Thomas Nichols

@ten_photos

Professor of Neuroimaging Statistics Oxford Big Data Institute Nuffield Department of Population Health University of Oxford

United Kingdom
Vrijeme pridruživanja: rujan 2013.

Tweetovi

Blokirali ste korisnika/cu @ten_photos

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

  1. 4. velj

    : Ten simple rules for documenting scientific software

    Poništi
  2. proslijedio/la je Tweet
    1. velj
    Poništi
  3. 1. velj

    Reference-rich thread on the eternal question: When comparing change between two groups, should you directly assess the change (paired t-test) or assess post baseline measurement adjusting for baseline (ANCOVA).

    Poništi
  4. 11. sij

    And according to this follow up article the problem is only getting worse.

    Prikaži ovu nit
    Poništi
  5. 11. sij

    This is absolutely nuts, wide open DICOM servers: Millions of Americans’ Medical Images and Data Are Available on the Internet. Anyone Can Take a Peek.

    Prikaži ovu nit
    Poništi
  6. proslijedio/la je Tweet

    What do the genetics of education tell us about learning? - a nice explainer by

    Poništi
  7. 3. pro 2019.

    "Ten common statistical mistakes" TL;DR (1) No control grp (2) Informal grp comparison (3) Ignoring dependence (4) Outliers in corr. (5) Small n (6) Circular inf. (7) p-hacking (8) Ignoring multiple testing (9) Interpreting p>0.05 (10) Causation vs. corr.

    Poništi
  8. proslijedio/la je Tweet
    2. pro 2019.

    In this contribution, we observe that the Harmonic-Mean-p-value is a promising multivariate (global) test, but possibly less promising than argued in the literature. Joint with and Jelle Goeman.

    Poništi
  9. proslijedio/la je Tweet
    2. pro 2019.
    Poništi
  10. proslijedio/la je Tweet
    27. stu 2019.

    Here is a new preprint with Shahin Tavakoli, Julia Brettschneider and . We use scalar-on-image regression models to get individual prediction intervals for chronological age from 3D brain scans.

    Poništi
  11. 27. stu 2019.

    "CV can be used to tune models and to estimate performance, but not on the same data." Such an important principle for ML, nicely laid out in this thread (with refs)

    Poništi
  12. proslijedio/la je Tweet
    22. stu 2019.

    Taking a longterm view - a lesson for research investment (thanks to et al): 4 yrs from proposal to 1st participant; 8 yrs to full enrolment; 13 yrs to 1st publications; ~20 yrs to an explosion of scientific discovery

    Poništi
  13. 16. stu 2019.

    Tickled to have been a part of this, with thanks to who lead our team, w/ .

    Prikaži ovu nit
    Poništi
  14. 16. stu 2019.

    The NARPS study has read-out... 70 independent analysis teams started with same fMRI data and reported on 9 precisely defined hypotheses. Remarkable variability in results, but the statistical maps are relatively consistent.

    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    13. stu 2019.

    FSLeyes 0.32.0 has been released, with BIDS/JSON metadata now displayed in the information panel, better Jupyter/IPython integration, and other small bells, whistles, and fixes. Get it from

    Poništi
  16. proslijedio/la je Tweet
    6. stu 2019.

    Just submitted our new automated meta-analysis tool for publication: I think that the tool is super useful. I dread that it might not be deemed neurosciency enough by peer review. In the mean time, use it to do good science!

    Poništi
  17. proslijedio/la je Tweet
    30. lis 2019.

    Introducing "BrainPower - resources for power analysis in neuroimaging". This is a community project we started as a hackathon at SIPS2019 and have been working on since then. Now we're inviting y'all to join and contribute! [thread 1/7]

    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    29. lis 2019.

    Save the date for in Montreal! Dates are June 26-30 (Friday-Tuesday). We're working to correct instances the dates were listed incorrectly, but please note the correct dates/days.

    Poništi
  19. proslijedio/la je Tweet

    - UK’s annual share of EU research funding has fallen by half a billion Euros since 2015 - Almost a 40% drop in UK applications to Horizon 2020 - 35% fewer scientists coming to the UK through key schemes Read the full analysis:

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    15. lis 2019.

    New preprint out! "Brain aging comprises multiple modes of structural and functional change with distinct genetic and biophysical associations", by Steve Smith, L. Elliott, , , G. Douaud, and - -

    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:

    ·