Sparse Data Analysis Strategy for Neural Spike Classification

Table 2

An example of calculation of the 𝐶_𝜆𝑚 criterion calculated from the matrix of the introductory example cf Figure 1(a).

Matrix of common neighbors	𝐵=200300010000003033200300003033003033
Parsimony level	𝜆≥1	𝜆≥2	𝜆≥3
Binary matrices of common neighbors	𝐵=100100010000001011100100001011001011	𝐵=100100000000001011100100001011001011	𝐵=100100000000001011100000001011001011
∑^6𝑖=1∑^5𝑗=1|𝑏^𝜆𝑚𝑖,𝑗−𝑏^𝜆𝑚𝑖,𝑗+1|	17	15	9
(calculus per line)	(3+2+3+3+3+3)	(3+0+3+3+3+3)	(0+0+3+0+3+3)
Sorted binary matrices	𝐵^sor1=110000110000001110001110001110000001	𝐵^sor2=110000110000001110001110001110000000	𝐵^sor3=111000111000111000000000000000000000
∑^6𝑖=1∑^5𝑗=1|(𝑏^𝜆𝑚𝑖,𝑗)sort−(𝑏^𝜆𝑚𝑖,𝑗+1)sort|	9	85	3
(calculus per line)	(1+1+2+2+2+1)	(1+1+2+2+2+0)	(1+1+11+0+0+0)
Criterion 𝐶𝜆𝑚	𝐶𝜆𝑚≥1=9/17=1.89	𝐶𝜆𝑚≥2=8/15=1.88	𝐶𝜆𝑚≥1=3/9=3.03