hcuve / asas Goto Github PK

the .acq file has triggers problems that need to be fixed manually. Seems like the digital messages hangout for too long create wrong trigger numbers. Not sure how this can be solved. potentially try to identify constants, and they will be either fixation valence or intensity ratings. I can use the reaction time data or timelogs from psychopy for this. Or try to find the stim by their duration (expected 6 seconds) but the latter might fail if the triggers shorted the duration fo the stim

so, in previous steps when using aggregated pupil, the dataset was created excluding pupil trials with excess loss. e.g. < 80% of valid gaze samples, so of we use this dataset db_full4new_stim_screen_pupil_nopract to create our database with the behavioral data to be merged with time series, we end up with trials with no self-report ratings. that's because tmp.df4_full_stim_downs_jun2021_60bins has all data.

two solutions here, find an earlier data frame that contains all behavioral data and tale the needed data from there, or just exclude the trials witha. lot of loss on the time-series datasets

ssid 606 has their ECG inverted, so the HR output is blank. Make sure to invert this before analyses of clinical data

1- clean and solve warnings in timecourse analyses
2- store estimates for all terms (arousal, interactions etc)
3 - smooth velocity because it's more jittery

• decide on rule for significance?
• flag 3 consecutive uncorrected ps at alpha < .05 and return the time min and max for geom_rect
• do the same for corrected
• make lmer pupil plots
• make timecourse plots for pupil and velocity
• make a panel for the paper
• have Joseph do ledalab for BIOPAC, gazepoint and heart rate, and self reports
• pre-process 345 and 356