Skip to content. | Skip to navigation

Personal tools
You are here: Home Forum Using data sets PFC7 data-Eye position tracking data

PFC7 data-Eye position tracking data

Up to Using data sets

PFC7 data-Eye position tracking data

Posted by Enrique Contreras Hernandez at October 16. 2019

Hello!! This question is respect the data provided by Hunt L.T., Malalasekera W.M.N and  Kennerley, S.W. (2018). Single neuron activity matrices are provided for each of the different stages of the task performed by the Monkey (fixationCenter, Cue1, etc). Each row corresponding to a single trial and each column the time in 1ms bins. So, the matrices have well defined sizes [#trials x 501 or 800] and are time locked to specific stages on the task. They also provided eye tracking analog data (X and Y Eye positions) that were recorded at the same time than the neuronal activity, however, the eye tracking data are not presented in a consistent manner, number of trials are different than the corresponding for neuronal data and recording times are variable (seems to be a continuous long recording along the duration of the trial ranging from tents to thousands of samples). Does it is possible to get the eye tracking data associated with the neuronal activity?


Thank you so much



Re: PFC7 data-Eye position tracking data

Posted by Laurence Hunt at October 25. 2019

Hi Enrique,

Sorry for the delay in getting back to you.

The reason why the trials don't match is most likely because the amount of trials that are in the matrices is preprocessed to specific trials, while this isn't the case in the eye tracking data. So you would need to run for each session the part of the preprocessing script that indexes which trials are going to be included in the neuronal matrices, i.e. this bit from the preprocessing script:


    %% isolate information gathering trials where error code doesn't indicate

    %  timeout etc., and therefore where subject completed the trial


    trial_error=BhvInfo.TrialError; % 0 = 'Correct Response' (chose side with higher expected value),

    % 6 = 'Incorrect Response' (chose side with lower expected value)

    % 1 = No Response, 2 = Late Response, 3 = Break Fixation, 4 = No Fixation

    % 5 = Early Response, 7 = Lever Break, 8 = Ignored, 9 =Aborted



    condition=BhvInfo.ConditionNumber(inf_responded_trials); %3 = option trial; 4 = attribute trial


The other reason, as you pointed out already, is that the eye tracking data is not epoched to the various event codes, so each trial is a [time X x_coord y_coord] matrix and needs to be aligned with the code times to be of use. Because we didn't directly use this in the paper we haven't done it ourselves, but it should be fairly simple by loading in the behavioural data and then checking these two together:



These can be then used by taking code events (CodeNumbers) for a particular task event, check the corresponding timings (codeTimes) and use those as an index of the eye tracking matrices. We are not certain as to whether there are any caveats to this approach or not (because we haven't actually done it ourselves).

Hope that helps!

Re: PFC7 data-Eye position tracking data

Posted by Enrique Contreras Hernandez at November 14. 2019

Hello Laurence, actually, as you mentioned I found the data I need in the zInfoGatheringInfo files with a single exception in the # of trials, not a problem. I'll use the other information you detail above to extract the time stamps and analyze the LFP also, thank you so much for you detailed answer.


Re: PFC7 data-Eye position tracking data

Posted by Enrique Contreras Hernandez at May 04. 2021

Hello Laurence, I'm writing you to ask about the LFPs data. Basically, the question is how to get the LFPs recorded on each channel along all the trials. On the F00X files, e.g. in the F002 file (corresponding to the session #2) I can find subfiles named F002_C1_U1, F002_C3_U1, etc. "C" corresponds to the channel. Within these files there is a variable named BhvInfo.PlexonFiledPotentials, Are the analog recordings stored there? Interestingly, the length of PlexonfiledPotentials is 724, that corresponds to the total number of neurons (from all the sessions, not only the session #2). Concretely, is it possible to get the LFPs recorded from each channel across all the good trials? Is there any information, like time stamps associated a Cue1, centerFix, cue2, etc as well as # of trials, that help me to extract such LFPs?

Thank you



Powered by Ploneboard
Document Actions