Journal of Stone, J M and Towse, J N 2015 A Working Memory Test Battery: Java-Based

open research software Collection of Seven Working Memory Tasks. Journal of Open Research Software,
3: e5, DOI: http://dx.doi.org/10.5334/jors.br

SOFTWARE METAPAPER

A Working Memory Test Battery: Java-Based Collection

of Seven Working Memory Tasks
James M. Stone1 and John N. Towse1
1
Department of Psychology, Lancaster University, UK
[email protected]
[email protected]

Working memory is a key construct within cognitive science. It is an important theory in its own right,
but the influence of working memory is enriched due to the widespread evidence that measures of its
capacity are linked to a variety of functions in wider cognition. To facilitate the active research environ-
ment into this topic, we describe seven computer-based tasks that provide estimates of short-term and
working memory incorporating both visuospatial and verbal material. The memory span tasks provided
are; digit span, matrix span, arrow span, reading span, operation span, rotation span, and symmetry span.
These tasks are built to be simple to use, flexible to adapt to the specific needs of the research design,
and are open source. All files can be downloaded from the project website http://www.cognitivetools.uk
and the source code is available via Github.

Keywords: Working Memory; Computerised Cognitive Testing; Cognitive Assessment; Verbal Working
Memory; Visuospatial Working Memory, Java; Tatool
Funding Statement: Funded via a PhD Studentship award from the ESRC.

(1) Overview languages (Shute, 1991), life event stress (Klein & Boals,
Introduction 2001), regulating emotion (Kleider, Parrott, & King, 2010),
Klatzky (1980) described the working memory system as and multitasking (Bühner, König, Pick, & Krumm, 2006;
being a mental workbench using the analogy of a carpen- Hambrick, Oswald, Darowski, Rench, & Brou, 2010).
ters workbench. As a carpenter will lay out the tools and
materials that they need for the job on their workbench so Complex span tasks
they are readily available, our mental workbench can hold Daneman and Carpenter (1980) reported a paradigm that
‘chunks’ of information that are required for our current was designed to capture the conceptual requirements of
cognitive goals. This workspace has limited capacity and simultaneous processing and memory operations thought
the critical factors that contribute to this capacity ‘limit’ to be inherent to working memory functioning. In their
are open to debate but it is clear that the working mem- reading span task participants were required to read aloud
ory system is a limited resource and that this limit varies sentences and attempt to remember the last word in each
across individuals. sentence. Administration consisted of three trials at set
Working memory ability has been shown to correlate sizes two through six. The simultaneous processing of
reliably with other cognitive abilities such as; fluid intel- information while needing to store information for recall
ligence (Conway, Cowan, Bunting, Therriault, & Minkoff, has become an integral part of working memory research.
2002), arithmetic (McLean & Hitch, 1999), the ability to Complex span tasks follow the paradigm of item storage
prevent mind wandering during tasks requiring focus with concurrent processing of a demanding task in which
(Kane et al., 2007), executive attention (Kane & Engle, there are a set number of item storage and cognitive pro-
2003), general learning disabilities (Alloway, 2009), and cessing events. The form of the to-be-remembered (TBR)
many more. In addition to its prominence within cogni- items and the processing task can take many forms. Turner
tive research there are a wide variety of other disciplines and Engle (1989) introduced the operation span task with
incorporating WM ability in to their research programmes two versions that differed in the TBR units. The process-
and assessing the impact of this cognitive system on their ing part of the operation span task involves presenting the
respective domains of study. Some examples of topics that participant with a mathematical operation (e.g. ‘(6/2) +
have seen measures of WM used as a predictor include 2 = 5’) to which the participant must assess whether or
depression (Arnett et al., 1999), learning computer not the printed answer is correct. In the ‘Operations Word’
Art. e5, p.  2 of 9 Stone and Towse: A Working Memory Test Battery

