Model comparison in psychophysics with multiple participants


New Member

I have a quite general methodological question about model comparison.
I ran a psychophysical experiment in which each participants performed ~500 trials of a perceptual decision task. My observation are reaction times.
I want to compare several possible model to find which one explains better the data.

I can use a criterium such as AIC or BIC to compare the models for each participant, but how should I then combine these individual scores to conclude which model explains best the data in the population?
Am I allowed, for instance, to take the mean of individual BIC scores to have a "population" score corresponding to each model?

Thank you for your help on this topic, I hope I was clear enough.