Uncovering Hidden Pathways: Structural Brain Networks Underpinning Connected Speech in Post-Stroke Aphasia

Uncovering Hidden Pathways: Structural Brain Networks Underpinning Connected Speech in Post-Stroke Aphasia | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Uncovering Hidden Pathways: Structural Brain Networks Underpinning Connected Speech in Post-Stroke Aphasia Ping-Jung Duh, Ioana Sederias, Vitor Zimmerer, Haya Akkad, Alex Leff, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5587181/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Individuals with post-stroke aphasia (PWA), an acquired language disorder, face significant communication challenges essential for daily life. Surprisingly, little is known about how brain damage following stroke disrupts the anatomical integration of language and multiple-demand brain networks required for speech production. To address this, we used a measure of brain structural network efficiency (SNE) to investigate the anatomical correlates of spoken language abilities in 36 chronic PWA. Connected speech abilities were correlated with higher SNE not only within the left (dominant) language network but also with bilateral language and multiple-demand networks. Post-hoc analyses found: 1) word-level language behaviours were associated solely with discrete left temporoparietal lesions, using voxel-based correlational methodology; 2) using canonical variate analysis, SNE was primarily sensitive to higher-level language behaviours, loading additional brain-behaviour variance beyond distributed voxels. Taken together, our findings illustrate how speech abilities in PWA rely on distributed bilateral networks, outlining higher-level language-cognition interplay. Health sciences/Diseases/Neurological disorders/Stroke Biological sciences/Neuroscience/Cognitive neuroscience/Language Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Aphasia is an acquired language disorder most commonly caused by a stroke affecting the dominant, typically left, hemisphere. While brain lesions causing aphasia primarily affect language functions, a growing body of research indicates that other cognitive domains crucial for daily communication, including working memory and cognitive flexibility are also impacted in post-stroke patients with aphasia (PWA) (Butler et al., 2014; Schumacher et al., 2019). Language and non-language cognitive functions are interrelated, with evidence suggesting bidirectional influences on speech communication abilities with language deficits disrupting non-language cognition performance (Keil & Kaszniak, 2010), and cognitive impairments impacting on functional language abilities (Heuer et al., 2017; Schumacher et al., 2019). For instance, working memory deficits frequently persist in PWA after recovery on standard language tests (Martin & Reilly, 2012), negatively impacting their ‘real-world’ daily abilities, and in some instances limiting treatment outcomes that rely on serial order memory for word re-learning (Murray, 2012; Salis et al., 2015). Current standardised measures of language ability, especially those used in routine clinical practice, primarily assess linguistic impairments in isolation and often fail to fully capture the integrated cognitive-linguistic deficits that contribute to real-world communication difficulties (Marini et al., 2011). For instance, preserved performance on a naming task may not capture a PWA’s semantic retrieval and/or verbal working memory impairments that could affect their ability to communicate more complex ideas or sequence information in conversation. Performance on more naturalistic, cognitively demanding verbal tasks such as composite picture description tasks or conversational data (Perkins, 1995) may better reflect this complexity. Composite picture description tasks are often preferred in clinical assessment of PWA as they allow for control of the target spoken language and enable comparisons across individuals and time points (Conroy et al., 2009). However, measuring PWA’s performance on this task by, for example, primarily counting the number of information carrying words produced, often fails to capture the richness of the speech data and reduces the sensitivity of the task to connected speech function i.e., it is akin to another (single-word) naming test. A proposed alternative method for quantitative speech production analysis is the Frequency in Language Analysis Tool (FLAT) (Zimmerer et al., 2018, 2020), an automated toolbox for quantification of language-based frequentist-use of spoken language features. By characterising connected speech samples at multiple linguistic levels, including frequency, novelty and connectivity of word combinations used, it aims to provide novel insights into more complex spoken language performance. It appears that familiarity of expressions is an important factor for organization of language in the brain. Surprisingly, analyzing connected speech from a usage-based or frequentist perspective has rarely been explored in the context of PWA. A few studies in PWA and those with Alzheimer's disease, have reported reduced linguistic creativity during speech, relying heavily on familiar phrases and fixed expressions (Van Lancker Sidtis & Postman, 2006; Zimmerer et al., 2018, 2020). For example, in non-fluent aphasia populations, bigram (i.e., two-word combination) frequency and collocation scores show significantly high values relative to controls, which can be associated with reduced lexical-semantic, grammatical, or working memory capacity, resulting in less flexible language processing (Zimmerer et