week 05/08

[ebridge2]

• [x] presentation rough draft completed. 05/04 goal. presentation
• [x] new discriminability package code. 05/04 goal. results
• [x] new schematic for upload workflow. Workflow is in the presentation.

@02agarwalt

Completed this week:

• [x] change job definition name to ndmg-fmri
• [x] change compute environment name to ndmg-fmri-env
• [x] change job queue name to ndmg-fmri-queue
• [x] restructure website: two separate tabs for submission and query
• [x] change job directory label to token
• [x] make none option for stc
• [x] fix security group/subnet ids
• [x] incorporate custom ami
• [x] incorporate data upload feature Deliverable: Demo

Left to do:

• [ ] improve website design
• [ ] p-value for mcflirt

[jovo]

• [x] compute p-value to determine whether preproc improves discriminability (@shangsiwang & @studiojlp) both know.
• [ ] slide 16, make fonts bigger.
• [ ] slide 24 is the schematic, put in the begging
• [ ] aim for 15 content slides (def < 20)
• [x] look up some papers on functional connectome and find figures of adjacency matrices and see if they look like anything
• [ ] show RAM consumption for different pipelines
• [x] compare other stuff, e.g., intuition of output folders (check to make sure compliant to BIDS)
• [ ] for task, use half the data per "task block"

[ebridge2]

http://journal.frontiersin.org/article/10.3389/fnhum.2015.00386/full figure 7 http://journal.frontiersin.org/article/10.3389/fnhum.2013.00063/full figures 1 and 2 etc

[ebridge2]

also, i have looked through the BIDs documentation pretty thoroughly, and see nothing about post-acquisition for derived organization.

[ebridge2]

http://ericwb.me/lectures/fngs_final.html#/

[jovo]

@ebridge2 compare to CPAC along multiple dimeniosn:

• CPU
• RAM
• time
• price
• discriminability
• ease of use
• QA

make a table with columns for features, rows for pipelines,