========================================================================

AFNI Digest, week of 2018/01/05.

00)  The AFNI Bootcamp cometh!
What better way to shake away the cold than to think about freshly
baked FMRI data, hot out of the afni_proc.py pipeline?  Registration
for the next AFNI Bootcamp, to be held at NIH in March (2018!) is now
open.  Please see the "Registration" link, as well other useful
information, here:
https://afni.nimh.nih.gov/bootcamp/

01)  ROI-based group analysis using Bayesian hierarchical modeling
The current practice of handling multiple testing through controlling
the overall FPR in neuroimaging under the null hypothesis significance
testing (NHST) paradigm excessively penalizes the statistical power
with inflated type II errors (“false negatives” or “low power”). An
ROI-based approach has been proposed to handling neuroimaging group
analysis using Bayesian hierarchical modeling (BHM) to achieve two
goals: 1) improving modeling efficiency via one integrative (instead
of many separate) model and dissolving the multiple testing issue, and
2) turning the focus of conventional NHST on FPR into quality control
by calibrating type S errors while maintaining a reasonable level of
inference efficiency. See more details here:
https://doi.org/10.1101/238998

02)  Startup tips in AFNI
The latest edition of the AFNI program now prints out a usage tip (to
the terminal window) when it finishes starting up. An example:

   ------------------------- AFNI Startup Tip (13/54)------------------
   You can run InstaCorr on several subjects at the same time, using
   multiple AFNI controllers opened with the 'New' button.
   --------------------------------------------------------------------

---------------------------------------------------------------------
+ This Digest is not a comprehensive list of code updates.
+ Please direct comments or questions to the AFNI Message Board:
 https://afni.nimh.nih.gov/afni/community/board/list.php?1
+ To unsubscribe or otherwise manage your Digest settings, click here:
 https://list.nih.gov/cgi-bin/wa.exe?SUBED1=afni_digest&A=1
---------------------------------------------------------------------
========================================================================

========================================================================

AFNI Digest, week of 2017/12/23.

00) Notes on temporal autocorrelation modelling in major FMRI packages.
Some new and interesting methods work comparing different programs, by
Olszowy et al.  https://arxiv.org/abs/1711.09877 From the abstract:
    "Though temporal autocorrelation modelling in AFNI was not
    perfect, its performance was much higher than the performance of
    temporal autocorrelation modelling in FSL and SPM. FSL and SPM
    could improve their autocorrelation modelling approaches for
    example adopting a noise model similar to the one used by AFNI."
AFNI models the temporal autocorrelation using an ARMA(1,1) model. The
feature that makes AFNI most distinct from FSL and SPM in this area is
that the autocorrelation model parameters are estimated separately for
each voxel time series (program 3dREMLfit), rather than globally (SPM)
or over regions (FSL).

01) A new option "-Tslices" in 3drefit.
This option allows a user to add slice time offsets directly to a
dataset. The purpose of this is to let a user (or script) set the
slice timing when the dataset creation process did not do so
correctly. AFNI allows arbitrary slice timing to be added to NIFTI
files, using the AFNI extension to the NIFTI header. These slice times
will be used by the "tshift" processing block in afni_proc.py. Details
of how to specify slice time offsets are described in the output of
3drefit -help.

02) Fixed "-substatpar" in 3drefit.
The -substatpar option in 3drefit did not always work properly for
some types of datasets. This problem has (hopefully) been fixed, so
that the statistical type and parameters associated with a dataset
sub-brick can now be changed. The main application for this is when
(say) 3dcalc is used to process a dataset: the statistical information
is usually removed from the output dataset, since 3dcalc cannot "know"
if the calculations affected the statistical distribution. Option
-substatpar would usually be combined with option -sublabel to change
the label of the same sub-brick, to match the new statistical type
(e.g., "Correlation"). Details of how to set statistical parameters
and type for a sub-brick are described in 3drefit -help.

03) A new AFNI driver command is now available.
The new command "DATASET#N" drives the Dataset#N "transformation" used
in graph viewers to overlay time series plots from more than one
dataset. This change allows a script to drive the creation of such
overlaid time series graphs and then save them to image files. The
details are described in the updated README.driver file. At this time,
this is the only AFNI plugin that can be controlled by driving.

04) Larger montages for AFNI viewer.
The maximum array size for an image viewer Montage has been increased,
significantly. This change is principally for driving the AFNI GUI to
produce image snapshots with lots of slices.

05) New recommended options in afni_proc.py.
Some of the more recent afni_proc.py option recommendations include:
      -volreg_align_to MIN_OUTLIER
      -tlrc_base MNI152_T1_2009c+tlrc (assuming one wants MNI space)
      -regress_motion_per_run
Please see "afni_proc.py -help" for details.