version each operation was followed by the presentation length, number of trials at each set size, randomised trial
of a word and these words made up the TBR array. Another order, control over what data is logged, and many more.
version used in their experiment was ‘Operations Digit’ The drawback to this paradigm is that there are likely
where the participant was required to recall the numbers countless versions of each ‘non-standardised’ working
that were given as answers to the operations (regardless memory task out there that have been developed for a
of whether the operation was true or false). Other variants specific research program. A large number of these could
of a verbal complex span task exist such as the counting have been reused in a way that would have generated a
span task (Case, Kurland, & Goldberg, 1982) where object huge saving in resources.
counting forms the processing and array totals provide There are some notable examples of computerised work-
the TBR items, which was designed to be appropriate for a ing memory tasks that have been published and made
wide developmental population. available online for other researchers to use. The attention
A popular thread of research in the working memory and working memory lab at Georgia Institute of Technology
literature relates to whether there is a separation of have made available versions of five (at the time of writ-
verbal and visuo-spatial domains in the WM system (as ing) complex span tasks using E-Prime software (Unsworth,
popularised by the multi-component model of working Heitz, Schrock, & Engle, 2005) also see Redick et al. (2012).
memory, Baddeley and Hitch (1974); Baddeley (1986)) or In addition to ‘normal’ length versions of the tasks, short-
if a domain-general pool of resources might be the driv- ened versions are also available (Foster et al., 2014). The
ing force behind WM performance. Therefore alongside availability of these tools is excellent given the extensive
verbal complex span tasks, a number of visuo-spatial com- research the group have put into validation (Redick et al.,
plex span tasks have developed over time. For example, 2012). Another freely available set of computerised tasks
Shah and Miyake (1996) introduced a ‘rotation span’ task. for assessment of working memory have been produced
This combined a processing phase which involved men- using Matlab (Lewandowsky, Oberauer, Yang, & Ecker,
tally rotating letters and judging whether or not they 2010). This battery consists of four tasks that the research-
were regular or mirror images with a storage phase that ers selected to be representative of the various facets of the
presented arrows in varying orientations and lengths. The WM construct and therefore provide a reliable and valid
symmetry span task (Kane et al., 2004) uses grid locations measure of WM ability.
in a 5x5 matrix as the storage units while the processing These tasks are presented in the form of scripts that
phase requires judgements on the symmetry of a pattern can be executed in their respective programs (E-Prime/
filled in an 8x8 matrix. Matlab). Therefore with some knowledge of the pro-
gramming in these frameworks one could modify them
Availability of computerised tasks to change elements of the tasks. However, the scripts
Given this widespread incorporation of working memory can only be executed on computers with the E-Prime/
measures in many domains of research, the availability Matlab installed and this involves expensive license fees.
of software to run computerised assessments of aspects Many Universities/Departments pay these license fees
of working memory ability is important for the ongoing and therefore may have computers with the programs
investigation into the construct itself and the relationship installed for researchers to use, or one may have to incor-
with other functions. porate the costs of the software into any grant proposal.
Currently, the choices available to a researcher wishing This issue can be a wider concern if your research involves
to use working memory tasks, are (a) to build a software going out and collecting data outside of the lab. For exam-
package from the ground up, or (b) to use a commer- ple, in developing a working memory training experiment
cial product, such as standardised test kits available that as part of a PhD thesis we conducted an experiment that
relate to working memory such as the automated working involved pupils at a number of schools carrying out work-
memory assessment (Alloway, 2007). The AWMA is sold ing memory tasks in a group setting. To do this we needed
primarily to the education sector to assess pupils work- to use the IT facilities that the school had. It would not
ing memory but is also used as a research tool to provide have been practical to use currently available systems to
measures of WM e.g. Holmes et al. (2010). One drawback collect such data.
to using such a tool might be cost, administering these A further tool which provides researchers with the means
packages to hundreds of participants would add a large to create computerised cognitive tasks is the Psychology
amount to the cost of an experiment. In terms of experi- Experiment Building Language (PEBL; Mueller & Piper,
mental design these tools might not be a good fit in that 2014). PEBL provides a framework for creating tasks but
they are administered in a specific way and there is no also includes a battery of commonly used tasks with the
room for modification. This is a necessary property for a install, some of which are working memory tasks. PEBL is
standardised tool so that one can assess scores compared an open source project. Of the seven tasks I present here,
to the normalised scores. there are versions of five of them in the PEBL battery (digit
In many cases it seems that when a researcher wants span, reading span, operation span, matrix span, and sym-
to use a measure of working memory (and indeed for metry span). There are subtle differences in the imple-
many other ‘non-standard’ tools for measuring cognition) mentation of these tasks between the versions in the PEBL
they produce a version of the task ‘in-house’. This allows battery and the versions presented here. The PEBL is an
researchers the flexibility of being able to make design excellent project and has evolved into an immensely use-
choices that fits their experimental design e.g. custom ful tool for researchers. There are many tasks available in
Stone and Towse: A Working Memory Test Battery Art. e5, p.  3 of 9

the PEBL battery beyond just the WM tasks we provide

here. With regards to the WM tasks I think using PEBL or
the software described here is a choice for each researcher
to make based on the computers you are testing on (PEBL
requires an install) and the ease with which the task can
be configured to the specification desired.

