Adaptive N-Back Task

Technical Manual

Script Author: Katja Borchert, Ph.D. (katjab@millisecond.com), Millisecond

Created: March 01, 2012

Last Modified: January 07, 2025 by K. Borchert (katjab@millisecond.com), Millisecond

Script Copyright © Millisecond Software, LLC

Background

This script implements a basic shape n-back procedure. The n-back task is a go/nogo working-memory performance task with increasing levels of difficulty.

The implemented procedure is based on:

Jaeggi, Susanne M.; Studer-Luethi, Barbara; Buschkuehl, Martin; Su, Yi-Fen; Jonides, John; Perrig, Walter J. (2010). The relationship between n-back performance and matrix reasoning - implications for training and transfer. Intelligence, 38, 625–635.

Millisecond thanks Dr. Susanne Jaeggi for kindly providing original instruction materials, stimuli, as well as assistance!

Adjustments to z-scores as recommended by: Gregg, A. & Sedikides, C. (2010). Narcissistic Fragility: Rethinking Its Links to Explicit and Implicit Self-esteem, Self and Identity, 9:2, 142-161 (p.148)

Duration

20 minutes

Description

In the 1-key single n-back task, participants are shown a sequence of stimuli (here: a sequence of yellow shapes on a black background) and are asked to indicate whether the currently presented stimulus fulfills the following criteria:

For N=0 trials: - is the shape the same as the shape '1.gif'? If so (it's a target), and press "A". If not, don't respond.

For N=1 trials: - is the shape the same as the one that preceded it? If so (it's a target), and press "A". If not, don't respond.

For N=2 trials: - is the shape the same as the one presented two trials before? If so (it's a target), and press "A". If not, don't respond.

For N=3 trials: - is the shape the same as the one presented three trials before? If so (it's a target), and press "A". If not, don't respond.

Etc. for increasing levels of N

Procedure

- After instructions, participants receive 10 trials of practice per level N tested (here: N = 2 to N = 4),
edit list.nextN_practice to change levels of N practiced

- once practice is done, participants get the option to repeat practice (see parameters.allowPracticeRepeat)

- After practice, participants receive 3 blocks per level of N tested. By default, the level of N increases from
1 to 3.
edit list.Nlevel to change a) N levels tested b) number of testblocks c)selection of N-levels

Block Information:
- Each single n-task block consists of 20+ trials using 8 different shapes.
=> + trials are the N trials that cannot display target shapes yet (=start trials). Their numbers vary depending on N.
The data collected during these start trials are NOT included in performance counts for Correct Rejections/False Alarms.
=> Of the actual 20 experimental trials, 6 present a target and 14 do not (3:7 ratio).
To change the ratio, you need to change the relevant code under block.Test
- The computer selects randomly
a) whether it is a target trial or not and
b) what shape to show if it is not a target trial

Adjustment of Level N:
(a) based on performance of participants either a higher, lower or the same level N is chosen for the following block
(after 15 blocks the session is over):
down: block acc performance < 75%correct
same level: 75%correct <= block acc performance < 90%
up: block acc performance >= 90%correct

(b) adaptive testing starts at level N = 1 in this script and can go up indefinitely
(c) participants receive performance feedback (% correct) after each block


Trial Information:
- Each trial presents the shape for 500ms and waits another 2500ms before presenting the next shape in the sequence.
The trialduration can be edited by experimenter.
- Participants have the entire 3000ms to respond by pressing "A" if they detect a target.

Stimuli

8 shapes, see section Editable Stimuli

Instructions

provided by Millisecond - can be edited under section Editable Instructions

Instructions are not original to the task. They are provided by Millisecond
as htm pages and simple page elements and can be edited either by changing
the provided html files or directly under Editable Instructions.

in order for the instructions to reflect alternative parameter settings,
some of the conditionalized code used in the htm files can be found in the script itself.

