Академический Документы
Профессиональный Документы
Культура Документы
ANNUAL
REVIEWS Further Rapid Automatized Naming
Click here for quick links to
Annual Reviews content online,
including:
(RAN) and Reading Fluency:
• Other articles in this volume
• Top cited articles
Implications for
Understanding and Treatment
• Top downloaded articles
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
of Reading Disabilities
by Harvard University on 08/04/14. For personal use only.
427
PS63CH17-Norton ARI 18 November 2011 11:41
to the sounds and structure of language from term “fluency” has been used to describe the
infancy (Minagawa-Kawai et al. 2011, Peña speed and quality of oral reading, often empha-
et al. 2003). In sharp contrast, each child sizing prosody, yet this definition does not en-
must develop reading skills using brain areas compass all the goals of reading or reflect the
that have evolved for other purposes, such as fact that most of our reading is done silently
language, vision, and attention (see Dehaene rather than aloud. We conceptualize fluency in
2009). Psychologist Steven Pinker (1997) a more comprehensive way. In this review, we
famously noted that children are born “wired” examine reading fluency in the sense of what has
for language, “but print is an optional accessory been called “fluent comprehension”: a manner
that must be painstakingly bolted on.” Indeed, of reading in which all sublexical units, words,
to be a successful reader, one must rapidly inte- and connected text and all the perceptual, lin-
grate a vast circuit of brain areas with both great guistic, and cognitive processes involved in each
accuracy and remarkable speed. This “reading level are processed accurately and automatically
circuit” is composed of neural systems that so that sufficient time and resources can be al-
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
memory, attention, motor movements, and tualizing fluency based largely on rapid word
higher-level comprehension and cognition. identification. How did we arrive at this more
As our reading abilities develop, each of encompassing conceptualization of fluency, and
these components works smoothly with both why is it important that educators and re-
accuracy and speed; the reader develops what searchers view reading in this way? This multi-
is called automaticity. As a cognitive process componential view of reading is based largely
becomes automatic, it demands less conscious on our understanding of the reading circuit
effort. Although at first the child experiences a in the brain. Additional research from many
laborious and slow process to decode a simple sources, including longitudinal, intervention,
word or sentence, most adult readers can’t help and cross-linguistic studies, supports this mul-
but instantaneously, effortlessly read almost ticomponential model of reading. However,
any word they perceive. The development of many current approaches to reading instruc-
automaticity at all the lower levels of reading tion, as well as methods for identifying children
represents the great apex of development that who are having reading difficulties or for pro-
provides us with the bridge to true reading with viding intervention struggling readers, do not
its capacity to direct cognitive resources to the reflect this more comprehensive view. If our
deepest levels of thought and comprehension. goal is to have children develop fluent compre-
When a child begins learning to read, many hension, then our instruction, assessment, and
assume that to accurately decode each word intervention must reflect these ideas.
of a simple story aloud represents reading. In A closely related aspect of our study that
reality, simply to translate printed words into has contributed greatly, albeit unexpectedly, to
a stream of speech is but the beginning step, our understanding of reading fluency involves
however necessary, of reading. Indeed, even the what is called rapid automatized naming,
initial comprehension that comes next is but a or RAN (Denckla & Rudel 1976b). The
second necessary step. Essentially, one must be seemingly simple task of naming a series of
able to comprehend the meaning of a text in familiar items as quickly as possible appears
order to go beyond what is on the page: mak- to invoke a microcosm of the later developing,
ing connections to existing knowledge, analyz- more elaborated reading circuit. Our ability to
ing the writer’s argument, and predicting the understand the multicomponential structure of
next twist in the story. It is here that the way RAN, therefore, has helped us to reconceptu-
we define successful reading is important. The alize the later development of reading fluency,
not as the simple consequence of accurate difficulty learning to read despite adequate
word recognition processes, but as an equally instruction, intelligence, and effort (Lyon et al.
complex circuitry of multiple components, 2003). There is no single test and no absolute
all of which contribute to the overall reading criteria for diagnosing dyslexia. This is in
fluency and comprehension of text. part due to the fact that there are so many
To be sure, the precise relationship be- processes in reading that can break down to
tween RAN and reading continues to elude re- cause reading failure. Inaccuracy at any level of
searchers, many of whom have sought to study language or processing or a lack of automaticity
and single out individual components of RAN, in connecting any of these circuits can lead to
such as visual or phonological processes. We poor reading. More than 100 years of research
have taken a different view, in which RAN is into developmental reading difficulties has
conceptualized as a microcosm or mini-circuit yet to reveal anything resembling one single
of the later-developing reading circuitry. There explanation for all the symptoms of dyslexia,
is an extensive body of research (described in yet such pursuits continue unabated today.
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
this review) that leads us and other researchers Given the multicomponential nature of
to consider RAN tasks as one of the best, per- reading, we begin with the premise that dyslexia
haps universal, predictors of reading fluency is not a simple thing. Taking the view that
by Harvard University on 08/04/14. For personal use only.
across all known orthographies (Georgiou et al. dyslexia is a heterogeneous disorder reflecting
2008b, Tan et al. 2005). Within this view, RAN difficulty with reading due to any number of
tasks and reading are seen to require many of sources is essential for successfully identifying
the same processes, from eye saccades to work- and remediating reading disabilities in children.
ing memory to the connecting of orthographic For too many years, schools have waited for
and phonological representations. Equally im- children to “grow up a bit” so that the reading
portantly, RAN tasks depend on automaticity troubles will disappear with time, or interven-
within and across each individual component tion has been provided that was insensitive to
in the naming circuit. It is within this context the individual child’s profile of strengths and
that Eden, Perfetti, and their colleagues refer weaknesses. These two mindsets can be deeply
to RAN as one of the universal processes that detrimental because the consequences of hav-
predict the young child’s later ability to connect ing unremediated reading difficulties can be se-
and automatize whole sequences of letters and vere and life-long. Children with dyslexia not
words with their linguistic information, regard- only show poorer academic performance, but
less of writing system (Tan et al. 2005). We con- also socioemotional and behavioral effects such
sider the ability to automate both the individual as lower self-esteem and higher rates of entry
linguistic and perceptual components and the in to the juvenile justice system (Grigorenko
connections among them in visually presented 2006, Humphrey & Mullins 2002, Svensson
serial tasks the major reason why RAN consis- et al. 2001).
tently predicts later reading. Our potential to ameliorate these outcomes,
The advancement of our knowledge of both on the other hand, is significant. Research
RAN and reading fluency has led us to a point shows that accurate early identification and
where we have the capacity to make great im- appropriate targeted intervention improve
provements in our ability to identify children reading ability as well as the other poten-
with reading difficulties early on and to provide tial negative effects associated with dyslexia
appropriate, effective intervention. Many (Foorman et al. 1997, Vellutino et al. 1998).
children develop accurate decoding with basic Thus, it has become crucial to identify dyslexia
instruction and then achieve automaticity with early and to characterize the precise strengths
time and practice. However, approximately and vulnerabilities of each child individually so
10% of children in the United States have that targeted intervention can be provided to
developmental dyslexia, defined as unexpected develop accuracy and then automaticity of each
aspect of the reading system. If risk for reading difficulties, Orton developed a theory in which
difficulties can be determined very early, the inappropriate cerebral dominance accounted
chances to improve reading skills are greater. for the reversed letters and words sometimes
RAN tasks have proven of great potential be- seen in children with reading difficulties (Orton
cause children can perform RAN tasks, naming 1925). Orton made several important obser-
familiar objects or colors, well before they are vations that influence our understanding and
able to read and because RAN is correlated with treatment of dyslexia today: he noted that many
reading ability in kindergarten and beyond. of the struggling readers he saw had average or
Indeed, research on longitudinal predictors above-average intellectual abilities; that per-
of reading has repeatedly shown that RAN is haps as many as 10% of children might suffer
one of the strongest predictors of later reading from reading difficulties; and that reading diffi-
ability, and particularly for reading fluency. culties were not likely due to a single brain ab-
normality. The latter conclusion was based on
the premise that the very complexity of reading
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
This measure was based on an array of 50 col- development, but elementary school children
ored squares arranged in a grid with five rows, were faster to name letters and numbers, which
where each of five familiar colors was repeated were learned later but enjoyed a greater degree
in random order. Geschwind suggested that the of automaticity. They were thoughtful in the
deficit in color naming displayed by this patient design of these tasks, for example, including
might also be due to loss of visual-auditory con- both the letters “p” and “d,” which if not fully
nections. He further speculated that “congen- automatized were easy to confuse with their
ital dyslexia,” what we now call developmental mirror-reversed counterparts. They also kept
dyslexia, might be due to an impairment in the the design and procedure of naming left-to-
visual-auditory pathways of the brain, especially right across rows, which parallels the motoric
in the angular gyrus. Geschwind also suggested and visual processes in reading. These early
that “it is conceivable that even the age of attain- studies showed that performance on RAN tasks
ment of color naming might be a significant clue differentiated children with reading difficulties
to the age at which reading can be acquired” from typical readers of the same age and
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
(1965, p. 283). Unlike many of the other theo- from children with other, nonlanguage-based
ries of developmental dyslexia, which focused learning disabilities (Denckla & Rudel 1976a).
on the surface level, this idea suggested that In a separate line of research investigating a
by Harvard University on 08/04/14. For personal use only.
there might be a deeper, more abstract ability possible speech-motor encoding deficit in boys
that supported reading. Geschwind didn’t be- with dyslexia, Spring & Capps (1974) had also
lieve that color naming was an aspect of reading, found similar group differences in serial object,
but rather that the neural processes supporting color, and digit naming.
rapid serial color naming might be similar to
those involved in reading.