Rationale for software described in this paper

There are a number of computerised tools available to a
researcher interested in measuring WM. However, each
one has its own limitations; either a rigid administration
or lack of flexibility in platform it can be used on. For these
reasons, we present our suite of computerised WM tasks.
Built using Tatool, a Java-based framework (von Bastian,
Locher, & Ruffin, 2013) and entirely open source. As the Figure 1: Illustration of the operation span task.
tasks are based on Java they are easily run on various oper-
ating systems providing the Java Runtime Environment ‘6 + 7 = 10’ and the participant must indicate if they think
(JRE) is installed. The JRE is often pre-installed on machines the given answer is correct or not. The digits and opera-
given the widespread use of Java and therefore no install tions are randomly generated each trial. The minimum
is usually required to run the application described in this and maximum digit can be set as an option in the mod-
paper. Should a JRE not already be present on the target ule file, the default is digits between 10 and 99. For each
machine it is freely and easily accessible online. operation there is a .5 probability that the operation will
We describe here a set of tasks commonly used tasks to be correct and the type of operation (multiplication,div
assess short-term/working memory, a number of complex ision,addition,subtraction) used has a .25 probability for
span tasks described above as well as simple span counter- each, this should provide a variety of types of operation
parts for both the verbal and visuo-spatial domain. Each requiring correct and incorrect responses.
task is independent and so there is no requirement to Processing-Storage order. It is worth mentioning
administer all of them as a package. Instead the pool of that the ‘traditional’ method of administering complex
tasks is made available so that anyone can select the most span tasks such as the operation span task involves using
appropriate based on your research needs. Measuring a processing-storage order of phases rather than storage-
working memory ability is best achieved through admin- processing as we have used. This method is rather curi-
istering a number of tasks and forming a composite score ous as the processing task serves the purpose of adding
from them (Kane, Hambrick, & Conway, 2005; Lane et al., to the cognitive demands of storage by requiring process-
2004). However, sometimes this is not possible. Perhaps ing of a task while trying to store stimuli. Therefore with
time spent with the participant is very limited or there are a processing-storage order of presentation it seems that
already a number of computerised tasks and fatigue is a the first processing phase of a trial has a different effect
concern. In these instances one can select the task(s) that to subsequent processing phases of the trial. In the first
most suit your needs based on your research goals. instance the participant is not holding any TBR stimuli
Throughout the rest of this paper we will outline the and thus this processing phase is not being carried out
tasks that are currently available, where to get them, how while task specific items are being stored in WM (for data
to use them, what we think they offer that is not already consistent with the notion that the first episode has dif-
readily available, and some commentary on the ongoing ferent requirements see Towse, Hitch, and Hutton (1998)).
nature of the project. That is not to say that the first processing element does
not add something to the cognitive requirements above
Tasks a simple span measure but that the effect it has should
Verbal WM tasks be considered differently to the subsequent processing
There are three verbal span tasks currently available, each phases. In this operation span task the similarity is clearly
a slight variation on the others. very high as the constituent parts of the processing tasks
Operation Span. The operation span is perhaps the are numbers, as are the TBR items. There is also then the
most commonly used verbal complex span task. Fig. 1 property of the participant being shown the final TBR item
shows a schematic view of how a trial in the operation immediately followed by the recall phase with no effective
span task is executed. Most complex span tasks can be retention interval. Therefore this item has not been sub-
broken down into a repeated cycle of memory and pro- ject to any additional processing decay. A span size two
cessing components. The current version of an operation trial illustrates the point most strikingly. With a processing-
span task consists of presenting the participant with an storage order the two TBR items bookend a processing
integer that needs to be stored and recalled at the end operation. The recall screen will appear and the second
of the trial in its correct serial position. For every storage TBR item will have only just been presented. If the par-
element (integer to remember) there is a processing phase ticipant can recall the first item that was one processing
immediately succeeding it. The processing phase presents phase ago then they can almost always be successful at
the participant with a mathematical operation such as these trials. It would seem that having this order could be
Art. e5, p.  4 of 9 Stone and Towse: A Working Memory Test Battery

