Format

Send to

Choose Destination
J Comput Neurosci. 2006 Dec;21(3):329-42. Epub 2006 Aug 14.

Robustness of the significance of spike synchrony with respect to sorting errors.

Author information

1
Neuroinformatics, Institute for Biology - Neurobiology, Free University, Königin-Luise Str. 28/30, 14195, Berlin, Germany. antonio.pazienti@neurobiologie.fu-berlin.de

Abstract

The aim of spike sorting is to reconstruct single unit spike times from extracellular multi-unit recordings. Failure in the identification of a spike (false negative) or assignment of a spike to a wrong unit (false positive) are typical examples of sorting errors. Their influence on cross-correlation measures has been addressed and it has been shown that correlation analysis of multi-unit signals may lead to incorrect interpretations. We formulate a model to study the influence of sorting errors on the significance of synchronized spikes, and thus are able to study if and how the significance changes in case of imperfect sorting. Here we explore the case of pairwise analysis of simultaneously recorded neurons. Interestingly, a decrease in the significance is observed in the presence of false positives, as well as for false negatives. Furthermore, false negative errors reduce the significance of synchronized spikes more strongly than false positives. Thus, conservative sorting strategies have a stronger tendency to lead to a loss of the significance of synchronization. We demonstrate that a detailed understanding of sorting techniques and their possible effects on subsequent data analyses is important in order to rule out inconsistencies in the interpretation of results.

PMID:
16927209
DOI:
10.1007/s10827-006-8899-7
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Springer
Loading ...
Support Center