Towards the Prediction of Dyslexia by a Web-based Game
with Musical Elements
Web Research Group Luz Rello
HCI Institute Ricardo Baeza-Yates
Web Research Group
Universitat Pompeu Fabra Carnegie Mellon University Universitat Pompeu Fabra
firstname.lastname@example.org email@example.com firstname.lastname@example.org
Music Technology Group
Universitat Pompeu Fabra
Current tools for screening dyslexia use linguistic elements,
since most dyslexia manifestations are related to diﬃculties
in reading and writing. These tools can only be used with
children that have already acquired some reading skills and;
sometimes, this detection comes too late to apply proper re-
mediation. In this paper, we propose a method and present
DysMusic, a prototype which aims to predict risk of hav-
ing dyslexia before acquiring reading skills. The prototype
was designed with the help of ﬁve children and ﬁve parents
who tested the game using the think aloud protocol and be-
ing observed while playing. The advantages of DysMusic
are that the approach is language independent and could be
used with younger children, i.e., pre-readers.
•Human-centered computing → Empirical studies in
accessibility; Accessibility design and evaluation meth-
ods; •Software and its engineering → Interactive games;
Dyslexia; Detection; Pre-Readers; Serious Games; Gamiﬁ-
cation; Web-based Assessment
The American Psychiatric Organization deﬁnes dyslexia
as a speciﬁc learning disorder which is caused by the ‘phono-
logical skills deﬁciencies associated with phonological cod-
ing deﬁcits’ [1, 16]. This means that people with dyslexia
have problems spelling words and decoding what they have
heard. Dyslexia does not aﬀect general intelligence, and it
is frequent: around 5% to 15% of the population has this
learning disorder .
Permission to make digital or hard copies of all or part of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed
for proﬁt or commercial advantage and that copies bear this notice and the full citation
on the ﬁrst page. Copyrights for components of this work owned by others than the
author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior speciﬁc permission
and/or a fee. Request permissions from email@example.com.
W4A 2017, April 02 - 04, 2017, Perth, Western Australia, Australia
© 2017 Copyright held by the owner/author(s). Publication rights licensed to ACM.
ISBN 978-1-4503-4900-0/17/04. . . $15.00
Jeffrey P. Bigham
HCI and LT Institutes
Carnegie Mellon University
The impact of dyslexia is related to diﬃculties with read-
ing and writing and for that reason current tools for screen-
ing dyslexia use linguistic elements. Examples of digital
screening tools for detecting dyslexia in English are Lex-
ercise Screener  and Nessy . Additionally, the com-
puter game Dytective  is available in English and Span-
ish with an accuracy of 83% for detecting if a person may
have dyslexia or not, according to a game that includes
recognition of linguistic elements, phonological awareness
and reading comprehension.
Reading and spelling tests need a minimum knowledge of
phonological awareness, grammar, and vocabulary of a child
to be able to predict dyslexia. This means that children can
be detected only after they begin to learn to read (generally,
during the ﬁrst year of school or later). This puts students
with dyslexia behind; therefore, new ways of detecting the
risk of having dyslexia are needed for pre-readers. Also, all
reading and spelling tests are language dependent.
The diﬃculty in detecting dyslexia before children learn
how to read and write is that the standard linguistic mani-
festations to detect dyslexia (diﬃculties in reading and writ-
ing) are missing. Hence, to detect dyslexia in a child before
they gain phonological awareness, new indicators of dyslexia
need to be discovered. Huss et al.  show that sound struc-
ture is related to the auditory perception of children with
dyslexia. Therefore, we aim to ﬁnd out if we can distinguish
two groups of pre-readers (with and without dyslexia) using
musical elements in a ﬁrst prototype of a game called Dys-
Music (see Figure 1). In this paper, we present the design of
the ﬁrst version of the game DysMusic and an usability test
with ﬁve children and ﬁve adults to discover problems that
could inﬂuence the prediction of the risk of having dyslexia
(ﬁrst prototype was presented in ). The rest of the paper
is organized as follow: next section presents related work;
Section 3 motivates the use of music for predicting dyslexia
while Section 4 presents the design of DysMusic. In Section 5
we present the usability test and the last sections give our
conclusions and future work.
2. RELATED WORK
We focus on examples for digital detection of dyslexia for
pre-readers. The tool DYSL-X aims to predict if a child
has dyslexia at the age of ﬁve [4, 15]. It contains three
mini-games for diﬀerent activities to measure indicators (e.g.
Figure 1: Example of the game DysMusic for the
ﬁrst two clicks on two sound cards (left) and then a
pair of equal sounds is found (right).
