Skip to main content

Table 1 For each type of test event, and for each of the mRFS, WSC, RFS, and PCA methods, we list the total number of scatterplots (on the denominator) and the number of plots that show four clear clusters (on the numerator)

From: Spike sorting by a minimax reduced feature set based on finite differences

Type

Analysis methods

mRFS

WSC

RFS

PCA

2SD

0/16

0/496

0/496

0/496

3SD

3/16

3/496

0/496

0/496

(1 vs. 3, 0.61)

(8 vs. 22, 0.59)

  

(1 vs. 1, 0.59)

(10 vs. 22, 0.58)

  

(3 vs. 2, 0.57)

(8 vs. 21, 0.57)

  

4SD

9/16

7/496

3/496

0/496

(1 vs. 3, 0.81)

(10 vs. 22, 0.80)

(17 vs. 19, 0.77)

 

(1 vs. 1, 0.80)

(10 vs. 21, 0.79)

(17 vs. 18, 0.55)

 

(3 vs. 2, 0.79)

(8 vs. 9, 0.78)

(17 vs. 20, 0.52)

 

(2 vs. 2, 0.78)

(9 vs. 11, 0.77)

  

(0 vs. 3, 0.71)

(8 vs. 21, 0.76)

  

(0 vs. 1, 0.66)

(8 vs. 22, 0.76)

  

(2 vs. 0, 0.62)

(9 vs. 21, 0.72)

  

(1 vs. 2, 0.59)

   

(3 vs. 0, 0.51)

   
  1. For each case with four clear clusters, we also list the rate of correct classifications by the fuzzy c-means clustering algorithm