pfc-2: possibly erroneous normed position?

Hi there,

First of all, a word of gratitude for making this dataset available! It's of fantastic scientific value, and I know how much work it can be to make it accessible and understandble for others, much appreciated!

I've been going through the mat-files of the various PFC-2 datasets (XXX_Behaviour.mat), and I think something went wrong with normalizing the (linearized) position (7th column of whlrl_ex). Below the a description of what I encounter, and two questions resulting from this. I hope you would have some time available.

In the documentation pdf of PFC-2, it says the position was normalized such that:

1 = the position at the time of the end of the last nose-poking

2 = the position at the time just before reward consumption

However, the normalized position at the end of the last nose poke as given in 2nd column of SessionNP, is often not 1, but various quite randomly between 1 and 2 (time converted to whlrl sample using FS=20khz/512).

When I plot the normalized position of each lap, together with the X and Y coordinates of the first LED (column 1-2 of whlrl_ex), and highlight the end of the last nose poke (column 2 SessionNP), it seems like the time in column 2 of SessionNP (i.e. end last nose poke, cyan line) most correctly reflects the position of the rat, but that the normalized position does not. I.e. in the example trials attached, the rat does not seem to be moving until the normed position is between 1-1.5.

(the attached examples come from session GG.069, but I see the same pattern for all the other sessions that have a XXX_Behavior mat-file)

My questions:

1. So, my first question would be, is this a simple scaling issue, and does the normalized position from the end of last nose poke (cyan line; 2rd column SessionNP) to position=2, reflect the full position range of the maze? I use the position of the rat to segment the data into sub-sections of the maze, so if this is true, I can just 'stretch' the position and work with that.
2. My second question relates to calculating the linearized distance using the X and Y position of the rat. I would love to be able to construct the linearized distance for the sessions without a mat-file, like EE.188. I assume this conversion depends on the hard limits of the coordinate space of the X/Y position? I tried to find a hint in the documentation of all the HC datasets, but couldn't find anything. It would be fantastic if you would have some pointers. 

Any help would be greatly appreciated,

All the best,

Roemer

-- 

Dear Roemer,

1. Normalized position is in the column 7 of 'whlrld', not 'whlrl_ex'. 

(I am sorry, I found a typo in the documentation; 'whlrl' on the second page should be 'whlrld').

2. For calculating the linearized distance, fist I made the mean trajectories for R and L trials in each session, and then all position points were projected to the mean trajectories. If you need 'whlrld' in your favorite sessions, I can send it to you. 

Best,

Shige

Dear Shige,

Much appreciated! Using whlrld(:,7) instead of whlrl_ex, the start of the trials all look well aligned to the position. 

Thanks for describing your method of linearizing the position. Much simpler than I assumed actually. I'll be able to construct the whlrld's now.

Thanks for taking the time to reply,

Roemer