considered a methodological choice that exacerbates the This processing element requires them to make a judge-
recency effect. ment of whether the presented pattern is symmetrical
Reading Span. The reading span task (Fig. 2) differed along the vertical axis or not using the left/right arrow
from the operation span task only in the processing ele- keys (8x8 grid used for presenting patterns).
ment. Rather than having to verify a mathematical opera- After the appropriate number of storage-processing
tion, the participants were presented with a sentence; elements have run for a trial the recall phase begins.
their task was to decide if it made sense or not. A note of Responses are recorded by presenting participants with
caution when using this task is that the sentences them- the 4x4 grid and allowing them to click the boxes in the
selves are defined in a stimuli file and any sentence is only order they recall seeing them. When a box is selected it
used once. Therefore if you require more trials than the turns blue so participants can keep track of their responses.
provided stimuli file can accommodate you would need The size of the grids (4x4 for storage and 8x8 for pro-
to update that file. It is likely that researchers may wish to cessing) is customisable in the module file as well as how
use their own sentences even in the case that the provided large they appear on screen.
stimuli file contains enough items. Matrix Span. The matrix span task is the memory span
Digit Span. The digit span task is the memory span equivalent of the symmetry span task. The procedure is
equivalent to the operation/reading span tasks. It is oper- the same as described for symmetry span except for the
ationally the same as these tasks but with no processing removal of the processing element.
phase therefore simply a stream of digits that must be Rotation Span. Fig. 4 shows a schematic representa-
remembered in serial position. tion of a rotation span trial showing the storage and pro-
cessing parts of the task. The to-be-remembered (TBR)
Spatial WM tasks stimuli in the rotation span task are images of arrows that
Symmetry Span. The symmetry span task is a spatial are differentiated in two characteristics. Any one arrow
complex span task. Participants are required to remem- can differ in its length (long or short), or it can differ in its
ber grid (4x4) locations presented to them in the correct angle of rotation (0°, 45°, 90°, 135°, 180°, 225°, 270°, or
serial order. Fig. 3 shows a schematic representation of 315°). Therefore the storage phase of this task is to remem-
this task. As is shown, participants are given a processing ber the arrows presented in their correct serial position.
operation to complete after each TBR grid is presented. The processing operation in this complex span task
presents participants with a letter (F, G, or R) that may be
standard or a mirror image. It may also be rotated at one
of the 45 degree rotations. The participant must mentally
rotate the image so that they can make a judgement on
whether the letter is a normal or mirror representation
using the left/right keys.
The recall screen presented the 16 possible arrows in a
2x8 grid where the top row of arrows were all the short
arrows and the bottom row were all the long arrows.
Participants used the mouse to select the arrows they
remembered seeing in the correct order.
Arrow Span. The arrow span task is the memory span
equivalent to the rotation span task. Therefore the pro-
cessing phase is dropped such that the arrow span task
is simply about remembering the arrows in correct serial
position.
Figure 2: illustration of the reading span task.

Figure 3: Illustration of the symmetry span task. Figure 4: Illustration of the rotation span task.
Stone and Towse: A Working Memory Test Battery Art. e5, p.  5 of 9

Implementation and architecture Instructions on using the tasks

The software is based on the Java programming language There are two components needed to run any of these
and therefore will be compatible with any computer with tasks and an optional third component that can help pro-
a Java runtime installed (currently tested and running cess the resultant data from using this package. The first
perfectly on the latest JRE which is version 8 update 31). component is the application itself. The application we
Therefore, the programs are accessible on multiple operat- provide is an executable jar and this jar file launches just
ing system platforms, including windows, OSX, and linux. like any other executable if an up to date Java runtime is
The framework for the tasks is provided by Tatool (von running on the operating system. The jar contains all of
Bastian et al., 2013). Tatool is an open source platform/ the Java dependencies as well as the Tatool libraries. In
library that provides much of the functionality required addition to these components it contains the executables
for creating computerised psychological tasks. Using we have written to run the tasks described above. The app
Tatool still involves programming a set of executables that contains all the source code but will do nothing without
run the type of task desired but Tatool provides a number the use of module files.
of functions to make this easier as well as providing a neat Tatool is designed to use xml files that specify (amongst
presentation of the program and storage of user informa- other things) for the application which executables to
tion and results. run in what order. What this means for a user is that after
Tatool and the working memory executables described launching the application first time a module file for each
here are also open source. This means that users are task needs to be loaded. I provide a default module file for
encouraged to engage with the source code and modify it each task. Inside the module file a number of values that
to their own specification if the functionality required is will affect the behaviour of the program can be changed.
not provided by default. For example, Tatool applications
and modules can be extended to interact with a web server Task flexibility
for remote administration of tasks. This is a built in Tatool As we have alluded to already everything is technically
feature and instructions on adding this functionality are flexible as the source code is provided and it is encouraged
included in the Tatool download (http://www.tatool.ch/ that users modify/improve where they see fit to achieve
download.htm). It is hoped that as users add functionality the results required.
these modifications can be added to the project. This is an The current suite of tasks have been designed and
important feature going forward but for the rest of this arranged to permit flexibility in their usage. Depending
paper we discuss the tasks as they are currently configured on hypotheses being tested and practical restrictions such
and specified. as duration constraints the deployment of a span task
The project website can be found at the following URL: may need to be altered from the default settings. By build-
http://www.cognitivetools.uk. The website hosts the ing the executables in a way that certain task behaviours
application as well as the supporting files all of which will can be easily altered by values in the xml the user is not
be discussed below. Using these tasks does not require restricted in their deployment. It is important to remind
any interaction with the source code. But we do provide users that a rigid administration of span measures may be
the source code in a github repository (DOI: 10.5281/ required for comparisons across studies.
zenodo.14825) and would welcome anyone to fork the Each task has a dedicated webpage, and one of the sec-
project and help improve and add to the project. This code tions on these pages outlines what variables can be cus-
needs to be added to a maven project with Tatool version tomised using values in the module files. If we take the
1.3.2 installed. symmetry span task as an example, Table 1 outlines the