06) A note on python2 and python3 "fun".
We are in the process of enhancing python programs to work with both
python2 and python3. Programs that should already be python3
compatible include afni_proc.py, afni_system_check.py, 1d_tool.py,
make_random_timing.py and timing_tool.py.

07) Enhancements in make_random_timing.py.
The program can now handle more complicated designs. One can define
stimulus and rest timing classes with their own probability
distributions, including limits. See "Advanced Usage" in the -help
output for details.

08) Intraclass Correlation (ICC) can improve your life.
Well, at least it might improve some of your analysis and results
interpretation, as described by G. Chen et al. in a new paper in HBM:
http://onlinelibrary.wiley.com/doi/10.1002/hbm.23909/full

-----------------------------
Updates 01-03 were suggested by our interactions with students during
our recent Bootcamps in Shenzhen and Guangzhou (China). Thanks go to
them for excellent hosting and such fruitful discussions!

---------------------------------------------------------------------
+ This Digest is not a comprehensive list of code updates.
+ Please direct comments or questions to the AFNI Message Board:
  https://afni.nimh.nih.gov/afni/community/board/list.php?1
+ To unsubscribe or otherwise manage your Digest settings, click here:
  https://list.nih.gov/cgi-bin/wa.exe?SUBED1=afni_digest&A=1
---------------------------------------------------------------------
========================================================================


========================================================================

AFNI Digest, week of 2017/11/22.

00) Job Opening in the AFNI group.
The Scientific and Statistical Computing Core (a.k.a. the AFNI group)
at the NIMH, NIH (Bethesda, MD, USA) is lookin' to hire a postdoc,
particularly someone with C and/or C++ experience, with Python and
OpenGL skills also being quite useful.

If you like problem solving and working with a fun group of people,
then check out the advert here for details:
https://afni.nimh.nih.gov/afni/community/board/read.php?1,156739,156739#msg-156739
(And please feel free to forward on.)

01) 3dLME: new fun option in the program.
Option "-ML" has been added in 3dLME, which can be invoked when
Maximum Likelihood is desirable, instead of the default method,
Restricted Maximum Likelihood (REML).

02) Have a good Thanksgiving!
And remember: "An optimist is a person who starts a new diet on
Thanksgiving Day." -- I. Kupcinet

---------------------------------------------------------------------
+ This Digest is not a comprehensive list of code updates.
+ Please direct comments or questions to the AFNI Message Board:
  https://afni.nimh.nih.gov/afni/community/board/list.php?1
+ To unsubscribe or otherwise manage your Digest settings, click here:
  https://list.nih.gov/cgi-bin/wa.exe?SUBED1=afni_digest&A=1
---------------------------------------------------------------------
========================================================================


========================================================================
AFNI Digest, week of 2017/09/26.

00) More FMRI cluster exploration.
A recent (brief) study by Mueller et al. found that SPM12 exhibits a
large, uniform dependence of FMRI clustering results on resampling
(i.e., voxel size, spatial resolution).  So, Bob checked what happens
in AFNI when resampling, to see if any such badness occurs.  In short
the effect is small (about 4x smaller), and not as uniform-- phew.
The MB post here:
https://afni.nimh.nih.gov/afni/community/board/read.php?1,156093,156093#msg-156093
contains a link to the AFNI work now available arXiv, as well as to
the Mueller et al. article.

01) New program: 3dBrainSync
This program “synchronizes” one dataset’s time series with another
dataset, in the sense of maximizing the average time series
correlation of the transformed dataset with the reference dataset. Two
different methods for the transformation are provided: orthogonal
matrix, and time index permutation. One goal is to allow direct
inter-run correlation between rs-FMRI datasets.  This program was
inspired by the nice poster on the BrainSync approach by Joshi et
al. at this past year's OHBM (also see here
http://neuroimage.usc.edu/neuro/Resources/BrainSync).


02) Bug Fix: 3dttest++ -ETAC
A bug was introduced when the multiple blur cases option "-ETAC_blur"
was added: runs without multiple blur cases would fail because the
intermediate filenames (used to communicate between 3dttest++ and
3dXClustSim) were formatted incorrectly. This bug has (fondly do we
hope) been patched.


03) Re. AFNI Bootcamp at NIH (October 2-6, 2017).
For those attending the AFNI bootcamp starting October 2nd, please get
your laptops ready.  Installation instructions are available at
https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/background_install/install_instructs/index.html.  
Once ready, please sent the output from: 
afni_system_check.py -check_all 
to afni.bootcamp@gmail.com with your first name in the
subject.  We will verify the output and help resolve any issues.  
(And please note that all spots for this Bootcamp have already been
filled.)


