Neuropsychology

Alterations of Procedural Memory Consolidation in Children With

Developmental Dyslexia
Martina Hedenius, Jarrad A. G. Lum, and Sven Bölte
Online First Publication, November 19, 2020. http://dx.doi.org/10.1037/neu0000708

CITATION
Hedenius, M., Lum, J. A. G., & Bölte, S. (2020, November 19). Alterations of Procedural Memory Consolidation in Children
With Developmental Dyslexia. Neuropsychology. Advance online publication. http://dx.doi.org/10.1037/neu0000708
 Neuropsychology
© 2020 American Psychological Association 2020, Vol. 2, No. 999, 000
ISSN: 0894-4105 https://doi.org/10.1037/neu0000708

Alterations of Procedural Memory Consolidation in Children With

Developmental Dyslexia
Martina Hedenius1, 2, Jarrad A. G. Lum3, and Sven Bölte2, 4
1
Unit for Speech and Language Pathology, Department of Neuroscience, University of Uppsala
2
Karolinska Institutet Center of Neurodevelopmental Disorders (KIND), Centre for Psychiatry Research, Department of Women’s and
Children’s Health, Karolinska Institutet, and Stockholm Health Care Services, Stockholm County Council, CAP Research Centre,
Stockholm, Sweden
3
Cognitive Neuroscience Unit, School of Psychology, Deakin University
4
Curtin Autism Research Group, School of Occupational Therapy, Social Work and Speech Pathology, Curtin University
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Objective: The objective of the present study was to determine if consolidation of procedural
memory in developmental dyslexia (DD) may be supported by prolonging the initial learning
session. Method: Children with DD (n ⫽ 31) and typically developing (TD) children (n ⫽ 34) were
given a procedural (implicit sequence) learning task over two sessions, separated by a 24-hr interval.
Session 1 was prolonged with 50% compared with previous studies of procedural memory consol-
idation in DD. Results: The amount of sequence learning was similar in the two groups at the end
of Session 1 (p ⫽ .797, ␩␳2 ⫽ .001). At the follow-up session after 24 hr, performance in the DD
group was poorer compared with the TD group (p ⫽ .003, ␩␳2 ⫽ .129). The amount of sequence
knowledge in Session 2 predicted unique variance in reading fluency (p ⫽ .024, ␩␳2 ⫽ 0.083),
independent of children’s phoneme awareness and inattention symptoms. Conclusion: The present
study shows that consolidation of procedural memory is impaired in DD, even when the initial
learning session is prolonged. The findings encourage further research on the mechanisms support-
ing procedural memory consolidation in children with DD, and how these may be strengthened.

Key Points
Questions: Is consolidation of procedural memory in children with developmental dyslexia
normalized when the initial learning session is prolonged? Findings: The present study found
alterations of procedural memory consolidation in children with developmental dyslexia despite
a prolonged initial learning session. Importance: These findings suggest that the procedural
learning problems previously observed in children with developmental dyslexia may be related
to the overnight consolidation phase rather than to the initial learning phase. Understanding
which aspects of learning that are particularly challenging in this condition is crucial to the
development of effective intervention. Next step: The findings encourage further research on the
mechanisms supporting memory consolidation in children with developmental dyslexia, and how
these may be strengthened.

Keywords: implicit sequence learning, procedural memory, consolidation, developmental dyslexia

consultant, or lecturer for Medice and Roche. He receives royalties

Martina Hedenius X https://orcid.org/0000-0002-4305-846X for text books and diagnostic tools from Hogrefe, Kohlhammer, and
Jarrad A. G. Lum X https://orcid.org/0000-0003-2098-2403 UTB.
This work was supported by grants from the Royal Swedish Academy of Martina Hedenius served as lead for conceptualization, formal analysis,
Sciences (Lennart “Aktiestinsen” Israelssons Stipendiefond), the Promobilia Foun- funding acquisition, investigation, methodology, project administration,
dation, and the Sunnerdahl Disability Foundation. The study would not have been and writing – original draft. Sven Bölte contributed lead to supervision and
possible without the help of the many families and schools that participated. In
resources. Jarrad A. G. Lum and Sven Bölte contributed to writing – review
addition, we gratefully acknowledge the generous and helpful advice from Pro-
fessor Lars Nyberg at Umeå University, and the invaluable assistance by Ineke and editing equally.
Samson at the Speech Language Pathology clinic at Danderyd’s hospital, and Correspondence concerning this article should be addressed to Martina
Anna-Karin Arnald at ABC Logopedtjänst, in the recruitment process. Hedenius, Unit for Speech and Language Pathology, Department of Neu-
The authors declare no direct conflict of interest related to this article. roscience, Uppsala University, Box 593, 751 24, Uppsala, Sweden. Email:
Sven Bölte discloses that he has in the last 3 years acted as an author, [email protected]

1
 2 HEDENIUS, LUM, AND BÖLTE

Developmental dyslexia (DD) is characterized by persistent and fixed sequence. Sequence learning on the task is examined by
age inappropriate difficulties with accurate and/or fluent word comparing the reaction times (RTs) of responses made to trials in
decoding that cannot be explained by insufficient general intellec- which the visual stimulus follows a sequence, to control trials in
tual abilities or a lack of educational opportunities (Lyon et al., which the visual stimulus appears randomly in one of the four
2003). The condition affects between 5–10% of all school-age locations on the computer display. If knowledge about the se-
children (Shaywitz et al., 1990). A relatively large proportion of quence has been acquired, responses are faster on trials comprising
children with DD also have co-occurring problems with inattention sequence stimulus presentations compared with random presenta-
and/or hyperactivity (Russell et al., 2015), and about 20 – 40% tions. The SRT task has been used to probe both learning and
fulfill criteria for attention deficit hyperactivity disorder (ADHD; consolidation/retention of procedural knowledge. Studies examin-
e.g., McGrath et al., 2011; Willcutt et al., 2010). There is clear ing consolidation/retention of the sequence, re-present the task to
evidence linking the reading difficulties in DD to impaired pho- participants following a delay of hours, days or even years (Ro-
nological processing ability, in particular to challenges with pho- mano et al., 2010; Song et al., 2007).
neme awareness (the ability to identify and manipulate speech Brain imaging studies have shown that the neural correlates of
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

sounds; e.g., Ehri et al., 2001; Hulme et al., 2012; Vellutino et al., sequence learning are modulated by the amount of practice.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

2004). This study examined a possible link between DD and Whereas early learning, and the fast acquisition phase, is charac-
alterations of procedural memory (Nicolson & Fawcett, 2011; terized by a widespread activation pattern that includes the medial
Ullman, 2004; Ullman et al., 2020). temporal lobe, prefrontal cortex, striatum, and cerebellum, studies
The procedural memory system underlies the implicit acquisi- including later learning and consolidation suggest an increasingly
tion, consolidation, and processing of skills and habits (Henke, important role for the striatum (Doyon et al., 2009; Pinsard et al.,
2010; Mishkin et al., 1984; Squire & Zola, 1996). It is supported 2019; Rieckmann et al., 2010; Simon et al., 2012). The striatal
by a network of brain structures in which cortico-striato-cerebellar regions involved in sequence learning are reactivated during sleep,
circuits are of pivotal importance (for a review see Doyon et al., and the extent of this reactivation has been shown to correlate with
2009). This system has long been implicated in the acquisition and off-line gains in performance (i.e., consolidation) between practice
use of motor skills (e.g., Mink, 1996). However, it is now widely sessions, at least for explicit motor sequence learning (Fogel et al.,
acknowledged that procedural memory supports a range of per- 2017). Notably, consolidation of implicit sequence learning ap-
ceptual, cognitive, and linguistic functions (Ullman et al., 2020), in pears to be less dependent on sleep (e.g., Nemeth et al., 2010).
particular when the information to be processed has a sequential, Poor procedural memory could contribute to the reading prob-
and possibly statistical, structure (e.g., Karuza et al., 2013; lems in DD in at least two ways. The procedural memory system
McNealy et al., 2010; Teichmann et al., 2006). is highly adept at learning and storing sequential and statistical
A fundamental aspect of learning supported by the procedural regularities in sensory stimuli, and may be implicated in the
memory system is that it occurs over an extended time-period that acquisition and processing of phonological skills (e.g., phonotac-
may be divided into distinct stages on the basis of both behavioral tics, Ullman et al., 2020). Thus, alterations of procedural memory
characteristics and neural correlates of performance (Debas et al., could affect reading indirectly by disrupting the development of
2010; Hauptmann & Karni, 2002; Hauptmann et al., 2005; Pinsard phonological skills that are necessary for reading development.
et al., 2019). The initial stage of learning is characterized by fast Another possibility is that impaired procedural memory has a
acquisition leading to rapid improvement in performance. Next, direct effect on reading by negatively impacting on the extraction
there is a gradual decrease in the learning rate and an asymptotic of orthographic sequential regularities and/or the automatization of
trend (Hauptmann & Karni, 2002; Hauptmann et al., 2005). The skills and knowledge (e.g., sound—letter knowledge) needed to
asymptotic flattening of the learning curve has been suggested to support fluent reading (e.g., Nicolson & Fawcett, 2011). Indeed,
reflect a saturation of learning that appears to be necessary for problems with reading fluency is a particularly persistent charac-
consolidation processes to occur (Hauptmann & Karni, 2002; teristic of DD even though reading accuracy may reach normal
Hauptmann et al., 2005; Karni et al., 1998). Consolidation is the levels (Torgesen & Hudson, 2006).
process by which an initially fragile memory trace becomes more Studies investigating implicit sequence learning in DD using
robust and resistant to interference and decay (Doyon et al., 2009; variants of the SRT task have yielded inconsistent results. Al-
Robertson et al., 2004). This process may be expressed as an though the majority of published studies have found a sequence
increase in performance between learning sessions, a phenomenon learning disadvantage in DD (e.g., Howard et al., 2006; Hsu &
referred to as “off-line” learning (Hauptmann et al., 2005; Song, Bishop, 2014; Jimenez-Fernandez et al., 2011; Vicari et al., 2003),
2009). With sufficient practice, procedural skills become autom- intact performance has also been reported (Deroost et al., 2010;
atized and require little attentional resources for successful execu- Kelly et al., 2002; Russeler et al., 2006). A meta-analysis of
tion (Seger & Spiering, 2011). published studies demonstrated significant sequence learning chal-
Procedural learning has been studied extensively using the serial lenges in the SRT task in both adults and children with DD (Lum
reaction time (SRT) task (Nissen & Bullemer, 1987). In this task, et al., 2013). Overall, while some studies report procedural learn-
a visual stimulus repeatedly appears in one of four locations on a ing is impaired in DD, this result has not been universally ob-
computer display. Following stimulus onset, participants only task served.
is to press one of four buttons that matches the location of the It has been suggested that poorer procedural memory in DD can
visual stimulus’ position. The task typically comprises hundreds of be accounted for by co-occurring attention problems (West et al.,
trials in which the visual stimulus appears which in turn prompts 2020). Thus, some of the inconsistencies of previous studies could
the participants to respond. In the implicit version of this task, possibly be explained by between-studies differences in sample
participants are not told that the location stimuli appear follows a characteristics (i.e., the extent of co-occurring ADHD symptoms in
 PROCEDURAL MEMORY AND DYSLEXIA 3

