I'm very sorry for my incorrect answer - it is of course the onset of thalamic input to the cortex and not the onset of the evoked cortical responses itself. This is why the default of 60 ms is a reasonable assumption for auditory evoked responses, such
as the N100 starting slightly later around 100 ms. If you assume earlier activity such as P50, then the input parameter should reflect this.
On 23 Jul2013, at 4:28 PM, Vladimir Litvak wrote:
Dear Laura and Martin,
Martin's answer is not quite correct (although close). Onset is the time of the thalamic input to the cortex so it should slightly precede the first cortical peak of the ERP. For auditory paradigms the default value is probably OK. sd is the input width
(input is parametrized as a Gaussian volley). If your input is a brief tone then again, keep the default. If you have a something longer it might make sense to adjust your input parameters to better match the time course of the input. It is also possible to
generate a more complicated input waveform by combining several Gaussians with possibly different widths (just specify several onsets and sd's).
Best,
Vladimir