Variable Potential Values Result

SpanTwoTrials– Any Integer Informs the program how many trials at each span size to run.
spanNineTrials
RandomisedTrials 0/1 If set to 0 then the trials will be given in ascending span size order.
If set to 1 then the order of trials is randomised.
GridDimension Any Integer (but be sensible…) The value (n) set here will be the size of the matrix presented to
participants. The default is 5 which produces a 5x5 grid.
ScalingFactor 1–100 Sets the size of the grid shown to participants. The program will
work out how much space there is available for the grid presenta-
tion. The scaling factor is then applied to determine the resultant
size of the display. For example if set to 50 then the grid size is half
the maximum.
MinSquares Integer < half the total number In conjunction with the ‘MaxSquares’ value, sets the minimum and
of squares maximum number of squares to fill when creating a pattern for
symmetry judgement.
MaxSquares Integer < half the total number See above.
of squares
Table 1: Properties that can be modified with simple value changes in the module file for the Symmetry Span task.
Art. e5, p.  6 of 9 Stone and Towse: A Working Memory Test Battery

information provided in the customisation section of the created and therefore much of its functionality is designed
symmetry span task page. Some of these variables are with that in mind. Therefore we must load a user profile to
common to many tasks such as the number of trials to begin serving the tasks to our participants. The user infor-
run at each span size, and whether or not you want these mation provided here is saved in every output file created
trials to run in order or randomised. If you elect to run from tasks administered to that user profile. Therefore it
them in order (non-randomised) then the program will is possible to create user accounts for participants and
run the specified amount of trials at each span size start- load up specific modules for them. This is not necessarily
ing at the lowest through to the highest, so if you have required though, as each task will ask the experimenter to
set three at each size then it will run three trials at span input a 5-digit participant code at the start of a trial. This
size two followed by three trials at span size three and so allows the experimenter to create one user account, load
forth. However, some researchers may wish to randomise the necessary modules for that testing session, and then
the trial order so that the participant isn’t aware on any supply different participant codes on each administration.
trial how many TBR items they are going to be presented For more complex designs the experimenter may wish to
with (e.g. Engle, Cantor, & Carullo, 1992). create different users for different participants, while for
As well as these attributes that are common across the more simple designs the experimenter can rely on the
tasks there are other attributes that can be altered that supplied participant codes.
only apply to a subsection of the tasks or are unique to Once a user has been created we are taken to the users
a task. An example of this is the ‘gridDimension’ variable module page, when we load a module file we add to
which controls the size of the matrix used to deliver grid the list of modules a user can access. To load a module
locations to participants and is present in the matrix and click the ‘Add’ button in the bottom-left corner and then
symmetry span tasks. select ‘Module from local file’. Navigate to the direc-
tory that stores the module file and select it. If there
Loading a module file are no errors in the module file then a new module will
Fig. 5 outlines the basic procedure for preparing an appear in the ‘My Modules’ pane (Fig. 5 box 3). Once
instance of one of the tasks for a participant. To load the module is loaded, changes to the module file are not
a module file one must first make a user profile in the reflected in the execution of that module. For example,
Tatool application. Tatool was created primarily to help if it is decided to modify the module file so as to admin-
researchers who conduct experiments that involves train- ister more trials then the updated module file needs to
ing (multi-session experiments) on the psychological tasks be reloaded.