‘letter knowledge, FM detection, end-phoneme recognition’
) for dyslexia and takes about one hour to complete. The
mini-games are designed with a high focus on the game play
to motivate the player.
The AGTB 5–12 is a computer-based memory test for
children of ﬁve to twelve years-old . The application
provides four tasks for the phonological working memory
processing, six tasks for the central working memory and
two tasks for the visual-spatial working memory. Every task
takes seven minutes for testing.
The Bielefelder Screening (BISC) is for children at
the last year of kindergarten . It contains nine tasks for
phonological perception, phonological working memory pro-
cessing, long-term memory, and visual attention. It clusters
the participants in risk groups for dyslexia and takes around
20 to 25 minutes.
In summary, the games described are designed with a high
focus on the game design and the suitable interaction for
pre-readers. These tools are mainly focused on having letter
knowledge and phonological awareness being AGTB 5–12
and BISC the only ones screening pre-readers. All these
tools require a minimum linguistic knowledge.
3. WHY MUSICAL ELEMENTS?
People with dyslexia have auditory and visual perception
diﬃculties that seem to be caused by diﬃculties with the
short-term memory . Johnson’s work found that 26 out
of 60 children with dyslexia had speech and language dif-
ﬁculties and that 75 out of 120 children with dyslexia had
problems in the auditory learning . This is applied es-
pecially to the phonemic discrimination and therefore the
perception of new words are problematic , as used in the
game Dytective .
The visual perception has been explored with various
games as described in the previous section. To explore
the auditory perception without directly referring to letter
knowledge or phonological awareness, music can be used as
the communication channel.
By musical elements we mean diﬀerent acoustic param-
eters of sound (i.e., frequency,duration, orintensity) which
relate with perceptual parameters (such aspitch,loudness,
ortimbre). Some musical elements for readers at the age
of 10 have been explored, and some diﬀerences have been
found for rise time, duration, intensity (related to loudness)
and frequency (related to pitch) between children with
and without dyslexia . The musical elements would be
especially helpful for children that have no knowledge of
The case of pre-readers is especially challenging, because
dyslexia is known for causing reading and writing problems.
As we mentioned, the auditory perception might be corre-
lated to the causes of dyslexia if the cause is phonological
. To make our detection game language independent, we
will use music (phonological grammar ), which is similar
in prosodic structure in language.
In the literature review we already found indicators in the
hearing perception to distinguish between readers with and
without dyslexia [5, 10]. However, the auditory perception
without using linguistic features for pre-readers has been not
studied and used for the prediction of dyslexia in a game.
4. CONTENT DESIGN
The game aims to detect diﬀerences in the perception of
auditory elements for children with and without dyslexia
caused by problems with the short-term memory. As is well-
known, a game is more fun than doing a test. We can think
that DysMusic is an adaptation of the already existing visual
Instead of ﬁnding the same picture under various analog
cards, the child will try to ﬁnd the same musical element
behind various digital cards. The game has four tasks and
they have to be played without interruptions. Each task
includes two subtasks. One subtask has four cards with two
sound ﬁles (see Figure 1) and the other one has six cards
with three sound ﬁles. The number of cards is reduced to
four and six cards for two reasons: (i) if there is a diﬀerence
in the perception, only two diﬀerent musical elements will
already show the diﬀerence; and (ii) increasing the number of
cards is a challenge in itself, especially for younger children.
To avoid external factors that could cause diﬀerences in
the perception (e.g., due to color blindness), no visual cues
(color, shape) where used to distinguish the cards. To avoid
random matches, no match is possible within the ﬁrst click of
every subtask. Sound cards for every subtask are in random
order and the tasks are counter-balanced with Latin Squares
All musical samples are generated by means of a simple si-
nusoidal waveform (tone) using the Audacity free software.2
Four musical parameters are modiﬁed, obtaining a set of
waveforms as the one shown in Figure 2 for Rhythm. The
exact details of each one are given in Table 1.3
5. USABILITY TEST
First, internal feedback from HCI researchers improved
the application and only minor changes on the game play
needed to be done. After that the usability test was con-
ducted by children and parents who are not the authors of
this paper and are not familiarized with the research.
1An example of a visual memory game can be found on
2Audacity is available at http://audacity.es/.
3The generated musical elements are available at http://bit.
Figure 2: The waveform of the musical element Rhythm.
