SSVEP response study for low semantic images