TABLE 1.

Algorithm outline for automatic detection of GCaMP6f signals.

Algorithm outline
Input: Aligned time sequence images I_j (j=1,2,⋯,J)
Output: GCaMP6f signals’ profile S_m (m=1,2,⋯,M) from activated
DRG neurons
1 Calculate the variance map V of the image stack I_j (j=1,2,⋯,J) along
the time dimension
2 Initialize the global threshold HG, the active neuron number P and neuron size
rang Rmin, Rmax
3 for n = 1: N (different iterations) with condition if identified neurons p <
estimation number P
4   Variance map VGlobalthresholdHG⇒ Binary image BWglobal
Variance map VAdaptivethresholdHA⇒ Binary image BWadaptive
5   BWcombined=BWglobal⋅BWadpative, where ‘⋅’ stands for point-wise
multiplication
6   Morphological closing operation: BWclose=(BWcombined⊕SE)⊖SE
where structuring element SE=[0,1,0;1,1,1;0,1,0]
7   Morphological opening operation: BWopen=(BWclose⊖SE)⊕SE
where use the same structuring element SE as above
8   BWopenRemovethesmallfeatures⇒BWopen_updated
9   BWopen_updatedClearthefeaturesatimageborder⇒BWupdated
10   BWupdatedCalculatethedistancematrix⇒Mdistance
11   MdistanceWatershedtransform⇒Lwatershed
12   LwatershedSelectneuronsizebetweenRminRmax⇒Lselected
13   Update the global threshold HG_updated=HG*α,where step size
α=0.75
14 end
15 for m = 1:M (different neuron regions)
16  for j = 1: J (different captured images)
17   Sm_j= Mean (Lselect_m⋅Ij) Calculate the average intensity profile
of the neuron
18  end
19 end