al., 2018). Overreliance on familiar combinations has been regarded a compensatory mechanism to mask linguistic impairments (Bridges & Van Lancker Sidtis, 2013; Wray, 2011; Zimmerer et al., 2016). With this interplay between language and cognitive control underpinning everyday communication behaviour, speech production likely involves coordinated activity across multiple brain regions i.e., both language-specific as well as multiple-demand regions in both hemispheres (Hickok & Poeppel, 2007; Kearney & Guenther, 2019). For most people, the left hemisphere plays a crucial role in language processing. Sentence retrieval relies on left hemisphere language areas, including the inferior frontal gyrus and superior temporal gyrus (Hartwigsen et al., 2017; Stefaniak et al., 2021). However, additional cognitive functions including holding the conversation topic in mind, organizing ideas sequentially, and flexibly shifting between themes likely engage bilateral regions classically implicated in non-verbal executive functions and working memory, such as the dorsolateral prefrontal and parietal cortices (Chen et al., 2013; Kalm et al., 2012). With right and bilateral temporo-parietal functional activation supporting residual language capacity after aphasia stroke (Turkeltaub et al., 2011). Consistently with this, our group found increased bilateral temporal lobe grey matter tissue density associated with improved auditory comprehension abilities in chronic PWA following behavioral training (Fleming et al., 2020). Suggesting that preserved functional, and by association anatomical connectivity between hemispheres may underpin language recovery after brain damage. The complex interactions between multiple brain regions that support language recovery leads us to consider a broader network perspective in language processing. Cognitive impairments, including those involved in speech production, are often described in terms of network disruptions (Fedorenko & Varley, 2016; Hartwigsen & Saur, 2019). Deficits may arise from anatomically intact but disconnected brain regions within a network (Billot et al., 2022a; Fridriksson et al., 2018; Marebwa et al., 2017; Price et al., 2017). For example, structurally intact regions may be functionally impaired due to white matter disconnection from key network nodes. Residual anatomical connections of these regions could provide a neural basis for compensatory cognitive functions to support speech production (Bassi et al., 2019; Guggisberg et al., 2019). To quantify anatomical damage associated with speech deficits in aphasia, researchers have traditionally used mass univariate approaches such as voxel-based lesion analyses. This method correlates variance in structural brain damage, e.g., from a T1 volumetric MRI brain scan, with variance in PWAs’ behavioural performance (Bates et al., 2003). However, a limitation of the voxel-based lesion method is that it treats each voxel independently and only considers regions with direct damage, which may not fully capture the distributed network disruptions underlying complex behaviors (Billot et al., 2022a; Carrera & Tononi, 2014). A proposed alternative and complementary approach is the analysis of brain structural connectivity, allowing for the examination of white matter pathways connecting various brain regions. Yet, this approach might overlook the complex interactions between different connectivity patterns (Foulon et al., 2018). To understand brain and behaviour function primarily in the healthy brain, many groups have used functional regions of interest (fROIs) defined by task-based fMRI activation patterns to parcellate language and multiple cognitive demand (MD) networks (Fedorenko et al., 2012; 2014). How residual structural connectivity between these functionally-defined nodes relates to post-stroke behavioral capacity remains unclear. Calculating the structural network efficiency (SNE) of fROIs using graph theory (Kaiser, 2011) may provide greater sensitivity and help bridge this gap by quantifying the efficiency of residual structural connections in the lesioned brain. The lesion quantification (Griffis et al., 2021) approach employs streamline-based tract disconnection severity, rather than voxel-wise tract lesion load (Hope et al., 2018). Tract lesion load estimates the proportion of damaged voxels within each tract, but it does not account for spatial distribution of lesions across streamlines. Thus, a small lesion volume could still substantially disrupt connectivity if distributed across streamlines (see Figure 1). The approach thereby directly captures tract disconnection and may better detect network disruptions relevant for brain function (Griffis et al., 2021). Building on this, in this paper we examine whether the SNE approach can effectively measure the integrity of the disrupted structural network underlying speech production abilities in a group of 36 PWA. Our approach diverges from previous studies (Cheng et al., 2019; Lawrence et al., 2014; Termenon et al., 2016) that looked at either whole-brain efficiency or local efficiency around a single region. To address the challenge of overfitted issues with our sample size, we reduced our brain variable dimensions, by focusing on two functional regions of interest networks(fROIs). These two fROIs were defined using the extensive database of normal healthy subjects’ functional activation patterns for language and multiple demand (MD) cognitive tasks available from Fedorenko and colleagues (Fedorenko, 2014; Fedorenko et al., 2010, 2012). These fROIs were chosen to capture the integrity of networks supporting language-specific and multiple-demand cognitive processes, which are often impaired in aphasia (Barbieri et al., 2019; Murray, 2012; Schumacher et al., 2019). The language fROI includes regions consistently activated during language tasks in healthy individuals, while the cognitive fROI comprises regions engaged in various demanding cognitive tasks, such as working memory and cognitive control. We then used SNE to quantify ipsilesional (left hemisphere), contralesional (right hemisphere) and bilateral connectivity with these same two functional brain networks. This enabled us to examine the impact of the left hemisphere lesions in our PWA sample (compared to the normative reference) on damaged dominant language and non-dominant (contralesional) language networks and bilateral multiple demand networks. We hypothesized that SNE captures brain-behaviour relationships at the network level. We predicted higher SNE values of bilateral language and MD brain networks would be associated with better spoken language abilities within the PWA group, complementing the localised lesion-deficit mappings from voxel-based method. We tested associations between variance in SNE and behavioral scores with bilateral, left hemisphere, and right hemisphere language and MD network measures, while controlling for lesion volume. To increase our sensitivity to connected speech production abilities, our analyses included 1) frequentist measures using FLAT analysis of PWAs’ speech samples from custom picture description tasks; and 2) anatomical measures of distributed brain network using SNE. Results Brain network correlations with connected (narrative) speech production In this group of PWA, better performance on connected speech measures was associated with greater structural efficiency in the left language network. Specifically, FLAT connectivity measure (r = 0.46, p = 0.013) and CAT standardised picture description scores (r = 0.37, p =0.029) showed positive correlations with left hemisphere language network SNE. Complementing these findings, FLAT measures of bigram frequency and bigram collocation – where lower scores reflects more creative, less formulaic language – were correlated with higher SNE in both left and bilateral language networks (r = -0.49, p = 0.003 and r = -0.54, p = 0.001 respectively). The FLAT bigram collocation performance measure was the only one negatively associated with bilateral SNE measures of the MD network (r = -0.42, p = 0.013). See Figure 2 for full reported results. Brain network correlations with sentence tasks but not word-level speech production Our analysis revealed a significant relationship between SNE and language behaviours at high-level but not at word-level language behaviours. There was a significant positive correlation between SNE of the bilateral language network and spoken sentence comprehension (r = 0.40, p = 0.018) in PWAs. The increased efficiency within residual bilateral hemisphere structural language network was associated with better spoken sentence comprehension. In contrast, word-level language processing including naming and spoken word comprehension, did not show significant correlations with SNE in any of the assessed networks. Post-hoc 1: Voxel-based lesion symptom mapping To investigate whether word-level language behaviours could be associated with focal lesions in specific brain regions, we conducted a post-hoc voxel-based correlational methodology (VBCM) (Akkad et al., 2023; Tyler et al., 2005), a variant of voxel-based lesion symptom mapping (VLSM; Bates et al., 2003). A voxel-wise F-test was used to compare the fuzzy lesion status between patients with their performance on each neuropsychological assessment. Total lesion volume was included as a covariate in the general linear model. Results are reported at p ≤ 0.001 voxel-level and p < 0.05 FWE corrected at cluster-level. All anatomical labels were based on the AAL atlas in MNI space. The analyses were run using SPM 12 with MATLAB 2022b. The results revealed neural correlates with word-level speech production deficits (Figure 3 and Table 1). Damage to left middle temporal gyrus was correlated with impairments in naming. Lesions in the left superior parietal gyrus were associated with deficits in spoken word comprehension. No significant neural correlates were found for sentence comprehension and connected speech measures. The results complement our main findings by demonstrating that while word-level language behaviours do not show a significant relationship with SNE measures, they are associated with focal lesions in specific brain regions. Table 1. Neural correlates of word-level speech production deficits. Behaviour Assessments Brain Location Extend Voxels Z MNI Coordinate x y z CAT Naming Left Middle Temporal Gyrus 149 3.31 -58 -8 -20 CAT Word Comprehension Left Superior Parietal Gyrus 240 3.14 -26 -56 48 *Only clusters with cluster-level FWE p < 0.05 are shown. Coordinates are in MNI space. Post-hoc 2: Canonical Variate Analyses To investigate whether SNE was primarily sensitive to high-level language behaviors and provided additional brain-behaviour information beyond distributed voxels per se, we conducted a post-hoc Canonical Variate Analysis (CVA) in SPM12 (Friston et al., 1995). CVA is a multivariate method that identifies linear combinations of variables that best explain variance in dependent variables. Here, CVA was used to quantify how much language behaviour variance could be accounted for by brain measures (SNE and voxels). The resulting canonical variates, ranked by most explained shared variance, represent dimensions of maximal covariation between the brain measures and language behaviours. Chi-square tests indicate the statistical significance of the amount of shared variance that can be explained. First, Singular Value Decomposition (SVD) was employed to reduce all voxels in language and MD ROIs to two factors, making them comparable to the two bilateral SNE measures. All variables were standardized to address scale differences. We performed three CVAs to examine reduced voxel factors plus SNE and language behaviours at different levels (Figure 4): The first CVA (Figure 4 first row) comprised of SNE and voxel factors with all language behaviours, which reached statistical significance (χ²=45.1, df=28, p = 0.021), was predominantly loaded on SNE and high-level language behaviours. The second CVA (Figure 4 second row) comprised SNE and voxel factors with high-level language behaviours. The result revealed a significant relationship (χ²=43, df=20, p = 0.0021), with SNE showing higher loading than voxel factors. The third CVA (Figure 4 third row) with SNE and voxel factors together with low-level language behaviours did not yield significant results (χ²=5.5, df=8, p = 0.708). SNE analyses revealed significant relationships between bilateral language and MD networks and spoken language abilities. Two post-hoc analyses utilizing (i) VBCM, identified focal left hemisphere lesion correlates of single-word processing deficits; and (ii) CVA, demonstrated that SNE bilateral network measures were more sensitive in capturing higher-level connected language behaviours than word-level behaviours. Discussion Our study aimed to investigate the relationship between SNE measures in bilateral language and MD networks and connected speech in PWA. Complementing localized lesion-deficit mappings derived from voxel-based methods we predicted a correlation between SNE measures in bilateral language and MD brain networks and spoken language abilities in PWA. The connected speech measure- bigram collocation, was associated with bilateral language and MD SNE brain measures. This suggests a complex interplay between linguistic and cognitive processes underlying speech function in PWA (Hartwigsen et al., 2024). In line with this result, sentence-level speech comprehension abilities and connected speech bigram frequency correlated with bilateral language SNE measures. These ‘higher-level’ speech production and comprehension abilities were demonstrated using SNE, as opposed to classic voxel-based methods, suggesting that network-based approaches may have a greater sensitivity when mapping more complex language and cognitive processes which may be more widely distributed in the damaged brain. This provides a valuable complement to the more focal, often word-level behavioral associations, established using VLSM to date. Connected speech production in language and multiple-demand networks Closer inspection of PWA's residual SNE brain measures and their speech performance found negative correlations between bilateral SNE measures and FLAT measures of bigram frequency and collocation. The bigram frequency negatively correlated with SNE measures in bilateral language networks suggests that generating less used lexical sequences (lower bigram frequency score) places greater demands on lexical processing, strengthening connections in bilateral language networks (Gow, 2012), while the bigram collocation measure also correlated with bilateral MD networks. Producing less common word combinations (lower bigram collocation score) likely recruits broader cognitive resources, engaging both language and MD cognitive processes. That the right hemisphere has a role to play in spoken language production is not new. Previous groups have proposed its role in maintaining connections between less semantically overlapping concepts and in processing less formulaic language (Van Lancker Sidtis & Postman, 2006). For PWA, given their language impairments and left hemisphere damage, they likely rely more heavily on residual bilateral cognitive control mechanisms to compensate for their deficits. The association of MD networks (Fedorenko & Thompson-Schill, 2014) with producing novel word combinations (as indexed by bigram collocation) suggests that this sample of PWA does likely recruit additional cognitive resources/networks to support flexible and creative spoken language use (Zimmerer et al., 2020). Notably, the CAT picture description scores and FLAT speech connectivity measure correlated exclusively with solely left, not bilateral, language network SNE measures. This is consistent with the core aspects of generating precise words (information carrying units) and grammatically correct speech still relying heavily on left-lateralized language networks (Bradshaw et al., 2017; Połczyńska et al., 2014), and why lesions here result in aphasia. Reflecting on those results, when describing complex scenes, which is likely harder for PWA due to left damaged networks may recruit bilateral and MD resources to support coherent speech production. Sentence speech comprehension in bilateral language networks In our PWA sample, efficient residual structural network connections between bilateral language regions correlated with better sentence-level comprehension as opposed to word comprehension abilities. This correlation likely reflects the increased cognitive demands associated with processing more complex linguistic structures, in line with the integration of language and working memory processes (Siegel et al., 2016). Sentence comprehension places significant demands on working memory components, such as maintaining multiple items and serial order information (Majerus, 2018). These working memory components are consistently engaged in demanding sentence-level language assessments, which require online maintenance of increasing semantic and syntactic information for comprehension (Leff et al., 2009; Majerus, 2018). For PWA, to support the maintenance and manipulation of linguistic information during sentence processing, efficient working memory integration may be particularly important, to compensate for their language deficits (Murray, 2012; Salis et al., 2015). The bilateral language network templates (fROIs) (Fedorenko et al., 2011) that we used here are large and arguably incorporate regions also associated with verbal working memory processes (Kalm & Norris, 2014; Majerus, 2018). This may be due to the nature of the language localizer tasks used to define the templates (Fedorenko et al., 2011, 2012) or the inherent integration of verbal working memory in language processing. Supporting the latter interpretation, a recent fMRI study showed working memory demands during naturalistic speech comprehension primarily engaged the same bilateral language-template regions (Shain et al., 2022). This bilateral engagement may provide a mechanism for compensation from acquired brain damage in PWA, allowing the less affected hemisphere to support speech comprehension (Fleming et al., 2020). Efficient bilateral connections potentially facilitate the exchange and integration of semantic and syntactic information, enhancing PWAs’ ability to understand meanings from complex inputs despite damage to specific language areas. Distributed and focal findings: SNE and Voxel-based analyses Post-hoc analyses revealed distinct neural substrates for basic and complex language skills. Through a series of VLSM and CVA analyses, we identified differing patterns of association between brain integrity and word-level and high-level language processes. SNE analyses were sensitive to high-level language behaviors compared to word-level ones, as evidenced by the significant high-level language CVA (p = 0.0021) versus the non-significant word-level language CVA (p = 0.708). SNE method’s particular utility appears to be in capturing distributed neural substrates associated with more complex language processes. This aligns with the notion that complex cognitive functions rely on distributed neural networks rather than isolated brain regions, echoing recent findings by Billot et al. (2022) on the effects of white matter disconnections in chronic PWA. The VLSM results identified specific regions in word comprehension and naming, which align with previous literature on the neuroanatomical basis of semantic and phonological processes (Butler et al., 2014; Schumacher et al., 2019). Unlike the higher-level language functions associated with our SNE factors, those word-level language abilities appear more closely tied to localized brain areas and focal damage. Taken together, these findings underscore the complementary nature of SNE and voxel-based approaches. While VLSM may be sensitive to focal lesions affecting more fundamental language functions, SNE appears to be a valuable measure for capturing the distributed nature of higher-level language processing. This dissociation highlights the importance of employing multiple analytical approaches in studying the neural basis of language function in the damaged brain. Methodological considerations In terms of the structural organization of post-stroke brain networks, our data suggest that the efficiency of residual path connections (possibly an index of plasticity) may be critical for cognitive function. This highlights the importance of considering structural network-level properties in understanding language performance after stroke. Our analytical framework is based on T1 volumetric brain images, which are widely available and routinely collected in clinical settings, making this approach especially useful and easy to apply. We used Fedorenko's functional ROIs, as they are currently from the largest openly available participant cognitive, including language, localizer dataset. Importantly, our framework can be tailored to incorporate any anatomical parcellation schemes and functional ROIs (we acknowledge there are many alternatives available), based on researcher’s specific needs. Future work could also investigate the impact of different parcellation schemes and thresholding approaches on SNE analyses. Clinically, our findings also emphasize the importance of assessing spoken language abilities using broader linguistic tools, such as FLAT, that integrate multiple cognitive functions and may be more sensitive to capturing more complex speech use and difficulties following aphasic stroke. We selected bigram patterns and syntactic coherence from FLAT; however, future studies incorporating additional linguistic measures will be instrumental in advancing our understanding of connected speech and its relationship with brain network metrics. Evaluating both language-specific and broader cognitive brain networks and behaviour will provide a more comprehensive understanding of an individual's language capabilities post-stroke. Conclusion In this study of chronic PWA, good speech performance was associated with efficient residual structural brain connections across bilateral language and multiple cognitive demand networks. Post-hoc analyses highlighted how SNE measures capture higher-level speech and cognitive integration, with analyses of focal lesion patterns most sensitive to word-level deficits. These results illustrate how