Task of verbal working memory (n-back) made by Michael Colwell ([email protected]).
Latest version of the task overhauls older versions (1.4 and earlier) with new stimuli, optimisations and two task versions;
- The scanner task version has been designed for use during fMRI acquisition
- The non-scanner task version can be completed standalone.
Both task versions have parity, therefore the non-scanner version of the task can be used to compliment the fMRI version for outside scanner practice.
Older versions of this task have been used within publications (e.g., https://www.sciencedirect.com/science/article/pii/S2772408523000686).
This version of the task is currently in use in the several projects in PERL Oxford (PEACE - NCT05849675; FEN&COG - NCT05026398; PROGRESS).
This release is open for free, academic use/modification. If you use this work in an academic paper/dissemination, Please cite the code below:
Colwell, Michael, Tagomori, Hosana, Martens, Marieke, Murphy, Susannah, & Harmer, Catherine. (2023). N-Back (Oxford) - fMRI and non-scanner release (Python). Zenodo. https://doi.org/10.5281/zenodo.8003407
Pre-processing scripts made by Hosana Tagomori ([email protected]) with optimisations by Michael Colwell.
Massive help from Marieke Martens ([email protected]) in helping to create a neuroimaging version of the task.
Original debuggers: Chloe Wigg, Amy Gillespie, Calum Guinea, & Tereza Ruzickova
This is a standard N-Back working memory task, with four subtypes (n-types): zero-back, one-back, two-back, & three-back.
Stimuli are displayed for 500ms, with a fixed ISI of 1500ms. The block design of the old n-back task on E-Prime has been repurposed here (16 blocks total, 10 stimulus per block), which was used in the following publication: https://www.nature.com/articles/mp2017104
There will only be 'target' stimuli around 2-3 times per each block (around 20-25%) as per standard n-back design.
Each block has been mixed so that there is a different rule each block (permenant set-shifting). Four new practice blocks has been added to this version, as well as two seperate versions of the task for pre/post design.
You may generate new random lists using a partially completed random array generation script provided in the main folder ('ArrayScriptBase.py').
If you are using this task in an fMRI scanner, you may find the python script in the 'Scripts' folder beneficial, as this allows automatic retrieval of relevant temporal information when creating contrasts/regressors in FSL.
This task has been tested on a 120hz monitor but should work fine on a 60hz monitor.
If you have any questions or encounter any bugs, please email [email protected]
- Added code to allow undertaking main task without practice, practice only, and both.
- Added a small dot fixation point (better than cross as it doesn't overlap letters) but does attract fixation to centre-screen
- Added requirement to press 'left' key when a letter is not 'n-back' (i.e., it is not a recurring letter or X). Previously (version 1.0-1.3) participants did not press a key here, but this approach likely undermeasured performance.
- Altered the letters displayed to deter phonological or visual strategies (only using b, B, d, D, g, G, p, P, t, T, v, V)
- Small error fixes
- Rebalanced blocks so there for each n-back level there are two 'two target' blocks and two 'three target' blocks. Rebalanced so there are 5 uppercase letters, 5 lowercase letters per block.
- Included a python file to extract scanner timings
- Download Psychopy (version v2020.2.8 and above) - https://www.psychopy.org/download.html
- Once installed, unzip the contents of this folder to a location on your computer
- Open the 'nback_2021_psychopy_version-main' folder and double click on 'N-Back__New_Version_2021_'
- Click 'run' experiment - You can choose between two task versions ('1' and '2') for PRE/POST design.
- Press run and follow onscreen instructions
- Once the experiment has ended, you will find your test data in the 'data' folder.