========================================================================
AFNI Digest, week of 2017/07/28.

00) ETAC notes.
A presentation by Bob describing philosophy+implementation of some new
clustering options in AFNI:
https://afni.nimh.nih.gov/afni/community/board/read.php?1,155528,155528#msg-155528

---------------------------------------------------------------------
+ This Digest is not a comprehensive list of code updates.
+ Please direct comments or questions to the AFNI Message Board:
 https://afni.nimh.nih.gov/afni/community/board/list.php?1
---------------------------------------------------------------------

========================================================================
AFNI Digest, week of 2017/07/21.

00) Dartmouth Bootcamp announcement:
    https://afni.nimh.nih.gov/afni/community/board/read.php?1,155427

01) 3dFWHMx will now output the mixed model ACF parameters by default,
    rather than the Gaussian FWHM ones.  This follows from recent
    discussion+changes related to FMRI clustering.  To discourage use,
    the Gaussian FWHM output row will show zeros (and will be reported
    by afni_proc.py as such, until the options are updated).


02) New program: 3dTto1D - a general program for collapsing time series data.
    The most useful computations are probably:

    enorm  - Euclidean norm of first differences 
             (of motion parameters or of EPI time series data, say)
    RMS    - Root Mean Square (a.k.a DVARS) of first differences
             (= enorm/sqrt(nvoxels))
    SRMS   - Scaled RMS, scaled by mean of input
             (= RMS/mean)
             (probably very good for censoring)
             (might be the same as DVARS censoring in FSL)

    Censoring with SRMS has not yet been incorporated into afni_proc.py.


03) Changes+problems with OpenGL on very recent versions MacOS (10.12.5).
    These changes will result in SUMA crashing under some circumstances,
    namely when the object controller panel (ctrl+s) is opened, then closed,
    then opened again;  something like this often occurs behind the scenes
    when scripts are driving SUMA.
    This new OpenGL behavior on Mac has been affecting many different
    softwares;  while we are looking into any fixes/workarounds, it is
    likely that a full solution is outside of our hands and will have to
    come from Mac.

========================================================================
AFNI Digest, week of 2017/07/07.

00) AFNI Bootcamp at NIH October 2-6, 2017: Registration open.
    Please see here:
    https://afni.nimh.nih.gov/bootcamp/
    ... for the registration page and for more information.

01) We will be retiring some of the AFNI packages.  Please let us know
    via the AFNI Message Board if there is a need to keep any of them.
    a) Linux_fedora_21_64.tgz     - not replaced
    b) macosx_10.8_gcc/icc.tgz    - not replaced
    c) macosx_10.10_gcc/icc.tgz   - probably replaced by a local version
    d) linux_xorg7/_64.tgz        - not replaced

02) New or planned AFNI binary packages include:
    a) new: linux_ubuntu_16_64.tgz
    b) new: linux_centos_7_64.tgz
    c) planned: macosx_10.10_local.tgz  (convenient for 10.10, 10.11)
    d) planned: macosx_10.12_local.tgz  (convenient for 10.12+)
    e) possibly: linux_fedora_25_64.tgz

 
 
---------------------------------------------------------------------
+ This Digest is not a comprehensive list of code updates.
+ Please direct comments or questions to the AFNI Message Board:
 https://afni.nimh.nih.gov/afni/community/board/list.php?1
---------------------------------------------------------------------

===========================================================================
AFNI Digest, week of 2017/06/23. 

00) Posters from the AFNI group for the OHBM (Organization for Human
Brain Mapping) conference are available online:

++ New FMRI clustering tools (more?!?), here with an emphasis on being
   "equitable" across the brain and in dependence on various
   parameters such as smoothing, voxelwise p-value, etc.:
   https://afni.nimh.nih.gov/pub/dist/OHBM2017/OHBM_2017_ETAC.pdf

++ New work on calculating intraclass correlation (ICC):
   https://afni.nimh.nih.gov/pub/dist/OHBM2017/OHBM_2017_ICC.pdf

++ New tools for FATCAT (whose online help is being written...)
   https://afni.nimh.nih.gov/pub/dist/OHBM2017/OHBM_2017_FATCAT.pdf

++ Comparisons of two FMRI task paradigms: "naturalistic" (=movie) and
   rest:
   https://afni.nimh.nih.gov/pub/dist/OHBM2017/OHBM_2017_ISC_and_RSFMRI.pdf

No doubt these make great plane reading for anyone flying to
Vancouver, or metro reading for any city.