Musical element Sound Properties
Always the same waveform: sinus, mono ﬁles
musical elements: 2 to 3 sound ﬁles
frequency : 440 Hz unless speciﬁed
Frequency 2 elements* : 440 Hz, 452.8929 Hz
(change of 3 elements** : 2 previous sounds
tone frequency) and 446.3998 Hz
fade in/out: 0.025s
Length 2 elements: 0.350s, 0.437s
(change of 3 elements: 2 previous sounds
tone length) and 0.525s
fade in/out: 0.025s
Rise Time 2 elements: 0.025s fade in, 0.250s
(change of fade in, both with fade out of
rise time) 0.025s
3 elements: 2 previous sounds and
0.025s fade in and 0.250s fade out
(change in rise time
2 elements: (i) musical events with
rise time equal to 100ms, 100ms
and 0.025s; and (ii) rise time equal
to 100ms, 0.025s and 100ms
3 elements: 2 previous sounds plus
one with rise time equal to 0.025s,
Table 1: Musical elements generated for the four
tasks in DysMusic. * 2 elements: (1/2 semitone -
50 cents interval); ** 3 elements: 3 sounds spaced
by 25 cents (quarter of a semitone) - 2 previous ones.
Design: Since DysMusic is a new application a ﬁve user
test  is conducted to discover (usability) problems which
could unintended inﬂuence the planned study for predicting
risk of having dyslexia. It should be mentioned that a ﬁve
user test is a preliminary test for ﬁnding major usability
problems and does not aim to ﬁnd all usability problems.
In a within-subject design, all participants played all four
tasks of the game DysMusic. Only parents entered addi-
tional details for the study (e.g., background information)
while using the think aloud protocol .
Participants: We recruited ten participants which were
ﬁve children (users) and ﬁve parents. Two female parents
(both with age 35) and three male parents (ages 35, 40, and
40) participated. Each parent had two children and ﬁve of
their children took part in the user testing of DysMusic.
Two female children (ages 3 and 8) and three male children
(ages 5, 9 and 9) participated. All participants were German
native speakers. Since there is no indication of signiﬁcant
diﬀerences in usability studies for people with or without
dyslexia, we did not diﬀerentiate the two groups for the us-
Procedure: All participants played DysMusic over the
website4 with the same tablet (Android Galaxy Tab A).
They choose if they wanted to use headphones or not while
playing (only one female parents used headphones). First
the parent read the study instructions and played with the
sound cards (see Figure 1) while using the think aloud pro-
tocol. Afterwards, the parent or the ﬁrst author ﬁlled in the
background details for the child and the child played with
the sound cards. After each subtask the ﬁrst author asked
the participant ‘How diﬃcult or easy was it to distinguish
between the sounds?’. At the end of game each participant
was asked if they had further comments on the interaction
design of the game or the musical elements.
We present now the results of the usability test and the
changes we made to DysMusic.
Wording and Text: Generally, the parents found the
text easy to read and to understand. They reported some
spelling mistakes which were mainly caused by the transla-
tion process from English to German e.g. study (English)
vs. Studie (German).
Also they mentioned the large amount of text for Online-
Consent and one parent suggested to only present the im-
portant information of the Online-Consent and give the pos-
sibility for further reading.
Interaction: All participants were able to play the mem-
ory game with the musical elements instantly and all became
faster after the ﬁrst tasks, independently of the musical ele-
ment. Only the youngest child (3 years old) had major prob-
lems with the amount of six sound cards for all musical ele-
ments and did not ﬁnd any sound pairs. We consider to only
use four cards when younger children play DysMusic. Some
participants suggested to include the button ‘let’s play’ into
the game summary to make the interaction more visible.
The ﬁrst author observed that the participants started to
play faster, especially after the ﬁrst task, and the delay of
releasing the sound cards for the next click helped to control
the speed of the game interaction without being annoyingly
User-Interface: In general, all participants liked the
structure, layout, and the game elements, e.g., story, and the
spoken motivation feedback ‘Yeah’. One participant com-
mented that the footer of the game was very visible (large)
4The demo of the prototype DysMusic is available at http:
and suggested to make it smaller and more conspicuous. We
did this change accordingly.
Musical Elements: All participants commented that
they had to listen and concentrate carefully to be able to
distinguish the musical elements. Participants had a dif-
ferent perception on how diﬃcult it was to distinguish the
sounds and ﬁnding the card pairs, depending on the musical
elements. But all participants mentioned the ﬁrst musical el-
ement of the ﬁrst subtask always as diﬃcult, independently
from the musical element (because of the counter-balanced
design, the musical element order changed). This seems to
be because it is the ﬁrst time they play. For the second
subtask, they were already familiarized with the parameter
and were able to name it. Two children and three par-
ents mentioned diﬃculties in recognizing the musical ele-
ment Length. One parent and one child of this group and
another parent described more diﬃculties with the musi-
cal element Frequency. Only three parents reported diﬃcul-
ties in distinguishing the musical element Rhythm and two
children mentioned diﬃculties for the musical element Rise
Functionality: The motivation sound between the exer-
cises were not always played on the tablet and needed to be
debugged for diﬀerent devices. Besides, the video sometimes
could not be played instantly which might have been caused
by bad Wi-Fi connection. A change of the video player, from
HTML 5: video-tag to YouTube: Iframe-tag, prevented the
Other Comments: In general, all participants found the
task easy to understand. The children expressed more fun
while playing than the parents, e.g., smiling or laughing.