spoken language research can be grounded in a broader approach that integrates cognitive behaviours and structural brain network theories. This framework signals a new perspective to long-term speech performance after aphasic stroke- it depends on efficient bilateral network structure and could extend beyond language to include other cognitive domains. This approach could potentially serve as a tool for assessing PWAs’ brain integrity to (a) predict individual behaviors using common SNE parameters, (b) explain individual differences via identified network sources (i.e. different functional regions), and (c) clarify recovery patterns and treatment response. Overall, the network perspective proposed here advances our understanding of PWA, shifting emphasis from regional damage to efficient residual circuits, providing insights to future brain research in understanding language recovery mechanisms. Methods Participants A total of 36 PWA in the chronic post-stroke phase took part. All were native English speakers with a single left-hemisphere stroke at least 12 months prior to taking part in the study. The key inclusion criteria were that each participant had 1) anomia as assessed by the naming subtest of the Comprehensive Aphasia Test (CAT; Swinburn, 2004), 2) relatively good single word repetition and comprehension as per the subtests of the CAT (Swinburn et al., 2004) and 3) no evidence of speech apraxia as assessed by the Apraxia Battery for Adults (Dabul, 2000). All had good functional hearing and visual acuity, no prior neurological or psychiatric disorders and no contraindications for MRI. The lesion overlap map illustrates the PWAs’ pattern of left hemisphere brain damage (Figure 5). The PWAs’ demographics are described in Table 2. Participants were recruited at University College London and the study was approved by the Central London Research Ethics Committee, UK. The PWA in the current sample are the same participants reported in the study of Akkad and colleagues (Akkad et al., 2023) that used voxel-based correlational methodology (VBCM) to investigate the neural correlates of language and domain-general cognitive deficits in PWA. Behavioural data All PWA completed a comprehensive assessment of spoken language abilities. This included a standardized assessment using subtests of the comprehensive aphasia test (CAT; Swinburn et al., 2004), as well as custom-made spoken picture description tasks. The custom tasks included a greater number of items and more complex scenes, compared to the CAT, allowing for larger speech samples to enable a comprehensive analysis of connected speech performance. Details of custom picture description tasks can be found in Supplementary Figure 1 – Figure 4. Connected speech - FLAT frequency, collocation and connectivity score To quantify PWAs’ connected speech performance, linguistic measures were computed utilizing the Formulaic Language Analysis Tool (FLAT, Zimmerer et al., 2017). FLAT is a software program that quantifies usage-frequency based language variables for word combinations. It extracts frequencies by looking up each unit within transcriptions from spoken language in the British National Corpus (BNC, 2007), which represents everyday communication. FLAT values cover several distinct linguistic dimensions such as: frequency, novelty, and connectivity. Here, we focused on the frequency of unique bigrams (i.e. 2 words combinations which only appear once per sample – no repetition). Lower frequency word combinations generally indicate greater lexical capacity (Wray, 2011). The novelty dimension reflects the linguistic creativity within the sample, measured by collocation strength. Collocation strength refers to the occurrence of two words together relative to how often each individual word appears. Collocation strength is computed as bigram t-scores. Stronger collocations are likely processed in a holistic, formulaic manner (i.e. retrieved as one unit; Wray, 2011). Formulaic language poses fewer demands to lexical and syntactic processes (Zimmerer et al., 2018). Overreliance on formulaic language can therefore suggest a diminished capacity to generate rare or novel expressions (Bello-Lepe et al., 2024), which hinders communication. FLAT measures connectivity of language production by computing the proportion of words occurring in grammatical trigrams (rather than single word or two-word chunks). The ability to produce higher proportions of grammatically meaningful word sequences reflects stronger skills in sentence formulation and syntactic processing (Zimmerer et al., 2016, 2020). The performance of PWA on standardised tests of CAT subtests and FLAT scores of our custom picture description task can be found in Table 3. Table 2. PWA demographic and clinical data ID Gender Age (years) Education (years) Time Post Stroke (years) Lesion Volume () 1 M 55 16 11 61.53 2 M 56 11 7 160.80 3 M 71 13 2 42.76 4 M 55 11 9 57.24 5 M 71 16 2 78.36 6 F 51 13 13 43.42 7 M 47 13 12 161.81 8 F 66 16 18 83.18 9 M 61 16 4 63.56 10 M 44 16 3 38.14 11 F 44 17 1 29.49 12 F 70 11 12 8.93 13 M 70 11 29 117.55 14 M 69 16 7 171.71 15 M 73 11 11 71.31 16 F 45 11 3 63.53 17 F 53 11 5 22.25 18 F 55 13 5 1.51 19 M 40 17 8 163.75 20 M 64 13 24 308.18 21 M 42 17 1 65.80 22 M 74 16 11 164.53 23 M 63 16 25 156.91 24 M 64 16 10 348.23 25 M 60 16 11 94.32 26 F 60 11 6 223.29 27 M 75 11 12 112.55 28 F 50 16 3 130.60 29 M 64 11 8 240.39 30 M 29 13 6 78.61 31 F 81 10 16 99.38 32 M 60 13 20 403.11 33 M 65 13 14 387.17 34 M 82 13 34 152.61 35 M 58 16 8 239.29 36 M 39 17 2 95.70 Avg(SD) 26 M/10 F 59 (12.51) 13.8 (2.38) 10.36 (8) 131.7 (79.89) Thirty six participants include 26 males(M) and 10 females (F): mean age = 59 years [standard deviation (SD) = 12.51], range = 29-82 years; mean years of education = 13.8 years (SD= 2.38) , range = 10-17 years. Lesion volume were estimated by the automated lesion identification method (Seghier et al., 2008) and mean time post stroke = 10.36 years (SD=8), range =1-34 years; and average lesion volume was 131.7. Table 3. Standardised assessment and FLAT connected speech scores of custom picture description task. CAT Spoken Comprehension CAT Speech Production FLAT scores of custom picture description ID Words Sentences Naming Picture Des Bigram Fre Bigram Col Connectivity 1 17 25 41 20 1973.69 15.01 0.47 2 26 19 13 14 2205.73 15.96 0.38 3 22 28 38 8 1058.86 9.06 0.51 4 30 18 26 8 1932.78 12.09 0.13 5 30 28 42 38 1269.23 10.96 0.68 6 26 28 38 40 1545.17 11.62 0.58 7 30 13 30 7 2375.66 13.97 0.34 8 22 16 30 19 1651.71 12.02 0.34 9 20 13 17 5 2409.99 21.43 0.44 10 28 18 34 20 1571.66 14.30 0.43 11 26 16 32 24 1600.73 13.45 0.52 12 25 25 35 26 1152.58 10.43 0.48 13 27 21 41 24 2277.86 17.99 0.57 14 30 30 29 13 1182.57 9.50 0.50 15 29 13 40 24 1773.48 15.75 0.51 16 29 23 41 22 1371.73 12.96 0.61 17 28 14 35 13 1620.55 13.10 0.47 18 30 24 39 32 1333.08 11.66 0.55 19 26 32 36 25 1228.18 11.83 0.55 20 24 21 29 11 1107.79 5.75 0.16 21 27 27 37 31 1386.84 10.85 0.57 22 28 22 38 13 2020.29 14.49 0.41 23 28 24 40 27 1264.18 10.41 0.65 24 29 18 40 18 562.60 1.72 0.51 25 29 24 40 13 204.75 3.19 0.43 26 26 23 43 35 1496.96 13.23 0.60 27 28 28 41 37 1344.78 10.97 0.69 28 30 25 39 20 1131.34 10.93 0.53 29 27 27 30 4 2847.93 20.82 0.37 30 22 18 33 32 1090.68 10.40 0.59 31 27 21 37 22 1641.05 12.41 0.56 32 24 20 21 1 3620.38 21.58 0.05 33 27 22 39 8 1810.27 13.71 0.15 34 29 27 43 N/A 1744.28 15.94 0.55 35 28 27 29 N/A 2061.85 14.57 0.26 36 29 32 48 11 1610.92 11.49 0.50 Avg(SD)range 26.75 (3.08) 17-30 22.5 (5.34) 13-32 35.11 (7.47) 13-48 19.56 (10.21) 1-40 1624.5 (620.19) 204-3620 12.65 (4.19) 1-21 0.46 (0.15) 0.05-0.69 Norm Avg(SD) range 29.15 (1.35) 25-30 30.17 (1.85) 26-32 46.37 (1.6) 42-48 52.20 (18.25) 33-87 1092.94 (356.10) 691 - 2293 10.02 (3.25) 6 - 21 0.79 (0.04) 0.7-0.88 cut off 25 27 43 33 1804 16.52 0.7 The table's bottom two rows present patient data and normative data for healthy individuals: average score (Avg), standard deviation (SD), and range. Normative data for CAT is sourced from the CAT manual (Swinburn et al., 2004), while FLAT normative data was obtained from 18 healthy participants completing identical composite picture tasks (see Supplementary S2). Bold values indicate scores below cutoff: for CAT, these are as suggested by the manual; for FLAT, bold values in frequency and collocation are above average plus two standard deviations, while for connectivity, they are below average minus two standard deviations. Abbreviation: Picture Des- picture description; Bigram Fre- bigram frequency; Bigram Col- bigram collocation; FLAT- Formulaic Language Analysis Tool (Zimmerer et al., 2017). N/A values were replaced with the mean for subsequent statistical analyses. MRI data processing – lesion quantification Structural T1-weighted whole brain MRI scans were acquired on a 3T Siemens TIM-Trio system at the Wellcome Centre for Human Neuroimaging. Images were registered into standard Montreal Neurological Institute (MNI) space using a modified unified segmentation–normalisation procedure (Seghier et al., 2008). Images were smoothed with an 8mm full-width at half-maximum (FWHM) Gaussian kernel. To match anatomical parcellations for later lesion quantification measurements, the lesion masks were binarised from these images, re-sliced to 1 mm isotropic voxel dimension and re-sampled to image dimensions of 181x217x181 in nifti file format. All pre-processed procedures were computed within Statistical Parametric Mapping software (SPM 12) running under MATLAB 2022a. The segmented lesion masks were then used as inputs in the Lesion Quantification Toolkit (Griffis et al., 2021). MRI data- structural network efficiency (SNE) To investigate the relationship between the lesioned brain’s residual structural connectivity and PWAs’ speech production performance, we used SNE to quantify the connectivity between regions within two functional brain networks supporting language-specific and multiple-demand cognitive processes. The two functional brain networks were selected based on the work by Fedorenko and colleagues (Fedorenko, 2014; Fedorenko et al., 2010, 2012). The data processing pipeline is outlined in Figure 6. First, the structural connectivity matrices were derived from lesion and functional regions of interest (fROI) masks (see Figure 6 a-b) using Automated Anatomical Labeling parcellations (AAL; Tzourio-Mazoyer et al., 2002) and HCP-842 tractography atlas (Yeh et al., 2018) in the Lesion Quantification Toolkit (Griffis et al., 2021). The AAL parcellation offers a well-established and widely adopted anatomical parcellation scheme (Lawrence et al., 2014; Liu et al., 2021; Rubinov & Sporns, 2010; Yeh et al., 2018), facilitating comparison across studies and promoting reproducibility. With 116 cortical regions, it provides a reasonable balance between anatomical detail and computational feasibility. The HCP-842 tractography atlas was chosen for its high-resolution and extensive coverage of white matter pathways (Griffis et al., 2021; Yeh et al., 2022). To maintain consistency with Fedorenko's lab's fROI definitions while working in AAL space, we quantified the overlap between their fROIs and AAL parcels, incorporating all overlapping regions