Home Cyberlaw Ars Technica Some types of brain studies need thousands of participants to be reliable