Figure 5: Flow diagram displaying the basic steps needed to launch a task.
Stone and Towse: A Working Memory Test Battery Art. e5, p.  7 of 9

Extracting the data A final note about the R scripts is that they have been
Each time a participant completes a trial the data is saved produced to work with the tasks as they are. This means
automatically. This data can be extracted in CSV format that if you change the source code to modify a task then
using the ‘Export’ button. The resultant CSV file will con- it is also possible you may need to change the R script to
tain information on all trials that the user account has analyse the output. Modifications might include the intro-
completed. As hinted above there are two likely methods duction of new variables or the renaming of existing vari-
that a researcher will have deployed these tasks; either a ables which would require an updated R script to process.
user account per participant with the necessary modules This issue relates to actual source code modification of the
loaded, or single account with each different participant executables, the R scripts are flexible to the various altera-
being identified by providing a 5-digit participant code at tions that can be made by the user within the module file.
the start of each administration. We provide two R (R Core
Team, 2014) scripts per task, one for each method. The Quality control
r-scripts are organised with the name format ‘x_span_ The software described in this paper has been used in two
process.r’ and ‘x_span_process_singleUser.r’ large scale experiments by the authors as well as two smaller
experiments as part of the lead authors PhD research.
Single account method
This simpler method will result in just one CSV file being (2) Availability
exported at the end of the data collection phase of your Operating system
experiment. Download the appropriate R-script from the OSX/Windows/Linux. Any OS with an up to date JRE
website (making sure to use the script with ‘_singleUser’ suf- installed.
fix). Simply open the script and change “datafile.csv” to point
to your data file; a) if the working directory of R is set to the Dependencies
location of the data file then simply put the file name here, or Using the application only requires a JRE installed. To
b) alternatively supply an absolute file path or a file path rela- engage with the source code a number of maven depend-
tive to the current working directory. Then execute the script encies are required (tatool version 1.3.2 and the depend-
and a data frame will be in the R workspace called ‘x.span. encies that is reliant). Using the provided pom.xml file
data’ which summarises each participants performance. in the repository and an IDE with maven functionality
should automatically download the required packages.
Multiple account method
This method will result in an output CSV file per partici- Software location
pant. Take all these files from the respective task and put The compiled ready-to-use software is hosted at the project
them in the same directory with no other files. The direc- website, the downloads page can be found at http://www.
tory that holds the data files needs to be set as the working cognitivetools.uk/cognition/downloads. The project web-
directory in the R environment and then the respective R site hosts the main application as well as the additional
script can be run; either by using the source() function, or default module files and R analysis scripts. The source
opening the script within R and highlighting and running code for the project is hosted in a GitHub repository.
the code. We provide links to useful resources on the web-
site for those who are unfamiliar with R. Once again after Archive and code repository
executing the script the result will be a new data frame GitHub
in the R workspace which summarises each participants
performance on that task. Name
cog-tasks: Working Memory Test Battery
General comments on data processing
The scripts are written to extract general summary Persistent identifier
information about performance on the respective tasks. 10.5281/zenodo.15488
Examples of measures calculated:
URI
• Number of trials successfully recalled at each span https://github.com/j-stone/cog-tasks.git
size.
• Accuracy of processing phase responses (complex Project homepage
span measures). http://www.cognitivetools.uk/
• Median response time for processing phase responses
(complex span measures). Licence
• Various ‘scores’ to reflect overall performance. GNU GPL v3.0

Each task webpage includes a data section which outlines Publisher

what all the variables produced by the script represent. James M Stone
The obtained data frame after executing the provided R
script can be analysed within R or extracted to be analysed Date published
in an alternate data analysis program. 08/02/2015
Art. e5, p.  8 of 9 Stone and Towse: A Working Memory Test Battery

