Journal ClubWe all know that toiling outside on a hot daycan conjure up all sorts of fantasies aboutlong, cool drinks, but does working up athirst actually affect our perceptions of reality?This is the question that Changizi and Hall posein a new paper in Perception[1]. They reasonedthat it would be highly adaptive for individualsto be biased toward perceiving water whenthey are thirsty; it is much less costly to seewater mistakenly in a place where there isn’tany, than to think that there isn’t any wateravailable when, in fact, there is. Changizi andHall therefore suggested that we should expectthirsty individuals to show perceptual biasestowards water-like properties of visual scenes.The specific attribute they chose to test wastransparency, because this is a typical propertyof water and also because there is no obviousassociation between thirst and transparency,so subjects would not realize what wasbeing tested and thus bias the results.Seventy-four subjects were divided intotwo groups: a ‘thirsty’ group, who ate a packetof salty crisps before the experiment, and a‘non-thirsty’ group, who drank water tosatiation. Subjects were then tested on oneof two stimulus sets. One set consisted of acircular pattern in which there were ‘definitely’transparent stimuli, ‘ambiguously’ transparentstimuli and ‘definitely not’ transparentstimuli. The second set of stimuli was of thesame format, except that the patterns wererectangular. Changizi and Hall predicted thatthirsty subjects would be more likely to classify‘ambiguous’ stimuli as transparent than thenon-thirsty subjects, but that they would showno difference in cases where the stimuli weredefinitely transparent or not transparent.Overall, for both stimulus sets, there wasa significant difference between groups inthe number of ambiguous stimuli perceivedto have a transparent surface: 58% of thirstysubjects perceived transparency, whereasonly 47% of non-thirsty subjects did so. Inaddition, the number of ‘pro-transparency’subjects (those who perceived more thanhalf of the ambiguous stimuli to betransparent) was greater for the thirsty groupthan the non-thirsty group – for all stimulicombined, only 14 of 37 non-thirsty subjectswere pro-transparency compared with 25 of37 thirsty subjects. As also predicted, therewas no difference between groups in howthey perceived the unambiguous stimuli.The results thus supported Changizi andHall’s hypothesis and, moreover, showed thata very small change in the subjects’ biologicalstate (small enough to be induced by a singlepacket of crisps) was sufficient to have asignificant influence on how they perceiveda visual scene. The authors conclude that,in real-world settings, such biases wouldundoubtedly contribute to individuals’subsequent behavioural responses,increasing their likelihood of making beneficialdecisions. And on that note, I’m off to the pub.