Evaluating single channel ionic current by fuzzy clustering with a partition validation index

Single channel ionic current at several conductance states has been evaluated by a machine learning approach based on fuzzy clustering. Current jumps values have been clustered into an optimal number of well separated partitions using an original fuzzy validation index. This optimal partitions scheme allows to discriminate single channel classes discarding multi channels and spike due cur-rent evaluating and using of the proper open single channel noise. This method may be applied to step-like or burst-like single ionic current jumps. Specifically, burst-like single ionic current jumps generated by tetanus toxin (TeTx) have been studied. This case study represents a typical methodological validation example due to the presence of sev-eral spikes, current multi-channels and different conductance.