(3) Reuse potential and summary ing memory capacity, short-term memory capacity,
Tasks that assess working memory performance are being processing speed, and general fluid intelligence. Intelli-
used widely across many domains of research. Often gence, 30(2): 163–183. DOI: http://dx.doi.org/10.1016/
research groups need to create their own version of S0160-2896(01)00096-4
described tasks due to a lack of freely available resources Daneman, M and Carpenter, P 1980 Individual differ-
that are flexible in their implementation and easy to ences in working memory and reading. Journal of ver-
deploy on a variety of platforms. In this paper we have bal learning and verbal behavior, 19(4): 450–466. DOI:
introduced open source simple/complex span tasks that http://dx.doi.org/10.1016/S0022-5371(80)90312-6
are built using Tatool, a Java-based platform, and will pro- Engle, R W, Cantor, J and Carullo, J J 1992 Individual dif-
vide researchers with measures of both verbal and visuos- ferences in working memory and comprehension: a test
patial working memory. of four hypotheses. Journal of Experimental Psychology:
These tasks can be downloaded and used without the need Learning, Memory, and Cognition, 18(5): 972–992. DOI:
for any code wrangling immediately by using the default http://dx.doi.org/10.1037/0278-7393.18.5.972
settings. Additionally, they can be altered via the built in Foster, J L, Shipstead, Z, Harrison, T L, Hicks, K L,
customisable options by simply changing values in the mod- Redick, T S and Engle, R W 2014 Shortened com-
ule files. And finally, as the source code for both Tatool (the plex span tasks can reliably measure working memory
framework) and our executables are open source any modi- capacity. Memory & Cognition, pp. 1–11. DOI: http://
fications that one wishes to make can be made. Additionally, dx.doi.org/10.3758/s13421-014-0461-7
data processing scripts are provided which will process the Hambrick, D, Oswald, F, Darowski, E, Rench, T and
resultant data using R. Brou, R 2010 Predictors of multitasking perfor-
mance in a synthetic work paradigm. Applied Cogni-
Competing interests tive Psychology, 24(8): 1149–1167. DOI: http://dx.doi.
The authors declare that they have no competing interests. org/10.1002/acp.1624
Holmes, J, Gathercole, S, Dunning, D, Hilton, K,
Acknowledgements Elliott, J, et al. 2010 Working memory deficits can
We would like to thank Claudia von Bastian and Andre be overcome: Impacts of training and medication on
Locher for their work developing Tatool and for their help- working memory in children with ADHD. Applied Cog-
ful suggestions on an earlier draft of this manuscript. nitive Psychology, 24(6): 827–836. DOI: http://dx.doi.
org/10.1002/acp.1589
References Kane, M J, Brown, L H, McVay, J C, Silvia, P J, Myin-
Alloway, T P 2007 Automated working memory assess- Germeys, I and Kwapil, T R 2007 For whom the mind
ment: Manual. Pearson. wanders, and when an experience-sampling study of
Alloway, T P 2009 Working memory, but not iq, predicts working memory and executive control in daily life.
subsequent learning in children with learning diffi- Psychological science, 18(7): 614–621. DOI: http://
culties. European Journal of Psychological Assessment, dx.doi.org/10.1111/j.1467-9280.2007.01948.x
25(2): 92–98. DOI: http://dx.doi.org/10.1027/1015- Kane, M J and Engle, R W 2003 Working-memory
5759.25.2.92 capacity and the control of attention: the contribu-
Arnett, P, Higginson, C, Voss, W, Wright, B, tions of goal neglect, response competition, and task
Bender, W, Wurst, J and Tippin, J 1999 Depressed set to stroop interference. Journal of experimental psy-
mood in multiple sclerosis: Relationship to capacity- chology: General, 132(1): 47–70. DOI: http://dx.doi.
demanding memory and attentional functioning. org/10.1037/0096-3445.132.1.47
Neuropsychology, 13(3): 434–446. DOI: http://dx.doi. Kane, M J, Hambrick, D and Conway, A 2005 Working
org/10.1037/0894-4105.13.3.434 memory capacity and fluid intelligence are strongly
Baddeley, A D 1986 Working memory. London: Oxford related constructs: comment on Ackerman, Beier, and
University Press. Boyle (2005). Psychological Bulletin, 131(1): 66–71.
Baddeley, A D and Hitch, G 1974 Working memory. DOI: http://dx.doi.org/10.1037/0033-2909.131.1.66
In G. H. Bower (Ed.), The psychology of learning and Kane, M J, Hambrick, D Z, Tuholski, S W,
motivation, pp. 47–89. New York: Academic Press. DOI: Wilhelm, O, Payne, T W and Engle, R W 2004 The
http://dx.doi.org/10.1016/S0079-7421(08)60452-1 generality of working memory capacity: a latent-
Bühner, M, König, C, Pick, M and Krumm, S 2006 Work- variable approach to verbal and visuospatial memory
ing memory dimensions as differential predictors of span and reasoning. Journal of Experimental Psychol-
the speed and error aspect of multitasking perfor- ogy: General, 133(2): 189–217. DOI: http://dx.doi.
mance. Human Performance, 19(3): 253–275. DOI: org/10.1037/0096-3445.133.2.189
http://dx.doi.org/10.1207/s15327043hup1903_4 Klatzky, R L 1980 Human memory: structures and processes.
Case, R, Kurland, D M and Goldberg, J 1982 Operational San Francisco: Freeman.
efficiency and the growth of short-term memory span. Kleider, H, Parrott, D and King, T 2010 Shooting
Journal of experimental child psychology, 33(3): 386–404. behaviour: How working memory and negative emo-
DOI: http://dx.doi.org/10.1016/0022-0965(82)90054-6 tionality influence police officer shoot decisions.
Conway, A R, Cowan, N, Bunting, M F, Therriault, D J Applied Cognitive Psychology, 24(5): 707–717. DOI:
and Minko, S R 2002 A latent variable analysis of work- http://dx.doi.org/10.1002/acp.1580
 Stone and Towse: A Working Memory Test Battery Art. e5, p.  9 of 9

