I Predict . . . Mindreading in the News

Note the difference below between the BBC headline and the next couple of sentences:

Scientists 'can predict memories'

Scientists say it may be possible to predict how well we will remember something before the event has even taken place.

By analysing scans, they discovered the brain must get into the 'right frame of mind' to store new information.

For top performance, the brain must mobilise its resources, not only at the moment we get new information, but also in the seconds before.

It's not that the scientists can actually read your mind and predict what you'll remember, but that they can predict how well you'll remember something. The paper is available online at Nature Neuroscience.

The authors (Otten, Quayle, Akram, Ditewig & Rugg) presented the participants with a list of words to memorize for a later memory test. Each word was preceded by a cue that told subjects either what task to perform with the word -- an orthographic or a semantic decision (Exp. 1) -- or whether the word would be presented in the auditory or visual modality (Exp. 2). EEG activity was recorded while the participants performed these experiments, and event-related brain potentials ("brain waves") were averaged separately for the cue words and the memorized words. The major new finding is that a late-onset, long-duration negative brain wave (akin to the CNV, or Contingent Negative Variation wave recorded in the interval between cue and target in S1-S2 tasks) was observed to cues preceding words that were later remembered, but not to cues preceding words that were later forgotten. Otten et al. didn't acknowledge that their negative wave resembled the CNV or related brain waves, and they discounted the explanation that greater attentional resources were marshalled to the cues preceding remembered words, instead arguing that the neural activity reflects the "adoption of a 'task set' optimal for semantically oriented processing." Hmm, that's not a very satisfying explanation, but it does account for the fact that they failed to observe the "prediction wave" for cues signaling that subjects should perform the orthographic task. What is not accounted for, however, is that the "prediction wave" was not seen to cues preceding auditory words that were later remembered in Exp. 2. But why?

SUMMARY from The Neurocritic: kinda interesting, but how did it get published in Nature Neuroscience??

ADDENDUM: The supplementary material for the paper indicates that the memory test was a surprise part of the experiment, so the participants weren't deliberately memorizing the words, merely performing the orthographic and semantic tasks blissfully unaware of the impending exam. SO the purpose here was to predict memory performance even when subjecs weren't trying to remember.