the DD group). Another possibility is that some of the inconsis- children with DD. For this purpose, we employed the same alter-
tency might be explained by differences in the time interval of nating SRT (ASRT) task as in the study by Hedenius et al. (2013),
learning that is examined, that is, whether the studies have been and prolonged the initial practice session by 50%. Based on
limited to initial learning or have also included later learning previous studies showing that saturation of learning may indeed be
stages and consolidation of sequence learning (Hedenius et al., obtained with longer practice time (e.g., Hauptmann et al., 2005),
2013; Orban et al., 2008). we hypothesized that children with DD would show equivalent
To the authors best knowledge only one study has examined levels of consolidation compared to their TD peers in the current
consolidation of sequence learning on the SRT task in children study. In addition, an exploratory arm of the study examined the
with DD. In a study by Hedenius et al. (2013), children engaged in relationship between consolidation of sequential knowledge and
a SRT task with an initial practice session that comprised 1,700 reading fluency while controlling for both phoneme awareness and
trials (200 repetitions of the eight-item sequence), and lasted about co-occurring ADHD symptoms.
20 min. After a delay of 24 hr, consolidation of sequence learning
was assessed by representing a shorter version of the task. No
Method
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

differences were found between the DD and typically developing

This document is copyrighted by the American Psychological Association or one of its allied publishers.

(TD) groups with respect to initial learning in Session 1. However,

consolidation of the sequence was found to be significantly poorer Ethics Statement
in the DD group compared with the TD group. In addition, chil-
dren’s performance at the 24-hr follow-up session, examining The study conformed to the regulations of the Declaration of
consolidation of sequence learning, predicted reading ability as a Helsinki, and was approved by the ethical review board of Upp-
continuous variable. sala, Sweden. All parents or legal guardians provided informed
The study by Hedenius et al. (2013) indicates that consolidation, written consent before their child participated in the study. Addi-
rather than initial learning of procedural knowledge, may be chal- tionally, all children provided informed written assent and received
lenging in DD. This has important clinical and educational impli- four cinema tickets as compensation for their efforts.
cations since successful consolidation is key to any longer-term
knowledge and skill development. However, an outstanding issue Participants
concerns whether individuals with DD are simply slower proce-
dural learners and thus require more initial practice time in order Thirty-one children with DD and 34 TD children participated in
to reach the saturation of learning, which, in turn, is necessary for the study. The two groups did not differ with respect to age (9 to
successful consolidation (e.g., Hauptmann et al., 2005). 13 years), sex, performance IQ, hyperactivity/impulsivity, or
broader language skills as assessed by the Test for Reception of
Grammar–2 (TROG–2, Bishop, 2009). There were significant
The Present Study
group differences with respect to word reading, spelling, reading
The aim of the present study was to examine if prolonged initial fluency, and inattention symptoms (see Table 1 for the final set of
practice in Session 1 would normalize procedural consolidation in participants included in statistical analyses).

Table 1
Participant Characteristics

DD (n ⫽ 30) TD (n ⫽ 34) Comparison

Variable M SD M SD t p

Age 11.6 0.8 11.7 1.3 0.30 .765

Sex 10/20 12/22 ␹2 ⫽ 0.03 .869
PIQ 102.6 13.3 106.1 10.4 1.17 .247
Phon dec 11.0 10.5 62.0 24.0 10.72 ⬍.001
Orto read 7.7 3.9 60.8 22.5 12.79 ⬍.001
Read fluency 2.2 0.9 5.2 1.3 10.70 ⬍.001
Spelling 11.3 8.0 72.5 21.6 14.62 ⬍.001
PA 19.6 8.0 36.6 7.6 8.73 ⬍.001
TROG–2 58.9 17.9 57.4 16.6 0.36 ⬍.714
IA 0.763 0.518 0.281 0.352 4.21 ⬍.001
H/I 0.337 0.390 0.200 0.320 1.49 .142
Note. DD ⫽ children with developmental dyslexia; TD ⫽ typically developing children; PIQ ⫽ standard
scores from WISC 4 performance IQ subtests (Wechsler, 2004); Phonological decoding ⫽ percentile scores from
the nonword reading subtest from LäST (Elwér et al., 2011); Orthographic reading ⫽ percentile scores from the
word reading subtest from LäST (Elwér et al., 2011); Spelling ⫽ percentile scores from the spelling subtest from
LäST (Elwér et al., 2011); Reading fluency ⫽ stanine scores from the reading fluency subtest from DLS
(Järpsten & Taube, 2010); PA ⫽ raw scores from the Paulin Phonological Awareness test (Andersson &
Berggren, 2013); TROG–2 ⫽ percentile scores from the Test for Reception of Grammar–2 (Bishop, 2009); IA ⫽
scores from the “inattention” subdomain of the Five-To-Fifteen (FTF) questionnaire (Kadesjö et al., 2004);
H/I ⫽ scores from the hyperactivity/impulsivity subdomain of the FTF (Kadesjö et al., 2004).
 4 HEDENIUS, LUM, AND BÖLTE

Recruitment and Screening Procedure For each domain/subdomain, a mean score ranging from 0 to 2 can
be calculated (with larger numbers reflecting larger problems).
Children with DD were recruited from speech and language The “executive functions” domain was developed to closely
pathology clinics in the Stockholm-Uppsala area. Inclusion criteria mirror the ADHD criteria in the Diagnostic and Statistical Manual
were a clinical diagnosis of dyslexia from a certified speech and of Mental Disorders–IV (American Psychiatric Association, 2000).
language pathologist (SLP). Exclusion criteria for the DD group It consists of 18 statements, nine of which targets the DSM–IV
were scores ⬎15th percentile on a standardized Swedish word criteria for inattention (inattention subdomain), and nine that tar-
reading test LäSt (Elwér et al., 2011), Wechsler Intelligence Scale gets the criteria for hyperactivity/impulsivity symptoms (hyperac-
for Children (WISC–IV) performance IQ (PIQ) scores ⬍80 tivity/impulsivity subdomain). A score of 2 on at least six of the
(Wechsler, 2004), any other known comorbid neuropsychiatric nine statements in the inattention subdomain indicate clinically
condition (as reported by parents) and broader language problems significant problems related to ADHD inattentive subtype. A score
as assessed by the TROG–2 (Bishop, 2009). The cutoff for of 2 on at least six of as the nine statements in the hyperactivity/
TROG–2 was set to ⬍ percentile 20 for both groups. impulsivity subdomain indicate ADHD hyperactive—impulsive
TD children were recruited from schools in the same area.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

