Miriam Klein-Flügge

@MKFlugge

Cognitive Neuroscientist interested in decisions, actions & mental health

Vrijeme pridruživanja: svibanj 2016.

Tweetovi

Blokirali ste korisnika/cu @MKFlugge

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @MKFlugge

  1. Prikvačeni tweet
    23. lis 2019.

    Very excited to share our new paper, out today in with Anna Shpektor, , and Matthew Rushworth

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    9. pro 2019.

    An ACC-basal forebrain circuit influence when to act and can be modulated using transcranial ultrasound stimulation (TUS). Fantastic collaboration with , , , , Matthew Rushworth, et al.

    Poništi
  3. proslijedio/la je Tweet
    29. stu 2019.

    Somewhat new paper... Textbooks happily ignore that the amygdala and prefrontal cortex are connected through two pathways, not one. To understand amyg-PFC interactions - so vital for psychiatry - we need to do better. Anatomy matters!

    Prikaži ovu nit
    Poništi
  4. 20. stu 2019.
    Poništi
  5. 20. stu 2019.

    Thanks for the nice summary, I also really enjoyed all the talks at

    Poništi
  6. 6. stu 2019.

    Beautiful new work from and Rogier Mars characterising tracts running from the amygdala to PFC

    Poništi
  7. 5. stu 2019.
    Poništi
  8. proslijedio/la je Tweet

    The high resource impact of reformatting requirements for scientific papers

    Poništi
  9. proslijedio/la je Tweet
    28. lis 2019.

    How our remember depends on how we learn Multiple networks in the brain store knowledge. Damage to one brain area will leave alternative mechanisms for learning. 🧠

    Prikaži ovu nit
    Poništi
  10. 23. lis 2019.

    We’d like to thank our reviewers who have made this a better paper! To support unthresholded fMRI maps are available on Neurovault () and data+code underlying all figures on

    Prikaži ovu nit
    Poništi
  11. 23. lis 2019.

    Why is this relevant? We have discovered that learnt knowledge is stored in different brain circuits depending on how the knowledge was acquired. This means, for example, that someone with a lesion in one part of the brain will still be able to use alternative learning mechanisms

    Prikaži ovu nit
    Poništi
  12. 23. lis 2019.

    As previously shown – ventral striatum shows prediction error like responses consistent with RL theory, here for the first element of the sequence ABCD. It does not show full knowledge of the RL-acquired sequence structure.

    Prikaži ovu nit
    Poništi
  13. 23. lis 2019.

    (3) Average (noncontingent) learning does not only happen in time (i.e. learning about stimuli that are temporally close by), as previously shown, but also in space and thus for stimuli that are nearby in space. The amygdala also mediates this type of association.

    Prikaži ovu nit
    Poništi
  14. 23. lis 2019.

    By contrast, we find that knowledge acquired via reinforcement in (1) is persistent and robust and slow to change.

    Prikaži ovu nit
    Poništi
  15. 23. lis 2019.

    (2) Statistical map-like learning helps people know spatial relationships between stimuli, but also which stimuli are likely followed by which other stimuli (transition frequency). This knowledge is found in mPFC and amyg/hippo, and it flexibly adjusts to changing task demands.

    Prikaži ovu nit
    Poništi
  16. 23. lis 2019.

    (1)Reward-learning leads to knowledge of a path to the goal (here reward). In our case four elements of a four-step sequence ABCD are learnt backwards by trial and error. The correctly ordered representation is present in temporal pole and posterior orbito-frontal cortex.

    Prikaži ovu nit
    Poništi
  17. 23. lis 2019.

    Here we don't focus on the learning process itself, but on the neural representations of the learned information. We show that different regions of the brain are important for storing the knowledge acquired through these three different learning mechanisms. In brief:

    Prikaži ovu nit
    Poništi
  18. 23. lis 2019.

    Other labs have of course invested great effort in studying some of these learning mechanisms already including Jill O’Reilly, , Philippe Tobler, Bolton Chau and many others...

    Prikaži ovu nit
    Poništi
  19. 23. lis 2019.

    (3) By contrast to RL, noncontingent learning does not link precise stimuli with precise outcomes but instead relies on simpler "average" mechanisms (if things are good in general, then presumably my recent choices were not that bad).

    Prikaži ovu nit
    Poništi
  20. 23. lis 2019.

    (1) Statistical learning is based on observing relationships in the world (e.g. slowly acquiring the map of a city). (2) RL refers to learning driven by reinforcement (i.e. incentives or goals, e.g. learning to turn on music using a CD player) and is a contingent form of learning

    Prikaži ovu nit
    Poništi
  21. 23. lis 2019.

    In this paper, we are looking at three ways in which humans learn information about the world and characterise their dissociable neural pathways and representations: (1) statistical learning, (2) reinforcement learning (RL) and (3) noncontingent learning.

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·