[Eeglablist] VEP P100 latencies ?

[Lester John]

Is there someone on the EEGLAB list who has some experience with VEP
P100 waveforms ?

We have an experimental setup using EPrime (PC based) to generate
reversing checkerboards, interfaced to a 128 channel EGI system (Apple
Mac based). The p100 waveforms are clearly visible, but there is a
variation in P100 latency between trials. The question is whether this
variation is acceptable for p100's or not. 

There are 3 subjects (1,2,3), each tested 5 times (trials 1-5), on 3
different days (a,b,c). Latency is in ms. 

trial no.	subject1a	subject1b	subject1c	

1		110		107		107	
2		103		106		109	
3		108		102		108	
4		109		107		104	
5		109		107		104	


trial no.	subject2 a	subject2b	subject2c	
1		104		104		105	
2		104		108		106	
3		104		103		108	
4		104		107		108	
5		102		107		107	


trial no.	subject3a	subject3b	subject3c
1		102		97		102
2		103		98		95
3		103		103		100
4		102		99		102
5		102		99		102

Is it acceptable to use a 5 trial mean, or mode ? We suspect that the
latency variation could be due to gaze wandering. The room is darkened,
and the subjects have a monocular eyepatch. The VEP is lowpass filtered
at 100Hz.

yours sincerely,

Lester, 
UCT (South Africa)