subtype. Children with six or more symptoms from each sub-

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Inclusion criteria were normal language, reading, and writing domain (i.e., at least 12 of the 18 statements in both subdomains)
development as reported by parents. Exclusion criteria were any are considered likely to meet diagnostic criteria for ADHD com-
known neurodevelopmental condition (as reported by parents), bined subtype. For the purpose of the present study, mean scores
WISC–IV performance IQ scores ⬍80 (Wechsler, 2004), and from the subdomains “inattention” and “hyperactivity/impulsiv-
word reading (Elwér et al., 2011), spelling (Elwér et al., 2011), or ity,” were extracted and used as a proxy for ADHD symptoms.
language comprehension (TROG–2) scores below the 20th per- Two children in the DD group met the FTF criteria for ADHD
centile. inattentive subtype and one child met the criteria for ADHD
Letters with information about the study were sent to about 400 combined subtype. Even with these children excluded, the inatten-
caregivers whose children had been assessed and diagnosed with tion scores were significantly higher in the DD group compared
DD by a certified SLP within the last year. In these letters, the with the TD group (see Table 1). By contrast, the two groups did
primary inclusion and exclusion criteria for the DD group (a not differ with respect to hyperactivity/impulsivity symptoms.
diagnosis of DD, and no other neurodevelopmental concern/diag-
nosis) were clearly stated. In addition, flyers containing the same
information were placed in the waiting rooms of SLP clinics. TD General Procedure
children were recruited by letters and flyers distributed by school
Participants were assessed by a trained experimenter, over two
teachers, and by advertisements in local newspapers. For the TD
sessions, on a battery of reading, writing, linguistic, and cognitive
group, the primary inclusion/exclusion criteria were typical devel-
tasks. The sessions were scheduled with a 24-hr intersession in-
opment with respect to reading and writing development, and no
terval, and each session lasted for about 2 hr, including breaks. The
other neurodevelopmental concern/diagnosis. A total of 144 chil-
accompanying parent completed the FTF questionnaire and re-
dren, all of which fulfilled the primary inclusion and exclusion
turned it at the second test session.
criteria for either the DD group (n ⫽ 74) or the TD group (n ⫽ 70),
were willing to participate in the study. Of these children, 43 were
excluded from the DD group, and 36 from the TD group, because Materials
they did not pass the additional exclusion criteria for the present
study (i.e., with respect to reading ability, PIQ, or broader lan- The Alternating Serial Reaction Time (ASRT) Task
guage skills). For the remaining children, possible unrecognized Sequence learning was examined using a version of the ASRT
ADHD, or subclinical ADHD symptoms, were controlled for by task (Howard & Howard, 1992; Howard et al., 2004). The ASRT
the inclusion of the Five-to-Fifteen parent questionnaire (see de- task has an important advantage over the SRT tasks used in most
scription below). previous studies, as in the ASRT task random trials are alternating
The Five-to-Fifteen (FTF) Questionnaire with trials following a fixed sequence throughout the task. The
alternating structure in this design allows for an assessment of
The Five-to-Fifteen (FTF) parent questionnaire (Kadesjö et al., sequence specific learning separate from general motor skill learn-
2004) was developed to target ADHD symptoms, as well as its ing (i.e., the overall improvement in response speed that is due to
common comorbidities, in children and adolescents between 5 and practice with the task). Moreover, it is possible to examine se-
15 years of age. The FTF has been shown to be a reliable and valid quence learning continuously throughout the task, rather than only
screening instrument (Kadesjö et al., 2004) that correlates signif- at a single point at the end of the task.
icantly with other ADHD questionnaires, as well as performance- On the ASRT task participants implicitly learn a visuospatial
based measures (Bohlin & Janols, 2004; Korkman et al., 2004). sequence. To administer the task, the participants were first seated
The FTF consists of 181 statements related to cognitive and in front of a computer display. On the display a visual stimulus
behavioral problems that can be endorsed as either “does not repeatedly appeared in one of four different horizontally arranged
apply” (0), “applies sometimes or to some extent” (1), or “defi- locations. The locations in which the visual stimulus appeared
nitely applies” (2). The 181 statements cover eight different do- were marked by four open circles (see Figure 1). The visual
mains (memory, learning, language, executive functions, motor stimulus was a picture of a dog. Participants were provided with a
skills, perception, social skills, and emotional/behavioral prob- response pad comprising four horizontally arranged buttons. Par-
lems), and most of these can be further divided into subdomains. ticipants’ task was to press the button on the response panel that
 PROCEDURAL MEMORY AND DYSLEXIA 5

Figure 1 skills. The phonemic decoding subtest assesses the ability to trans-
Computer Display for the Alternating Serial Reaction Time late letters to sounds and words, and the word recognition subtest
(ASRT) Task measures orthographic decoding skills, that is, the ability to rec-
ognize a word as fast as possible (Torgesen et al., 1999). In both
subtests, children were given 45 seconds in which to read as many
nonwords/real words as possible from the provided list. Two
equivalent forms of the test, Forms A and B, are used for both the
phonemic decoding and the word recognition subtests. That is,
participants were given two equivalent lists for phonemic decod-
Note. A picture of a dog appeared sequentially in one of the four open ing, and two equivalent lists for word recognition, each with a 45
circles, and the participant was required to respond by pressing the
second time limit. The test score for each subtest was the total sum
corresponding button on a response pad, as fast and accurately as possi-
of successfully read nonwords and words.
ble. Participants were not informed that the dog appeared according to an
eight element sequential structure. Labeling the left most circle as 1, and Reading Fluency. The ability to read with speed, accuracy,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

the right most circle 4, the sequence was 1r2r4r3r. In this structure “r” and preserved text comprehension is called reading fluency (Wag-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

denotes that the visual stimulus appeared randomly in any of the four ner, 2011). This capacity is thought to tap the degree of automa-
circles. See the online article for the color version of this figure. tization of reading skills because intact text comprehension re-
quires that sufficient attentional resources may be used to process
the content of what is read. Reading fluency was assessed with
matched the location of the visual stimulus on the screen. Specif- DLS Läshastighet (Järpsten & Taube, 2010), which is a standard-
ically, they were asked to “catch the dog” as quickly and accu- ized Swedish silent reading fluency test with Stanine norms for
rately as possible by pressing the button corresponding to the circle ages 9 to 13. On the DLS, participants read a continuous text with
in which the dog appeared (Hedenius et al., 2011). Subjects were 36 blanks where words are missing. Next to the blanks are three
instructed to use the middle and index fingers of both hands, and suggested words, and the participants are instructed to underline
responses were collected with a PST Serial Response Box. Stimuli the word that is most suitable in the context. The task is timed, and
were presented (on a LCD screen) and response times were ac- participants have 4 min for completion. This test was included to
quired with E-Prime Version 2.0. allow for an examination of the relationship between procedural
Unbeknownst to the children, the location that the visual stim- memory and reading skill as a continuous variable.
ulus appeared followed an eight-element structure comprising Phoneme Awareness. The Paulin’s Test of Phonological
fixed and random locations. Labeling the left most position on the Awareness (Andersson & Berggren, 2013) was applied. The test
computer display that the visual stimulus could appear as 1, and has a maximum score of 50 and consists of five parts: phoneme
the right most position 4, the sequence was 1r2r4r3r. In this deletion, deletion of a sequence of phonemes, reversed phoneme
structure “r” denotes that the visual stimulus appeared in a random sequences in words, reversed phoneme sequences in nonwords,
position. That is, the dog appeared in position 1 (the left-most and spoonerisms (Chard & Dickson, 1999; Kamhi & Catts, 2012).
circle), then randomly in any of the four circles, then in position 2, Language Comprehension. The Swedish version of the
and so on. On random trials, the events were sampled from a TROG–2 (Bishop, 2009) was used to estimate language compre-
uniform distribution such that the four locations were equally hension. TROG–2 is a multiple-choice sentence picture-matching
likely to appear. task. Participants listen to a target sentence read by an experi-
Each stimulus presentation and response constituted one trial. menter and are asked to identify, from a choice of four, the picture
Trials were organized into blocks of 85 trials. The first five trials that corresponds to the heard sentence. Items are presented in 20
of each block were “warm-up” trials in which the visual stimulus blocks, each with four items focusing on a particular grammatical
appeared in random positions. Data from these trials were not used structure. Participants are required to pass all four items within
in the analyses. These warm-up trials were followed by 10 repe- each block and testing is discontinued after five failed consecutive
titions of the eight-item sequence (Howard et al., 2004). Short blocks. The maximum score is 20.
breaks (around 5 to 10 min) were offered between blocks, during Performance IQ. Nonverbal IQ was assessed with the per-
which participants were encouraged to for example, have a small formance IQ subtests from the WISC–IV (Wechsler, 2004): ma-
snack or to stretch their legs. All instructions and feedback were trices, block design, and picture categories.
displayed visually on the screen as well as read aloud to the
participants. The task was self-paced, such that the correct button
had to be pressed before a new stimulus would appear on the ASRT Data Reduction
screen. However, the experimenter controlled the beginning of Following previous ASRT studies, sequence specific learning
each new block. RT was measured from target onset to the first was examined with “triplet” based analyses (for a detailed descrip-
response. The next stimulus followed correct responses after a tion, see Howard et al., 2004). All trials were categorized into
fixed 120-ms delay. either high or low frequency triplets. High frequency (HF) triplets
are those in which the trial is the third in a sequence in which the
Reading and Cognitive Tasks
first and third items are consistent with pattern (fixed) items in the
Word Reading. The phonemic decoding and sight word rec- sequence (1r2r4r3r). High frequency triplets thus consist of all
ognition efficiency subtests from LäSt (Elwér et al., 2011), which triplets ending on a pattern trial as well as those random trials that
is a Swedish version of the Test of Word Reading Efficiency by chance form a structure-consistent triplet (e.g., the four possible
(TOWRE, Torgesen et al., 1999), were used to assess word reading 1r2 triplets, namely 112, 122, 132, 142). Low frequency (LF)
 6 HEDENIUS, LUM, AND BÖLTE

