No abstract

SummaryData analysis workflows in many scientific domains have become increasingly complex and flexible. To assess the impact of this flexibility on functional magnetic resonance imaging (fMRI) results, the same dataset was independently analyzed by 70 teams, testing nine ex-ante hypotheses. The flexibility of analytic approaches is exemplified by the fact that no two teams chose identical workflows to analyze the data. This flexibility resulted in sizeable variation in hypothesis test results, even for teams whose statistical maps were highly correlated at intermediate stages of their analysis pipeline. Variation in reported results was related to several aspects of analysis methodology. Importantly, meta-analytic approaches that aggregated information across teams yielded significant consensus in activated regions across teams. Furthermore, prediction markets of researchers in the field revealed an overestimation of the likelihood of significant findings, even by researchers with direct knowledge of the dataset. Our findings show that analytic flexibility can have substantial effects on scientific conclusions, and demonstrate factors related to variability in fMRI. The results emphasize the importance of validating and sharing complex analysis workflows, and demonstrate the need for multiple analyses of the same data. Potential approaches to mitigate issues related to analytical variability are discussed.

Recent findings show that preferences for food items can be modified without external reinforcements using the cue-approach task. In the task, the mere association of food item images with a neutral auditory cue and a speeded button press, resulted in enhanced preferences for the associated stimuli. In a series of 10 independent samples with a total of 255 participants, we show for the first time that using this non-reinforced method we can enhance preferences for faces, fractals and affective images, as well as snack foods, using auditory, visual and even aversive cues. This change was highly durable in follow-up sessions performed one to six months after training. Preferences were successfully enhanced for all conditions, except for negative valence items. These findings promote our understanding of nonreinforced change, suggest a boundary condition for the effect and lay the foundation for development of novel applications.Behavioral change is an essential tool to improve health and quality of life, from treating addictions to eating and mood disorders [1][2][3] . Scientific research on behavioral change mainly focused on the effects of external reinforcements 2,4-7 or altering the presentation of the decision problem [8][9][10] . Recently, the cue-approach training (CAT) 11 paradigm was introduced as a successful method for enhancing preferences for food items, without external reinforcement, context change or self-control. Here, we test multiple hypotheses that are aimed to shed light on this mechanism by studying its generalizability to multiple stimuli and cues, as well as the long-term durability of the effect. In the cue-approach task, participants initially indicated their preferences for a set of snack-food items by specifying their willingness to pay for each item in an auction procedure. Then, in the CAT, some of the items were consistently associated with a neutral auditory cue and a speeded button press response (these stimuli were termed 'Go items'), whereas other stimuli were presented without a cue ('No-Go items'). In the following probe phase, participants were asked to choose a snack they would like to eat at the end of the experiment. Each probe-choice comprised of a pair of items with similar initial values, in which only one of the two snacks was a Go item. Results showed that the mere association of snack-food images with a neutral auditory cue and a speeded button press, resulted in enhanced preferences for Go items over No-Go items. This preference change effect varied across different value categories -resulting in enhanced preferences for snack-food items of initial high-value, yet significantly less prominent change in preferences for low-value items. The effect was maintained two-months following training 11 . Additional studies with the cue-approach task 12 found that for the behavioral change to take place, CAT required the presence of both a speeded button press response and a cue; i.e. CAT had no effect when training was conducted with an early cue onset which was followed with ...