At the International Conference on Memory last month, I presented some new work from my lab, a 21 participant, fMRI-scanned, memory experiment. We imaged people’s brains as they underwent a procedure that generates sensations likened to déjà vu (based on Josie Urquhart’s procedure, published in 2014, that you can find here). What makes this work particularly exciting is that, to our knowledge, this is the first time people undergoing an experimental analogue of déjà vu have been imaged. It lead to some pretty neat results.

Memory conflict-related brain regions that track deja vu reports
Memory conflict-related brain regions that track deja vu reports

The findings were picked up by New Scientist and are summarised in the piece below:
https://www.newscientist.com/article/2101089-mystery-of-deja-vu-explained-its-how-we-check-our-memories/

Embedded within that article is the following video, which distills the essence of what we’re excited about – brain regions associated with memory conflict, rather than false memory, appear to be driving the déjà vu experience. This is consistent with our idea of deja vu as the conscious awareness of a discrepancy in memory signals being corrected. This in turn sheds some light on why déjà vu occurrence appears to decline with age despite the fact that memory errors tend to increase with age. If it’s not an error, but the prevention of an error, this makes a lot more sense.

[17/08/2016 UPDATE]:

A few other news organisations have since reported the story:

BBC World Service Newshour (interview, audio below)


Absolute Radio
 (interview, audio below: 42.55 – 48.02)


IFLScience
 (text, rejigged NS article)

New York Magazine (text, neat explanation of the paradigm)

[18/08/2016 UPDATE]:

Digital Trends (text, one of the only online news organisations to speak to me in person)

Daily Mail (text, unnecessarily scary headline, lots of lovely comments :/ )

Gizmodo (text, more hyperbole)

[19/08/2016 UPDATE]:

Medical Daily (text)

New.com.au (text, emphasises importance of peer review to come)

5 thoughts on “This is your brain on… déjà vu

  1. Error correcting codes (ECC) are an interesting and difficult topic. I was thinking about them in relation to the Hopfield network where the attractor states tend to be rather shallow. By adding in ECC you could improve that situation a lot. However I haven’t actually tried that in practice yet. Maybe sometime. What I did try however was replacing the simple computational neurons with decision tree neurons. That is a very interesting route to go. Worthwhile things happen when you do that.
    It would certainly be worth investigating in more detail. What is the decision tree equivalent of a biological neuron? Do small groups of biological neurons form decision trees systems with some kind of simple greedy learning algorithm?

    Reply
  2. Dear Mr. O’Connor,

    this is just to let you know that the German edition of Motherboard—Vice’s tech and science magazine—will publish an article on your fascinating research today. I’m not sure how well you can read German, but I promise there is no unneccasary scary headline. We were fighting about who could write about it! Just a quick question: Could you grant us permission to use the picture on this site („Memory conflict-related brain regions that track deja vu reports) to illustrate our piece?

    You’ll find out article here: http://motherboard.vice.com/de/

    Have a great day.
    Best regards,

    Theresa Locker
    Editor @ Motherboard Germany

    Reply
  3. Obviously the Barkhausen criterion is met in the human brain, otherwise people wouldn’t have brain waves. The other possible effect of positive feedback is hysteresis which could explain short-term memory. The brain is an analog system striving to be digital it seems.

    Reply
  4. Hi Dr O’Connor,

    I’m writing an article on deja vu for Dose/OMGFacts and would like to ask you some questions via email. What’s the best way to reach you?
    Best,

    Josh O’Connor

    Reply

Leave a reply

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong> 

required


*