

In [1]:

    
clear all



In [2]:

    
import nibabel as nib
import os
import numpy as np
import scipy.io as sio
import scipy.optimize
from Tkinter import Tk
from tkFileDialog import askdirectory
import libtiff 
import matplotlib.pyplot as plt
%matplotlib inline

Get folder



In [3]:

    
Tk().withdraw() # we don't want a full GUI, so keep the root window from appearing
foldername = askdirectory() # show an "Open" dialog box and return the path to the selected file
print(foldername)









    



/media/sophie/TestTrigger/Data100298



In [4]:

    
path=foldername



In [5]:

    
A=foldername.split('/')
Dataname=A[-1]



In [6]:

    
Dataname









    Out[6]:





'Data100298'



In [7]:

    
from PIL import Image

Open the images



In [8]:

    
im = Image.open(path+'/'+Dataname+'_1.tif')
tt = np.array(im)

S=tt.shape
data=np.zeros([S[0],S[1],len(os.listdir(path))])
#for i in range(1,15000):
for i in range(len(os.listdir(path))):       
#for fn in os.listdir(path):
    im = Image.open(path+'/'+Dataname+'_'+str(i+1)+'.tif')
    tt = np.array(im)
    data[:,:,i]=tt[:][:]
    i=i+1



In [9]:

    
S=data.shape
S









    Out[9]:





(4, 2048, 7506)

Find end of onset of light and begining of offset (to align to behavior)

Calculate average time series



In [10]:

    
M=np.mean(np.mean(data,0),0)
Mav=M.mean()
M=M/Mav



In [11]:

    
Mav









    Out[11]:





1613.2637699085524



In [12]:

    
plt.plot(M,'+')
#plt.axis([-1,200,0,1.5])









    Out[12]:





[<matplotlib.lines.Line2D at 0x7f07dcd35e90>]

Get approxiamte on and off times



In [13]:

    
liston=[i for i in range(len(M)) if M[i]>0.5]
liston[0]









    Out[13]:





52

Model for fitting onset and offset



In [14]:

    
def model(x,a,b,c,d):
    if x<a:
        return b
    elif x<c:
        return b+(x-a)*d
    else:
        return (c-a)*d+b

Model onset and find precise onset time



In [15]:

    
Ms=M[range(liston[0]-8,liston[0]+8)]



In [16]:

    
def Sq(X):
    return sum([(model(i,X[0],X[1],X[2],X[3])-Ms[i])**2 for i in range(len(Ms))])



In [17]:

    
liston[0]-8









    Out[17]:





44



In [18]:

    
res = scipy.optimize.minimize(Sq,x0=[7.6,0.3,9.1,0.9])



In [19]:

    
ON=liston[0]-8+res.x[2]
print(ON)









    



51.996600864



In [20]:

    
ONint=np.int(np.ceil(ON))
#ONint=1
print(ONint)



In [21]:

    
plt.plot(np.squeeze(M[range(liston[0]-8,liston[0]+8)]),'+')
plt.plot(np.arange(0,len(Ms),0.1),[model(i,res.x[0],res.x[1],res.x[2],res.x[3]) for i in np.arange(0,len(Ms),0.1)])
plt.show()

Model offset and find precise offset time



In [22]:

    
Ms=M[range(liston[len(liston)-1]-6,liston[len(liston)-1]+6)]



In [23]:

    
def Sq(X):
    return sum([(model(i,X[0],X[1],X[2],X[3])-Ms[i])**2 for i in range(len(Ms))])



In [24]:

    
res = scipy.optimize.minimize(Sq,x0=[6,3,8,-1])



In [25]:

    
OFF=liston[len(liston)-1]-6+res.x[0]
#OFF=liston[len(liston)-1]
print(OFF)
OFFint=np.int(np.floor(OFF))
print(OFFint)









    



7472.00059572
7472



In [26]:

    
plt.plot(np.squeeze(Ms),'+')
plt.plot(np.arange(0,len(Ms),0.1),[model(i,res.x[0],res.x[1],res.x[2],res.x[3]) for i in np.arange(0,len(Ms),0.1)])
plt.show()

Open images times



In [32]:

    
TimeFile='/home/sophie/Downloads/'+''.join([Dataname[i] for i in range(10)])+'_.csv'



In [33]:

    
print(TimeFile)









    



/home/sophie/Downloads/Data100298_.csv



In [34]:

    
Listfile = open(TimeFile, 'r')
ListTime = [line.split('\n')[0] for line in Listfile.readlines()]



In [35]:

    
Timespl=[float(ListTime[i].split(',')[2]) for i in range(1,len(ListTime))]



In [36]:

    
Ttest=np.array(Timespl[2:(len(Timespl))])-np.array(Timespl[1:(len(Timespl)-1)])
plt.plot(Ttest)
#plt.axis([10000,11000,0,0.6])









    Out[36]:





[<matplotlib.lines.Line2D at 0x7f07da204910>]

Get times corresponding to images during light on (excitation light completely on : t=0)



In [50]:

    
print(ONint)
print(OFFint)
print(ON)
print(OFFint-ONint)
print(OFF-ON)









    



52
7472
51.996600864
7420
7420.00399486



In [38]:

    
TimeOn=[Timespl[i] for i in range(ONint,(OFFint+1))]



In [39]:

    
Tinit=(ON-(ONint-1))*(Timespl[ONint]-Timespl[ONint-1])+Timespl[ONint-1]



In [48]:

    
Toff=(OFF-OFFint)*(Timespl[OFFint+1]-Timespl[OFFint])+Timespl[OFFint]



In [49]:

    
Toff-Tinit









    Out[49]:





148.57789400080947



In [42]:

    
Timespl[ONint]-Timespl[ONint-1]









    Out[42]:





0.020027000000000017



In [43]:

    
TimeOn[0]-Tinit









    Out[43]:





6.8074495863656992e-05



In [44]:

    
import numpy as np



In [45]:

    
TimeOnFinal=np.array(TimeOn)-Tinit



In [46]:

    
sio.savemat('/home/sophie/Desktop/'+Dataname+'TimeFluoOn.mat', {'TimeFluoOn':TimeOnFinal})

Keep only the frames for which the excitation is on and save



In [48]:

    
D4=np.transpose(data[:,:,:,range(ONint,(OFFint+1))],(2,1,0,3))
nim=nib.Nifti1Image(D4,np.eye(4))
nib.save(nim,'/home/sophie/Desktop/'+Dataname+'on.nii.gz')



In [ ]:



In [ ]: