Brain Struct Funct DOI 10.1007/s00429-013-0666-8

ORIGINAL ARTICLE

Investigating the ventral-lexical, dorsal-sublexical model of basic reading processes using diffusion tensor imaging Jacqueline Cummine • Wenjun Dai • Ron Borowsky Layla Gould • Claire Rollans • Carol Boliek

•

Received: 10 May 2013 / Accepted: 21 October 2013 Ó Springer-Verlag Berlin Heidelberg 2013

Abstract Recent results from diffusion tensor imaging (DTI) studies provide evidence of a ventral-lexical stream and a dorsal-sublexical stream associated with reading processing. We investigated the relationship between behavioural reading speed for stimuli thought to rely on either the ventral-lexical, dorsal-sublexical, or both streams and white matter via fractional anisotropy (FA) and mean diffusivity (MD) using DTI tractography. Participants (N = 32) overtly named exception words (e.g., ‘one’, ventral-lexical), regular words (e.g., ‘won’, both streams), nonwords (‘wum’, dorsal-sublexical) and pseudohomophones (‘wun’, dorsal-sublexical) in a behavioural lab. Each participant then underwent a brain scan that included a 30-directional DTI sequence. Tractography was used to extract FA and MD values from four tracts of interest: inferior longitudinal fasciculus, uncinate fasciculus, arcuate fasciculus, and inferior fronto-occipital fasciculus. Median reaction times (RTs) for reading exception words and regular words both showed a significant correlation with the FA of the uncinate fasciculus thought to underlie J. Cummine (&)
W. Dai
C. Boliek Department of Speech Pathology and Audiology, Faculty of Rehabilitation Medicine, University of Alberta, 2-70 8205 114st, Edmonton, AB T6G 2G4, Canada e-mail: [email protected] J. Cummine
C. Boliek Centre for Neuroscience, University of Alberta, Edmonton, Canada R. Borowsky
L. Gould Department of Psychology, University of Saskatchewan, Saskatoon, Canada C. Rollans Department of Cognitive Science, McGill University, Montreal, Canada

the ventral processing stream, such that response time decreased as FA increased. In addition, RT for exception and regular words showed a relationship with MD of the uncinate fasciculus, such that response time increased as MD increased. Multiple regression analyses revealed that exception word RT accounted for unique variability in FA of the uncinate over and above regular words. There were no robust relationships found between pseudohomophones, or nonwords, and tracts thought to underlie the dorsal processing stream. These results support the notion that word recognition, in general, and exception word reading in particular, rely on ventral-lexical brain regions. Keywords Diffusion tensor imaging
Dorsalsublexical processes
Ventral-lexical processes
Reading
Tractography

Introduction Today’s society is highly dependent on written communication. As such, impairments in reading could have a major impact on daily interactions and the quality of life. Welldeveloped models of reading are important for furthering our understanding of reading processes. While there has been a good deal of research investigating the functional framework of basic reading (for a review see Price 2012; Cummine et al. 2010, 2013; Borowsky et al. 2012; Hickok and Poeppel 2004), there have been relatively few studies exploring the structural framework (for a review see Richardson and Price 2009; Vandermosten et al. 2012). Understanding the relationships between structural measurements of the brain and reading processes will aid in the development of more comprehensive models of reading and reading impairment.

123

Brain Struct Funct

Much of what we know about current models of reading, and the lexical/sublexical distinction, is based on word type effects while individuals read aloud or make button responses to individual monosyllabic letter strings (e.g., the regularity effect; Besner and Smith 1992; Cummine et al. 2013; Gould et al. 2012; see also Balota et al. 2004 for a review of speeded visual word recognition). Reading of different word types is thought to rely on different processing streams (Coltheart et al. 2001). As such, a systematic manipulation of letter strings representing lexical and sublexical processing would be particularly valuable to advance our understanding of dorsal-sublexical and ventral-lexical processing models. For example, exception words (e.g., one) could not be pronounced using sublexical analysis because the pronunciations of these words do not follow the regular letter-to-sound correspondences found in English. Thus, reading of exception words necessitates orthographic lexically-based reading using the ventral processing stream, where the orthographic lexical representation of the whole-word is used to retrieve the phonological lexical representation from memory. Reading of regular words, which possess typical grapheme-to-phoneme correspondences, likely involves the activation of both processing streams depending on the frequency of their occurrence. More frequently occurring regular words (e.g., won) would more likely be stored in orthographic lexical memory due to their high frequency of occurrence and accessed through the ventral stream. Conversely, regular words that occur infrequently (e.g., wisp), or are new to the reader, would likely be analysed sublexically by the dorsal stream because they are less likely to be stored in the memory-based orthographic ventral-lexical processing stream. Nonwords (e.g., wum) have familiar graphemic and phonemic representations while lacking whole-word lexical representations and thus require the sublexical analysis thought to be performed by the dorsal processing stream, as they are unfamiliar letter strings that need to be analytically sounded out. Like nonwords, pseudohomophones (e.g., wun) have the familiar graphemic and phonemic representations, but these stimuli also sound like real words (i.e., they have phonological lexical representations). However, pseudohomophones lack a whole-word orthographic lexical representation. As such, reading of pseudohomophones should require the sublexical analysis performed by the dorsal processing stream because they need to be sounded out, and additionally, would activate phonological lexical representations as their phonology is familiar. Manipulating the type of letter strings can optimize reliance on a particular reading stream (i.e., exception words should optimize ventral-lexical reading, whereas nonwords or pseudohomophones should optimize dorsal-sublexical reading; Borowsky et al. 2006) and thus has the potential to

123

isolate the relationships across representative white matter fibre tracts and reading processes. Structural framework of basic reading Diffusion Tensor Imaging (DTI) is a Magnetic Resonance Imaging (MRI) technique used to measure the diffusion of water molecules in brain tissue, as well as the magnitude and direction of diffusion in three dimensions (Basser et al. 1994; Le Bihan et al. 2001; Hunsche et al. 2001; Bammer et al. 2002; Beaulieu 2002). In white matter, water diffuses more readily along the orientation of axonal fibres than in other directions due to restriction from structural components of the fibres such as the myelin sheath and the axonal membrane (Moseley et al. 1990; Beaulieu 2002; Niogi and McCandliss 2006). Diffusion anisotropy is a measure that quantifies the degree of directionality of diffusion in the three dimensions (van Gelderen et al. 1994; Basser 1995; Basser and Pierpaoli 1996; Conturo et al. 1996). Fractional anisotropy (FA) is a normalized measure of diffusion anisotropy that can range in value between zero and one; high values of FA suggest the presence of highly directional diffusion (Deutsch et al. 2005; Niogi and McCandliss 2006; Odegard et al. 2009). This information can be used to infer the microstructure of white matter in vivo, with higher values of FA indicating better structural integrity (Klingberg et al. 2000; Deutsch et al. 2005; Niogi and McCandliss 2006). Mean diffusivity (MD) is the average of the eigenvalues of the diffusion tensor and as such, it reflects the magnitude of diffusion. The MD measure is invariant with respect to the orientation of diffusion tensor (Ben-Shachar et al. 2007). Importantly, FA and MD values can be correlated with behavioural performance to provide information about the relationship between underlying white matter tracts and cognitive tasks. Using this approach, a ventral/dorsal structural model of basic language processing has emerged in the literature. The ventral-lexical stream The ventral stream encompasses the inferior longitudinal fasciculus (ILF), inferior fronto-occipital fasciculus (IFOF) and the uncinate fasciculus (UF) (see Papagno et al. 2011 for a review; Duffau et al. 2009; Lebel et al. 2013; Mandonnet et al. 2007) and is hypothesized to play a role in lexical retrieval and semantic processing (Dick and Tremblay 2012; Mahoney et al. 2013; Menjot de Champfleur et al. 2013; Shinoura et al. 2013; Yeatman et al. 2013, Wilson et al. 2011). Most of what we know about white matter fibre tracts representing the ventral stream and their relation to single word reading comes from studies involving participants varying in reading proficiency from severely disordered to skilled. For example, using diffusion

Brain Struct Funct

images, Epelbaum et al. (2008) observed the progressive and selective degeneration of the left ILF in a patient who developed pure alexia (i.e., able to perform letter-by-letter reading, but not whole-word reading) following a small surgical lesion that damaged the left ILF just behind the putative visual word form area (VWFA). The VWFA, located in the left occipito-temporal region, is a part of the left ventral visual stream involved in whole-word identification (Szwed et al. 2011; Epelbaum et al. 2008). Similarly, Wilson et al. (2011) reported that reduced FA values in the UF were predictive of deficits in lexical processing (e.g., single word comprehension) in individuals with primary progressive aphasia (PPA). A recent study by Lebel et al. (2013), found that lexical reading scores (i.e., word identification scores from the Woodcock–Johnson test) were uniquely positively correlated with FA values in the IFOF. Together, these studies support the notion that ventral white matter tracts including the IFOF, UF and ILF are related to lexical/semantic processing in general and lexical reading in particular. The dorsal-sublexical stream The dorsal stream encompasses the arcuate fasciculus and superior longitudinal fasciculus (AF/SLF) and has been hypothesized to play a specific role in phonological and articulatory processes (Dick and Tremblay 2012; Menjot de Champfleur et al. 2013; Shinoura et al. 2013; Wilson et al. 2011). White matter pathways in the dorsal stream have been linked to reading ability (Richardson and Price 2009). Similar to literature on white matter tracts of the ventral stream, the relationship between dorsal tracts and reading has been reported from participants having a range of reading profiles. A landmark study by Klingberg et al. (2000) found that adults with a history of developmental dyslexia demonstrated decreased FA in the left hemisphere temporo-parietal region when compared to adults with no known reading difficulties. DTI analysis showed that axons in the temporo-parietal region were predominately anterior-posterior in direction; the location and the orientation of the fibres suggest that the white matter is composed of the left SLF (Klingberg et al. 2000). Steinbrink et al. (2008) reported similar findings. Their results indicated that adults with a history of developmental dyslexia and concomitant lower accuracy and reading speed for real words and nonwords when compared to controls, exhibited decreased FA in regions associated with the left SLF, ILF and IFOF compared to controls. Steinbrink et al. (2008) also found that higher anisotropy values in the frontotemporal part of the SLF were associated with faster pseudoword reading, which is in line with the notion that the AF/SLF plays a role in phonology. A study by Vandermosten et al. (2012) found that adults with dyslexia had

significantly reduced FA in the left AF when compared to adults with no reading difficulties. Vandermosten et al. (2012) also found a relationship between phoneme awareness and the FA of the left AF, as well as a relationship between orthographic processing and FA in the left IFOF. Overall, these results support the notion of a dorsal-sublexical stream involving the AF/SLF for reading. Much of the work to date that investigates the FAreading relationship has focused primarily on reading impairments. However, understanding the FA-reading relationship in skilled readers is as important for developing and extending neuroanatomical models of reading. For example, a study by Gold et al. (2007) explored whether the reaction time (RT) of a lexical decision task (using lowfrequency words and orthographically legal nonwords, which are thought to activate the dorsal-sublexical stream) correlated with white matter FA values for healthy young adults. The results indicated that lexical decision RT was negatively correlated with FA of left hemisphere white matter in inferior parietal and frontal language regions; regions that are implicated in the mapping of letter-tosound correspondences (Gold et al. 2007). These white matter regions were associated with the left SLF. To date, no DTI study has yet incorporated a systematic manipulation of word type and behavioural measures of overt reading speech and accuracy to explore the ventral-lexical/ dorsal-sublexical structural model that is emerging in the literature. Summary The purpose of this study was to investigate the relationship between lexical and sublexical reading and ventral and dorsal white matter tracts, respectively, in the brains of healthy adult participants. Specifically, we examined whether there was a negative relationship between the FA of white matter tracts hypothesized to underlie the ventrallexical and dorsal-sublexical processing streams and the reaction times for overt reading of exception words, regular words, nonwords and pseudohomophones. Based on the dual route model of reading and previous research on the connection between the left hemisphere white matter tracts and reading, it is hypothesized that there should be a negative relationship between RT for reading stimuli that typically activate the ventral-lexical stream (exception words) and measures of the underlying white matter tracts in the ventral region (the left IFOF, ILF, and UF), and that there should be a negative relationship between RT for reading stimuli that typically activate the dorsal-sublexical stream (pseudohomophones, nonwords) and measures of the underlying white matter tracts in the dorsal region (the left AF/SLF). Regular word reading RT, which can rely on either or both systems, was hypothesized to be more

123

Brain Struct Funct

weakly associated with measures of both dorsal and ventral white matter tracts. Furthermore, multiple regression of FA scores on the reaction times (RTs) of various letter string types was used to explore whether significant variance in FA scores can be uniquely attributed to lexicality in order to further identify relationships between FA and lexical versus sublexical reading processes.

Method Participants A total of 32 volunteers (21 female, 9 male) participated in this study. Participants included undergraduate and graduate students at the University of Alberta as well as individuals from the community. The participants ranged in age from 18 to 30 years (mean = 22.61 ± 3.02), and 28 of the participants were right handed1. Inclusion criteria consisted of normal or corrected to normal vision, no diagnosed language impairments, and English as a first language. Consent was obtained from the participants and the experiment was performed in compliance with institutional guidelines and with the approval of the University of Alberta Health Research Ethics Board. Stimuli A total of 756 stimuli organized by stimulus type, were presented. Stimuli consisted of 252 pronounceable nonwords (Mean length = 4.5; orthographic neighbourhood = 5.1; bigram mean = 1,376), 252 pseudohomophones (Mean length = 4.5; orthographic neighbourhood = 3.6; bigram mean = 1,198), 126 exception words (Mean length = 4.5; log HAL frequency = 9.7; orthographic neighbourhood = 6.0; bigram mean = 1,537), and 126 regular words (Mean length = 4.5; log HAL frequency = 9.5; orthographic neighbourhood = 6.9; bigram mean = 1,617). The nonwords were created by changing one letter of the exception and regular words. Stimuli were all monosyllabic and matched for initial onset and length. Materials Stimuli were presented on a computer monitor using EPrime software (Psychology Software Tools, Inc., http:// 1

Handedness was assessed by asking the participant about their hand preference. The pattern of results with and without the left-handed participants included did not change substantially (e.g., the exception words RT—uncinate FA correlation with left-handed participants was -0.428 and without was -0.418; similarly, the regular word RT— uncinate FA correlation with left-handed participants was -0.358 and without was -0.349).

123

www.pstnet.com). Voice onset was coded via a microphone and the experimenter used a button response to code accuracy on each trial. Procedure Participants were tested individually in a normally lit room. Letter strings were presented to the centre of a computer screen. Participants were instructed to name aloud the presented stimuli as quickly and as accurately as possible. The stimuli were presented in pure blocks (i.e., the participants named all the exception words in one block, all the regular words in another, etc.). Stimuli were randomly presented in each block and the order of blocks was randomized across participants. Participants had to initiate their overt response of each stimulus within a maximum of 1,800 ms after which time the stimulus was removed. The experimenter used a mouse to code correct and incorrect responses on all trials. Diffusion tensor image data acquisition All DTI imaging was conducted using a 1.5 T Siemens Sonata MRI scanner and scans were acquired using a dual spin-echo single-shot echo-planar imaging sequence. Thirty non-collinear directions of diffusion-sensitizing gradients were acquired, diffusion sensitivity b = 1,000 s/mm2, TR = 6,900 ms, TE = 100 ms. Forty, 3 mm thick axial slices (no inter-slice gap; 2 9 2 9 4) were obtained with an image matrix of 128 9 128 with 75 % phase partial Fourier zero-filled to 256 9 256. Raw images were visually examined and no motion artefacts were found. Tractography analysis of FA and RT correlation Tractography is a method of analyzing diffusion tensor data that uses the bulk-averaged tissue properties in each voxel and mathematical modelling to infer the dominant fibre orientation in each voxel (Wakana et al. 2004; Mukherjee et al. 2008). Tractography remains the only method of investigating white matter pathways and brain connectivity in vivo (Mori et al. 1999; Richards et al. 2008; Ben-Shachar et al. 2007). Three-dimensional tract reconstruction (tractography) for the tracts of interest in the left hemisphere were conducted for each participant on DTIstudio, using a multiple region of interest (ROI) approach and referencing the protocols outlined in Wakana et al. (2007), Zisner and Phillips (2009a, b, c, d). Multiple ROI approach utilizes existing anatomical knowledge of tract trajectories in order to reconstruct said tracts. Fibres that penetrate the manually defined ROIs are assigned to the specific tracts associated with those ROIs (Agosta et al. 2010; Wakana et al. 2007). Tractography was performed using raw DTI images that were not spatially normalized, smoothed, or manipulated in

Brain Struct Funct

Fig. 1 Representative tractography results for a arcuate fasciculus, b inferior fronto-occipital fasciculus, c inferior longitudinal fasciculus, d uncinate fasciculus

order to avoid potential artefacts or loss of sensitivity. The primary ventral tracts of interest were the left ILF, IFOF, and UF (see Fig. 1b–d). The primary dorsal tract of interest was the left AF/SLF (see Fig. 1a). An FA threshold of 0.25 was used to initiate and continue tracking; the maximum angle threshold was 70 degrees. With reference to the protocol outlined in Wakana et al. (2007), three types of ROI operations within DTIstudio were used for fibre tracking: ‘‘OR’’, ‘‘AND’’, and ‘‘NOT’’. The ‘‘OR’’ function selects all the fibres that penetrates a particular ROI, the ‘‘AND’’ function selects only those fibres that penetrate the defined ROIs, and the ‘‘NOT’’ function removes any fibres that penetrate the defined ROI. After the reconstruction of the tract, DTIstudio software computed the mean FA and mean diffusivity (MD) value of each tract for each individual, which was then correlated with reading reaction time. Reliability of tractography analysis Seven participants were randomly selected from the pool of participants and DTI tractography was performed again to check the intra-rater reliability. Overall, it was found that the FA value obtained from the second tractography analysis correlated strongly with the FA value from the first tractography analysis: ILF (r (5) = 0.996, p \ 0.001, twotailed), IFOF (r (5) = 0.997, p \ 0.001, two-tailed), UF (r (5) = 0.969, p \ 0.001, two-tailed), and AF (r (5) = 0.991, p \ 0.001, two-tailed).

Results Median correct reading RTs are displayed in Table 1. The RTs (ms) ranged from 550.32 ms for regular words to 635.62 ms for pseudohomophones2. 2

Accuracy for each word type was greater than 90 %, thus precluding any analysis with this measure given near ceiling performance and the limited variability [exception words: 94.7 % (standard deviation, SD = 3.0 %, regular words: 98.5 % (SD = 1.5 %); nonwords: 97.5 % (SD = 3.1 %), pseudohomophones 91.2 % (SD = 5.2 %))].

Table 1 Average median reaction time (RT) and interquartile range for naming exception words (EXC), regular words (REG), nonwords (NW), and pseudohomophones (PH) Average median RT (ms)

RT interquartile range

EXC

568.79

126.00

REG NW

550.32 618.92

143.00 143.50

PH

635.62

122.50

Table 2 Pearson’s correlation coefficients of the fractional anisotropy (FA) and mean diffusivity (MD) of the dorsal white matter tract [arcuate fasciculus (AF)] derived from tractography and median reaction times for naming regular words (REG), pseudohomophones (PH), and nonwords (NW) REG

PH

NW

Left AF FA

-0.173

-0.065

0.016

Left AF MD

-0.107

-0.111

-0.172

Tractography analysis: dorsal white matter tract Correlational analyses found no significant relationship for the mean FA or MD derived from tractography with the left AF and RT for reading of the letter string types that can involve sublexical reading processes: regular words, pseudohomophones or nonwords (see Table 2). Tractography analysis: ventral white matter tracts Correlational analyses found no relationship for the mean FA or MD obtained from tractography of the left ILF or IFOF thought to underlie the ventral processing stream and RT for reading of the letter strings that can involve lexical reading processes: exception words or regular words (see Table 3). A significant negative relationship was found for FA of the UF and the RT for overt reading of exception words (see Fig. 2). Converging with this finding, a positive relationship was found for MD of the UF and the RT for overt reading of exception words (one-tailed) (Fig. 2).

123

Brain Struct Funct Table 3 Pearson’s correlation coefficients of the fractional anisotropy (FA) and mean diffusivity (MD) of the ventral white matter tracts [inferior longitudinal fasciculus (ILF), inferior fronto-occipital fasciculus (IFOF), and uncinate fasciculus (UF)] derived from tractography and median reaction times for naming exception words (EXC) and regular words (REG) EXC

REG

Left ILF FA

-0.055

-0.090

Left IFOF FA

-0.198

-0.230

Left UF FA

-0.428*

Left ILF MD

0.023

-0.025

Left IFOF MD

0.268

0.283

Left UF MD

0.345?

0.309?

-0.358* (p = 0.051)

* Significant correlation, p \ 0.05 ?

Significant correlation, p \ 0.10

Fig. 3 Scatter plot of the correlational relationship between the mean fractional anisotropy (FA) of the left uncinate fasciculus (UF) derived from tractography and the median reading reaction time (RT) for regular words (REG; r (30) = -0.358, p = 0.051)

Fig. 2 Scatter plot of the correlational relationship between the mean fractional anisotropy (FA) of the left uncinate fasciculus (UF) derived from tractography and the median reading reaction time (RT) for exception words (EXC; r (30) = -0.428, p \ 0.05)

Similarly, a significant negative relationship was found for FA of the UF and the RT for overt reading of regular words (see Fig. 3). In addition, a positive relationship was found for MD of the UF and the RT for overt reading of regular words (one-tailed) (Fig. 3). Regression analyses Linear regression analyses were conducted on each participant’s correct median RTs, separately for lexical (exception words and regular words) and sublexical (regular words, pseudohomophones and nonwords) stimuli, and each tract (ILF, IFOF, UF, AF), where tract FA scores were regressed on median RTs. This was done in order to

123

explore the variance that could be attributed to orthographic lexical processing separately from variance that could be attributed to sublexical processing. Only the results that are statistically significant, or approaching significance, are reported. The overall regression model for UF FA scores regressed onto exception word and regular word median RT approached significance, R2 = 0.19, F(2,28) = 3.20, p = 0.056. The standardized regression coefficient (b) for exception words (-0.50) was significant, t(28) = -1.74, p = 0.04 (one-tailed), suggesting that as the UF FA scores increase, median RT decreases (see Fig. 4). The standardized regression coefficient (b) for regular words (0.09) was not significant, t(28) = 0.32, p = 0.75 (see Fig. 5).

Discussion Using tractography, we found a negative correlation between median RT for reading exception words and the mean FA of the left hemisphere UF, a ventral white matter tract that is part of the ventral-lexical reading stream. Consistent with this finding, we also found a positive relationship (one-tailed) between mean MD of the left hemisphere UF and median RT for reading exceptions words. In addition, the relationship between median RT for reading regular words and the mean FA of the left hemisphere UF approached significance (p = 0.051). Interestingly, a multiple regression analysis including both exception word and regular word median RT found that

Brain Struct Funct

Fig. 4 Partial regression plot of UF FA scores on EXC median reading RT after partialling out REG word median reading RT (r (30) = -0.312, p = 0.04, one-tailed)

Fig. 5 Partial regression plot of UF FA scores on REG median reading RT after partialling out EXC median reading RT (r (30) = 0.061, p [ 0.05)

only exception words accounted for any unique variance in mean FA of the UF. These results support the hypothesis that there is a relationship between orthographic lexical processing in the ventral stream and the underlying white matter, specifically the UF. We did not find any relationships between stimuli that typically activate the ventral stream and the IFOF and ILF, nor were there relationships between stimuli that typically activate the dorsal stream and the left AF. Overall, our results are consistent with previous research showing that the UF is important for basic language processes (Papagno et al. 2011), and suggest that the UF may play a particularly central role in orthographic lexical processing.

Although tractography had been used extensively to discover the origin, pathway, and termination of specific white matter tracts (Turken and Dronkers 2011; Bernal and Altman 2010; Yeatman et al. 2012) and to map tracts passing through specific ROIs (Beaulieu et al. 2005; Niogi and McCandliss 2006; Colnat-Coulbois et al. 2010), studies that have directly used tractography to examine the relationship between white matter and reading processes are limited (Lebel et al. 2013; Yeatman et al. 2011; Vandermosten et al. 2012). Here we present one of the first studies to explore the ventral-lexical, dorsal-sublexical structural model of basic language processes using overt reading in healthy adults; a necessary step in the development of a comprehensive reading model. In line with previous research, our data support the notion that the ventral stream plays a role in processing lexical stimuli (Dick and Tremblay 2012; Mahoney et al. 2013; Menjot de Champfleur et al. 2013; Shinoura et al. 2013; Yeatman et al. 2013, Wilson et al. 2011). The UF is a tract that interconnects the anterior temporal lobe to the orbitofrontal cortex (Catani et al. 2002; Schmahmann et al. 2007). Our finding of a relationship between FA of the UF and response time for reading exception words is consistent with the hypothesis that the UF plays a role in processing lexical stimuli, semantic associations, and aspects of reading (Grossman et al. 2004; Lu et al. 2002; Marchina et al. 2011). Interestingly, although the relationship between FA of the UF and response time for reading regular words also approached significance, our regression analysis underscores that there is something particularly unique about exception words and this ventral white matter tract. Therefore, it is not just the familiarity of the stimulus that is driving the relationship between FA and RT in the UF, but the inherent lexicality of the stimulus. Current structural models of language processing also have attributed the IFOF and ILF to the ventral-lexical stream. The IFOF is a large white matter tract that connects the occipital and frontal cortex via the extreme capsule (Catani et al. 2007; Yeatman et al. 2012; Vandermosten et al. 2012) and has been reported to be involved in reading and writing (Catani and Mesulam 2008; Steinbrink et al. 2008). The ILF is a ventral associative bundle, which joins the occipital and anterior temporal lobe (Catani et al. 2003; Steinbrink et al. 2008; Agosta et al. 2010; Yeatman et al. 2012) and is thought to play an important role in visual object recognition and in linking object representations to their lexical labels by mediating the fast transfer of visual signals to anterior temporal regions (Mummery et al. 1999; Steinbrink et al. 2008). Surprisingly, we did not find any relationships between these ventral tracts and overt reading of familiar exception words or regular words. However, we did find an association that approached significance between pseudohomophone RT and the IFOF. As such, this

123

Brain Struct Funct

tract may be particularly sensitive to lexical phonology but further work that manipulates that degree of lexical phonology would be needed to test this hypothesis. The SLF is one of the major cortical association fibre pathways in the human brain that underlies the dorsal processing stream (Wakana et al. 2004; Makris et al. 2005; Gold et al. 2007; Agosta et al. 2010). It can be divided into four components; the most important portion for language processing is thought to be the AF (Bernal and Altman 2010; Catani et al. 2005; Geschwind 1970; Makris et al. 2005; Parker et al. 2005; Petrides and Pandya 1984; Powell et al. 2006; Schmahmann et al. 2007; Selnes et al. 2002). In recent structural language models, the AF has been hypothesized to play a specific role in phonological and articulatory processes (Bernal and Ardila 2009; Dick and Tremblay 2012; Marchina et al. 2011; Menjot de Champfleur et al. 2013; Shinoura et al. 2013; Wilson et al. 2011). In the current study, we did not find any associations between overt reading RT and FA of the AF. Much of the previous work on the AF has been done in the area of aphasia (Mahoney et al. 2013; Marchina et al. 2011; Wilson et al. 2011; see Berthier et al. 2012 for a review) and many of these studies have used auditorily presented stimuli (e.g., Wilson et al. 2011). It may be that the AF is particularly sensitive to phonological and articulatory processes (Shuren et al. 1995; Vandermosten et al. 2012; Yeatman et al. 2011) when the task is completed in the auditory domain. These processes are different from that used during overt reading and may have made a difference in detecting a correlation between reading processes and FA. In the current study, tractography analysis derived the mean FA and MD from the entire reconstructed tract. Considering the fact that the IFOF, ILF, and AF are all large white matter tracts, it is possible that by examining the entire tract, any relationships with small portions of white matter within the reconstructed tracts are obscured (Makris and Pandya 2009). As such, the significant relationships found in the present study between exception word reading RT and the UF FA and MD are particularly revealing about the general aspects of the entire UF tract. The results of this study aid our understanding of the relationship between white matter and reading and expands upon previous DTI research examining the relationship between the FA of white matter tracts and reading (Klingberg et al. 2000; Deutsch et al. 2005; Beaulieu et al. 2005; Niogi and McCandliss 2006; Gold et al. 2007; Richards et al. 2008; Qiu et al. 2008; Steinbrink et al. 2008; Odegard et al. 2009; Rollins et al. 2009; Carter et al. 2009; Palmer et al. 2010; Rimrodt et al. 2010; Yeatman et al. 2011; Vandermosten et al. 2012). Much of the previous research focused on reading ability and used scores from standardized tests such as the Woodcook Reading Mastery

123

Test (Klingberg et al. 2000; Beaulieu et al. 2005; Niogi and McCandliss 2006;), the Woodcook–Johnson Tests of Academic Achievement (Palmer et al. 2010), or a battery of standardized reading measures (Steinbrink et al. 2008; Carter et al. 2009; Odegard et al. 2009; Rollins et al. 2009; Deutsch et al. 2005; Yeatman et al. 2011). In contrast, the present study used RT from overt reading tasks as measures of word recognition, which has been previously established in the literature (Borowsky et al. 2012; Besner and Smith 1992; Cummine et al. 2010, 2013). It is likely that the RT for overt reading captures different aspects of the reading process than the standardized measures previously used, and is a likely reason we find varying results to those previously reported. In addition, we manipulated the types of words (exception and regular) and pseudowords (nonword and pseudohomophone) so as to further isolate components of lexical and sublexical processing, which the standardized tests of word and letter string reading do not do beyond the separation of nonwords (e.g., Word Attack in the WRMT) and words (e.g., Word Identification in the WRMT).

Conclusion In this study, it was hypothesized that there should be a relationship between stimuli that typically activated the ventral stream and the underlying white matter tracts in the ventral region and that there should be a relationship between stimuli that typically activate the dorsal stream and the underlying white matter tracts in the dorsal region. This hypothesis was formed based on a dual route model of reading as it relates to functional neuroimaging (e.g., Cummine et al. 2010, 2013; Borowsky et al. 2012) and on previous research examining the connection between structural integrity of the left hemisphere white matter and reading. Our DTI tractography analysis provided support to suggest that the left UF played a role in ventral-lexical orthographic processing. We conclude that relationships between reading processes and the underlying white matter tracts are highly complex and may involve far more than the four tracts measured in the current study. Overall, the findings of this study have important implications for our understanding of the relationship between data on white matter microstructure and reading processes.

References Agosta F, Henry RG, Migliaccio R, Neuhaus J, Miller BL, Dronkers NF et al (2010) Language networks in semantic dementia. Brain 133(Pt 1):286–299

Brain Struct Funct Balota DA, Cortese MJ, Sergent-Marshall SD, Spieler DH, Yap MJ (2004) Visual word recognition of single-syllable words. J Exp Psychol Gen 133:283–316 Bammer R, Auer M, Keeling SL, Augustin M, Stables LA, Prokesch RW et al (2002) Diffusion tensor imaging using single-shot SENSE-EPI. Magn Reson Med 48(1):128–136 Basser PJ (1995) Inferring microstructural features and the physiological state of tissues from diffusion-weighted images. NMR Biomed 8(7–8):333–344 Basser PJ, Pierpaoli C (1996) Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. J Magn Reson Ser B 111(3):209–219 Basser PJ, Mattiello J, LeBihan D (1994) MR diffusion tensor spectroscopy and imaging. Biophys J 66(1):259–267 Beaulieu C (2002) The basis of anisotropic water diffusion in the nervous system: a technical review. NMR Biomed 15(7–8): 435–455 Beaulieu C, Plewes C, Paulson LA, Roy D, Snook L, Concha L et al (2005) Imaging brain connectivity in children with diverse reading ability. NeuroImage 25(4):1266–1271 Ben-Shachar M, Dougherty RF, Wandell BA (2007) White matter pathways in reading. Curr Opin Neurobiol 17(2):258–270 Bernal B, Altman N (2010) The connectivity of the superior longitudinal fasciculus: a tractography DTI study. Magn Reson Imaging 28(2):217–225 Bernal B, Ardila A (2009) The role of the arcuate fasciculus in conduction aphasia. Brain 132:2309–2316 Berthier ML, Ralph MAL, Pujol J, Green C (2012) Arcuate fasciculus variability and repetition: the left sometimes can be right. Cortex 48:133–143 Besner D, Smith MC (1992) Models of visual word recognition: when obscuring the stimulus yields a clearer view. J Exp Psychol Learn Mem Cogn 18:468–482 Borowsky R, Cummine J, Owen WJ, Friesen CK, Shih F, Sarty GE (2006) FMRI of ventral and dorsal processing streams in basic reading processes: insular sensitivity to phonology. Brain Topogr 18(4):233–239 Borowsky R, Esopenko C, Gould L, Kuhlmann N, Sarty G, Cummine J (2012) Localization of function for noun and verb reading; converging evidence for shared processing from fMRI activation and reaction time. Lang Cognitive Process. doi:10.1080/ 01690965.2012.665466 Carter JC, Lanham DC, Cutting LE, Clements-Stephens AM, Chen X, Hadzipasic M et al (2009) A dual DTI approach to analyzing white matter in children with dyslexia. Psychiatry Res 172(3):215–219 Catani M, Mesulam M (2008) The arcuate fasciculus and the disconnection theme in language and aphasia: history and current state. Cortex 44(8):953–961 Catani M, Howard RJ, Pajevic S, Jones DK (2002) Virtual in vivo interactive dissection of white matter fasciculi in the human brain. NeuroImage 17(1):77–94 Catani M, Jones DK, Donato R, Ffytche DH (2003) Occipitotemporal connections in the human brain. Brain 126(Pt 9):2093–2107 Catani M, Jones DK, ffytche DH (2005) Perisylvian language networks of the human brain. Ann Neurol 57(1):8–16 Catani M, Allin MP, Husain M, Pugliese L, Mesulam MM, Murray RM et al (2007) Symmetries in human brain language pathways correlate with verbal recall. Proc Natl Acad Sci USA 104(43):17163–17168 Colnat-Coulbois S, Mok K, Klein D, Penicaud S, Tanriverdi T, Olivier A (2010) Tractography of the amygdale and hippocampus: anatomical study and application to selective amygdalohippocampectomy laboratory investigation. J Neurosurg 113(6):1135–1143

Coltheart M, Rastle K, Perry C, Langdon R, Ziegler J (2001) DRC: a dual route cascaded model of visual word recognition and reading aloud. Psychol Rev 108:204–256 Conturo TE, McKinstry RC, Akbudak E, Robinson BH (1996) Encoding of anisotropic diffusion with tetrahedral gradients: a general mathematical diffusion formalism and experimental results. Magn Reson Med 35(3):399–412 Cummine J, Sarty GE, Borowsky R (2010) Localizing the frequency x regularity word reading interaction in the cerebral cortex. Neuropsychologia 48:2147–2157 Cummine J, Gould L, Zhou C, Hrybouski S, Siddiqi Z, Chouinard B, Borowsky R (2013) Manipulating instructions strategically affects reliance on the ventral-lexical reading stream: converging evidence from neuroimaging and reaction time. Brain Lang. doi:10.1016/j.bandl.2012.04.009 Deutsch GK, Dougherty RF, Bammer R, Siok WT, Gabrieli JD, Wandell B (2005) Children’s reading performance is correlated with white matter structure measured by diffusion tensor imaging. Cortex 41(3):354–363 Dick AS, Tremblay P (2012) Beyond the arcuate fasciculus: consensus and controversy in the connectional anatomy of language. Brain 135:3529–3550 Duffau H, Gatignol P, Moritz-Gasser S, Mandonnet E (2009) Is the left uncinate fasciculus essential for language? A cerebral stimulation study. J Neurol 256(3):382–389 Epelbaum S, Pinel P, Gaillard R, Delmaire C, Perrin M, Dupont S et al (2008) Pure alexia as a disconnection syndrome: new diffusion imaging evidence for an old concept. Cortex 44(8):962–974 Geschwind N (1970) The organization of language and the brain. Science 170:940–944 Gold BT, Powell DK, Xuan L, Jiang Y, Hardy PA (2007) Speed of lexical decision correlates with diffusion anisotropy in left parietal and frontal white matter: evidence from diffusion tensor imaging. Neuropsychologia 45(11):2439–2446 Gould L, Cummine J, Borowsky R (2012) The cognitive chronometric architecture of reading aloud: semantic and lexical effects on naming onset and duration. Front Hum Neurosci (in press) Grossman M, McMillan C, Moore P, Ding L, Glosser G, Work M (2004) What’s in a name: voxel-based morphometric analyses of MRI and naming difficulty in Alzheimer’s disease, frontotemporal dementia and corticobasal degeneration. Brain 127:628–649 Hickok G, Poeppel D (2004) Dorsal and ventral streams: a framework for understanding aspects of the functional anatomy of language. Cognition 92(1–2):67–99 Hunsche S, Moseley ME, Stoeter P, Hedehus M (2001) Diffusiontensor MR imaging at 1.5 and 3.0 T: initial observations. Radiology 221:550–556 Klingberg T, Hedehus M, Temple E, Salz T, Gabrieli JD, Moseley ME et al (2000) Microstructure of temporo-parietal white matter as a basis for reading ability: evidence from diffusion tensor magnetic resonance imaging. Neuron 25(2):493–500 Le Bihan D, Mangin JF, Poupon C, Clark CA, Pappata S, Molko N et al (2001) Diffusion tensor imaging: concepts and applications. J Magn Reson Imaging 13(4):534–546 Lebel C, Shaywitz B, Holahan J, Shaywitz S, Marchione K, Beaulieu C (2013) Diffusion tensor imaging correlates of reading ability in dysfluent and non-impaired readers. Brain Lang 125:215–222 Lu LH, Crosson B, Nadeau SE, Heilman KM, Gonzalez-Rothi LJ, Raymer A et al (2002) Category-specific naming deficits for objects and actions: semantic attribute and grammatical role hypotheses. Neuropsychologia 40(9):1608–1621 Mahoney CJ, Malone IB, Ridgway GR, Buckley AH, Downey LE, Golden HL, Warren JD et al (2013) White matter tract signatures of the progressive aphasias. Neurobiol Again 34:1687–1699

123

Brain Struct Funct Makris N, Pandya DN (2009) The extreme capsule in humans and rethinking of the language circuitry. Brain Struct Funct 213:343–358 Makris N, Kennedy DN, McInerney S, Sorensen AG, Wang R, Caviness Jr VS et al (2005) Segmentation of subcomponents within the superior longitudinal fascicle in humans: a quantitative, in vivo, DT-MRI study. Cereb Cortex (New York, N.Y.: 1991) 15(6):854–869 Mandonnet E, Nouet A, Gatignol P, Capelle L, Duffau H (2007) Does the left inferior longitudinal fasciculus play a role in language? A brain stimulation study. Brain 130(Pt 3):623–629 Marchina S, Zhu LL, Norton A, Zipse L, Wan CY, Schlaug G (2011) Impairment of speech production predicted by lesion load of the left arcuate fasciculus. Stroke 42:2251–2256 Menjot de Champfleur N, Maldonado IL, Moritz-Gasser S, Machi P, Le Bars E, Bonafe A, Duffau H (2013) Middle longitudinal fasciculus delineation within language pathways: a diffusion tensor imaging study in human. Eur J Radiol 82:151–157 Mori S, Crain BJ, Chacko VP, van Zijl PC (1999) Three-dimensional tracking of axonal projections in the brain by magnetic resonance imaging. Ann Neurol 45(2):265–269 Moseley ME, Cohen Y, Kucharczyk J, Mintorovitch J, Asgari HS, Wendland MF et al (1990) Diffusion-weighted MR imaging of anisotropic water diffusion in cat central nervous system. Radiology 176(2):439–445 Mukherjee P, Berman JI, Chung SW, Hess CP, Henry RG (2008) Diffusion tensor MR imaging and fiber tractography: theoretic underpinnings. Am J Neuroradiol 29(4):632–641 Mummery CJ, Patterson K, Wise RJ, Vandenberghe R, Price CJ, Hodges JR (1999) Disrupted temporal lobe connections in semantic dementia. Brain 122(Pt 1):61–73 Niogi SN, McCandliss BD (2006) Left lateralized white matter microstructure accounts for individual differences in reading ability and disability. Neuropsychologia 44(11):2178–2188 Odegard TN, Farris EA, Ring J, McColl R, Black J (2009) Brain connectivity in non-reading impaired children and children diagnosed with developmental dyslexia. Neuropsychologia 47(8–9):1972–1977 Palmer SL, Reddick WE, Glass JO, Ogg R, Patay Z, Wallace D et al (2010) Regional white matter anisotropy and reading ability in patients treated for pediatric embryonal tumors. Brain Imaging Behav 4(2):132–140 Papagno C, Miracapillo C, Casarotti A, Romero Lauro LJ, Castellano A, Falini A et al (2011) What is the role of the uncinate fasciculus? Surgical removal and proper name retrieval. Brain 134(Pt 2):405–414 Parker GJ, Luzzi S, Alexander DC, Wheeler-Kingshott CA, Ciccarelli O, Lambon Ralph MA (2005) Lateralization of ventral and dorsal auditory-language pathways in the human brain. NeuroImage 24(3):656–666 Petrides M, Pandya DN (1984) Projections to the frontal cortex from the posterior parietal region in the rhesus monkey. J Comp Neurol 228:105–116 Powell HWR, Parker GJM, Alexander DC, Symms MR, Boulby PA, Wheeler-Kingshott CAM et al (2006) Hemispheric asymmetries in language-related pathways: a combined functional MRI and tractography study. NeuroImage 32(1):388–399 Price C (2012) A review and synthesis of the first 20 years of PET and fMRI studies of heard speech, spoken language and reading. NeuroImage 62:816–847 Qiu D, Tan LH, Zhou K, Khong PL (2008) Diffusion tensor imaging of normal white matter maturation from late childhood to young adulthood: voxel-wise evaluation of mean diffusivity, fractional anisotropy, radial and axial diffusivities, and correlation with reading development. NeuroImage 41(2):223–232