Neurologist Martha Denckla then explored Toward a Multicomponential View
the idea of a relationship between naming and of Reading and Reading Disability
reading, testing boys with reading difficulties Also in the early 1970s, notions of reading flu-
on a speeded naming task. As her mentor ency were developing in parallel. LaBerge &
Geschwind had done with patients, Denckla Samuels (1974) proposed a model of reading
used an array of 50 colored squares arranged in that was one of the first to emphasize what we
five rows. Though color-naming ability wasn’t now know as “fluency”: the idea that successful
considered to be generally impaired in chil- reading depends on not only accuracy but au-
dren with dyslexia, in studying color naming tomaticity of multiple cognitive and linguistic
in a large group of kindergarteners, Denckla processes, requiring minimal conscious effort.
(1972) discovered five boys who had dyslexia Similar ideas were presented by Perfetti (1986)
and were particularly slow and inconsistent in in his verbal efficiency theory of reading, where
serial color naming for their age, despite typical he noted that reading comprehension was asso-
intelligence and color vision. ciated with accuracy as well as speed of single-
Together with Rita Rudel, Denckla created word identification.
three other versions of the speeded serial Another more widely known line of research
naming test, using objects, letters, and num- was unfolding regarding another possible core
bers as stimuli. They coined the term “rapid deficit in dyslexia: difficulty with phonological
automatized naming” to describe these tasks awareness (PA), which involves the explicit abil-
that were designed to measure the speed of ity to identify and manipulate the sound units
naming familiar items (1976b). They found that comprise words. Isabelle Liberman pro-
that RAN latencies were not related to how moted the idea that reading development de-
early certain stimuli were learned, but instead pends on an explicit awareness of the sounds
how “automatized” the naming process was; of language and that perhaps the greatest chal-
object names were learned much earlier in lenge facing young readers is learning to match
the phonemes of speech with the graphemes ization of reading disabilities. By studying large
that represent them in print (Liberman 1971). samples of children with reading disabilities in
This work was extended to show that chil- the United States and Canada, they found that
dren with reading difficulties had trouble with phonological awareness and RAN contributed
phonological awareness (e.g., Bradley & Bryant separately to reading ability. In an attempt to
1978, Wagner & Torgesen 1987). show the importance of both sets of processes,
The field generally now agrees that PA is Wolf & Bowers (1999) proposed the double
a crucial precursor to reading acquisition in deficit hypothesis (DDH) as a way to show how
alphabetic languages and that many, if not children can be characterized in various sub-
most, children with dyslexia have PA deficits groups according to their performances on each
(Morris et al. 1998, Natl. Inst. Child Health set of processes. According to this hypothesis,
Human Dev. 2000). Though the exact nature of a deficit in either phonological awareness or
the deficit continues to be specified with some naming speed (as measured by RAN tasks) can
debate about differences across writing systems cause reading difficulties, with RAN deficits in-
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
with more regular orthographies, the fact that dicating weakness in one or more of the un-
phonological deficits can cause reading diffi- derlying fluency-related processes, not simply a
culty has been extensively researched and well naming speed deficit. In addition, these deficits
by Harvard University on 08/04/14. For personal use only.
accepted. Indeed, many of the most studied and can co-occur, and children with a double deficit
successful reading instruction and intervention in PA and RAN characterize the most severely
programs are centered around this approach. impaired readers. Wolf and Bowers developed
The fact that a deficit in phonological aware- the DDH as a first step toward a multidi-
ness can cause dyslexia does not mean, how- mensional understanding of reading difficulties,
ever, that a phonological deficit is the single intending it to promote further research and
and universal cause of dyslexia, a view espoused discussion on the variety of impairments that
by many researchers and clinicians. Many chil- can cause developmental dyslexia. Researchers
dren have difficulty reading despite intact PA around the world have taken up this challenge;
and decoding skills. As a result, those children both the DDH and the relationship between
with a reading difficulty not due to phono- rapid naming and reading have been studied
logical awareness and decoding are less likely extensively over the past decade. These studies
to be identified as having a reading disability have suggested that 60% to 75% of individu-
on traditional single-word decoding tests. Fur- als with reading or learning disabilities exhibit
ther and more importantly, they will be less RAN deficits (Katzir et al. 2008, Waber et al.
likely to benefit from standard instruction or 2004, Wolf et al. 2002).
intervention that focuses only on phonological
deficits. Beginning with research from Orton to
Geschwind, and continuing with the increas- DEFINING THE RAN TASKS
ingly expanding research from neuroimaging,
Basic Structure of RAN Tasks
we know that the reading circuit is intrinsi-
cally complex and that a lack of accuracy or Most RAN tasks appear very similar to the
automaticity at one of any number of levels original tasks developed nearly 40 years ago by
can cause reading difficulties. Any single-deficit Denckla and Rudel. These tasks have been de-
view, however important individually, is at odds scribed in the literature using slightly different
with a multicomponential conceptualization of terms, such as rapid serial naming, serial visual
reading. naming, continuous rapid naming, rapid nam-
An understanding of the complexity of the ing, and naming speed. In this review, we use
reading circuit and its multiple processes un- “RAN” to mean generally any rapid automa-
dergirds the efforts by Wolf & Bowers (1999) tized naming task or process. Essentially, a task
to move beyond a unidimensional conceptual- falls into the broader category of a RAN task
if it involves timed naming of familiar stimuli standard score and percentile rank, which
presented repeatedly in random order, in left- provides information about how the child
to-right serial fashion. In some uses of the RAN performed relative to others of the same age or
task, self-corrections and errors are noted for grade level. Self-corrections and errors can be
the purposes of qualitative observations, but the noted for qualitative interpretation but do not
key dependent variable is the total time taken factor into the scores. This is not to say that
to name the items. It is crucial that the items to these do not affect the score at all, as errors
be named, whether objects, colors, letters, or and corrections are often related to a lack of
numbers, are sufficiently familiar to the exami- fluency and, as a result, increase the time it
nee. For this reason, as in Denckla and Rudel’s takes to complete the task.
original studies, most rapid naming tests begin
with practice or pretest trials asking examinees RAN-RAS Tests. The published RAN-RAS
to name each of the stimulus items individu- Tests include the four classic subtests used in
ally to ensure that they are named accurately in Denckla and Rudel’s original RAN measures:
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
Published Standardized items each. The five different token items for
Measures of RAN each subtest are pseudorandomized, with no
The two most widely used standardized tests item appearing consecutively on the same line
of RAN in the United States are the Rapid Au- (Figure 1). Age- or grade level–based standard
tomatized Naming-Rapid Alternating Stimulus scores and percentiles are calculated based on
(RAN-RAS) Tests developed by Denckla and the total naming time (latency) for each sub-
expanded by Wolf & Denckla (2005; published test. Norms are available for individuals age 5
by Pro-Ed), and the rapid naming subtests of through 18.
the Comprehensive Test of Phonological Pro- The RAN-RAS tests are unique in their in-
cessing (CTOPP), by Wagner and colleagues clusion of rapid alternating stimulus, or RAS
(1999; published by Pro-Ed). The CTOPP subtests. The RAS was first developed in the
uses a briefer format that is considered by 1980s by Wolf as a way to incorporate pro-
its authors to measure phonological retrieval. cesses involved in switching and disengaging at-
Both of these measures are standardized and tention to rapid-naming tests (Wolf 1986). The
normed on large, nationally representative RAS is structured analogously to the RAN, with
samples in the United States and have been two or three types of items repeated alternately
used in many research studies. A child’s raw throughout the card, reflecting the demands of
score on these tests can be used to derive a shifting attention and processing between sets
of different stimuli. The RAN-RAS Tests in-
clude a two-set RAS composed of alternating
letters and numbers and a three-set RAS with
alternating letters, numbers, and colors.
CTOPP rapid naming subtests measure rapid Literacy Skills (DIBELS) contains several
object, color, digit, and letter naming. The test “fluency” subtests, including letter-naming
is normed for individuals ages 5 through 24. fluency, but this test uses all the upper and
For each subtest, there are six token items, and lowercase letters in one array and scores the
the task is divided into two parts, with the items number of letters correctly identified in one
arranged in two arrays on separate pages. Each minute, a procedure that differs significantly
of the two arrays includes 4 rows of 9 items, for from classic RAN tasks.
