Does cognition-specific computer training have better clinical outcomes than non-specific computer training? A single-blind, randomized controlled trial