123

Richards T, Stevenson J, Crouch J, Johnson LC, Maravilla K, Stock P et al (2008) Tract-based spatial statistics of diffusion tensor imaging in adults with dyslexia. Am J Neuroradiol 29(6):1134–1139 Richardson FM, Price CJ (2009) Structural MRI studies of language function in the undamaged brain. Brain Struct Funct 213:511–523 Rimrodt SL, Peterson DJ, Denckla MB, Kaufmann WE, Cutting LE (2010) White matter microstructure differences linked to left perisylvian language network in children with dyslexia. Cortex 46:739–749 Rollins NK, Vachha B, Srinivasan P, Chia J, Pickering J, Hughes CW et al (2009) Simple developmental dyslexia in children: alterations in diffusion-tensor metrics of white matter tracts at 3 T. Radiology 251(3):882–891 Schmahmann JD, Pandya DN, Wang R, Dai G, D’Arceuil HE, de Crespigny AJ et al (2007) Association fibre pathways of the brain: parallel observations from diffusion spectrum imaging and autoradiography. Brain 130(Pt 3):630–653 Selnes OA, van Zijl PCM, Barker PB, Hillis AE, Mori S (2002) MR diffusion tensor imaging documented arcuate fasciculus lesion in a patient with normal repetition performance. Aphasiology 16:897–901 Shinoura N, Midorikawa A, Onodera T, Tsukada M, Yamada R, Tabei Y, et al (2013) Damage to the left ventral, arcuate fasciculus and superior longitudinal fasciculus-related pathways induces deficits in object naming, phonological language function and writing, respectively. Int J Neurosci. doi:10.3109/ 00207454.2013.765420 Shuren JE, Schefft BK, Yeh HS, Privitera MD, Cahill WT, Houston W (1995) Repetition and the arcuate fasciculus. J Neurol 242:596–598 Steinbrink C, Vogt K, Kastrup A, Muller HP, Juengling FD, Kassubek J et al (2008) The contribution of white and gray matter differences to developmental dyslexia: insights from DTI and VBM at 3.0 T. Neuropsychologia 46(13):3170–3178 Szwed M, Dehaene S, Kleinschmidt A, Eger E, Valabregue R, Amadon A et al (2011) Specialization for written words over objects in the visual cortex. NeuroImage 56(1):330–344 Turken AU, Dronkers NF (2011) The neural architecture of the language comprehension network: converging evidence from lesion and connectivity analyses. Front Syst Neurosci 5:1 van Gelderen P, de Vellschouwer MHM, DesPres D, Pekar J, van Ziji PCM, Moonen CTW (1994) Water diffusion and acute stroke. Magn Reson Med 31(2):154–163 Vandermosten M, Boets B, Ghesquiere P (2012) A qualitative and quantitative review of diffusion tensor imaging studies in reading and dyslexia. Neurosci Biobehav Rev 36:1532–1552 Wakana S, Jiang H, Nagae-Poetscher LM, van Zijl PC, Mori S (2004) Fiber tract-based atlas of human white matter anatomy. Radiology 230(1):77–87 Wakana S, Caprihan A, Panzenboeck MM, Fallon JH, Perry M, Gollub RL et al (2007) Reproducibility of quantitative tractography methods applied to cerebral white matter. NeuroImage 36(3):630–644 Wilson SM, Galantucci S, Tartaglia MC, Rising K, Patterson DK, Henry ML, Gorno-Tempini ML et al (2011) Syntactic processing depends on dorsal language tracts. Neuron 72:397–403 Yeatman JD, Dougherty RF, Rykhlevskaia E, Sherbondy AJ, Deutsch GK, Wandell BA, Ben-Shachar M (2011) Anatomical properties of the arcuate fasciculus predict phonological and reading skills in children. J Cognitive Neurosci 23:3304–3317 Yeatman JD, Rauschecker AM, Wandell BA (2012) Anatomy of the visual word form area: adjacent cortical circuits and long-range white matter connections. Brain Lang (in press)

Brain Struct Funct Yeatman JD, Rauschecker AM, Wandell BA (2013) Anatomy of the visual word form area: adjacent cortical circuits and long-range white matter connections. Brain Lang 125:146–155 Zisner A, Phillips O (2009) Identifying the arcuate fasciculus (AF) in DTIStudio using a multiple ROI approach. Retrieved July/20, 2011, from http://www.loni.ucla.edu/*narr/protocol.php?q=dti_ilf Zisner A, Phillips O (2009) Identifying the inferior fronto-occipital fasciculus (IFO) in DTIStudio using a multiple ROI approach. Retrieved July/20, 2011, from http://www.loni.ucla.edu/*narr/ protocol.php?q=dti_ifo

Zisner A, Phillips O (2009) Identifying the inferior longitudinal fasciculus in DTIStudio using a multiple ROI approach. Retrieved July/20, 2011, from http://www.loni.ucla.edu/*narr/ protocol.php?q=dti_ilf Zisner A, Phillips O (2009) Identifying the uncinate fasciculus (UF) in DTIStudio using a multiple ROI approach. Retrieved July/20, 2011, from http://www.loni.ucla.edu/*narr/protocol.php?q=dti_ uf

123