Summary Data

File Name: singlenbacktaskadaptive_shapes_summary*.iqdat

Data Fields

NameDescription
inquisit.version Inquisit version number
computer.platform Device platform: win | mac |ios | android
startDate Date the session was run
startTime Time the session was run
subjectId Participant ID
groupId Group number
sessionId Session number
elapsedTime Session duration in ms
completed 0 = Test was not completed
1 = Test was completed
totalBlocks The total number of experimental blocks run
parameters.startN The first level of N tested
parameters.minN The lowest possible level of N that can be tested
meanLevelN The mean level N run
medianLevelN The median level N run
maxLevelN The maximum level N run
minLevelN The minimum level N run
the following summary variables are only based on performance on trial.target/trial.nontarget
(excludes trial.start even if parameters.excludeStartTrialfromPerformanceMeasure = false)
propCorrect Overall proportion correct (across all test blocks)
meanHitRT Overall mean correct response time (in ms); is equal to the overall mean hit response time
z-score calculations: adjustments (see Gregg & Sedikides, 2010, p.148)
If the hit rate FA rate is 0 => 0.005 is used instead (aka 0.005 is added to the hitFA rate)
IF the hit rate FA rate is 1.0 => 0.995 is used instead (aka 0.005 is subtracted from the hitFA rate)
list.hitsOverall.itemCount The absolute number of target trials (across all test blocks)
hitRateOverall Overall hit rate (across all test blocks) = hitting 'A' for targets
missRateOverall Overall miss rate
hitRTOverall Overall mean hit response time (in ms; across all test blocks)
list.commissionsOverall.itemCount The absolute number of nontarget trials (across all test blocks)
faRateOverall Overall False Alarm/Commission Error Rate (across all test blocks) = hitting 'A' for nontargets
crRateOverall Overall correct rejection rate
faRT Overall mean False Alarm response time (in ms; across all test blocks)
zHitRateOverall The z-score of the overall hit rate
zFaRateOverall The z-score of the overall False Alarm rate
dPrimeOverall Computes d' (parametric measure of sensitivity)
The higher the value, the better targets were overall correctly distinguished
from nontargets (d' = 0: chance performance; negative values: nontargets were treated as targets
and targets were treated as nontargets)
cOverall C-criterion in signal detection:The absolute value of c provides an indication of the strength of
the response bias/response style
negative: participant more likely to report that signal (go stims) is present (liberal response style)
may favor faster responding in speed-accuracy trade-off response paradigms
positive: favoring caution (conservative response style)

Raw Data

File Name: singlenbacktaskadaptive_shapes_raw.iqdat

Data Fields

NameDescription
build Inquisit version number
computer.platform Device platform: win | mac |ios | android
date Date the session was run
time Time the session was run
subject, group, With the current subject/groupnumber
session Session number
blockCode Name of the current block
blockNum Number of the current block
trialCode Name of the current trial
trialNum Number of the current trial
totalBlocks The total number of experimental blocks run
n The current level N tested
startTrialCounter Keeps track of how many start trials have been run
stimulusItem.1 The shape presented
stimulusNumber.1 The item number of the presented shape
currentTarget The item number of the current target
response The Response of the participant (scancode of response button)
0 = no response
31 = 'A' key (positive response)
responseCategory "Hit" vs. "Omission Error" vs. "CorrReject" vs. "Commission Error"
correct The correctness of the response (1 = correct; 0 = otherwise)
latency How fast a participant responded within the given timeframe, if at all (in ms)
no responses = the entire trialduration
list.blockAcc.mean Proportion correct during the current block (start trials are excluded from block acc assessment by default, editable)

Parameters

The procedure can be adjusted by setting the following parameters.

NameDescriptionDefault
allowPracticeRepeat True: practice can be repeated as often as indicated (default)
false: practice cannot be repeated
excludeStartTrialfromPerformanceMeasure True (1): performance during start trials that cannot be targets yet is not considered
for Block level performance analyses, e.g. for feedback purposes and N-level adjustments (default)
false (0): performance during start trials that cannot be targets yet IS considered for block performance analyses (default)
numberNBackTasks The number of n-back tasks to perform for adaptive testing 15
moveUpCriterium Sets the minimum proportion correct in a block to move up to the next level N
excludes start trials by default
0.9
moveDownCriterium Sets the minimum proportion correct in a block to stay on the same level N
(anything lower and participant moves down a level) (default: 0.75); excludes start trials by default
startN Sets the starting N 1
lowestN Sets the lowest possible N to run 1
soa Stimulus Onset Asynchrony 3000ms
stimulusPresentationTime The presentation time (in ms) of the stimuli 500ms
debugmode Debugmode = 1: targetalerts are shown on screen,
debugmode = 0, no targetalerts are shown (default)
0