a total of 72 items. The examiner determines a
score by adding the total number of seconds to
complete both arrays, and this raw score can be Subcomponents of the RAN Task
used to determine age- and grade level–based Like reading, performing a RAN task requires
percentiles and standard scores. The CTOPP a synchronization and integration across a
raw scores can also be used to derive composite wide range of processes. Wolf and colleagues
scores based on multiple subtests. (Wolf & Bowers 1999, Wolf & Denckla 2005)
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
similarities, the two measures differ slightly (a) attentional processes to the stimulus;
in their format, reflecting different theoretical (b) bihemispheric visual processes responsible
viewpoints in the field about the relationship for initial feature detection, visual discrimina-
of rapid naming to other cognitive processes. tion, and pattern identification; (c) integration
The RAN-RAS tests treat rapid naming as a of visual features and pattern information with
cognitive ability that includes phonology but stored orthographic representations; (d) inte-
also other linguistic and visual processes; fur- gration of visual and orthographic informa-
thermore, the collective processes underlying tion with stored phonological representations;
RAN are conceptualized as contributing inde- (e) access and retrieval of phonological labels;
pendent variance to the prediction of reading (f ) activation and integration of semantic and
skills, particularly reading fluency. In contrast, conceptual information with all other input;
the CTOPP was designed on the basis of a and (g) motoric activation leading to articula-
model of overall phonological processing that tion. (Wolf & Denckla 2005, p. 2)
includes phonological awareness, phonological
memory, and rapid naming as related subcom- Several factors, such as the exact items to
ponents. These theoretical differences and evi- be named and the precise number of rows and
dence for a model where naming speed is sepa- columns, have varied between the many exper-
rate from phonological processes are discussed iments that have investigated RAN. Even with
below. deviations from the traditional RAN, the strong
relationship with reading seems to be preserved
Other criterion-based measures of nam- as long as the factors that underlie the theo-
ing speed. Several other psychoeducational retical link between RAN and reading are in-
assessment tests include RAN subtests, tact, including naming in a serial, left-to-right
such as the Kaufman Test of Educational fashion, and sufficient familiarity of items to be
Achievement-II, Clinical Evaluation of named. For example, an “alternate” version of
Language Fundamentals-4, and Process As- the RAN used in the Colorado Learning Dis-
sessment of the Learner; however, in most abilities Research Project contained 13 rows of
cases, the RAN measures are not fully normed, 5 items each, with some consecutively repeated
and only criterion scores are given (e.g., per- items, and in which examinees are instructed
formance is categorized only as normal versus to name as many items as possible in 15 sec-
nonnormal). The Dynamic Indicators of Basic onds. This RAN task showed relationships to
reading ability similar to those of a traditional changes in time to name each row of stimuli
RAN task and actually predicted more of the on a standard 50-item RAN task (as in Denckla
variance in reading ability than did traditional & Rudel 1976b and the published RAN-RAS
RAN in slower namers and children whose tests, Wolf & Denckla 2005). This allowed
naming ability was influenced by attention is- them to examine various factors related to initi-
sues (Compton et al. 2002). ating the task (e.g., retrieval of item names) ver-
In order to investigate which aspects of sus continuously operating processes (such as
rapid naming might drive the relationship with executive functioning or sustaining item names
reading, researchers have broken down the in working memory). They found that individ-
RAN task into component parts. At the sur- uals who were slower namers overall tended to
face level, one can consider the amount of time take longer to name subsequent rows, whereas
taken to articulate each item’s name versus the row time was more stable in faster namers. The
amount of time taken for processing between time for the first row was also slower in the
items (often called pause time). Several studies overall slower namers. Children with dyslexia
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
have found that articulation time itself is not were slower to name the first row of stimuli
strongly associated with reading in the same than were slightly younger typically develop-
manner as are overall RAN scores (Clarke et al. ing readers, suggesting that the slow naming
by Harvard University on 08/04/14. For personal use only.
2005, Cutting & Denckla 2001, Georgiou et al. times seen in dyslexia might be related to au-
2006, Neuhaus et al. 2001, Obregon 1994). In- tomaticity of retrieval or a difficulty sustaining
stead, it seems that the interitem processing or processes needed for retrieval.
pause time may reflect the components of RAN
that drive their close association with reading.
In considering pause and articulation times, RAN Differentiated
Neuhaus and colleagues (2001) found that the from Similar Tasks
two were not strongly related to each other and Single-item naming. It has been thought that
that pause time, especially on the RAN letters timed single-item naming and serial naming
task, predicted both single-word reading and would be closely related. However, the added
reading comprehension in first- and second- demands of serial naming in RAN render it
graders. Georgiou and colleagues (2006) found quite different from single-item naming. Across
that pause times at the end of kindergarten were several studies, single-item and serial naming
significantly correlated with reading accuracy have been found to be only moderately cor-
and fluency in first grade. In contrast, Clarke related, with correlation coefficients of about
and colleagues (2005) found that pause time 0.5 (see Logan et al. 2009). The added de-
was not correlated with reading single words mands associated with the continuous, serial
or nonwords, though their sample was small nature of RAN make it a better predictor of
(n = 30), and their RAN measure included reading than is single-item naming (Bowers &
many more different token items (10 digits Swanson 1991, Meyer et al. 1998). Logan
and 25 different letters) than are typically used. and colleagues showed that single-item nam-
Although these findings give us some insight as ing does explain any variance in reading beyond
to how the component parts of RAN relate to that of PA and RAN and may even be a sup-
reading, the overall RAN time is much easier pressor of serial naming (Logan et al. 2009).
to measure than pause time and shows similar Further, in their longitudinal analysis from
patterns of correlation with reading outcomes kindergarten through second grade, single-
(Georgiou et al. 2006). item naming and serial naming speed grew at
Another dimension of the RAN that has different rates as children got older, supporting
been considered in research is the differences the notion that RAN is not a simple permuta-
between each row of stimuli. Berninger and tion of single-item naming, nor are both gov-
colleagues (Amtmann et al. 2007) examined erned by an underlying system (as considered in
the global processing speed model of explaining of speed and fluency), these results underscore
RAN, below). the ideas of Wolf & Bowers (1999) that RAN
builds on the existing architecture for more
Stroop tasks. Some characteristics of RAN general speeded processing. The slow nam-
tasks bear resemblance to the classic Stroop ing speed observed in many individuals with
color-word interference task developed in the dyslexia might occur at a level higher than sim-
1930s, in which participants name the color of ple processing speed; for example, it may occur
the ink rather than the name of a printed color in the connections between visual and speech
word. The Stroop task is designed to take ad- circuits in the brain.
vantage of the relatively greater automaticity
for word reading than color naming, requiring
the examinee to inhibit reading the word and in- Independence of RAN and
stead attend to naming the color. Studies of the Phonological Awareness
Stroop task in relation to reading show several A crucial question for our understanding of
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
patterns of association similar to the RAN and reading is the relationship between RAN and
reading (MacLeod 1991). The RAN has been phonological awareness. These two constructs
studied more extensively in relation to reading, have been perhaps the most widely studied and
by Harvard University on 08/04/14. For personal use only.
however, because it removes the extra executive consistently implicated in predicting reading
function demands of the Stroop task. ability. Some controversy has existed in the
field regarding whether rapid naming should
General processing speed. Researchers in- be considered a subskill related to phonological
cluding Kail & Hall (1994) have argued that processing or whether RAN is a separate
RAN should be considered one facet of general process and should be so considered. A major
or global processing speed. Global processing argument that has been made for including
speed deficits have been associated with other RAN as a part of a larger phonological con-
developmental difficulties, including general struct is that rapid naming tasks depend on the
learning disabilities and attention deficit hy- retrieval of phonological codes (e.g., Torgesen
peractivity disorder (Willcutt et al. 2005). The et al. 1997). To subsume rapid naming tasks
majority of studies using alphanumeric RAN under phonological processing for this reason
(that is, rapid naming of letters or numbers) find alone would, however, be inaccurate. Consider
that processing speed does not account for the tests of vocabulary, where an examinee is
RAN–reading relationship (though see Catts asked to name or provide information about
et al. 2002, who found that nonalphanumeric a word. These responses require retrieval of
RAN did not account for variance in reading phonological information just as rapid naming
beyond the contribution of general process- does, yet a vocabulary task would never be
ing speed). In a large study using structural considered a subcomponent of phonology.
equation modeling, Powell et al. (2007) found At least three areas of research provide
that although children with slower RAN had evidence against considering RAN as a subset
slightly slower global processing speed than did of phonology. These notions are each reviewed
matched peers, RAN made a significant contri- in an earlier paper (Wolf et al. 2000), so we
bution to reading after processing speed was summarize previous findings focusing on more
controlled for. Similarly, Cutting & Denckla recent data that add to these discussions.
(2001) found that in a path analysis, RAN First, RAN and phonological processing are
and other reading-related skills contributed not strongly correlated. A comprehensive
to the understanding of word reading after meta-analysis of the relationship of PA and
general processing speed was controlled for. RAN confirms that these two abilities are only
Although general processing speed certainly af- moderately correlated, with an overall corre-
fects both RAN and reading (especially in terms lation coefficient of r = 0.38 (Swanson et al.
