In the encoding of narrative episodes, the hippocampus exhibits memory-predictive activity time-locked to stimulus offset. In real life, however, events usually occur in succession, raising the question of how the immediate offline processing of one event is affected by presentation of another. To address this issue, participants were presented with brief narrative movie clips in a functional magnetic resonance imaging scanner. Each clip was immediately followed by an additional, unrelated, clip; by a visually scrambled clip with background auditory noises; or by a fixation cross. Memory for the gist of the clips was tested outside the scanner in a cued-recall test 20 min after termination of the study session. The hippocampus responded at the offset of each clip, even when a second clip was presented in immediate succession, suggesting that the hippocampus processes each brief clip as a discrete event. Presentation of a second narrative clip, and to a lesser degree of a scrambled clip, retroactively interfered with memory for the first clip. In parallel, the offline response of the posterior hippocampus to the first movie was reduced. In the anterior hippocampus, presentation of a second clip did not reduce the overall offline response but significantly reduced the difference in activity between remembered and forgotten clips. These findings are in line with the proposition that immediate offline hippocampal activity reflects registration of episodes to memory and suggest a potential brain correlate of retroactive interference.

Recent findings suggest that novel associations can be learned during sleep. However, whether associative learning during sleep can alter later waking behavior and whether such behavioral changes last for minutes, hours, or days remain unknown. We tested the hypothesis that olfactory aversive conditioning during sleep will alter cigarette-smoking behavior during ensuing wakefulness. A total of 66 human subjects wishing to quit smoking participated in the study (23 females; mean age, 28.7 Ϯ 5.2 years). Subjects completed a daily smoking diary detailing the number of cigarettes smoked during 7 d before and following a 1 d or night protocol of conditioning between cigarette odor and profoundly unpleasant odors. We observed significant reductions in the number of cigarettes smoked following olfactory aversive conditioning during stage 2 and rapid eye movement (REM) sleep but not following aversive conditioning during wakefulness (p Ͻ 0.05). Moreover, the reduction in smoking following aversive conditioning during stage 2 (34.4 Ϯ 30.1%) was greater and longer lasting compared with the reduction following aversive conditioning during REM (11.9 Ϯ 19.2%, p Ͻ 0.05). Finally, the reduction in smoking following aversive conditioning during sleep was significantly greater than in two separate control sleep experiments that tested aversive odors alone and the effects of cigarette odors and aversive odors without pairing. To conclude, a single night of olfactory aversive conditioning during sleep significantly reduced cigarette-smoking behavior in a sleep stage-dependent manner, and this effect persisted for several days.

Social chemosignaling is a part of human behavior, but how chemosignals transfer from one individual to another is unknown. In turn, humans greet each other with handshakes, but the functional antecedents of this behavior remain unclear. To ask whether handshakes are used to sample conspecific social chemosignals, we covertly filmed 271 subjects within a structured greeting event either with or without a handshake. We found that humans often sniff their own hands, and selectively increase this behavior after handshake. After handshakes within gender, subjects increased sniffing of their own right shaking hand by more than 100%. In contrast, after handshakes across gender, subjects increased sniffing of their own left non-shaking hand by more than 100%. Tainting participants with unnoticed odors significantly altered the effects, thus verifying their olfactory nature. Thus, handshaking may functionally serve active yet subliminal social chemosignaling, which likely plays a large role in ongoing human behavior.DOI: http://dx.doi.org/10.7554/eLife.05154.001

Yahoo's native advertising (also known as Gemini native) serves billions of ad impressions daily, reaching a yearly run-rate of many hundred of millions USD. Driving the Gemini native models that are used to predict both click probability (pCTR) and conversion probability (pCONV) is Offset -a feature enhanced collaborativefiltering (CF) based event prediction algorithm. Offset is a one-pass algorithm that updates its model for every new batch of logged data using a stochastic gradient descent (SGD) based approach. Since Offset represents its users by their features (i.e., user-less model) due to sparsity issues, rule based hard frequency capping (HFC) is used to control the number of times a certain user views a certain ad. Moreover, related statistics reveal that user ad fatigue results in a dramatic drop in click through rate (CTR). Therefore, to improve click prediction accuracy, we propose a soft frequency capping (SFC) approach, where the frequency feature is incorporated into the Offset model as a user-ad feature and its weight vector is learned via logistic regression as part of Offset training. Online evaluation of the soft frequency capping algorithm via bucket testing showed a significant 7.3% revenue lift. Since then, the frequency feature enhanced model has been pushed to production serving all traffic, and is generating a hefty revenue lift for Yahoo Gemini native. We also report related statistics that reveal, among other things, that while users' gender does not affect ad fatigue, the latter seems to increase with users' age.

No abstract