TrialRecord.User not recorded in data in .bhv2 log file or mlread cannot read

Hi!
I am using ML2 build 199, a conditions file and scene framework with "MultiTarget" adaptor. I want to record the monkey's choice from "ChosenTarget" into a User defined structure "TrialRecord.User.choices".

choice = MultiTarget(eye_);
scene2 = create_scene(choice,[TARGET_LEFT TARGET_RIGHT]);
run_scene(scene2);
if choice.Success
        trialerror(0);  % successful fixation
        goodmonkey(100,'juiceline',1,'numreward', 3,...
            'pausetime', 300);
        TrialRecord.User.choices = choice.ChosenTarget;
        dashboard(1, sprintf('TaskObject#%d chosen', choice.ChosenTarget));
end

The dashboard showed well during the task. But when I load the behavior log with mlread,

[data,MLConfig,TrialRecord,filename,varlist] = mlread

the data(1). UserVars only contains "SkippedFrameTimeInfo". The TrialRecord has choices in the User field, but it only has the information from the last trial.
Could you help me out? Thanks!

I don't know why you think that the variables assigned to TrialRecord.User will be saved in the data file. See the bhv_variable function.

Thank you, that worked perfect.
I was following this example earlier...
example

I don't see any description or indication there that the variables will be saved to the data file.

Hi, I'd like to ask a follow up question here:

I understand that the recommended way to save trial variables is with bhv_variable, but is there a way to access the history of the variables saved this way from TrialRecord?

E.g., if I want to save a boolean value for each trial and on each trial I wanted to print a running "percent true" of that boolean variable, how would I do that? For now, all I can think to do is to save the variable with bhv_variable AND save a running list of the values with TrialRecord.User (or a running percent and number that can be updated analytically).

Thanks!

No, there isn't. Your computer has limited memory, so it is not reasonable to track the changes of all saved variables. Your example is just one boolean, but some people save huge variables for thousands trials.