Synaptic Current Analysis with Axobasics -question

Matt Jones jonesmat at
Fri Aug 25 14:43:12 EST 1995

In article <41kqlj$h3m at> Andras Fancsik,
fancsik at writes:
>I would like to know if anyone has this same problem (I am sure, there
houndreds out 
>there analizing PSC-s) and what software they use. I would prefer the
solutions with 
>programs written on the basis of AXOBASICS, because that way we would
not have to 


An alternative to using threshold-crossing of the current is to use
threshold crossing of the derivative of current. Try making a trace in
Axobasic that is the derivative of a current record, then display it over
the current. You'll see that while the current noisily oscillates around
threshold, the derivative makes a sharp spike during the rising phase of
the PSC. This spike is a lot easier to detect. The height of the spike
depends on the speed of rise of the current, which is handy because it
allows you to exclude any signals that are slow and potentially filtered
by the cell's cable properties. The main problems now become how to
decide on the best threshold conditions, how to deal with overlapping
minis, etc...

Here's the code I use for making the derivative trace:
Say you have raw data in integer locations 0 thru 1023, and some extra
space for the derivative trace in 1024 thru 2047. 

movl 0, 1024, 1024							'copy the data trace into spare buffer space
																 						'you can filter the data here, if you want, or
just leave it.
subl 1025, 1024, 1024  			'get the point-to-point difference
mulln srate%, 1024, 1024  	'multiply by the sampling rate. Locations
1024-2048 now 
																						'contain the derivative of the raw data. 
disl 0,1024													'display the raw data
disl 1024, 1024										'display the derivative

Then you can detect a peak using any number of methods. A really simple
(but slow) one is to
have a FOR/NEXT loop that goes through the derivative one point at a time
and looks for values greater than a threshold. 

Have fun,
Matt Jones

More information about the Neur-sci mailing list