Three users commented that the game was fun and all users
reacted positive on the spoken feedback ‘Well done’ when it
was played. We included more game sound elements, e.g.,
after each found pair we added a spoken feedback like ‘Great’
6. CONCLUSIONS AND FUTURE WORK
The main advantages of DysMusic are that it is language
independent and that could be used by pre-readers. Re-
garding the user test, all participants understood the game
easily and played with no interruptions. The next step is to
include the visual elements into the game. After that it will
be possible to perform a set of experiments with 30 partic-
ipants to ﬁnd how musical and visual elements can distin-
guish between a person with or without dyslexia. To study
the dependency on diﬀerent languages, experiments will be
conducted in English, Spanish, and German. The best sce-
nario would be to prove that the approach is truly language
independent and only the interface needs to be translated.
The contents of this paper were partly funded by grants
from the National Science Foundation (NSF) and the Na-
tional Institute on Disability, Independent Living, and Reha-
bilitation Research (NIDILRR). We deeply thank all parents
and children for playing DysMusic (German version). Spe-
cial thanks to Hendrik Witzel for his advice during the de-
velopment of DysMusic and suggestions on fonts and color.
We thank Samsung Spain for donating the tablet used in the
 American Psychiatric Association. Diagnostic and
Statistical Manual of Mental Disorders. American
Psychiatric Association, may 2013.
 H. Brau and F. Sarodnick. Methoden der Usability
Evaluation. Verlag Hans Huber, Bern, 2 edition, 2006.
 A. P. Field and G. Hole. How to design and report
experiments. SAGE Publications, 2003.
 L. Geurts, V. Vanden Abeele, V. Celis, J. Husson,
L. Van den Audenaeren, L. Loyez, A. Goeleven,
J. Wouters, and P. Ghesqui`ere. DIESEL-X: A
Game-Based Tool for Early Risk Detection of Dyslexia
in Preschoolers. In Describing and Studying
Domain-Speciﬁc Serious Games, pages 93–114.
Springer International Publishing, 2015.
 M. Huss, J. P. Verney, T. Fosker, N. Mead, and
U. Goswami. Music, rhythm, rise time perception and
developmental dyslexia: Perception of musical meter
predicts reading and phonology. Cortex,
47(6):674–689, jun 2011.
 D. J. Johnson. Persistent auditory disorders in young
dyslexic adults. Bulletin of the Orton Society,
30(1):268–276, jan 1980.
 Lexercise. Dyslexia Test - Online from Lexercise.
[Online; accessed 07-September-2016].
 Nessy. Dyslexia Screening - Nessy UK. https:
2011. [Online; accessed 07-September-2016].
 J. Nielsen. Why You Only Need to Test with 5 Users.
Jakob Nielsens Alertbox, 19(September 23):1–4, 2000.
 K. Overy. Dyslexia, Temporal Processing and Music:
The Potential of Music as an Early Learning Aid for
Dyslexic Children. Psychology of Music,
28(2):218–229, oct 2000.
 R. F. Port. Meter and speech. Journal of Phonetics,
 M. Rauschenberger. DysMusic: Detecting Dyslexia by
Web-based Games with Music Elements. In The Web
for All conference Addressing information barriers –
W4A’16, Montreal, Canada, 2016. ACM Press.
 L. Rello, M. Ballesteros, A. Ali, M. Serra, D. Alarc´on,
and J. P. Bigham. Dytective: Diagnosing risk of
dyslexia with a game. https://dytectivetest.org/, 2016.
[Online; accessed 06-January-2017].
 C. Steinbrink and T. Lachmann.
orung. Springer Berlin Heidelberg,
Berlin, Heidelberg, 2014.
 L. Van den Audenaeren, V. Celis, V. Vanden Abeele,
L. Geurts, J. Husson, P. Ghesqui`ere, J. Wouters,
L. Loyez, and A. Goeleven. DYSL-X: Design of a
tablet game for early risk detection of dyslexia in
preschoolers. In Games for Health, pages 257–266.
Springer Fachmedien Wiesbaden, Wiesbaden, 2013.
 F. R. Vellutino, J. M. Fletcher, M. J. Snowling, and
D. M. Scanlon. Speciﬁc reading disability (dyslexia):
what have we learned in the past four decades?
Journal of Child Psychology and Psychiatry,
45(1):2–40, jan 2004.