2003), and that these load on separate factors underpinnings of these abilities. Though
in an exploratory factor analysis. Based on data research has yet to directly compare RAN with
from the norming of the Comprehensive Test phonological tasks, functional brain imaging
of Phonological Processing (Wagner et al. studies of the two tasks show some shared
1999), the rapid naming components of the test regions, as would be expected with their similar
were moderately correlated with phonological task demands, yet also separate areas of process-
awareness and phonological memory, r = 0.46 ing. These studies are discussed further in the
and 0.45, respectively, for children ages 5–6; section titled Contributions of Neuroscience
r = 0.38 and 0.38 for ages 7–24. By com- and Genetics to Understanding RAN and
parison, the other aspects of phonological Fluency.
processing, PA and phonological memory,
were strongly correlated at r = 0.88 for ages
5–6 and r = 0.85 for ages 7–24. CHARACTERISTICS AND
Second, regression and structural equation PREDICTIVE VALUE OF RAN
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
2006). Models that treated RAN as a sepa- dictors of reading ability that can be assessed
rate latent variable from phonological aware- before children learn to read and thus can be
ness and memory provided a better fit to the used as early indicators of risk for reading diffi-
data, and confirmatory factor analysis studies culties. Published measures of RAN are normed
suggest that different underlying factors sup- to provide standard scores and percentile ranks
port RAN and PA (Powell et al. 2007). These for children beginning at age 5 years. Impor-
relationships may change somewhat with age; tantly, most 5-year-old children in the United
Wagner and colleagues (1997) found that RAN States are very familiar with the common ob-
contributed to the variance in reading skill af- jects and colors presented on rapid naming
ter PA was controlled for only until third grade tests, yet many are still learning the numbers
(although measures in subsequent grades in and alphabet. As a result, 5- and 6-year-olds
their longitudinal study controlled for ear- often name the color and object stimuli more
lier reading ability, which depends on RAN). quickly than letters and numbers. With more
Furthermore, RAN varies independently from practice and exposure to letters and numbers,
several potential sources of covariance with the alphanumeric stimuli become much more
phonology. In a recent review, Kirby and col- automatic. At this point, alphanumeric stimuli
leagues (2010) point out that RAN retains its re- are named faster and alphanumeric RAN be-
lationship with reading even after a host of pos- comes more strongly associated with reading
sible explanatory factors have been accounted ability (Meyer et al. 1998, Wolf et al. 1986).
for. These include verbal and nonverbal IQ, These differences underscore the importance
prior reading ability, attention deficit disor- of considering alphanumeric RAN separately
der, socioeconomic status, articulation rate, from nonalphanumeric RAN stimuli. It is also
speed of processing, phonological short-term important to consider the predictive ability of
memory, morphological awareness, and ortho- RAN across groups, as research suggests that its
graphic processing (see Kirby et al. 2010 for predictive value may be different for poor than
references). for typical readers. The study design and type
Third, genetic and neuroimaging studies of reading outcome may affect these findings, as
find different biological bases for RAN and research studies have found that RAN–reading
PA abilities. In the past decade, substantial relationships are stronger in poor than in typ-
advancements have occurred in this area, ical readers (Frijters et al. 2011, Meyer et al.
allowing us to identify the genetic and neural 1998, Scarborough 1998).
How does RAN change into adolescence shifts accordingly to accommodate different
and beyond? Published tests of RAN con- emphases in different orthographies. That said,
tain norms for people through the late teens we should be better able to understand dyslexia
by Harvard University on 08/04/14. For personal use only.
to early twenties (age 18 for the RAN-RAS when we know what types of deficits account
and age 22 for the CTOPP, described previ- for reading failure across various languages.
ously). Differences in RAN ability persist be-
tween young adults with and without dyslexia
through age 25 (Vukovic et al. 2004). Van den Shallow Orthographies
Bos and his Dutch colleagues (2002) studied Much of the research regarding reading is
how RAN changes with age and its relation- conducted in English, although English is sub-
ship with reading. Their cross-sectional study stantially different from many other languages.
included groups of Dutch children ages 8, 10, Alphabetic languages can be considered as
12, and 16, and a group of adults ages 36 to 65. falling along a continuum based on the com-
They found that the developmental trajectory plexity of the mapping between sounds and
of alphanumeric RAN reached an asymptote af- letters, or phonology and orthography. The
ter age 16 but that RAN latencies for colors and orthography of English is considered very deep
objects continued to decrease through adoles- or opaque because the correspondences from
cence and adulthood. The correlations between phonemes to graphemes are not consistent.
alphanumeric RAN and reading are also signif- On the other hand, many other alphabetic
icant through adulthood, at r = 0.53 in adults. languages such as German, Spanish, and Greek
The adults were considered a single group; it is have what is called a shallow or transparent
unclear whether there are slight differences in orthography, where grapheme-phoneme cor-
RAN or its relationship with reading associated respondences are highly predictable. As a result,
with aging. learning sound-to-letter correspondences and
decoding is more straightforward in these
orthographically shallow languages. Because
CROSS-LINGUISTIC STUDIES there are fewer rules to learn, children who
OF RAN AND FLUENCY speak these languages usually master accurate
RAN and its relationship to reading have decoding by the end of first grade (Seymour
now been studied in many of the world’s et al. 2003), whereas children learning deep
languages. This growing list includes, to our orthographies take longer at a proportion
knowledge, Arabic, Chinese, Dutch, Finnish, based on the opacity of the language.
French, German, Greek, Hebrew, Hungarian, Several recent studies have compared the
Italian, Korean, Japanese, Norwegian, Persian, effects of orthographic depth on reading
processes (Vaessen et al. 2010, Ziegler et al. readers differently depending on the orthogra-
2003). Overall, it appears that PA is important phy of their language. Cross-linguistic research
early in reading acquisition but that as children suggests that similar proportions of RAN, PA,
essentially reach ceiling in their ability to de- and double-deficits exist in other European
code words accurately, a shift occurs in which languages (Ramus et al. under review) and
the relationship between RAN and reading be- in Hebrew (Shany & Share 2011), which is
comes much stronger. The orthographic depth consistent with the double-deficit hypothesis.
of the language dictates when this shift from Again, these findings underscore that a variety
reliance on phonology to fluency-related skills of deficits can cause reading difficulties but that
occurs; children reading more transparent lan- these factors interact depending on language.
guages shift away from phonology earlier in
schooling (Vaessen et al. 2010).
A current project that has exciting poten-
tial to answer more questions in this area is Nonalphabetic Orthographies
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
the NeuroDys consortium project in Europe. Whereas English is considered a rather deep
This group is studying the longitudinal course orthography, nonalphabetic languages, such
of reading and dyslexia across six languages in as Chinese and Japanese orthographies, are
by Harvard University on 08/04/14. For personal use only.
eight countries, using a large sample of about composed of thousands of characters that are
2,000 children with and without dyslexia. The essentially unrelated or much less related to
first set of results from their research suggest phonemes. At the syllable level, Chinese words
that orthographic complexity affects the rela- share many similar syllables, with each syllable
tionship of PA and reading ability but that the represented by many different characters (Tan
relationship of RAN and reading is essentially et al. 2005). Phonological decoding plays a
consistent across languages (Ramus et al. 2011). much more minor role in reading standard
The measures that are used to define reading Chinese and Japanese, although somewhat
ability are also important. Across languages, PA more phonologically based systems (e.g.,
was a stronger predictor than RAN for untimed Chinese Pin-yin, Japanese Kana) do exist
word-reading measures, but RAN was stronger for introducing children to reading in these
than PA for timed reading. A study by Georgiou languages. As one would expect, phonological
and colleagues (2008b) corroborates these find- awareness is a weaker predictor of timed read-
ings. They studied typically developing chil- ing in Chinese; in a regression model, PA did
dren who spoke English, Greek, and Chinese not account for significant variance in timed
and found that the relationships between RAN single-word reading when RAN was controlled
and reading fluency were similar across lan- for (Tan et al. 2005). One might imagine that
guages. Similar to now extensive findings in the orthographic knowledge accounts for much of
field, they reported that the correlation of RAN the variance in Chinese reading ability because
with fluency measures was stronger than its cor- of the many characters that must be learned and
relation with reading accuracy measures. recognized. However, RAN is strongly corre-
Patterns of fluency and naming speed are lated with reading in Chinese and accounts for
also similar in poor readers across languages. additional variance after writing (orthographic)
Overall, children who are poor readers in ability is controlled for. In several cases, cor-
shallow orthographies do exhibit lower phono- relations reported between RAN and reading
logical awareness scores than those of both in Chinese and Japanese are even greater than
age-matched and younger reading-matched those reported in Swanson and colleagues’
peers (Landerl et al. 1997, Ziegler et al. 2003). (2003) meta-analysis of English (Georgiou et al.
However, there is also evidence that, as in 2008a, Kobayashi et al. 2005, Tan et al. 2005).
English, multiple deficits can cause dyslexia and This may reflect the powerful contribution of
that difficulties with PA versus RAN will affect visual processes also measured within RAN
to reading the logosyllabaries of China and IFG has been implicated in a wide variety of
Japan. reading and language-related functions, from
Underscoring the fact that some factors may semantic search to working memory. The
be language specific and others may be more temporoparietal aspect of the reading circuit
general, McBride-Chang and colleagues (2011) includes areas of posterior temporal cortex as
studied Chinese-English bilinguals who had well as the angular gyrus and supramarginal
reading difficulties in one language or in both. gyrus. These regions are classic “association
Individuals who had difficulty reading both areas” as described by Geschwind, responsible
Chinese and English were significantly slower for the integration of information across visual
namers than were peers who struggled in just and auditory modalities. The occipitotemporal
one of their languages or who were typical region includes the fusiform gyrus and inferior
readers. Furthermore, this effect was stable temporal gyrus and is most often implicated in
in children who were followed longitudinally orthographic processing.
from ages 5 through 9. Overall, the differences In people with dyslexia relative to controls,
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
in RAN across languages and orthographies are the most consistent finding is an underrecruit-
small in comparison with the many similarities. ment (hypoactivation) of left temporoparietal
We have seen no evidence of a language and left occipitotemporal areas (Maisog et al.
by Harvard University on 08/04/14. For personal use only.
in which RAN has not been shown to be 2008, Richlan et al. 2009). Functional brain dif-
important for reading. ferences in both of these areas are thought to
be related to the etiology of dyslexia, rather
than absolute level of reading ability, because
CONTRIBUTIONS OF younger children matched for ability to dyslexic
NEUROSCIENCE AND readers do not show hypoactivation of these
GENETICS TO areas (Hoeft et al. 2007). In addition to these
UNDERSTANDING RAN areas of the reading circuit that show reduced
AND FLUENCY activation in dyslexia, many individual studies
Perhaps the greatest advances in our under- have identified areas of the right frontal and
standing of reading disabilities over the past temporal lobes that show greater activation in
decade have come from neuroimaging studies. people with dyslexia relative to controls. These
Developments in magnetic resonance imaging are thought to represent compensatory mecha-
(MRI) technology have progressed such that nisms or effortful processing, as they are some-
it can be used easily with children to address times engaged in younger relative to older typ-
questions about the brain structures and ically developing readers (Hoeft et al. 2007).
associated functions involved in reading. In Several studies have reported cerebellar differ-
addition, recent genetic and twin studies have ences associated with dyslexia, but these have
produced results that give first-time insights to varied widely and were not significant in meta-
the biological mechanisms that underlie brain analyses of imaging studies (Maisog et al. 2008,
and behavioral differences in dyslexia. Richlan et al. 2009).
