chirp_probing/util.py

143 lines
4.9 KiB
Python

from typing import ValuesView
import numpy as np
import scipy.signal as sig
from numpy.lib.function_base import iterable
from numpy.lib.index_tricks import diag_indices
def despine(axis, spines=None, hide_ticks=True):
def hide_spine(spine):
spine.set_visible(False)
for spine in axis.spines.keys():
if spines is not None:
if spine in spines:
hide_spine(axis.spines[spine])
else:
hide_spine(axis.spines[spine])
if hide_ticks:
axis.xaxis.set_ticks([])
axis.yaxis.set_ticks([])
def gaussKernel(sigma, dt):
""" Creates a Gaussian kernel with a given standard deviation and an integral of 1.
Args:
sigma (float): The standard deviation of the kernel.
dt (float): The temporal resolution of the kernel, given in seconds.
Returns:
numpy.ndarray : the kernel in the range -4 to +4 sigma
"""
x = np.arange(-4. * sigma, 4. * sigma, dt)
y = np.exp(-0.5 * (x / sigma) ** 2) / np.sqrt(2. * np.pi) / sigma
return y
def extract_am(signal):
"""Extract the amplitude modulation from a signal using the Hilbert transform. Performs padding to avoid artefacts at beginning and end.
Args:
signal (np.ndarray): the signal
Returns:
np.ndarray: the am, i.e. the absolute value of the Hilbert transform.
"""
# first add some padding to both ends
front_pad = np.flip(signal[:int(len(signal)/100)])
back_pad = np.flip(signal[-int(len(signal)/100):])
padded = np.hstack((front_pad, signal, back_pad))
# do the hilbert and take abs, cut away the padding
am = np.abs(sig.hilbert(padded))
am = am[len(front_pad):-len(back_pad)]
return am
def firing_rate(spikes, duration, sigma=0.005, dt=1./20000.):
"""Convert spike times to a firing rate using the kernel convolution with a Gaussian kernel
Args:
spikes (iterable): list of spike times, times should be in seconds
duration (float): duration of the trial in seconds
sigma (float, optional): standard deviation of the Gaussian kernel. Defaults to 0.005s.
dt (float, optional): The stepsize of the trace. Defaults to 1./20000.s.
Returns:
np.ndarray: the firing rate
"""
binary = np.zeros(int(np.round(duration/dt)))
indices = np.asarray(np.round(spikes / dt), dtype=np.int)
binary[indices[indices < len(binary)]] = 1
kernel = gaussKernel(sigma, dt)
rate = np.convolve(kernel, binary, mode="same")
return rate
def spiketrain_distance(spikes, duration, dt, kernel_width=0.001):
"""Calculate the Euclidean distance between spike trains. Firing rates are estimated using the kernel
convloution technique applying a Gaussian kernel of the given standard deviation.
Args:
spikes (list of iterable): list of spike trains. event times are given in seconds.
duration (float): duration of a trial given in seconds.
dt (float): stepsize of the recording, given in seconds.
kernel_width (float, optional): standard deviation of the Gaussian kernel used to estimate the firing rate. Defaults to 0.001.
Returns:
np.ndarray: the distances
"""
# perform some checks
if not isinstance(spikes, list):
raise ValueError("spikes must be a list of spike trains, aka iterables of spike times.")
if len(spikes) > 1 and not isinstance(spikes[0], iterable):
raise ValueError("spikes must be a list of spike trains, aka iterables of spike times.")
rates = np.zeros((len(spikes), int(duration/dt)))
for i in range(len(spikes)):
rates[i,:] = firing_rate(spikes[0], duration, kernel_width, dt)
distances = np.zeros((len(spikes), len(spikes)))
for i in range(len(spikes)):
for j in range(len(spikes)):
if i < j:
distances[i, j] = np.sqrt(np.sum((rates[i,:] - rates[j,:])**2))
distances[j, i] = distances[i, j]
elif i == j:
distances[i, j] = 0.0
else:
break
return distances
def across_group_distance(rates1, rates2, axis=0):
if axis == 1:
rates1 = rates1.T
rates2 = rates2.T
distances = np.zeros((rates1.shape[axis], rates2.shape[axis]))
for i in range(distances.shape[0]):
for j in range(distances.shape[1]):
distances[i, j] = np.sqrt(np.sum((rates1[i,:] - rates2[j,:])**2))/rates1.shape[1-axis]
return distances
def within_group_distance(rates, axis=0):
distances = np.zeros((rates.shape[axis], rates.shape[axis]))
if axis == 1:
rates = rates.T
for i in range(distances.shape[0]):
for j in range(distances.shape[1]):
if j < i:
distances[i, j] = np.mean(np.sqrt(np.sum((rates[i,:] - rates[j,:])**2)))/rates.shape[1-axis]
distances[j, i] = distances[i, j]
elif i == j:
distances[i, j] = 0.0
else:
break
return distances