PFC-2 datasets: behavioral trials with errors

Hello everyone,

let me start by thanking the people behind this website/project. It is really a precious resource. 

I have a question concerning the PFC-2 dataset. 

I would like to compare the structure of prefrontal activity during behavioral trials in which a mouse performs the task correctly, with those in which the mouse commits a mistake. 

The documentation describes a "SessionNP" variable that contains the timing information but only for trials in which the mouse performs the task correctly. On top of the time information, in the same variable (and in others, such as "whlrl"), the direction in which the mouse turned is also encoded (right=1, left=2). While in the documentation there is nothing on trials in which the mouse commits an error, I have noticed that there is also a "Session" variable, which contains information about a larger amount of trials than "SessionNP". The direction of these trials is then generally encoded as 4, 5, or 6. 

After this long introduction, my questions are: are the "extra" trials encoded in "Session" the ones in which the mouse commits a mistake? If this is not the case, is there any way to extract information about the trials in which the animal commits a mistake? What does the 4,5,6 direction encoding stay for? 

Thank you in advance to anyone that will help!

Hi Mattia,

Thank you for visiting the site and using the dataset.

Yes, 'Session' contains all trials including error. The meanings of the numbers are as follows:

1 = Right-correct,

2 = Left-correct,

3 = Error (choice R, reward L)

4 = Error (choice L, reward R)

5 = Time-out (choice R)

6 = Time-out (choice L)

(Time-out means that the rat took more that 10 sec before reaching the goal (e.g., stopped running and started grooming, etc.) I did not include these trials for the analysis.)

I also attached 'SessionNPAll' that is SessionNP including all trials.

Note that the rat performed this task very well so that we did not have many true errors (i.e., '3' or '4'). In EE.188, there is only one error trials... 

Best,

Shige

Hi Shige,

first of all sorry for the late reply, I am only reading this now. Also, thanks for the precise answer. Very clear.

If you don't mind, I would have an additional question.

If I am reading the description of the dataset correctly, there should be 13 sessions in which the mice performed a working memory task. Again, if I am not mistaken, each of these sessions should have a "XXX_Behavior.mat" file (which I found) that contains a series of variables that are described in the accompanying pdf. 

Of these, I find that "spiket" and "spikeind" are missing from several files (EE088, EE196, FF103, FF130, FF167, GG097, GG149). In two other files (EE188, EE210), "Session" and "SessionNP" are missing.

I can retrieve the info about spikes from the "filebase.res" and "filebase.clu" files, but I first wanted to ask you whether I am missing something or whether spike info is genuinely missing in the "XXX_Behavior.mat" files and/or I am interpreting something incorrectly.

The missing "Session" and "SessionNP" files are a bit more tricky, as I don't see how to easily retrieve the missing info here. Would it perhaps be possible to have these files?

I apologize for the lengthy question. Thank you in advance for your time.

Best,

Mattia

Hi Mattia,

For 'spiket', 'spikeind', and 'SessionNP':

  I remember that I had similar questions previously and I prepared these data at that time. How can I send you the data?

For 'Session':

  I can also send you the 'Session' for EE188. Sorry, but I was not able to find 'Session' of EE210...

Best,

Shige

Thank again, I answered you via PM