nan preassignment failed under caser 'threshold'

update from Mensen

update from Sinergia-BMZ

excludes time from start to first stage from timing calculations | first wake stage considered as lights off and latencies calculated according to this time and not the recording start | added function to calculate the sleep transition probability matrix

added parameter to indicate with which canonical wave (mean, primary, or all) the channel correlation should be performed with

added functions to plot the average slow wave found as well as extracting a single example from the raw traces

added function to perform a spectral analysis on the data, with and without the detected spindles to get a rough impression of detection success

save output using the newer matlab 7.3 flag for larger files | fixed bug with scoring but no marked arousals

added a maximum amplitude criteria to ignore clear artefacts | small changes for analysis other data types

update to spindle default, change back to median deviation | check for spindle type in the wave summary

find spindle frequency peak in ref | added and changed default channel spindle detection to power (pwelch) based method | adjusted explorer and plotting according to new names

new soft amplitude threshold to calculate real start of spindle after hard detection default(50% of hard threshold) | improved double spindle detection between canonical waves

new wavelet function for spindle detection | find frequency peak using power spectrum for each spindle