Memory Guided Saccade Encode and Recall

Initially a single psychopy saccade and recall paradigm, now this repo contains a small suit of saccade tasks.

Usage

1. install psychopy
2. run mgs_enc_mem.py
3. run mgs_recall.py

Windows install

additionally a library to toggle volume so monitor re/connect notifications are not audible:

python -m pip --user install https://github.com/AndreMiras/pycaw/archive/master.zip

Additional Tasks

• vgs_anti.py - eeg anti or vgs (not mixed)
• ring_reward.py- behavioral "rewarded" antisaccade task
• eeg_eyecal.py- eeg eog eye position calibration task

MGS Timing

stimtime/00_mktime.bash uses genTaskTime to generate distribution and duration of events.

• Cue, VGS, and MGS each have a 2s duration
• Dly 15 @ 6.00s, 7 @ 8.00s, 2 @ 10.00s
• ITI varies by modality.

Modality settings are also, redundantly, set in mgs_enc_mem.py

• number of runs: {'mri': 3, 'eeg': 4, 'practice': 1, 'behave': 2}
• number of trials per run: {mri: 24, eeg: 24, practice: 20 (@ faster pace), behave: 16}
• each run length in seconds: {'mri': 420, 'eeg': 358, 'practice': 65, 'behave': 240}

For a quick check, compare the mktime outputs:

for f in stims/*/*/cue.1D; do task=$(basename $(dirname $(dirname $f))); tr ' ' '\n' < $f|sed 's/.*://;/^$/d'|sort |uniq -c|sed s/^/$task/; done|sort -u|column -t

MODALITY   NTRIAL CUEDUR
behave     16     2.00
eeg        24     2.00
mri        24     2.00
practice   20     1.00

Eye tracking

• Eye tracking interfaces with ViewPoint EyeTracker software (via Arrington Research and Avotec). see MR_note.txt
• EEG eye tracking using EOG after collecting calibration with eeg_eyecal.py
• ASL eye tracking triggers can be sent over parallel port same as EEG TTL

TTL Parallel Port Triggers

• inpout32.dll should be in directory where python runs from psychopy import parallel
• used to mark task events EEG or ASL eye tracking

Images

Selection made with img_pick.py:

# r,g,b all .1 from equal. percieved brightness around 1std of the mean
q = d.query('abs(r-.33) < .1 and abs(g-.33) < .1 and abs(b-.33) <.1 and abs(p-116) < 10 ')

after imgedit.py (help from SUN/iminfo.py) converted images into 225x225 circle pngs.

More images

2017-12-07: identified 635 more browsing in alphabetical order 2018-11-27: removed some w/help from anna,remove_c.bash from (list)

image source

Images are from SUN dataset.

• all images (37Gb!, 38 uncompressed)
• 3 level hierarchy annotation (<1Mb)

Data analysis

are ttl triggers reporting resonable times? eeg/read.py