The tasks used in nearly all brain imaging
studies to date have focused on accuracy
Functional Brain Networks rather than fluency. One recent study to focus
in Reading and Dyslexia on fluency had typical adult readers read
Brain activation for reading-related tasks sentences presented at rates slower than, equal
has been consistently found in three main to, and faster than their normal reading speed
areas of the left hemisphere: the inferior (Benjamin & Gaab 2011). As compared to
frontal gyrus (IFG), temporoparietal area, and a letter-reading baseline task, the posterior
occipitotemporal area (see meta-analyses by middle temporal gyrus was engaged at all read-
Maisog et al. 2008, Richlan et al. 2009). The ing speeds, whereas areas of the left IFG and
occipitotemporal region were more active at several posterior areas in the occipital and
both slow and fast, but not normal, speeds. parietal regions bilaterally more than did the
These findings suggest that when the auto- group with dyslexia (shown in red, Figure 2b),
maticity of normal reading is disrupted, activa- whereas the adults with dyslexia (shown in blue)
tion in reading-related regions changes, consis- showed greater activity than did controls in a
tent with a multicomponential view of fluency. variety of bilateral temporal, motor, and left
Several important questions remain to supramarginal gyrus (part of the temporopari-
be answered, including whether readers with etal area). These results suggest that readers
different subtypes of dyslexia use different areas with dyslexia are employing a more distributed
of the brain in reading and how activation for network that may represent compensatory
timed reading might differ from untimed accu- mechanisms for performing RAN tasks.
racy measures. However, we are starting to gain
some insight into the brain processes that sup-
port RAN. There is some evidence that phono- Timing of Brain Processes
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
logical and RAN or fluency abilities may have in Reading and Dyslexia
separate neural substrates. Eden and colleagues Functional MRI studies have provided us with
(Turkeltaub et al. 2003) examined correlations a clearer picture of what happens in the brain
by Harvard University on 08/04/14. For personal use only.
between activation for an fMRI implicit reading while we read, but what do we know at this
task and behavioral measures of RAN, phono- juncture about the timing aspect of reading that
logical awareness, and working memory. They is so important for fluency? Electroencephalog-
found that patterns of correlations with brain raphy (EEG) allows us to examine the precise
activation were spatially distinct for each task, timing of neural processes, which can comple-
suggesting that each of these processes may ment information obtained about the location
tap separate aspects of the reading network. of processes determined by fMRI. EEG records
To our knowledge, the brain basis of RAN the electrical activity of the brain from the scalp,
tasks has been examined in only two studies. so researchers can present stimuli and analyze
Misra and colleagues (2004) and Christodoulou the response, called an event-related potential
and colleagues (2011, Lymberis et al. 2009) had (ERP), to each type of stimulus. From EEG re-
adults name stimuli, as in a traditional RAN task search we know that different aspects of words
(5 × 10 matrix), on a screen during fMRI scan- are processed along a timeline. For example,
ning. Both studies found that for letter naming initial visual processing occurs within the first
contrasted with fixation, the RAN task engaged 50 milliseconds after a word is presented.
the left inferior frontal gyrus, left posterior mid- Word-specific orthographic processing begins
dle frontal gyrus, and bilateral inferior occipi- around 150 msec and executive and attention
tal areas (Figure 2a). Misra et al. (2004) found processes at about 200 msec, with phonological
additional activation in left parietal and right processes between 150 and 300 msec, followed
frontal areas, although their statistical thresh- by semantic and comprehension processes
olds were much more liberal. These areas are (Wolf 2007). Ongoing debate, however, con-
consistent with areas involved in the reading cerns whether phonological processing occurs
network as well as for tasks that require eye well before other linguistic processes, perhaps
saccades. in an interactive mode with orthographic pro-
Christodoulou and colleagues (2011, cesses. As we have noted, a lack of automaticity
Lymberis et al. 2009) also compared in-scanner in any one of these areas can cause a delay that
RAN performances of typical adult readers leads to less time available for comprehension.
and adults with dyslexia who were matched on Indeed, research finds that individuals with
age and IQ. The adults with dyslexia had lower dyslexia show later peak responses for several
standardized RAN scores and lower in-scanner of these different components during word
performance. The typical controls engaged reading (see Shaul 2008 for a review). Not
surprisingly, the peak of each of the ERP scores. On the other hand, a separate set of
components involved in rapid naming was anatomical predictors related to the size and
delayed in adults with dyslexia relative to symmetry of the planum temporale (part of
controls (Breznitz 2005). the temporoparietal area implicated in success-
Because EEG systems are relatively inex- ful reading) has been related to word reading
pensive and portable as compared to MRI, EEG and comprehension. However, conflicting re-
research regarding early indicators of reading sults as to the lateralization of the asymmetry
disability is especially promising. In particular, have arisen from postmortem anatomy studies
the mismatch negativity (MMN) ERP compo- and in vivo MRI studies (Leonard et al. 2006).
nent, which is a preattentive response to a dif- It may be the case that extreme asymmetries of
ference within a series of auditory stimuli, has the planum temporale in either direction may
been studied as a possible correlate of automatic induce risk for dyslexia. Pernet and colleagues
language processing. The MMN response is (2009) also found that structural volumes either
a significant predictor of reading outcomes, much larger or smaller than those of controls
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
even better than a combination of behavioral were associated with atypical reading. In their
assessments in children (Maurer et al. 2009), sample, 100% of adults could be accurately clas-
and differs among infants with and without a sified as typical or dyslexic on the basis of the
by Harvard University on 08/04/14. For personal use only.
family history of reading disability (Leppänen volumes of the right cerebellar declive and left
et al. 2002). Recently, we found that the MMN lentiform nucleus (part of the basal ganglia).
response in children was significantly corre- Their findings also suggested that the concept
lated with RAN, timed single-word reading, of a U-shaped curve, in which extreme values
and timed connected text reading, but not with on either the high or low end can cause a dis-
PA or untimed reading (Norton et al. 2011), order, could also help explain the conflicting
suggesting that it might reflect processes im- findings of asymmetry noted above. In partic-
portant for the rapid processing of stimuli nec- ular, smaller volumes of the cerebellar declive
essary for fluent reading. Further research in were associated with more severe phonologi-
this area has great potential to help us under- cal deficits. Although the precise role of the
stand the relationship between automaticity of cerebellum relating to PA is not entirely clear,
language processing and reading fluency. notable research has implicated the lentiform
nucleus in the automaticity for automatic, se-
rial processing of language, such as is required
Brain Structure and Connectivity for rapid naming (Smits-Bandstra & De Nil
Differences in Dyslexia 2007). Although anatomical differences relat-
Researchers have also used structural MRI to ing to RAN have been less studied, a few dif-
look for an anatomical basis of reading and lan- ferences have been reported, including greater
guage disorders. In a series of studies, Leonard, rightward asymmetry of pars triangularis of left
Eckert, Berninger, and colleagues (e.g., Eckert IFG and right cerebellum associated with lower
et al. 2003, Leonard et al. 2006) have exam- RAN scores (Eckert et al. 2003).
ined the brain structure differences associated Because RAN and fluency depend on the
with RAN, single-word reading, and reading speed and integration of multiple processes
comprehension. Children with dyslexia showed throughout the brain, the extent and quality of
smaller volumes of the pars triangularis area white matter pathways may play a substantial
of the IFG bilaterally as well as an area of the role in helping us to understand the biologi-
right cerebellum. On the basis of these anatom- cal basis of fluency-related processes. A newer
ical markers, more than 80% of the subjects type of MRI scan, called diffusion tensor imag-
could be correctly classified as dyslexic or typ- ing (DTI), has allowed researchers to look at
ical readers. These anatomical measurements white matter pathways of the brain. Studies sug-
were also significantly correlated with RAN gest that white matter differences exist between
typical and dyslexic readers in reading-related measures, diagnostic criteria, and methods
regions including IFG, temporoparietal, and used, but the concordance of dyslexia is con-
occipitotemporal areas (Rimrodt et al. 2010); sistently reported to be higher in monozygotic
white matter characteristics in these areas than in dizygotic twins (Scerri & Schulte-Körne
were also correlated with speeded word-reading 2010). Several studies have examined the rela-
ability. tionship between RAN and PA and whether
Furthermore, one of the first and most they are based on shared or unique genetic
striking insights into the fluency circuits in the factors. Several researchers have reported that
brain came from research on a rare genetic there is a set of common genetic influences that
brain malformation known as periventricular affect PA, RAN, and reading (that is, they are all
nodular heterotopia (PNH), in which neurons affected by some common genes) but that there
migrate into the ventricles of the brain to form are also separate genetic influences on PA and
nodules in various areas both posterior and RAN (Byrne et al. 2005, Compton et al. 2001,
anterior. Subjects with PNH all demonstrate Petrill et al. 2006).
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
specific deficits in reading fluency despite At least nine major candidate genes for
intact IQ and single-word reading ability and susceptibility to dyslexia have been identified,
despite great diversity in where the nodules located on eight different chromosomes (Scerri
by Harvard University on 08/04/14. For personal use only.
formed across individuals (Chang et al. 2007). & Schulte-Körne 2010). Most of these are
In people with PNH, RAN letters and numbers related to neuronal migration and axon growth
were strongly correlated (r = 0.78 and 0.91, in utero (Galaburda et al. 2006). Indeed, the
respectively) with a DTI measure of white importance of neuronal migration for dyslexia
matter quality called fractional anisotropy (FA). is echoed in findings of PNH as well as in
DTI scans also revealed that white matter tracts Galaburda’s and Geschwind’s earlier studies
were disorganized around areas where nodules of postmortem brains that showed abnormal
occurred in each individual. This unique disor- migration, especially between cortical layers.
der provides further evidence that reading can It will be essential in future research to link
be disrupted at the fluency level only and that findings from structural and functional MRI,
the connectivity of various regions in the brain DTI, EEG, and genetics, to learn how biology
may play a strong part in determining fluency. and behavior interact to affect reading ability.
Researchers including Hoeft, Gaab, and
Gabrieli have begun work in this area.
Genetics of RAN and Fluency
Although researchers have long recognized that
dyslexia is heritable, the leap from genes to be- IMPLICATIONS OF RAN
havior in a process that is not genetically dic- AND FLUENCY FOR
tated (like vision or language) is likely to be IDENTIFYING READING
extraordinarily complex. Our relatively recent DIFFICULTIES, INSTRUCTION,
ability to compare genetic samples from twins AND INTERVENTION
or groups of different reading abilities and to
scan the genome for markers associated with
Identification and Assessment
behavioral variables allows us another window Although the relationships between rapid
into the processes of the reading circuit and un- naming ability and reading abilities have been
derlying causes of dyslexia. studied extensively, there remains insufficient
Heritability estimates for dyslexia range understanding of its clinical uses among
widely, from 0.3 to 0.7 (a trait that was 100% some practitioners. It is our assessment that
determined by genetics would measure 1.0). RAN tasks can be best used by educators
The precise level of heritability is difficult and psychologists as part of a clinical assess-
to ascertain because of the different reading ment to identify risk for reading and learning
difficulties and as a measure of the development prove reading fluency, and whether one can
and efficiency of processes related to word improve RAN ability, is much more difficult.
retrieval and reading fluency (Wolf & Denckla First, the RAN task itself is a surface indica-
2005). tor of the efficiency of the underlying processes
RAN tasks take only a few minutes to ad- shared by naming and reading. There have been
minister and require only modest training to no large-scale, well-controlled studies that have
administer and score. It is essential that RAN tried to explicitly train naming speed. Here, a
and other fluency measures be included in psy- gap in the literature is not a bad thing—most
choeducational assessment batteries. For early researchers would agree that training students
screening for potential reading difficulties, we on a RAN task would not be the optimal way
presented evidence from multiple longitudinal to improve their reading fluency. RAN seems
studies that show that RAN is one of the most to be related to individual developmental pro-
robust early indicators of potential reading dif- cesses; RAN times improve with age, but indi-
ficulties, along with phonological skills and viduals seem to be relatively consistent in their
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
letter name and sound knowledge. Using pub- overall naming ability across time, relative to
lished normed measures, examiners can deter- peers. In terms of assessment, we would expect
mine how a child’s RAN ability compares with to see raw scores for RAN change as children
by Harvard University on 08/04/14. For personal use only.
what is typical for a given age or grade. A second develop and become more automatic, but an in-
important reason for assessing RAN and other dividual’s standard score based on age would be
fluency issues is that speed and automaticity are more consistent. Our own studies have shown
essential components of what it means to be that although our best interventions can im-
a good reader, yet we tend to measure read- prove most reading and language variables, the
ing too often only in terms of accuracy. Myriad RAN changes little from pre- to posttreatment,
studies have shown that one can be an accu- indicating that RAN taps a more basic index of
rate reader without being a fluent reader (see processing.
Breznitz 2006). Often, children who have an How, then, do we promote reading flu-
“invisible” speed deficit are not identified until ency and provide intervention for students who
later in school, and they may start to suffer the struggle with this skill? How do we train this
negative effects of having a reading difficulty, system that seems inherently untrainable? One
such as poorer academic performance in other technique that has been widely used as a pur-
subjects. For this reason, fluency measures that ported way to improve fluency is repeated read-
take into account speed and comprehension ing. In this technique, a student reads a passage
should be included in reading assessments. multiple times, with increasing speed. After re-
peated reading, students show some generaliz-
ible increases in speed and accuracy of decoding
Interventions for Fluency (see Meyer & Felton 1999 for a review). How-
A question that naturally follows from these ever, these results and the entire approach of re-
findings is, can we train children to improve peated reading measures yield changes in speed
their RAN ability and thus impact their read- that may not be related to improvements in our
ing skills? Children with phonological weak- sine qua non of reading, fluent comprehension.
nesses who receive high-quality phonological Whereas we know that fluent comprehension
interventions tend to improve both their PA depends on accuracy and automaticity at every
skills and decoding ability (Torgesen 2004). level of language, few intervention programs
A host of well-designed, structured, multisen- reflect this. There are numerous programs de-
sory phonology programs exist, and they are signed to address phonological decoding skills,
indeed effective in remediating phonological but few programs explicitly address multiple
deficits. However, the question of how to im- components of language, such as orthography,
morphology, syntax, and semantics, with the and also on vocabulary measures, both post-
goal of improving fluent comprehension. treatment and at one-year follow-up. These
Few random-assignment treatment-control results highlight the importance of explicitly
studies examine the effects of different read- addressing the multiple levels of language
ing intervention programs. One such study, and multiple cognitive processes involved in
led by Lovett, Morris, and Wolf, examined reading.
the impact of intervention on 279 students The present review of the fluency research
with reading difficulties (Morris et al. 2011). highlights the need for multicomponential
Students were randomly assigned to one of interventions, such as PHAST, RAVE-O,
four different intervention programs designed and Language!, especially for students with
to contrast different types of instruction: RAN or double deficits whose weaknesses are
(a) study skills and math instruction (no not adequately addressed by a phonological
reading instruction), (b) PHAB + study skills, decoding program. As we better understand
a phonological program plus study skills each child’s ability, we can better tailor in-
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
instruction, (c) PHAST, a multicomponential struction to benefit each child. Children whose
word-identification strategy and phonological teachers were trained in individualizing literacy
program, or (d ) PHAB+RAVE-O, a multi- instruction (including more emphasis at the
by Harvard University on 08/04/14. For personal use only.
componential program designed to address subword and word levels versus connected
each level of reading (Wolf et al. 2009) and a text comprehension) during first grade had
phonology program. Students were matched better literacy outcomes than those of matched
for IQ, race, and socioeconomic status among classrooms without individualized instruction
groups, and each group received 70 hours of (Connor et al. 2009). Ultimately, our goal
small-group instruction. should be to understand the abilities of all
Results showed that children who received children and to provide the types of instruction
multicomponential interventions (PHAST or that best addresses their needs.
PHAB+RAVE-O) had significantly greater
growth than did other intervention groups
on timed and untimed word and nonword CONCLUSION
reading and passage comprehension. The The field of reading research has come a long
multicomponential groups also maintained way toward understanding the complex set of
these levels of growth at follow-up one year skills that allow fluent comprehension of text.
after intervention. In terms of fluency, which Research across the globe studying individuals’
is notoriously difficult to improve, children in brains and whole classrooms’ development has
the multicomponential groups again outper- shown that RAN is deeply linked with read-
formed the other interventions, with only the ing processes. Slowly but surely, the field is
RAVE-O group gaining more than six stan- moving from narrow, polarized views on the
dard score points on the Gray Oral Reading best ways to teach reading and conceptualize
Quotient (Morris et al. 2011). In sum, the two dyslexia to multicomponential frameworks for
multicomponential interventions significantly assessment and intervention. Even the long-
improved children’s reading accuracy and flu- held ideal that fluency is mostly reflected in the
ent comprehension relative to closely matched quality of prosody in oral reading is changing
programs that included phonology-only or (Kuhn et al. 2010), so that fluency is understood
general academic instruction, and RAVE-O, as the crux of when many processes at multiple
which targeted the most components, had levels integrate seamlessly to promote the com-
the best results for fluent comprehension prehension of text.
SUMMARY POINTS
1. Rapid automatized naming (RAN) measures act as a microcosm of the reading system,
providing an index of one’s abilities to integrate multiple neural processes.
2. RAN and phonological awareness are both robust early predictors of reading ability, and
one or both are often impaired in people with dyslexia. Longitudinal, cross-linguistic,
genetic, and neuroimaging studies suggest that these two crucial reading-related pro-
cesses should be considered distinct constructs rather than subcomponents of a single
construct.
3. It is advantageous to conceptualize fluent reading as a complex ability that depends on
automaticity across all levels of cognitive and linguistic processing that are involved in
reading, allowing time and thought to be devoted to comprehension.