triplets, by contrast, are those in which the trial is the third in a with average RT as the dependent variable, to examine consolida-
sequence in which this is not the case, and thus the triplet is tion of general skill learning in Session 2.
“structure-inconsistent” (e.g., the 16 r2r triplets). Repetitions (e.g., The relationship between procedural memory and reading abil-
111) and trills (e.g., 121) were excluded from analysis (Howard et ity as a continuous variable was first examined in a simple regres-
al., 2004). sion analysis with the normalized RT difference between HF and
The median RTs of correct responses were calculated separately LF triplets in Session 2 as the independent variable and DLS
for HF triplets and LF triplets, for each block, and for each reading fluency stanine scores as the dependent variable. Next, in
participant. The mean of these block medians was then computed order to examine if procedural learning is incrementally predictive
for each epoch (five blocks), that is, for epochs 1 through 6 in of reading ability, in addition to phoneme awareness, the normal-
Session 1, and for the single epoch in Session 2, again separately
ized RT difference between HF and LF triplets in Session 2, and
for HF and LF triplets. Sequence learning was defined as increas-
phoneme awareness scores, were entered as independent predictors
ingly faster responses to HF triplets, as compared with LF triplets,
in a general linear regression analysis, again with DLS reading
over epochs, reflecting increasing sensitivity to the statistical struc-
fluency scores as the dependent variable.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ture of the sequence. Finally, because using the simple RT differ-

In addition, a mediation analysis was conducted in which an
This document is copyrighted by the American Psychological Association or one of its allied publishers.

ence between LF and HF triplets may result in participants (and

groups) with overall longer response times falsely displaying more indirect effect of procedural memory on reading ability, with
learning, the difference between LF and HF RTs was normalized phonological awareness as a mediator, was tested directly. For this
(Faust et al., 1999; Hedenius et al., 2011). This was done by purpose, we employed the PROCESS macro for SPSS (Hayes,
dividing, for each participant, the RT difference between LF and 2013). In PROCESS, a bootstrap confidence interval for the indi-
HF triplets for each epoch by the average RT (that is, over both LF rect effect is calculated, and the effect is considered statistically
and HF triplets) for this same epoch: (LF mean RT for epoch X ⫺ significant if the confidence interval does not include zero. In the
HF mean RT for epoch X)/((LF mean RT for epoch X ⫹ HF mean present mediation analysis, model number 4 (testing for indirect
RT for epoch X)/2). Thus, sequence learning was operationalized effects with one mediator variable) was used, with 95% confidence
as an increase over epochs in the normalized RT difference be- intervals, and the number of bootstrap estimates set to 5,000. The
tween LF and HF triplets. DLS reading fluency scores was entered as the dependent variable,
the normalized RT difference between HF and LF triplets in
Session 2 was entered as the independent variable, and phonolog-
Statistical Analyses ical awareness scores was entered as the mediating variable.
Initial sequence learning (i.e., over all six epochs in Session 1) An outlier was observed in the DD group Session 2 data (3.4 SD
was examined with a mixed 2 (between-subjects; group: DD vs. from the mean), and was excluded from the parametric analyses.
TD) ⫻ 6 (within-subjects; epoch: 1 through 6) factorial design Nonparametric analyses, including the outlier, were also per-
ANOVA with the normalized RT difference between HF and LF formed and are presented at the end of the Results section.
trials as the dependent variable. Consolidation of sequence learn- Finally, in order to exclude a confounding effect of inattention
ing between Session 1 and Session 2 was examined with a mixed- symptoms, control analyses were performed in which three DD chil-
design 2 (group: DD vs. TD) ⫻ 2 (epoch: final epoch, Session 1 vs. dren with significant ADHD symptoms on the FTF were excluded.
single epoch, Session 2) ANOVA. Group differences in overall RT Because the group difference in the inattention subdomain was sig-
and/or accuracy were investigated with independent sample t tests nificant even with these three children excluded, FTF inattention
for the initial learning and consolidation sessions. scores were also included as a covariate in these control analyses. The
General skill learning, (i.e. the sequence-independent RT de- alpha-level for all statistical analyses was set to p ⬍ .05.
crease across epochs) was investigated in an analogous manner
using a 2 (between-subjects; group: DD vs. TD) ⫻ 6 (within-
Results
subjects; epoch: 1 through 6) ANOVA, with average RT across
Initial General Skill Learning
both HF and LF trials as the dependent variable, to examine
learning in Session 1; and a 2 (group: DD vs. TD) ⫻ 2 (epoch: No group differences emerged in average RT or accuracy in any of
final epoch, Session 1 vs. single epoch, Session 2) ANOVA, again the two sessions (see Table 2). In addition, there was no difference

Table 2
Reaction Times and Accuracy for Session 1 and 2

DD (n ⫽ 30) TD (n ⫽ 34) Comparison

Variable M SD M SD t p

Reaction time session 1 462 68 435 82 1.37 .174

Accuracy session 1 88.4 4.1 90.1 3.6 1.71 .092
Reaction time session 2 403 38 385 78 1.19 .235
Accuracy session 2 90.5 4.4 90.7 4.2 0.23 .819
Note. Reaction times are reported in milliseconds. Accuracy is reported in percent. DD ⫽ children with
developmental dyslexia; TD ⫽ typically developing children.
 PROCEDURAL MEMORY AND DYSLEXIA 7

between the two groups in the rates of gradual response time decrease, difference emerging between epoch 1 and epoch 6 (epoch 1: M ⫽
across both HF and LF trials, in the initial session (i.e., general skill 0.007, SD ⫽ 0.040; epoch 6: M ⫽ 0.038, SD ⫽ 0.032, F(1, 63) ⫽
learning). The 2 (group: DD vs. TD) ⫻ 6 (epoch: 1 through 6) 33.8, p ⬍ .0001, ␩␳2 ⫽ .349). No group effects were observed (main
ANOVA, with average RT as the dependent variable, revealed a main effect of group: F(1, 62) ⫽ .748, p ⫽ .390, ␩␳2 ⫽ .012; group ⫻
effect of epoch, F(5, 310) ⫽ 106.6, p ⬍ .001, with a large effect size epoch interaction, F(5, 310) ⫽ 1.04, p ⫽ .395, ␩␳2 ⫽ .016). This
(␩␳2 ⫽ .632). No group effects or interactions were observed (main indicates that both groups became increasingly sensitive to the statis-
effect of group: F(1, 62) ⫽ 1.896, p ⫽ .174, ␩␳2 ⫽ .030; Group ⫻ tical structure of the sequence over epochs, and that the amount and
Epoch interaction F(5, 310) ⫽ .687, p ⫽ .633, ␩␳2 ⫽ .011, see pattern of learning were similar in the two groups.
Figure 2).
Overnight Consolidation of Sequence Specific
Consolidation of General Skill Knowledge Knowledge

Consolidation of general skill knowledge also occurred to a The 2 (group: DD vs. TD) ⫻ 2 (epoch: final epoch, Session 1 vs.
single epoch, Session 2) ANOVA, examining consolidation of
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

similar degree in both groups (see Figure 3). The 2 (group: DD vs.
TD) ⫻ 2 (epoch: final epoch, Session 1 vs. single epoch, Session sequence specific knowledge, produced a main effect of group,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

