How to analyze a sample with few data points per participant in a familiarization paradigm in artificial grammar learning?