Klein, K and Boals, A 2001 The relationship of life event Shah, P and Miyake, A 1996 The separability of working
stress and working memory capacity. Applied Cogni- memory resources for spatial thinking and language
tive Psychology, 15(5): 565–579. DOI: http://dx.doi. processing: an individual differences approach. Journal
org/10.1002/acp.727 of Experimental Psychology: General, 125(1): 4–27. DOI:
Lewandowsky, S, Oberauer, K, Yang, L and Ecker, U http://dx.doi.org/10.1037/0096-3445.125.1.4
K 2010 A working memory test battery for matlab. Shute, V 1991 Who is likely to acquire programming
Behavior Research Methods, 42(2): 571–585. DOI: skills? Journal of Educational Computing Research,
http://dx.doi.org/10.3758/BRM.42.2.571 7(1): 1–24. DOI: http://dx.doi.org/10.2190/VQJD-
McLean, J F and Hitch, G J 1999 Working memory T1YD-5WVB-RYPJ
impairments in children with specific arithmetic learn- Towse, J N, Hitch, G J and Hutton, U 1998 A
ing difficulties. Journal of Experimental Child Psychol- re-evaluation of working memory capacity in children.
ogy, 74(3): 240–260. DOI: http://dx.doi.org/10.1006/ Journal of memory and language, 39(2): 195–217. DOI:
jecp.1999.2516 http://dx.doi.org/10.1006/jmla.1998.2574
Mueller, S T and Piper, B J 2014 The psychology experi- Turner, M and Engle, R 1989 Is working memory ca-
ment building language (PEBL) and PEBL test battery. pacity task dependent? Journal of Memory and
Journal of neuroscience methods, 222: 250–259. DOI: Language, 28(2): 127–154. DOI: http://dx.doi.
http://dx.doi.org/10.1016/j.jneumeth.2013.10.024 org/10.1016/0749-596X(89)90040-5
R Core Team 2014 R: A language and environment for Unsworth, N, Heitz, R, Schrock, J and Engle, R 2005
statistical computing [Computer software manual]. An automated version of the operation span task.
Vienna, Austria. Retrieved from http://www.R-project.org Behavior Research Methods, 37(3): 498–505. DOI:
Redick, T S, Broadway, J M, Meier, M E, Kuriakose, P S, http://dx.doi.org/10.3758/BF03192720
Unsworth, N, Kane, M J and Engle, R W 2012 Measur- von Bastian, C C, Locher, A and Ruffin, M 2013 Tatool:
ing working memory capacity with automated complex A java-based open-source programming framework
span tasks. European Journal of Psychological Assessment, for psychological studies. Behavior research methods,
28(3): 164–171. DOI: http://dx.doi.org/10.1027/1015- 45(1): 108–115. DOI: http://dx.doi.org/10.3758/
5759/a000123 s13428-012-0224-y

How to cite this article: Stone, J M and Towse, J N 2015 A Working Memory Test Battery: Java-Based Collection of
Seven Working Memory Tasks. Journal of Open Research Software 3:e5, DOI: http://dx.doi.org/10.5334/jors.br

Published: 05 June 2015

Copyright: © 2015 The Author(s). This is an open-access article distributed under the terms of the Creative Commons
Attribution 3.0 Unported License (CC-BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium,
provided the original author and source are credited. See http://creativecommons.org/licenses/by/3.0/.

Journal of Open Research Software is a peer-reviewed open access journal published by

Ubiquity Press
OPEN ACCESS

View publication stats