2) ANOVA produced a large main effect of epoch, F(1, 62) ⫽ F(1, 62) ⫽ 5.12, p ⫽ .027, ␩␳2 ⫽ .076 as well as a group ⫻ epoch
17.7, p ⬍ .001, ␩␳2 ⫽ .222 but no effects of group (main effect of interaction, F(1, 62) ⫽ 4.97, p ⫽ .029, ␩␳2 ⫽ .074. There was
group, F(1, 62) ⫽ 2.14, p ⫽ .149, ␩␳2 ⬍ .033; group ⫻ epoch no main effect of epoch, F(1, 62) ⫽ .029, p ⫽ .864, ␩␳2 ⬍ .001.
interaction, F(1, 62) ⫽ .679, p ⫽ .413, ␩␳2 ⫽ .011). That is, the Follow-up analyses on the group ⫻ epoch interaction showed
general task-related performance was increased in both groups that the two groups did not differ in the amount sequence
when reassessed after a 24-hr interval. learning in the final epoch of Session 1 (DD epoch 6: M ⫽
0.0373, SD ⫽ 0.0324; TD epoch 6: M ⫽ 0.0394, SD ⫽ 0.0312,
F(1, 62) ⫽ .067, p ⫽ .797, ␩␳2 ⫽ .001). By contrast, as shown
Initial Sequence Specific Learning
in Figure 5, there was a large difference between the two groups
The sequence specific learning curves of the two groups are in the follow-up session after 24 hr (DD epoch 7: M ⫽ 0.0260,
shown in Figure 4. The 2 (group: DD vs. TD) ⫻ 6 (epoch: 1 SD ⫽ 0.0368; TD epoch 7: M ⫽ 0.0525, SD ⫽ 0.0329, F(1,
through 6) ANOVA, with the normalized RT difference between 62) ⫽ 9.21, p ⫽ .003, ␩␳2 ⫽ .129). Within-groups analyses
HF and LF triplets as the dependent variable, revealed a main revealed that whereas the TD group showed a large increase in
effect of sequence learning, over both groups (main effect of sequence knowledge between the two sessions, F(1, 62) ⫽ 5.34,
epoch: F(5, 310) ⫽ 6.14, p ⬍ .001), with a medium effect size p ⫽ .027, ␩␳2 ⫽ .139, there was a descriptive loss of knowledge
(␩␳2 ⫽ .090). Planned comparisons from the main effect of epoch in the DD group in the same period, F(1, 62) ⫽ 1.34, p ⫽ .256,
showed larger normalized RT differences between HF and LF triplets ␩␳2 ⫽ .044.
(i.e., sequence specific learning) in epochs 2, 3, 4, 5, and 6, compared In order to exclude a confounding effect of inattention symp-
to epoch 1 (all p values ⱕ .012; all ␩␳2 ⱖ .098) with the largest toms on the group difference in Session 2, a one-way ANCOVA

Figure 2
General Skill Learning in Session 1

Note. The dashed line represents children with developmental dyslexia

(DD) and the solid line represents typically developing (TD) children. The
figure displays the means and standard errors, for each epoch, of the
average (over both high and low frequency triplets) RT in milliseconds.
See the online article for the color version of this figure.
 8 HEDENIUS, LUM, AND BÖLTE

Figure 3
Overnight Consolidation of General Skill Learning Between
Sessions 1 and 2
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Note. The dashed line represents children with developmental dyslexia

(DD) and the solid line represents typically developing (TD) children. The
figure displays the means and standard errors, for the final epoch of
Session 1 and the single epoch of Session 2, of the average (over both
high and low frequency triplets) RT in milliseconds. See the online article
for the color version of this figure.

was performed, in which the three DD children with scores indic- showed that phonological awareness did not act as a mediator
ative of clinically significant problems, were excluded, and FTF for the effect of procedural memory on reading fluency (indirect
inattention scores were included as a covariate. This ANCOVA effect: b ⫽ 6.69, 95% CI [⫺1.09, 15.27]).
produced a significant effect of group, F(1, 62) ⫽ 6.35, p ⫽ .015,
␩␳2 ⫽ .105 and no effect of FTF inattention symptoms, F(1, 62) ⫽ Nonparametric Analysis Including Outlier in Session 2
0.430, p ⫽ .515, ␩␳2 ⫽ .008.
Mann–Whitney U tests of the group difference in the final epoch
in Session 1 and the single epoch in Session 2 confirmed the
Consolidation of Sequence Learning as a Predictor of pattern of results described above. Whereas the groups did not
Reading Fluency differ in the amount of sequence learning at the end of practice in
Session 1 (p ⫽ .901), the DD group showed poorer performance in
The simple regression analysis revealed a modest correlation
Session 2 (p ⫽ .016).
between sequence learning in Session 2 and children’s reading
fluency ability (r ⫽ .334, p ⫽ .007, r2 ⫽ .112). The general
Discussion
linear regression analysis, with sequence learning in Session 2
and phoneme awareness scores as predictors, produced a large The present study examined implicit learning and consolidation
effect of phoneme awareness, F(1, 60) ⫽ 59.5, p ⬍ .0001, of a visuomotor sequence in children with DD. The hypothesis
␩␳2 ⫽ 0.498, together with a medium sized effect of sequence tested in this study was that consolidation of sequential knowledge
knowledge in Session 2, F(1, 60) ⫽ 5.40, p ⬍ .024, ␩␳2 ⫽ in DD may be improved by increasing the amount of initial
0.083. The independent effect of sequence learning in Session practice in Session 1, thus allowing for saturation of learning also
2 on reading fluency held also when inattention symptoms were in children who may be slower learners. The results did not support
controlled for by excluding the three children with significant our hypothesis. Critically, despite showing equivalent levels of
ADHD symptoms, and including FTF inattention scores as a learning to their TD peers in Session 1, children with DD were
covariate, in the general linear regression analysis. This anal- significantly impaired at the 24-hr follow-up test session (Session
ysis produced a medium sized effect of sequence knowledge in 2), examining consolidation of sequence knowledge. The observed
Session 2, F(1, 60) ⫽ 4.85, p ⫽ .032, ␩␳2 ⫽ 0.085, a large consolidation problem does not appear to be explained by group
effect of phoneme awareness, F(1, 60) ⫽ 32.02, p ⬍ .001, differences in general attentional mechanisms, processing speed,
␩␳2 ⫽ 0.381 and no effect of FTF inattention scores, F(1, 60) ⫽ or motor abilities (Hari & Renvall, 2001; West et al., 2019;
0.825, p ⫽ .368, ␩␳2 ⫽ .016. There was no correlation between Willcutt et al., 2010), as no group differences were observed in
consolidation of sequence learning and phoneme awareness average RT, accuracy, or general visuomotor skill learning and
(r ⫽ .176, p ⫽ .165, r2 ⫽ .031) nor between consolidation of consolidation.
sequence learning and FTF inattention scores (r ⫽ ⫺.007, p ⫽ Our results are in line with previous studies that have shown a
.614, r2 ⫽ .005). The mediation model set up in PROCESS link between implicit sequence learning challenges and DD (e.g.,
 PROCEDURAL MEMORY AND DYSLEXIA 9

Figure 4
Sequence Specific Learning in Session 1
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Note. The dashed line represents children with developmental dyslexia

(DD) and the solid line represents typically developing (TD) children. The
figure displays the means and standard errors, for each epoch, of the
normalized RT difference between high and low frequency triplets, which
indicates sequence learning. See the online article for the color version of
this figure.

Hedenius et al., 2013; Howard et al., 2006; Hsu & Bishop, 2014). present results are not in accordance with research showing that
They extend previous findings by showing that the core procedural the link between procedural memory problems and poor reading
memory problem in DD may indeed be related to the consolida- can be explained by co-occurring difficulties with inattention
tion, rather than learning, of implicit sequential knowledge. The (West et al., 2020). In the current study, the correlation between

Figure 5
Overnight Consolidation of Sequence Specific Learning Be-
tween Sessions 1 and 2

Note. The dashed line represents children with developmental dyslexia

(DD) and the solid line represents typically developing (TD) children. The
figure displays the means and standard errors, for the final epoch of
Session 1 and the single epoch of Session 2, of the normalized RT
difference between high and low frequency triplets, which indicates
sequence learning. See the online article for the color version of this
figure.
 10 HEDENIUS, LUM, AND BÖLTE

procedural memory consolidation and reading fluency held also the development of phonological skills necessary for reading,
when ADHD symptoms were controlled for, and there was no and/or directly, by impacting on the acquisition and automatization
correlation between inattention symptoms and procedural memory. of reading-related skills and knowledge (such as sound–letter
Possible explanations for this inconsistency might include differ- knowledge and the extraction of orthographic regularities) that
ences in the ways attention capacity was assessed, and that the underlie fluent reading (e.g., Nicolson & Fawcett, 2011; Ullman,
alternating and self-paced design of the ASRT task might make it 2004; Ullman et al., 2020). Although the correlational nature of
less sensitive to the effects of lapses in attention compared with the our study precludes any claims about cause and effect, the corre-
SRT task used in the study by West et al. (2020). lation between sequence learning in Session 2, and reading flu-
Additional research is needed to understand the underlying ency, is in line with the notion of a direct effect of procedural
mechanisms of poorer memory consolidation in children with DD. memory on reading ability. By contrast, we found no support for
One possibility is that children’s performance in Session 2 was an indirect effect of procedural memory on reading, via pho-
influenced by group differences in sleep patterns. Few studies to nological skills, as the mediating effect of phonological aware-
date have investigated sleep patterns in children with DD, but
ness was nonsignificant. However, it should be noted that
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

those that have, suggest that the condition may indeed be associ-
phoneme awareness represents a type of phonological skill that
This document is copyrighted by the American Psychological Association or one of its allied publishers.

