A Framework for the Comparative Assessment of Neuronal Spike Sorting Algorithms towards More Accurate Off-Line and On-Line Microelectrode Arrays Data Analysis

Table 3

Properties of the implemented and evaluated clustering algorithms.

	Input	Percentage of publications; (with respect to Table 1)	Automaticity	Parametric	Need of training before on-line clustering
-means	Spike features	30%	Yes	Yes	—
Fuzzy--means (FCM)	Spike features	25%	Yes	Yes	Yes
Density-based (DBC)	Spike features	4%	Yes	No	Yes
O-sort	Spikes	11%	Yes	Yes	No
Other methods	—	30%	—	—	—

The “percentage of publications (with respect to Table 1)” is the ratio between the number of publications dealing with a certain clustering method and the total number of analyzed publications (reported in Table 1). “Automaticity” refers to the possibility not to define a number of clusters a priori. “Parametric” refers to the need to set one or more threshold values for parameters involved in the algorithm. The “need of training before on-line clustering” column refers to the necessity of an off-line phase to train the algorithms on data before an on-line classification can be performed. The optimization of a proper index during training phase is needed. Including nonautomatic methods (e.g., manual methods) and methods not suitable for on-line mode.