Detection ofSpontaneous Synaptic Eventswithan Optimally ScaledTemplate

Spontaneous synaptic events can bedifficult todetect whentheir amplitudes areclosetothebackground noise level. Herewe report a sensitive new technique forautomatic detection ofsmallasynchronous events.A waveformwiththe time courseofatypical synaptic event(atemplate) isslid alongthecurrent orvoltage traceandoptimally scaled tofit thedata ateachposition. A detection criterion iscalculated basedon theoptimumscaling factor andthequality ofthefit. Anevent isdetected whenthis criterion crossesa threshold level. Thealgorithm automatically compensatesforchangesinrecording noise. Thesensitivity andselectivity ofthemethodwere testedusingrealandsimulated data,andtheinfluence ofthe template parameter settings was investigated. Itsperformance was comparable tothatobtained byvisual eventdetection, anditwas more sensitive thanpreviously described threshold detection techniques. Undertypical recording conditions, all fast synaptic events withamplitudes ofatleast three times thenoise standard deviation (3a) couldbedetected, ascould75% ofevents withamplitudes of2o-. Thescaled template technique isimplemented within a commercial dataanalysis application andcan beapplied tomany standard electrophysiological datafile formats.