ated with sleep disturbances (Bruni et al., 2009; Carotenuto et al.,

is relatively dependent on explicit learning and processing
2016). Although sleep appears to play a limited role for consoli-
mechanisms, and it is possible that the inclusion of more
dation of the type of implicit learning studied here (e.g., Nemeth et
implicit forms of phonological skills would yield different
al., 2010), the evidence to date is not sufficient for such an effect
to be ruled out. results (Ullman et al., 2020).
Another possibility is that the observed group difference in Even though the results from the present study show a link
procedural memory consolidation stems from developmental dif- between procedural memory and reading ability, the unique
ferences in the neurobiological circuitry that underlie consolida- amount of variance in reading fluency accounted for by sequence
tion of procedural learning. Although there is substantial overlap in learning in Session 2, when controlling for phoneme awareness, is
the neural correlates of initial procedural learning and the subse- relatively small (8.5%). By contrast, the amount of variance ex-
quent consolidation of acquired procedural knowledge, they are plained by phoneme awareness was 38.1%. Therefore, the contri-
not identical. Whereas early learning has been shown to engage a bution of procedural memory consolidation to reading ability is
wide-spread network of brain structures, including the medial perhaps better understood in light of the multiple deficit model
temporal lobe, studies suggest an increasingly important role for (MDM, e.g., McGrath et al., 2020; Pennington et al., 2012) rather
the striatum in later learning stages, including memory consolida- than a model in which procedural memory is considered a core
tion (Doyon et al., 2009; Pinsard et al., 2019; Rieckmann et al., deficit in DD (e.g., Nicolson & Fawcett, 2011; Ullman, 2004;
2010; Simon et al., 2012). Ullman et al., 2020). According to the MDM, the etiology of DD
The gradual reorganization of the network supporting proce- (as well as several other neurodevelopmental conditions) is
dural learning and memory suggests that different stages in the multifactorial, and stems from the interaction of multiple cog-
learning process may be differentially sensitive to neural dysfunc- nitive challenges that probabilistically lead to reading disabil-
tion in any of the underlying brain structures. For example, the ity. In this view, a weakness in procedural memory could be one
relatively more wide-spread activation pattern of early learning of several potential risk factors, possibly shared with other
may entail that a dysfunction in anyone of the supporting struc- neurodevelopmental conditions, that influence the total risk of
tures might have less of an impact compared to later stages in developing reading problems (Hedenius, 2013; McGrath &
which a smaller set of structures appears to be involved. It has been Stoodley, 2019).
suggested that DD is characterized, not only by difficulties in The present study has limitations that may be addressed by
procedural learning, but also by intact learning in the declarative future studies. As noted above, our study did not include informa-
memory (medial temporal lobe-based) system (Ullman & Pullman,
tion about children’s sleep patterns. Such information may be
2015). In light of this hypothesis, it may be that early learning
important for the interpretation of the observed between-groups
allows for a compensatory role of the medial temporal lobe to a
difference in memory consolidation, and would be an interesting
larger degree compared to later stages involving consolidation and
extension of future work in this field. Future studies would also
automatization. Interestingly, compensatory medial temporal lobe
benefit from including larger sample sizes, a wider range of
activation during sequence learning has previously been observed
in older compared to younger adults (Rieckmann et al., 2010). phonological skills, and a longitudinal design, in order to better
Such compensatory mechanisms could possibly help explain the understand the relationship between procedural memory, phono-
observed pattern of the learning curve in the present study, as well logical development, and reading, as well as ensure the reproduc-
as in previous studies of procedural consolidation in DD and in ibility of the findings. In addition, although previous research has
developmental language disorder (Desmottes et al., 2016; Hede- linked activation of procedural memory brain structures to ASRT
nius et al., 2013; Hedenius et al., 2011). However, hypotheses learning in adults (Bo et al., 2011), there is a need for neuroim-
about the neural basis of the observed consolidation deficit in the aging data to examine brain activation related to learning and
present study are speculative and further research employing neu- consolidation of this task in children. Finally, an important avenue
roimaging methods will be required to advance knowledge of for future research may be to investigate the effects of additional
underlying brain-behavior relationships. variables that could have an effect on procedural memory consol-
As noted in the Introduction, alterations of procedural memory idation in children with DD (e.g., spaced vs. massed learning,
could influence reading development both indirectly, by affecting Gerbier & Toppino, 2015).
 PROCEDURAL MEMORY AND DYSLEXIA 11

Conclusion children learn to read: Evidence from the National Reading Panel’s
meta-analysis. Reading Research Quarterly, 36(3), 250 –287. https://doi
The present study shows that procedural memory, specifically .org/10.1598/RRQ.36.3.2
consolidation of implicit sequence knowledge, is impaired in DD, Elwér, Å., Fridolfsson, I., Samuelsson, S., & Wiklund, C. (2011). LäSt:
and that this impairment may impact on the development of Test i läsförståelse, läsning och stavning [Test of Reading Comprehen-
reading fluency. Considering the paramount importance of mem- sion, Reading, and Spelling]. Hogrefe förlag.
ory consolidation for enabling any longer term knowledge and Faust, M. E., Balota, D. A., Spieler, D. H., & Ferraro, F. R. (1999).
skill development, the present findings encourage further research Individual differences in information-processing rate and amount: Im-
on the mechanisms supporting memory consolidation in children plications for group differences in response latency. Psychological Bul-
letin, 125(6), 777–799. https://doi.org/10.1037/0033-2909.125.6.777
with DD, and how these may be strengthened.
Fogel, S., Albouy, G., King, B. R., Lungu, O., Vien, C., Bore, A., Pinsard,
B., Benali, H., Carrier, J., & Doyon, J. (2017). Reactivation or transfor-
References mation? Motor memory consolidation associated with cerebral activa-
tion time-locked to sleep spindles. PLoS ONE, 12, e0174755. https://doi
American Psychiatric Association. (2000). Diagnostic and statistical man-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

.org/10.1371/journal.pone.0174755
ual of mental disorders (4th ed., text rev.). American Psychiatric Asso-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Gerbier, E., & Toppino, T. C. (2015). The effect of distributed practice:

