-
Notifications
You must be signed in to change notification settings - Fork 2
Expand file tree
/
Copy pathDecoding_main.py
More file actions
132 lines (105 loc) · 5.2 KB
/
Decoding_main.py
File metadata and controls
132 lines (105 loc) · 5.2 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Wed Jan 29 14:36:51 2020
@author: numata
"""
import numpy as np
from SignalProcessing.preprocess_signal import Prep_signal
import matplotlib.pyplot as plt
from matplotlib import gridspec
from Utils.utils import Utilfunc, import_config
from Model.Decoding import Model
from FileIO.fileio import FileIO
from mpl_toolkits.axes_grid1.axes_divider import make_axes_locatable
from mpl_toolkits.axes_grid1.colorbar import colorbar
import sys
## import analysis config file
config = import_config()
## set analysis finger
finger_id = 4
## set subject number
subj_num =0
if __name__ == "__main__":
###### generate instances ######
prep = Prep_signal(config=config)
uti = Utilfunc(config)
Decoder = Model(config)
fio = FileIO(config)
###### load data ######
data = fio.loadBCI4()[subj_num]
##### Preprocess digit movement signal ######
resampled_dg = prep.Rectify(prep.downsample_sig(data['train_dg']), freqs=[1,10], btype='band', gaussian_pram=[200,500])
##### Preprocess ECoG signals ######
resampled_ecog = prep.downsample_sig(data['train_data'])
if (resampled_ecog.shape[-1] < 64) and (resampled_ecog.shape[-1] >48):
resampled_ecog = np.append(resampled_ecog, np.zeros([resampled_ecog.shape[0],2]),axis=1)
F_value = prep.Feature_Ext_filt(resampled_ecog, standardization=True, smoothing=True)
##### Set channel labels #####
chan = [str(i+1) for i in range(resampled_ecog.shape[1])]
finger = ['Thumb','Index','Middle','Ring','Pinky']
##### create event signals from digit movements #####
# This function is only for use BCI comp 4. (Dataset no.4)
# If you want to use your custom dataset included event signal, substitute it to "trigger".
trigger =prep.CreateTriggerBCI4(data['train_dg'], threshhold=0.5)
event = trigger[:,finger_id][:,np.newaxis].T
###### create feature Epochs ######
epR = uti.makeEpochs(resampled_ecog.T, event, chan).get_data()[:,0:-1,:]
ep_dg = uti.makeEpochs(resampled_dg.T, event, finger).get_data()[:,0:-1,:]
F_data = np.zeros((len(config['feature_freqs'].keys()),epR.shape[0],epR.shape[1],epR.shape[2]))
for i in range(len(config['feature_freqs'].keys())):
F_data[i,:,:,:] = uti.makeEpochs(F_value[:,:,i].T, event, ch_info=chan,
reference_type='Average').get_data()[:,0:-1,:]
Ep_dg = ep_dg.transpose(0,2,1)[:,:,finger_id]
F_data = np.reshape(F_data.transpose(1,2,0,3),[F_data.shape[1],
F_data.shape[0]*F_data.shape[2], F_data.shape[3]])
##### Set training and test dataset for decoding analysis.######
data_len= int(F_data.shape[0]*4/5)
train_ecog = F_data[0:data_len,:,:]
test_ecog = F_data[data_len:-1,:,:]
train_dg = Ep_dg[0:data_len,:]
test_dg = np.ravel(Ep_dg[data_len:-1,:])
##### Run decoding analysis ######
weight = Decoder.Fit(train_ecog.transpose(1,0,2), train_dg, key='PLS', PLS_components=1)
reconst_dg = Decoder.runReconst(test_ecog.transpose(1,0,2), weight =weight)
##### Reconstructiuon score evaluation #####
reconst_dg = uti.Zscore(reconst_dg)
test_dg = uti.Zscore(test_dg)
pad_len = int(config['Decoding']['sliding_step'] *config['Decoding']['sample_points'])
cc = np.round(np.corrcoef(reconst_dg[pad_len:-1],test_dg[pad_len:-1])[0,1],3)
##### calculate feature score ######
freqs= [config['feature_freqs'][list(config['feature_freqs'].keys())[i]] for i in range(len(config['feature_freqs'].keys()))]
spacial_weight = np.reshape(weight[0:-1],[len(chan),len(config['feature_freqs'].keys())])
freqs_domein= np.mean(spacial_weight,axis=0)
spatial_domein = np.mean(spacial_weight,axis=1)
freqs_score = (freqs_domein/np.sum(freqs_domein))*100
spacial_score = (spatial_domein/np.sum(spatial_domein))*100
reshape_score = np.reshape(spacial_score,[int(len(spacial_score)/8),8])
#plot
fig = plt.figure(figsize = (12,8))
gs = gridspec.GridSpec(2,2, width_ratios=[1,1])
ax0 = plt.subplot(gs[0,0:2])
ax0.plot(reconst_dg[pad_len:-1], label='Estimate from ECoG',linewidth =1.5)
ax0.plot(test_dg[pad_len:-1], label = 'Actual Digit movement',linewidth =1.5)
ax0.set_title('Subject no.'+str(subj_num+1)+'. Finger: '+finger[finger_id])
ax0.set_ylabel('Finger flection (Zscore)')
ax0.grid()
ax0.legend()
ax1 = plt.subplot(gs[1,0])
ax1.plot(np.arange(len(freqs)),freqs_score, color = 'dimgray', linewidth =2.0)
ax1.grid()
ax1.set_title('Frequency domein contribution ratio')
ax1.set_ylabel('Contribution ratio [%]')
ax1.set_xlabel('Frequency [Hz]')
ax1.set_xticklabels(freqs,rotation=60, fontsize=10)
ax2 = plt.subplot(gs[1,1])
im=ax2.pcolormesh(reshape_score,cmap='jet')
ax2_divider = make_axes_locatable(ax2)
cax2 = ax2_divider.append_axes("right", size="7%", pad="2%")
cb2 = colorbar(im, cax=cax2)
ax2.set_yticks(np.arange(len(spacial_score)/8)+0.5)
ax2.set_yticklabels(np.arange(len(spacial_score)/8)+1)
ax2.set_xticks(np.arange(8)+0.5)
ax2.set_xticklabels(np.arange(8)+1)
ax2.set_title('Spacial Contribution ratio [%]')
plt.show()