Checklist

[aforren1]

• [x] Record absolute times only -- figure out relative times when needed
  • easier to document (what is relative to what is moot)
• [x] Drop relative presentation times in the timed response, do absolute times & convert to frames
  • Will stop consistently late presentation
• [x] TR object
  • Keyboard- or serial-based?
  • Starters exist for both
• [x] Add MRI experiment
• [x] Add MRI vigilance type Nixing for stronger data elsewhere
  • Use hands as dummy stimuli
• [x] Find the 200ms chaff in TimedResp
• [x] Explicit go-nogo fMRI
• [x] Feedback in the scanner & better view for experimenters
  • [x] related: try getting multi-screen presentations right

Technical ToDo:

• [x] Replace numeric id with string based on tfile directory
• [x] Save data automatically (based on tfile name + directory + timestamp)
• [x] Save data on error and/or key combination (e.g. esc, which is already checked by TR class)
• [ ] Transition with black screen in the scanner (Nice-to-have)
• [x] Timed response feedback based on peak force, not movement onset (makes feedback feel more natural)
• [x] Snappier free response feedback? It feels a bit mushy
• [x] Pre-generate files for 3 subjects
• [x] Feedback about peak force & push duration (post trial)
• [ ] calibrate?
• [x] Randomize initial training split of symbols independently of practice split of symbols
• [x] Seed generation of tgtfiles
• [x] Correctness feedback in the scanner? Quick 100-200ms flash on the push feedback would do

Design ToDo:

• [x] Finalize # of Training Blocks (2 split / 2 interleaved + top-up if below 95% accuracy)
• [x] Finalize # of TR Blocks after 1st scan (6 runs, 60/symbol) and time window (0.3 to 0.8) before last beep)
• [x] Finalize # of practice blocks per day (10 practice + 1 or 1/2 block non-practiced).
  • Possibly at a random time during practice
• [ ] Finalize switches ([1 3 5] or [2 4]?) Do we also switch non-practiced symbols? (Yes, but different fingers)
  • not same fingers for practiced/unpracticed
  • e.g. [2 4] vs [3 5](not the thumb). Counterbalance which ones are permuted in practice set.
• [x] Figure out exactly WHEN ForcedRT will be run
  • After initial training on Day 1
  • After scanning on Day 5
• [ ] Figure out how many switch-training trials prior to TR (200?)

[aforren1]

Is it sufficient to seed at the top of MakeAllTFiles? I already seed WriteTrTgt and WriteRtTgt within their calls, which I would remove (it would mess with "randomness" anyway? All subjects end up with the same trial order)

[adrianhaith]

We definitely don't want all subjects to end up with the same trial order. So should seed based on their subject number.