ciation. Neuroscience, cognition and education. Trends in Neuroscience and
Andersson, J., & Berggren, M. (2013). Paulin test av fonologisk medveten- Education, 4(3), 49 –59. https://doi.org/10.1016/j.tine.2015.01.001
het: En normering samt utvärdering av testet och faktorer som kan Hari, R., & Renvall, H. (2001). Impaired processing of rapid stimulus
påverka deltagarnas resultat [Paulin Test of Phonemic Awareness: A
sequences in dyslexia. Trends in Cognitive Sciences, 5(12), 525–532.
standardization and evaluation of the test and factors that may influence
https://doi.org/10.1016/S1364-6613(00)01801-5
results]. Magisteruppsats i logopedi, Lunds universitet, Institutionen för
Hauptmann, B., & Karni, A. (2002). From primed to learn: The saturation
logopedi och foniatri.
of repetition priming and the induction of long-term memory. Cognitive
Bishop, D. V. M. (2009). TROG-2, Test for Reception of Grammar,
Brain Research, 13(3), 313–322. https://doi.org/10.1016/S0926-
Version 2 – manual. Swedish version. Pearson Assessment.
6410(01)00124-0
Bo, J., Peltier, S. J., Noll, D. C., & Seidler, R. D. (2011). Symbolic
Hauptmann, B., Reinhart, E., Brandt, S. A., & Karni, A. (2005). The
representations in motor sequence learning. NeuroImage, 54(1), 417–
predictive value of the leveling off of within session performance for
426. https://doi.org/10.1016/j.neuroimage.2010.08.019
procedural memory consolidation. Cognitive Brain Research, 24(2),
Bohlin, G., & Janols, L. O. (2004). Behavioural problems and psychiatric
181–189. https://doi.org/10.1016/j.cogbrainres.2005.01.012
symptoms in 5–13 year-old Swedish children-a comparison of parent
Hayes, A. F. (2013). Introduction to mediation, moderation, and condi-
ratings on the FTF (Five to Fifteen) with the ratings on CBCL (Child
tional process analysis: A regression based approach. Guilford Press.
Behavior Checklist). European Child & Adolescent Psychiatry, 13(3),
Hedenius, M. (2013). Procedural and declarative memory in developmen-
14 –22. https://doi.org/10.1007/s00787-004-3003-1
tal disorders of language and literacy. (Published Doctoral Thesis,
Bruni, O., Ferri, R., Novelli, L., Terribili, M., Troianiello, M., Finotti, E.,
Uppsala University, Sweden). Retrieved from http://uu.diva-portal.org/
Leuzzi, V., & Curatolo, P. (2009). Sleep spindle activity is correlated
with reading abilities in developmental dyslexia. Sleep, 32(10), 1333– smash/get/diva2:638188/FULLTEXT01.pdf
1340. https://doi.org/10.1093/sleep/32.10.1333 Hedenius, M., Persson, J., Alm, P. A., Ullman, M. T., Howard, J. H., Jr.,
Carotenuto, M., Esposito, M., Cortese, S., Laino, D., & Verrotti, A. (2016). Howard, D. V., & Jennische, M. (2013). Impaired implicit sequence
Children with developmental dyslexia showed greater sleep disturbances learning in children with developmental dyslexia. Research in Develop-
than controls, including problems initiating and maintaining sleep. Acta mental Disabilities, 34(11), 3924 –3935. https://doi.org/10.1016/j.ridd
Paediatrica, 105(9), 1079 –1082. https://doi.org/10.1111/apa.13472 .2013.08.014
Chard, D. J., & Dickson, S. V. (1999). Phonological awareness: Instruc- Hedenius, M., Persson, J., Tremblay, A., Adi-Japha, E., Verissimo, J., Dye,
tional and assessment guidelines. Intervention in School and Clinic, C. D., Per Alma, P., Jennischea, M., Tomblin, J. B., & Ullman, M. T.
34(5), 261–270. https://doi.org/10.1177/105345129903400502 (2011). Grammar predicts procedural learning and consolidation deficits
Debas, K., Carrier, J., Orban, P., Barakat, M., Lungu, O., Vandewalle, G., in children with specific language impairment. Research in Develop-
Tahar, A. H., Bellec, P., Karni, A., Ungerleider, L. G., Habib Benal, H., mental Disabilities, 32(6), 2362–2375. https://doi.org/10.1016/j.ridd
& Doyon, J. (2010). Brain plasticity related to the consolidation of motor .2011.07.026
sequence learning and motor adaptation. Proceedings of the National Henke, K. (2010). A model for memory systems based on processing
Academy of Sciences of the United States of America, 107(41), 17839 – modes rather than consciousness. Nature Reviews Neuroscience, 11(7),
17844. https://doi.org/10.1073/pnas.1013176107 523–532. https://doi.org/10.1038/nrn2850
Deroost, N., Zeischka, P., Coomans, D., Bouazza, S., Depessemier, P., & Howard, D. V., & Howard, J. H. (1992). Adult age differences in the rate
Soetens, E. (2010). Intact first- and second-order implicit sequence of learning serial patterns: Evidence from direct and indirect tests.
learning in secondary-school-aged children with developmental dys- Psychology and Aging, 7(2), 232–241. https://doi.org/10.1037/0882-
lexia. Journal of Clinical and Experimental Neuropsychology, 32(6), 7974.7.2.232
561–572. https://doi.org/10.1080/13803390903313556 Howard, D. V., Howard, J. H., Japikse, K., DiYanni, C., Thompson, A., &
Desmottes, L., Meulemans, T., & Maillart, C. (2016). Later learning stages Somberg, R. (2004). Implicit sequence learning: Effects of level of
in procedural memory are impaired in children with specific language structure, adult age, and extended practice. Psychology and Aging, 19(1),
impairment. Research in Developmental Disabilities, 48, 53– 68. https:// 79 –92. https://doi.org/10.1037/0882-7974.19.1.79
doi.org/10.1016/j.ridd.2015.10.010 Howard, J. H., Jr., Howard, D. V., Japikse, K. C., & Eden, G. F. (2006).
Doyon, J., Bellec, P., Amsel, R., Penhune, V., Monchi, O., Carrier, J., . . . Dyslexics are impaired on implicit higher-order sequence learning, but
Benali, H. (2009). Contributions of the basal ganglia and functionally not on implicit spatial context learning. Neuropsychologia, 44(7), 1131–
related brain structures to motor learning. Behavioural Brain Research, 1144. https://doi.org/10.1016/j.neuropsychologia.2005.10.015
199(1), 61–75. https://doi.org/10.1016/j.bbr.2008.11.012 Hsu, H. J., & Bishop, D. V. (2014). Sequence-specific procedural learning
Ehri, L. C., Nunes, S. R., Willows, D. M., Schuster, B. V., Yaghoub-Zadeh, deficits in children with specific language impairment. Developmental
Z., & Shanahan, T. J. (2001). Phonemic awareness instruction helps Science, 17(3), 352–365. https://doi.org/10.1111/desc.12125
 12 HEDENIUS, LUM, AND BÖLTE

Hulme, C., Bowyer-Crane, C., Carroll, J. M., Duff, F. J., & Snowling, M. J. Mishkin, M., Malamut, B., & Bachevalier, J. (1984). Memories and habits:
(2012). The causal role of phoneme awareness and letter-sound knowl- Two neural systems. In G. Lynch, J. L. McGaugh, & N. W. Weinburger
edge in learning to read: Combining intervention studies with mediation (Eds.), Neurobiology of learning and memory (pp. 65–77). Guilford
analyses. Psychological Science, 23(6), 572–577. https://doi.org/10 Press.
.1177/0956797611435921 Nemeth, D., Janacsek, K., Londe, Z., Ullman, M. T., Howard, D. V., &
Järpsten, B., & Taube, K. (2010). DLS: För skolår 4 – 6. Handledning. Howard, J. H., Jr. (2010). Sleep has no critical role in implicit motor
Hogrefe Psykologiförlaget [Diagnostic Test of Reading and Spelling. sequence learning in young and old adults. Experimental Brain Re-
Manual]. search, 201(2), 351–358. https://doi.org/10.1007/s00221-009-2024-x
Jimenez-Fernandez, G., Vaquero, J. M., Jimenez, L., & Defior, S. (2011). Nicolson, R. I., & Fawcett, A. J. (2011). Dyslexia, dysgraphia, procedural
Dyslexic children show deficits in implicit sequence learning, but not in learning and the cerebellum. Cortex, 47(1), 117–127. https://doi.org/10
explicit sequence learning or contextual cueing. Annals of Dyslexia, .1016/j.cortex.2009.08.016
61(1), 85–110. https://doi.org/10.1007/s11881-010-0048-3 Nissen, M. J., & Bullemer, P. (1987). Attentional requirements of learning:
Kadesjö, B., Janols, L. O., Korkman, M., Mickelsson, K., Strand, G., Evidence from performance measures. Cognitive Psychology, 19(1),
Trillingsgaard, A., & Gillberg, C. (2004). The FTF (Five to Fifteen): The 1–32. https://doi.org/10.1016/0010-0285(87)90002-8
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

development of a parent questionnaire for the assessment of ADHD and Orban, P., Lungu, O., & Doyon, J. (2008). Motor sequence learning and
This document is copyrighted by the American Psychological Association or one of its allied publishers.

