notebook.community

Edit and run



In [1]:

    
# block_path = '/mnt/cube/ephys-example-data/Pen01_Lft_AP2500_ML1350__Site10_Z2026__B997_cat_P01_S10_Epc10/'
block_path = '/mnt/cube/nyoni-raw/test_data/'
# block_path = '/mnt/cube/ephys-example-data/Pen04_Lft_AP2555_ML500__Site02_Z2688__st979_cat_P04_S02_1/'

First, let's load the clusters.



In [4]:

    
from ephys import core



In [6]:

    
clusters = core.get_clusters(block_path)
print clusters.head()









    



---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-6-43f5c2e22007> in <module>()
----> 1 clusters = core.get_clusters(block_path)
      2 print clusters.head()

AttributeError: 'module' object has no attribute 'get_clusters'

Let's check the distribution of cluster values.



In [4]:

    
clusters.quality.value_counts()









    Out[4]:





Good        28
unsorted    24
MUA         17
Noise        2
Name: quality, dtype: int64



In [5]:

    
%matplotlib inline
import seaborn as sns

sns.countplot('quality',data=clusters)









    Out[5]:





<matplotlib.axes._subplots.AxesSubplot at 0x7fe1720bbd50>

Sweet! Let's get a new dataframe without the noise or unsorted clusters



In [6]:

    
neurons = (
    clusters[clusters.quality.isin(['Good','MUA'])]
    .sort_values(['quality','cluster'])
    .reset_index()
)
print neurons.quality.unique()









    



['Good' 'MUA']

Nice! Now, let's compute some metrics.



In [7]:

    
from ephys import clust

neurons['width'] = neurons['cluster'].map(
    lambda clu: clust.get_width(block_path,clu)
)

print neurons.head()









    



   index  cluster quality     width
0      2        3    Good  0.000546
1      3        4    Good  0.000432
2      7        9    Good  0.000202
3     10       13    Good  0.000698
4     11       15    Good  0.000389



In [8]:

    
sns.distplot(neurons['width'])









    Out[8]:





<matplotlib.axes._subplots.AxesSubplot at 0x7fe1bc2dbf50>



In [9]:

    
neurons['x_probe'] = neurons['cluster'].map(
    lambda clu: clust.get_cluster_coords(block_path,clu,weight_func=clust.mean_masks_w)[0]
)
neurons['y_probe'] = neurons['cluster'].map(
    lambda clu: clust.get_cluster_coords(block_path,clu,weight_func=clust.mean_masks_w)[1]
)



In [10]:

    
import matplotlib.pyplot as plt
plt.scatter(neurons['x_probe'],neurons['y_probe'])
plt.axis('equal')









    Out[10]:





(-0.0060000000000000001, 0.0060000000000000001, -100.0, 800.0)



In [11]:

    
from ephys import viz
f,ax = plt.subplots(1,1,figsize=(8,16))
viz.plot_all_clusters(block_path,neurons,scale_factor=0.05)



In [19]:

    
viz.plot_cluster_locations(block_path)



In [18]:

    
sns.set_style('white')