, alpha=0.9, ecolor='k', elinewidth=2.0)
plt.xlabel('Frequency (Hz)')
plt.ylabel('Normalized LR')
leg = custom_legend([COLOR_BLUE_LFP, COLOR_YELLOW_SPIKE], ['LFP', 'Spike'])
plt.legend(handles=leg, fontsize='x-small')
plt.title(pdata['aprop'])
plt.axis('tight')
if pdata['aprop'] != 'category':
plt.ylim(0, 5)
else:
plt.axis('tight')
figsize = (24, 9)
multi_plot(plist, _plot_freqs, nrows=2, ncols=3, hspace=0.45, wspace=0.45, facecolor='w', figsize=figsize)
fname = os.path.join(get_this_dir(), 'perf_by_freq.svg')
plt.savefig(fname, facecolor='w', edgecolor='none')
t_c = 0.5 * (1.0 / t_freq) - 0.010
strf = onset_strf(t,
f,
t_freq=t_freq,
t_phase=np.pi,
f_c=f_c,
f_sigma=1000.0,
t_sigma=t_sigma,
t_c=t_c)
title = '$f_c$=%dHz, $\sigma_t$=%dms, $f_t$=%dHz' % (f_c, t_sigma *
1e3, t_freq)
onset_plist.append({'strf': strf, 'title': title})
multi_plot(onset_plist,
plot_strf,
nrows=len(onset_f_c),
ncols=len(onset_t_sigmas))
#build harmonic stack STRFs
stack_t_sigma = 0.005
stack_f_sigma = 1500
stack_f_c = np.linspace(300.0, 8000.0, 10)
stack_f_freq = np.linspace(1e-4, 7e-4, 5)
stack_t_freqs = np.array([20.0, 15.0, 10.0, 5.0])
stack_plist = list()
for f_c in stack_f_c:
for f_freq in stack_f_freq:
strf = checkerboard_strf(t,
f,
#build onset STRFs of varying center frequency and temporal bandwidths
onset_f_sigma = 500
onset_f_c = np.linspace(300.0, 8000.0, 10)
onset_t_sigmas = np.array([0.005, 0.010, 0.025, 0.050])
onset_t_freqs = np.array([20.0, 15.0, 10.0, 5.0])
onset_plist = list()
for f_c in onset_f_c:
for t_sigma,t_freq in zip(onset_t_sigmas, onset_t_freqs):
t_c = 0.5*(1.0 / t_freq) - 0.010
strf = onset_strf(t, f, t_freq=t_freq, t_phase=np.pi, f_c=f_c, f_sigma=1000.0, t_sigma=t_sigma, t_c=t_c)
title = '$f_c$=%dHz, $\sigma_t$=%dms, $f_t$=%dHz' % (f_c, t_sigma*1e3, t_freq)
onset_plist.append({'strf':strf, 'title':title})
multi_plot(onset_plist, plot_strf, nrows=len(onset_f_c), ncols=len(onset_t_sigmas))
#build harmonic stack STRFs
stack_t_sigma = 0.005
stack_f_sigma = 1500
stack_f_c = np.linspace(300.0, 8000.0, 10)
stack_f_freq = np.linspace(1e-4, 7e-4, 5)
stack_t_freqs = np.array([20.0, 15.0, 10.0, 5.0])
stack_plist = list()
for f_c in stack_f_c:
for f_freq in stack_f_freq:
strf = checkerboard_strf(t, f,
t_freq=10.0, t_phase=0.0,