comorbid conditions. European Child & Adolescent Psychiatry, developmental dyslexia. Annals of the New York Academy of Sciences,
13(Suppl 3), 3–13. https://doi.org/10.1007/s00787-004-3002-2 1145(1), 151–172. https://doi.org/10.1196/annals.1416.016
Kamhi, A., & Catts, H. (2012). Language and reading disabilities. 3rd Pennington, B. F., Santerre-Lemmon, L., Rosenberg, J., MacDonald, B.,
edition. Pearson. Boada, R., Friend, A., Leopold, D. R., Samuelsson, S., Byrne, B.,
Karni, A., Meyer, G., Rey-Hipolito, C., Jezzard, P., Adams, M. M., Turner, Willcutt, E. G., & Olson, R. K. (2012). Individual prediction of dyslexia
R., & Ungerleider, L. G. (1998). The acquisition of skilled motor by single versus multiple deficit models. Journal of abnormal psychol-
performance: Fast and slow experience-driven changes in primary motor ogy, 121(1), 212–224. https://doi.org/10.1037/a0025823
cortex. Proceedings of the National Academy of Sciences of the United Pinsard, B., Boutin, A., Gabitov, E., Lungu, O., Benali, H., & Doyon, J.
States of America, 95(3), 861– 868. https://doi.org/10.1073/pnas.95.3 (2019). Consolidation alters motor sequence-specific distributed repre-
.861 sentations. eLife, 8, e39324. https://doi.org/10.7554/eLife.39324
Karuza, E. A., Newport, E. L., Aslin, R. N., Starling, S. J., Tivarus, M. E., Rieckmann, A., Fischer, H., & Backman, L. (2010). Activation in striatum
& Bavelier, D. (2013). The neural correlates of statistical learning in a and medial temporal lobe during sequence learning in younger and older
word segmentation task: An fMRI study. Brain and Language, 127(1), adults: Relations to performance. NeuroImage, 50(3), 1303–1312.
46 –54. https://doi.org/10.1016/j.bandl.2012.11.007 https://doi.org/10.1016/j.neuroimage.2010.01.015
Kelly, S. W., Griffiths, S., & Frith, U. (2002). Evidence for implicit Robertson, E. M., Pascual-Leone, A., & Miall, R. C. (2004). Current
sequence learning in dyslexia. Dyslexia, 8(1), 43–52. https://doi.org/10 concepts in procedural consolidation. Nature Reviews Neuroscience,
.1002/dys.208 5(7), 576 –582. https://doi.org/10.1038/nrn1426
Korkman, M., Jaakkola, M., Ahlroth, A., Pesonen, A. E., & Turunen, Romano, J. C., Howard, J. H., Jr., & Howard, D. V. (2010). One-year
M. M. (2004). Screening of developmental disorders in five-year-olds retention of general and sequence-specific skills in a probabilistic, serial
using the FTF (Five to Fifteen) questionnaire: a validation study. Euro- reaction time task. Memory, 18(4), 427– 441. https://doi.org/10.1080/
pean Child & Adolescent Psychiatry, 13(Suppl. 3), 31–38. https://doi 09658211003742680
.org/10.1007/s00787-004-3005-z Russeler, J., Gerth, I., & Munte, T. F. (2006). Implicit learning is intact in
Lum, J. A., Ullman, M. T., & Conti-Ramsden, G. (2013). Procedural adult developmental dyslexic readers: Evidence from the serial reaction
learning is impaired in dyslexia: Evidence from a meta-analysis of serial time task and artificial grammar learning. Journal of Clinical and
reaction time studies. Research in Developmental Disabilities, 34(10), Experimental Neuropsychology, 28(5), 808 – 827. https://doi.org/10
3460 –3476. https://doi.org/10.1016/j.ridd.2013.07.017 .1080/13803390591001007
Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A. (2003). A definition of Russell, G., Ryder, D., Norwich, B., & Ford, T. (2015). Behavioural
dyslexia. Annals of Dyslexia, 53(1), 1–14. https://doi.org/10.1007/ difficulties that co-occur with specific word reading difficulties: A U.K.
s11881-003-0001-9 population-based cohort study. Dyslexia, 21(2), 123–141. https://doi
McGrath, L. M., Pennington, B. F., Shanahan, M. A., Santerre-Lemmon, .org/10.1002/dys.1496
L. E., Barnard, H. D., Willcutt, E. G., Defries, J. C., & Olson, R. K. Seger, C. A., & Spiering, B. J. (2011). A critical review of habit learning
(2011). A multiple deficit model of reading disability and attention- and the basal ganglia. Frontiers in Systems Neuroscience, 5, 66. https://
deficit/hyperactivity disorder: searching for shared cognitive deficits. doi.org/10.3389/fnsys.2011.00066
Journal of Child Psychology and Psychiatry, and Allied Disciplines, Shaywitz, S. E., Shaywitz, B. A., Fletcher, J. M., & Escobar, M. D. (1990).
52(5), 547–557. https://doi.org/10.1111/j.1469-7610.2010.02346.x Prevalence of reading disability in boys and girls: Results of the Con-
McGrath, L. M., Peterson, R. L., & Pennington, B. F. (2020). The multiple necticut Longitudinal Study. Journal of the American Medical Associ-
deficit model: Progress, problems, and prospects. Scientific Studies of ation, 264(8), 998 –1002. https://doi.org/10.1001/jama.1990.03450
Reading, 24(1), 7–13. https://doi-org.ezproxy.its.uu.se/10.1080/ 080084036
10888438.2019.1706180 Simon, J. R., Vaidya, C. J., Howard, J. H., Jr., & Howard, D. V. (2012).
McGrath, L. M., & Stoodley, C. J. (2019). Are there shared neural corre- The effects of aging on the neural basis of implicit associative learning
lates between dyslexia and ADHD? A meta-analysis of voxel-based in a probabilistic triplets learning task. Journal of Cognitive Neurosci-
morphometry studies. Journal of Neurodevelopmental Disorders, 11(1), ence, 24(2), 451– 463. https://doi.org/10.1162/jocn_a_00116
Article 31. https://doi.org/10.1186/s11689-019-9287-8 Song, S. (2009). Consciousness and the consolidation of motor learning.
McNealy, K., Mazziotta, J. C., & Dapretto, M. (2010). The neural basis of Behavioural Brain Research, 196(2), 180 –186. https://doi.org/10.1016/
speech parsing in children and adults. Developmental Science, 13(2), j.bbr.2008.09.034
385– 406. https://doi.org/10.1111/j.1467-7687.2009.00895.x Song, S., Howard, J. H., & Howard, D. V. (2007). Sleep does not benefit
Mink, J. W. (1996). The basal ganglia: Focused selection and inhibition of probabilistic motor sequence learning. The Journal of Neuroscience,
competing motor programs. Progress in Neurobiology, 50(4), 381– 425. 27(46), 12475–12483. https://doi.org/10.1523/JNEUROSCI.2062-07
https://doi.org/10.1016/S0301-0082(96)00042-1 .2007
 PROCEDURAL MEMORY AND DYSLEXIA 13

Squire, L. R., & Zola, S. M. (1996). Structure and function of declarative four decades? Journal of Child Psychology, 45(1), 2– 40. https://doi.org/
and nondeclarative memory systems. Proceedings of the National Acad- 10.1046/j.0021-9630.2003.00305.x
emy of Sciences of the United States of America, 93(24), 13515–13522. Vicari, S., Marotta, L., Menghini, D., Molinari, M., & Petrosini, L. (2003).
https://doi.org/10.1073/pnas.93.24.13515 Implicit learning deficit in children with developmental dyslexia. Neu-
Teichmann, M., Dupoux, E., Kouider, S., & Bachoud-Levi, A. (2006). The ropsychologia, 41(1), 108 –114. https://doi.org/10.1016/S0028-
role of the striatum in processing language rules: Evidence from word 3932(02)00082-9
perception in Huntington’s disease. Journal of Cognitive Neuroscience, Wagner, R. K. (2011). Relations among oral reading fluency, silent reading
18(9), 1555–1569. https://doi.org/10.1162/jocn.2006.18.9.1555 fluency, and reading comprehension: A latent variable study of first-
Torgesen, J. K., & Hudson, R. (2006). Adolescents and adults with autism grade readers. Scientific Studies of Reading, 15(4), 338 –362. https://doi
spectrum disorders. In S. Samuels & A. Farstrup (Eds.), What research .org/10.1080/10888438.2010.493964
has to say about reading instruction (pp. 130 –158). International Read- Wechsler, D. (2004). The Wechsler Intelligence Scale for Children-fourth
edition. Pearson Assessment
ing Association.
West, G., Clayton, F., Shanks, D., & Hulme, C. (2019). Procedural and
Torgesen, J., Rashotte, C., & Wagner, R. (1999). TOWRE: Test of word
declarative learning in dyslexia. Dyslexia, 25(3), 246 –255. https://doi-
reading efficiency. Psychological Corporation Toronto.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

org.ezproxy.its.uu.se/10.1002/dys.1615
Ullman, M. T. (2004). Contributions of memory circuits to language: The
West, G., Shanks, D., & Hulme, C. (2020). sustained attention, not proce-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

declarative/procedural model. Cognition, 92(1–2), 231–270. https://doi

dural learning, is a predictor of reading, language and arithmetic skills in
.org/10.1016/j.cognition.2003.10.008 children. Scientific Studies of Reading. Advance online publication.
Ullman, M. T., Earle, F. S., Walenski, M., & Janacsek, K. (2020). The https://doi.org/10.1080/10888438.2020.1750618
neurocognition of developmental disorders of language. Annual Review Willcutt, E. G., Betjemann, R. S., McGrath, L. M., Chhabildas, N. A.,
of Psychology, 71, 389 – 417. https://doi.org/10.1146/annurev-psych- Olson, R. K., DeFries, J. C., & Pennington, B. F. (2010). Etiology and
122216-011555 neuropsychology of comorbidity between RD and ADHD: The case for
Ullman, M. T., & Pullman, M. Y. (2015). A compensatory role for multiple-deficit models. Cortex, 46(10), 1345–1361. https://doi.org/10
declarative memory in neurodevelopmental disorders. Neuroscience and .1016/j.cortex.2010.06.009
Biobehavioral Reviews, 51, 205–222. https://doi.org/10.1016/j
.neubiorev.2015.01.008 Received April 5, 2020
Vellutino, F. R., Fletcher, J. M., Snowling, M. J., & Scanlon, D. M. (2004). Revision received September 9, 2020
Specific reading disability (dyslexia): What have we learned in the past Accepted September 11, 2020 䡲