This respository contains data collected from two monkeys performing a delayed center-out reaching task with uncertain target information. The data was collected by Alexa Riehle's lab at the CNRS Marseille, France. The data includes single unit recordings from the motor cortex of the monkeys, as well as the behavioral data from the monkeys performing the task and the target information provided to the monkeys during the task.
This repository is used to aquire a DOI. Here we provide the data in the pickle format. Please cite this dataset via the aquired DOI. Additionally we provide the data in csv and hdf5 format as a derived format. Please have a look at https://gin.g-node.org/nawrotlab/delayed_center-out_uncertainty_Riehle_derived, if you prefer to use these fileformats.
The monkey experiments were conducted in Alexa Riehle's lab at the CNRS Marseille, France, in accordance to European and French government regulations. Two monkeys, one male aged three to four years (M1) and one female aged four to five years (M2) performed a delayed center-out reaching task, which involved three different task conditions that differed in the amount of initial target information available to the monkey as illustrated in 1,2. The monkey was seated in front of a panel featuring a hexagonal array of touch-sensitive target LEDs and a central LED indicating the starting position. The monkey initiated a trial by touching the central LED (trial start, TS). During the 1 s delay period starting at t=500 ms the preperatory signal (PS) provided either complete or incomplete information about the final movement target and consisted of either a single target LED (Condition 1), two adjacent target LEDs (Condition 2), or three adjacent target LEDs (Condition 3) that lit up in green. At t=1500 ms the response signal (RS) appeared and one of the green target LEDs turned red. This indicated the final movement target and prompted the monkey to move his hand to that target. In Conditions 2 and 3 the final target was randomly chosen among the PS-cued targets, while the other target LEDs went dark. The times of movement onset (MO) and movement end (ME) were recorded and if the monkey touched the correct target LED, the trial was registered as successful and a drop of juice was given as a reward. A premature onset of the behavior before RS led to the abortion of the trial and no reward was given. Only successful trials are included in this dataset.
The task conditions of one, two or three possible targets presented during the 1 s preparatory period were executed in blocks and the order of these blocks were randomized across recordings sessions. In each block, 150 trials with randomized target directions were carried out so that each of the directions appeared on average 25 times per condition. Note that to obtain the same number of possible trial types in all conditions, not all possible combinations of directions for the preparatory stimulus were used in Conditions~2 and~3. Since six combinations are possible for Condition~1, only the pairs 1-2, 3-4 and 5-6 were used in Condition 2 and for Condition 3, only two cases occurred (6-1-2, 3-4-5).
Extracellular recordings were obtained with a multielectrode microdrive (Reitböck system; Thomas Recording) to insert transdurally seven independently movable electrodes within the area of the recording chamber, which was positioned over the motor cortex close to the central sulcus in monkey M1 and slightly more anterior in monkey M2 covering in parts primary and dorsal premotor cortex2. Online spike sorting resulted in up to seven simultaneously recorded single-unit spike trains1. Not on all experimental days all three conditions could be successfully completed.
Following event markers were recorded:
Data was obtrained from two monkeys, Joe and Lili. Each monkey performed the task in multiple sessions. Spike trains are stored per unit in each session. We provide the data in pickle format.
pickle contains the data in pickle protocol 0 format.The filenames encode the monkey, the session and the unit. The filename is of the form
<monkey><session>_<unit>.<format>.
Each .pkl contrains a dictionary with the following keys:
eventtimes : A Trials x 7 numpy-array of behavioral and experimental events in milliseconds (relative to the start of the trial).event_names : Column names of the events.spiketimes : A 3xN numpy-array. The first row contains the spike times in milliseconds,
the second row contains the trial number and the third row contains the direction of RS.gdf_file : Name of original gdf file, containing the raw data.column_names : Row names of spiketimes.The pickle files were created with the pickle protocol 0. To load the data use the following code:
import numpy as np
import pickle
with open('<monkey><session>_<unit>.pkl', 'rb') as f:
data = pickle.load(f, encoding='latin1')
print(data.keys())
The folder contains a file named toc. It is a product of our analysis and contains metadata about the files to speed
up the filtering of the data.
The data can be either downloaded via the command line or via the browser.
git clone https://gin.g-node.org/nawrotlab/delayed_center-out_uncertainty_Riehle
Download the latest release as a zip file by clicking on Releases on the main page at https://gin.g-node.org/nawrotlab/delayed_center-out_uncertainty_Riehle. This zip file contains all files.