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
FUTURE ISSUES
1. To better understand RAN and fluency as behavioral predictors and outcome measures,
based on longitudinal studies incorporating brain imaging and/or genetics (such work
is underway among the Neurodys consortium in Europe and by our colleagues Nadine
Gaab and John Gabrieli in Boston).
2. To determine the most appropriate instruction and intervention techniques for certain
profiles of readers or subtypes of dyslexia, especially those with fluency and naming
deficits who may not benefit from traditional phonologically based interventions.
3. To research how reading in new and electronic media (e.g., on the Internet or from an
e-reader) affects automaticity and fluent comprehension.
DISCLOSURE STATEMENT
The authors are unaware of any affiliation, funding, or financial holdings that might be perceived
as affecting the objectivity of this review.
LITERATURE CITED
Amtmann D, Abbott RD, Berninger V. 2007. Mixture growth models of RAN and RAS row by row: insight
into the reading system at work over time. Read. Writ. 20:785–813
Benjamin C, Gaab N. 2011. What’s the story? The tale of reading fluency told at speed. Hum. Brain Mapp. In
press
Bowers PG, Swanson LB. 1991. Naming speed deficits in reading disability: multiple measures of a singular
process. J. Exp. Child Psychol. 51:195–219
Bradley L, Bryant PE. 1978. Difficulties in auditory organisation as a possible cause of reading backwardness.
Nature 217:746–47
Breznitz Z. 2005. Brain activity during performance of naming tasks: comparison between dyslexic and regular
readers. Sci. Stud. Read. 9:17–42
versions of alphanumeric rapid automatized naming and word level reading skills. Sci. Stud. Read. 6:343–68
Connor CM, Piasta SB, Fishman B, Glasney S, Schatschneider C, et al. 2009. Individualizing student instruc-
tion precisely: effects of child × instruction interactions on first graders’ literacy development. Child Dev.
by Harvard University on 08/04/14. For personal use only.
80:77–100
Cutting LE, Denckla MB. 2001. The relationship of rapid serial naming and word reading in normally
developing readers: an exploratory model. Read. Writ. 14:673–705
Dehaene S. 2009. Reading in the Brain: The Science and Evolution of a Human Invention. New York: Viking
Denckla MB. 1972. Color naming deficits in dyslexic boys. Cortex 8:164–76
Denckla MB, Rudel RG. 1976a. Naming of objects by dyslexic and other learning disabled children. Brain
Lang. 3:1–15
Denckla MB, Rudel RG. 1976b. Rapid automatized naming (R.A.N): dyslexia differentiated from other learn-
ing disabilities. Neuropsychologia 14:471–79
Eckert MA, Leonard CM, Richards TL, Aylward EH, Thomson J, Berninger VW. 2003. Anatomical correlates
of dyslexia: frontal and cerebellar findings. Brain 126:482–94
Foorman BR, Francis DJ, Shaywitz SE, Shaywitz BA, Fletcher JM. 1997. The case for early reading interven-
tion. In Foundations of Reading Acquisition and Dyslexia: Implications for Early Intervention, ed. B Blachman,
pp. 243–64: London: Psychol. Press
Frijters JC, Lovett MW, Steinbach KA, Wolf M, Sevcik RA, Morris RD. 2011. Neurocognitive predictors of
reading outcomes for children with reading disabilities. J. Learn. Disabil. 44:150–66
Galaburda AM, LoTurco J, Ramus F, Fitch RH, Rosen GD. 2006. From genes to behavior in developmental
dyslexia. Nat. Neurosci. 9:1213–17
Georgiou GK, Parrila R, Kirby J. 2006. Rapid naming speed components and early reading acquisition.
Sci. Stud. Read. 10:199–220
Georgiou GK, Parrila R, Kirby JR, Stephenson K. 2008a. Rapid naming components and their relation-
ship with phonological awareness, orthographic knowledge, speed of processing, and different reading
outcomes. Sci. Stud. Read. 12:325–50
Georgiou GK, Parrila R, Liao CH. 2008b. Rapid naming speed and reading across languages that vary in
orthographic consistency. Read. Writ. 21:885–903
Geschwind N. 1965. Disconnexion syndromes in animals and man. Brain 27:237–94
Geschwind N, Fusillo M. 1966. Color-naming defects in association with alexia. Arch. Neurol. 15:137–46
Grigorenko EL. 2006. Learning disabilities in juvenile offenders. Child Adolesc. Psychiatr. Clin. N. Am. 15:353–
71
Hallahan DP, Mercer CD. 2002. Learning disabilities: historical perspectives. In Identification of Learn-
ing Disabilities: Research to Practice, ed. R Bradley, LC Danielson, DP Hallahan, pp. 1–67. Mahwah,
NJ: Erlbaum
Hoeft F, Meyler A, Hernandez A, Juel C, Taylor-Hill H, et al. 2007. Functional and morphometric brain
dissociation between dyslexia and reading ability. Proc. Natl. Acad. Sci. USA 104:4234–9
Humphrey N, Mullins PM. 2002. Self-concept and self-esteem in developmental dyslexia. J. Res. Spec. Educ.
Needs 2:1–13
Kail R, Hall LK. 1994. Processing speed, naming speed, and reading. Dev. Psychol. 30:949–54
Katzir T, Kim Y, Wolf M, Morris R, Lovett MW. 2008. Comparing subtypes of children with dyslexia at
letter, word, and connected text levels of reading. J. Learn. Disabil. 41:47–66
Katzir T, Kim Y, Wolf M, O’Brien B, Kennedy B, et al. 2006. Reading fluency: the whole is more than the
parts. Ann. Dyslexia 56:51–82
Kirby JR, Georgiou GK, Martinussen R, Parrila R, Bowers P, Landerl K. 2010. Naming speed and reading:
from prediction to instruction. Read. Res. Q. 45:341–62
Kobayashi MS, Haynes CW, Macaruso P, Hook PE, Kato J. 2005. Effects of mora deletion, nonword rep-
etition, rapid naming, and visual search performance on beginning reading in Japanese. Ann. Dyslexia
55:105–28
Kuhn M, Schwanenflugel P, Meisinger E. 2010. Aligning theory and assessment of reading fluency: auto-
maticity, prosody, and definitions of fluency. Read. Res. Q. 45:230–51
LaBerge D, Samuels SJ. 1974. Toward a theory of automatic information processing in reading. Cogn. Psychol.
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
6:293–323
Landerl K, Wimmer H, Frith U. 1997. The impact of orthographic consistency on dyslexia: a German-English
comparison. Cognition 63:315–34
Leonard C, Eckert M, Given B, Virginia B, Eden G. 2006. Individual differences in anatomy predict reading
by Harvard University on 08/04/14. For personal use only.
Natl. Inst. Child Health Human Dev. 2000. Report of the National Reading Panel. Teaching children to
read: an evidence-based assessment of the scientific research literature on reading and its implications for
reading instruction: reports of the subgroups. NIH Publ. No. 00-4754. Washington, DC: US Gov. Print.
Off.
Neuhaus G, Foorman BR, Francis DJ, Carlson CD. 2001. Measures of information processing in rapid
automatized naming (RAN) and their relation to reading. J. Exp. Child Psychol. 78:359–73
Norton ES, Eddy MD, Perrachione T, Cyr AB, Wolf M, et al. 2011. Mismatch negativity predicts reading fluency
in young children. Presented at Cogn. Neuro. Soc. Annu. Meet. 18th, San Francisco, CA
Obregon M. 1994. Exploring Naming Timing Patterns by Dyslexic and Normal Readers on the Serial RAN Task.
Medford, MA: Tufts Univ. Press
Orton ST. 1925. “Word-blindness” in school children. Arch. Neurol. Psychiatry 14:581–615
Orton ST. 1939. A neurological explanation of the reading disability. Educ. Rec. 20:58–68
Peña M, Maki A, Kovacić D, Dehaene-Lambertz G, Koizumi H, et al. 2003. Sounds and silence: an optical
topography study of language recognition at birth. Proc. Natl. Acad. Sci. USA 100:11702–5
Pennington BF, Lefly DL. 2001. Early reading development in children at family risk for dyslexia. Child Dev.
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
72:816–33
Perfetti CA. 1986. Continuities in reading acquisition, reading skill, and reading disability. Rem. Spec. Educ.
7:11–21
Pernet CR, Poline JB, Demonet JF, Rousselet GA. 2009. Brain classification reveals the right cerebellum as
by Harvard University on 08/04/14. For personal use only.
Swanson HL, Trainin G, Necoechea DM, Hammill DD. 2003. Rapid naming, phonological awareness, and
reading: a meta-analysis of the correlation evidence. Rev. Educ. Res. 73:407–40
Tan LH, Spinks JA, Eden GF, Perfetti CA, Siok WT. 2005. Reading depends on writing, in Chinese.
Proc. Natl. Acad. Sci. USA 102:8781–85
Torgesen JK. 2004. Avoiding the devastating downward spiral: the evidence that early intervention prevents
reading failure. Am. Educator 28:6–19
Torgesen JK, Wagner RK, Rashotte CA, Burgess S, Hecht S. 1997. Contributions of phonological awareness
and rapid naming ability to the growth of word reading skills in second-to-fifth-grade children. Sci. Stud.
Read. 1:161–85
Torppa M, Lyytinen P, Erskine J, Eklund K, Lyytinen H. 2010. Language development, literacy skills, and
predictive connections to reading in Finnish children with and without familial risk for dyslexia. J. Learn.
