If there is a 'noteworthy' energy cost when the brain is more active (if indeed this equates to 'thinking hard'), presumably this should result in an increase in overall metabolic rate, which could be measured, for instance, by respirometry. But I'm yet to find any papers that report on this...