---------------------------------------------------------------------
+ This Digest is not a comprehensive list of code updates.
+ Please direct comments or questions to the AFNI Message Board:
  https://afni.nimh.nih.gov/afni/community/board/list.php?1
---------------------------------------------------------------------

===========================================================================
AFNI Digest, week of 2017/06/09. 

00) Updated N27 FreeSurfer surfaces are available in MNI and Talairach spaces.
These were remade with FS 6.0 and the updated "@SUMA_MakeSpecFS -NIFTI
...".  The surfaces are available at their standard address:
https://afni.nimh.nih.gov/pub/dist/tgz/suma_TT_N27.tgz
https://afni.nimh.nih.gov/pub/dist/tgz/suma_MNI_N27.tgz
 
01) AFNI now usable on Windows (who'd have thought??).
Via the "Bash on Ubuntu on Windows 10" interface, AFNI and SUMA can
now be run directly on Windows systems.  Setup instructions are
provided here:
https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/background_install/install_instructs/steps_windows10.html
 
02) New AFNI binary package for Linux Ubuntu 16.
We now distribute a set of new precompiled binaries for Ubuntu 16 (and
potentially for later systems), streamlinining installation.  Setup
instructions are provided here:
https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/background_install/install_instructs/steps_linux_ubuntu16.html
 
03) SUMA and AFNI in concert at the Kennedy Center.
Internationally famous soprano, Renee Fleming, performed with her own
functional results overlaid on the aforementioned N27 brain
surfaces. Dr. David Jangraw, a post-doc in Peter Bandettini's group,
used AFNI+SUMA for the analysis and display. Ms. Fleming and SUMA
presented the "high notes" of the evening:
https://www.facebook.com/KennedyCenter/photos/a.10155104350680865.1073741876.73250630864/10155104351155865/?type=3&theater
Associated article by Dr. Francis Collins and Ms. Fleming:
http://jamanetwork.com/journals/jama/fullarticle/2630954
 
---------------------------------------------------------------------
+ This Digest is not a comprehensive list of code updates.
+ Please direct comments or questions to the AFNI Message Board:
 https://afni.nimh.nih.gov/afni/community/board/list.php?1
---------------------------------------------------------------------

===========================================================================
AFNI Digest, week of 2017/05/26.

00) AFNI is starting a new “Digest” email distribution-- welcome!

01) FMRI Clustering has been a widely discussed topic this year. These
recent papers by the AFNI group discuss the work by Eklund et
al. (with useful summaries of their results) and discuss both old and
new AFNI clustering functionality.
+ Full Brain Connectivity paper:
  http://online.liebertpub.com/doi/full/10.1089/brain.2016.0475
+ And its “Executive Summary”:
  https://drive.google.com/file/d/0B6Sn11ZTBrhvR2VYWmhvYmpZU2M/view
+ PNAS reply Letter:
  http://www.pnas.org/content/114/17/E3370.full
+ Presentation slides:
  https://drive.google.com/file/d/0B6Sn11ZTBrhvMi1TRDdKRUlMaEk/view

02) In afni_proc.py, blur estimation and clustering are now done using
the ACF method by default, replacing Gaussian FWHM (as of Aug. 15,
2016). Further progress in this area will be debuted at the OHBM
meeting in Vancouver BC (end of June).

03) The AFNI team has a new member, Justin "DiscoRaj" Rajendra.
Please direct your most complicated inquiries directly to him.

04) For Mac users, a note about AFNI compiling, installation, and
updates: For relative ease, please use the "local" Mac version now
(easier dependencies), described in Step #2 here:
  https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/background_install/install_instructs/steps_mac.html

05) The command line form of dcm2niix is now distributed with AFNI, as
"dcm2niix_afni" (renamed to avoid potential filename conflicts for
anyone with it already installed from elsewhere):
   https://afni.nimh.nih.gov/pub/dist/doc/program_help/dcm2niix_afni.html
Thanks very much to Chris Rorden!

06) For convenient referencing+remembering, the AFNI version number is
now stored in the header of data files, as part of the history
(viewable with 3dinfo).  For example,
   ----- HISTORY -----
   [user@computer: Mon May 22 17:21:52 2017] {AFNI_17.1.06:linux_openmp_64} 3dmask_tool -input ...

---------------------------------------------------------------------
+ This Digest is not a comprehensive list of code updates.
+ Please direct comments or questions to the AFNI Message Board:
  https://afni.nimh.nih.gov/afni/community/board/list.php?1
+ To unsubscribe or otherwise manage your Digest settings, click here:
  https://list.nih.gov/cgi-bin/wa.exe?SUBED1=afni_digest&A=1
---------------------------------------------------------------------