Disabil. 43:308–21
Turkeltaub PE, Gareau L, Flowers DL, Zeffiro TA, Eden GF. 2003. Development of neural mechanisms for
reading. Nat. Neurosci. 6:767–73
Vaessen A, Bertrand D, Tóth D, Csépe V, Faı́sca L, et al. 2010. Cognitive development of fluent word reading
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
does not qualitatively differ between transparent and opaque orthographies. J. Educ. Psychol. 102:827–42
van den Bos KP, Zijlstra BJH, Spelberg HC. 2002. Life-span data on continuous-naming speeds of numbers,
letters, colors, and pictured objects, and word-reading speed. Sci. Stud. Read. 6:25–49
Vellutino FR, Scanlon DM, Tanzman MS. 1998. The case for early intervention in diagnosing specific reading
by Harvard University on 08/04/14. For personal use only.
b
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
by Harvard University on 08/04/14. For personal use only.
Figure 2
fMRI brain activations for a RAN letters task from Christodoulou et al. (2011). (a) Whole-brain activations
for RAN letters >visual fixation. N = 18 typical adults. Activations significant at height threshold of
p < 0.05, FWE (family-wise error) corrected, k > 10 voxels. (b) Whole brain differences for RAN letters
>visual fixation in typical adult readers (N = 9) versus adults with dyslexia (N = 9). Activations significant
at height threshold of p < 0.05, FDR (false discovery rate) corrected, k > 10 voxels.
Annual Review of
Psychology
Prefatory
Working Memory: Theories, Models, and Controversies
Alan Baddeley ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 1
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
Developmental Psychobiology
Learning to See Words
Brian A. Wandell, Andreas M. Rauschecker, and Jason D. Yeatman ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !31
by Harvard University on 08/04/14. For personal use only.
Memory
Remembering in Conversations: The Social Sharing
and Reshaping of Memories
William Hirst and Gerald Echterhoff ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !55
Judgment and Decision Making
Experimental Philosophy
Joshua Knobe, Wesley Buckwalter, Shaun Nichols, Philip Robbins,
Hagop Sarkissian, and Tamler Sommers ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !81
Brain Imaging/Cognitive Neuroscience
Distributed Representations in Memory: Insights from Functional
Brain Imaging
Jesse Rissman and Anthony D. Wagner ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 101
Neuroscience of Learning
Fear Extinction as a Model for Translational Neuroscience:
Ten Years of Progress
Mohammed R. Milad and Gregory J. Quirk ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 129
Comparative Psychology
The Evolutionary Origins of Friendship
Robert M. Seyfarth and Dorothy L. Cheney ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 153
Emotional, Social, and Personality Development
Religion, Morality, Evolution
Paul Bloom ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 179
vi
PS63-FrontMatter ARI 10 November 2011 9:52
Social Neuroscience
Mechanisms of Social Cognition
Chris D. Frith and Uta Frith ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 287
by Harvard University on 08/04/14. For personal use only.
Personality Processes
Personality Processes: Mechanisms by Which Personality Traits
“Get Outside the Skin”
Sarah E. Hampson ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 315
Work Attitudes
Job Attitudes
Timothy A. Judge and John D. Kammeyer-Mueller ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 341
The Individual Experience of Unemployment
Connie R. Wanberg ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 369
Job/Work Analysis
The Rise and Fall of Job Analysis and the Future of Work Analysis
Juan I. Sanchez and Edward L. Levine ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 397
Education of Special Populations
Rapid Automatized Naming (RAN) and Reading Fluency:
Implications for Understanding and Treatment of Reading Disabilities
Elizabeth S. Norton and Maryanne Wolf ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 427
Human Abilities
Intelligence
Ian J. Deary ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 453
Research Methodology
Decoding Patterns of Human Brain Activity
Frank Tong and Michael S. Pratte ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! 483
Contents vii
PS63-FrontMatter ARI 10 November 2011 9:52
Indexes
Errata
viii Contents
ANNUAL REVIEWS
It’s about time. Your time. It’s time well spent.
Editor: Frederick P. Morgeson, The Eli Broad College of Business, Michigan State University
The Annual Review of Organizational Psychology and Organizational Behavior is devoted to publishing reviews of
the industrial and organizational psychology, human resource management, and organizational behavior literature.
Annu. Rev. Psychol. 2012.63:427-452. Downloaded from www.annualreviews.org
Topics for review include motivation, selection, teams, training and development, leadership, job performance,
and development, gender and diversity, statistics and research methodologies, and other emerging topics.
by Harvard University on 08/04/14. For personal use only.
TABLE OF CONTENTS:
An Ounce of Prevention Is Worth a Pound of Cure: Improving Perspectives on Power in Organizations, Cameron Anderson,
Research Quality Before Data Collection, Herman Aguinis, Sebastien Brion
Robert J. Vandenberg Psychological Safety: The History, Renaissance, and Future
Burnout and Work Engagement: The JD-R Approach, of an Interpersonal Construct, Amy C. Edmondson, Zhike Lei
Arnold B. Bakker, Evangelia Demerouti, Research on Workplace Creativity: A Review and Redirection,
Ana Isabel Sanz-Vergel Jing Zhou, Inga J. Hoever
Compassion at Work, Jane E. Dutton, Kristina M. Workman, Talent Management: Conceptual Approaches and Practical
Ashley E. Hardin Challenges, Peter Cappelli, JR Keller
The Contemporary Career: A Work–Home Perspective,
Dean Tjosvold, Alfred S.H. Wong, Nancy Yi Feng Chen
Coworkers Behaving Badly: The Impact of Coworker Deviant The Fascinating Psychological Microfoundations of Strategy
Behavior upon Individual Employees, Sandra L. Robinson, and Competitive Advantage, Robert E. Ployhart,
Wei Wang, Christian Kiewitz Donald Hale, Jr.
Delineating and Reviewing the Role of Newcomer Capital in The Psychology of Entrepreneurship, Michael Frese,
Organizational Socialization, Talya N. Bauer, Berrin Erdogan
Emotional Intelligence in Organizations, Stéphane Côté The Story of Why We Stay: A Review of Job Embeddedness,
Employee Voice and Silence, Elizabeth W. Morrison Thomas William Lee, Tyler C. Burch, Terence R. Mitchell
Intercultural Competence, Kwok Leung, Soon Ang, What Was, What Is, and What May Be in OP/OB,
Mei Ling Tan Lyman W. Porter, Benjamin Schneider
Learning in the Twenty-First-Century Workplace, Where Global and Virtual Meet: The Value of Examining
Raymond A. Noe, Alena D.M. Clarke, Howard J. Klein the Intersection of These Elements in Twenty-First-Century
Pay Dispersion, Jason D. Shaw Teams, Cristina B. Gibson, Laura Huang, Bradley L. Kirkman,
Personality and Cognitive Ability as Predictors of Effective Debra L. Shapiro
Performance at Work, Neal Schmitt Work–Family Boundary Dynamics, Tammy D. Allen,
Eunae Cho, Laurenz L. Meier
Access this and all other Annual Reviews journals via your institution at www.annualreviews.org.
well as all scientists and users of statistics about major methodological advances and the computational tools that
by Harvard University on 08/04/14. For personal use only.
and bioinformatics, economics, machine learning, psychology, sociology, and aspects of the physical sciences.
TABLE OF CONTENTS:
What Is Statistics? Stephen E. Fienberg High-Dimensional Statistics with a View Toward Applications
A Systematic Statistical Approach to Evaluating Evidence in Biology, Peter Bühlmann, Markus Kalisch, Lukas Meier
from Observational Studies, David Madigan, Paul E. Stang, Next-Generation Statistical Genetics: Modeling, Penalization,
Jesse A. Berlin, Martijn Schuemie, J. Marc Overhage, and Optimization in High-Dimensional Data, Kenneth Lange,
Marc A. Suchard, Bill Dumouchel, Abraham G. Hartzema, Jeanette C. Papp, Janet S. Sinsheimer, Eric M. Sobel
Patrick B. Ryan Breaking Bad: Two Decades of Life-Course Data Analysis
The Role of Statistics in the Discovery of a Higgs Boson, in Criminology, Developmental Psychology, and Beyond,
David A. van Dyk Elena A. Erosheva, Ross L. Matsueda, Donatello Telesca
Brain Imaging Analysis, F. DuBois Bowman Event History Analysis, Niels Keiding
Statistics and Climate, Peter Guttorp
Climate Simulators and Climate Projections, Christopher D. Steele, David J. Balding
Jonathan Rougier, Michael Goldstein Using League Table Rankings in Public Policy Formation:
Probabilistic Forecasting, Tilmann Gneiting, Statistical Issues, Harvey Goldstein
Matthias Katzfuss Statistical Ecology, Ruth King
Bayesian Computational Tools, Christian P. Robert Estimating the Number of Species in Microbial Diversity
Bayesian Computation Via Markov Chain Monte Carlo, Studies, John Bunge, Amy Willis, Fiona Walsh
Dynamic Treatment Regimes, Bibhas Chakraborty,
Build, Compute, Critique, Repeat: Data Analysis with Latent Susan A. Murphy
Variable Models, David M. Blei Statistics and Related Topics in Single-Molecule Biophysics,
Structured Regularizers for High-Dimensional Problems: Hong Qian, S.C. Kou
Statistical and Computational Issues, Martin J. Wainwright Statistics and Quantitative Risk Management for Banking
and Insurance, Paul Embrechts, Marius Hofert
Access this and all other Annual Reviews journals via your institution at www.annualreviews.org.