except those with minimal overlap (less than 10% of the total area). Details of AAL fROI nodes and their corresponding proportions can be found in Supplementary Figure 5 and Supplementary Table 7. The residual structural connectivity matrix comprised intact white matter connections obtained from the residual tract disconnection between each AAL brain parcel (Figure 6c). After constructing the residual structural connectivity matrices, brain structural parcellations were designated as nodes, with white matter connections exceeding a 70% threshold binarised to 1 and utilised as edges. The threshold was selected based on prior studies (Griffis et al., 2021; Wang et al., 2015; Zalesky et al., 2010) as it balances network sparsity, reducing weak or spurious links while preserving the brain network's core structure. Nodes and edges within functional regions of interest (fROIs) formed network graphs (Figure 6d and 6e). The shortest path is the minimum structural connection(s) needing traversal from one region to another. Graph theory defines efficiency as the inverted shortest path length between two nodes (Kaiser, 2011). Thus, efficiency ranges from 0 to 1, with higher values indicating greater efficiency. We implemented the breadth-first search algorithm (Rubinov & Sporns, 2010) to calculate shortest path. Structural network efficiency, SNE, was defined as the mean efficiency across all nodes within the fROI. Figure 7 illustrates a graph including 3 nodes and its SNE. The residual structural connectivity matrix was calculated by Lesion Quantification Toolkit (Griffis et al., 2021) running under MATLAB 2019b and the SNE was measured by Networkx package running under Python 3.9 (see Supplementary S4). Spearman correlation To investigate whether higher residual SNE within fROIs correlated with better preserved behavioural functions in PWA at the group level, we examined the correlation between individual behavioural scores and SNE in both lateralised and bilateral fROIs. Given the PWA brain data did not always satisfy the assumption of normal distribution we used two-tailed Spearman's rank correlation coefficients consistent with prior brain-behaviour mapping studies (Bertoux et al., 2020; Schumacher et al., 2019). For greater lesion-behavior statistical power (Rorden et al., 2007) we used p-value correction based on 10,000 random permutation tests and reported at p < 0.05 and p < 0.01 alpha threshold. Lesion volume, estimated by the automated lesion identification method (Seghier et al., 2008), was included as a covariate in the statistical models (Price et al., 2017). Analyses were run with SciPy under Python 3.9. Declarations Data availability The data described in this study is available to accredited researchers from J. C. - on request. Acknowledgements We are grateful to Karl Friston and Peter Zeidman for their valuable guidance during the SPM Methods Clinic meeting, with regard to the Canonical Variates Analysis (CVA) methodologies. Authors contributions P. J. D. and J. C. performed the research; P. J. D. conducted brain image quantification and analysis of SNE; I. S. conducted FLAT connected speech quantification; P. J. D. and J. C. wrote the original draft; V. C. Z., T. H., and A. L. reviewed methods and results; H. A. preprocessed brain image data; all authors reviewed and approved the final manuscript. Funding This research was funded in part, by the Wellcome Trust [203147/Z/16/Z and106161/Z/14/Z J.C]. For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. The funders had no participation in the design and results of this study. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References Akkad H, Hope TMH, Howland C, Ondobaka S, Pappa K, Nardo D, Duncan J, Leff AP, Crinion J (2023) Mapping spoken language and cognitive deficits in post-stroke aphasia. NeuroImage: Clin 39:103452 Barbieri E, Mack J, Chiappetta B, Europa E, Thompson CK (2019) Recovery of offline and online sentence processing in aphasia: Language and domain-general network neuroplasticity. Cortex 120:394–418 Bassi MS, Iezzi E, Gilio L, Centonze D, Buttari F (2019) Synaptic Plasticity Shapes Brain Connectivity: Implications for Network Topology. Int J Mol Sci 20(24):6193 Bates E, Wilson SM, Saygin AP, Dick F, Sereno MI, Knight RT, Dronker NF (2003) Voxel-based lesion–symptom mapping. Nat Neurosci 6(5):448–450 Bello-Lepe S, Mahmood S, Varley R, Zimmerer V (2024) Speech pauses in speakers with and without aphasia: A usage-based approach. Cortex 178:287–298 Bertoux M, Duclos H, Caillaud M, Segobin S, Merck C, De La Sayette V, Belliard S, Desgranges B, Eustache F, Laisney M (2020) When affect overlaps with concept: emotion recognition in semantic variant of primary progressive aphasia. Brain 143(12):3850–3864 Billot A, Thiebaut de Schotten M, Parrish TB, Thompson CK, Rapp B, Caplan D, Kiran S (2022) Structural disconnections associated with language impairments in chronic post-stroke aphasia using disconnectome maps. Cortex 155:90–106 Bradshaw AR, Thompson PA, Wilson AC, Bishop DVM, Woodhead ZV (2017) J. Measuring language lateralisation with different language tasks: A systematic review. PeerJ 10:e3929 Bridges KA, Van Sidtis L (2013) Formulaic language in Alzheimer’s disease. Aphasiology 27(7):799–810 Butler RA, Ralph MAL, Woollams AM (2014) Capturing multidimensionality in stroke aphasia: mapping principal behavioural components to neural structures. Brain 137(12):3248–3266 Carrera E, Tononi G (2014) Diaschisis: past, present, future. Brain 137(9):2408–2422 Chen AC, Oathes DJ, Chang C, Bradley T, Zhou ZW, Williamsa LM, Glover GH, Deisseroth K, Etkin A (2013) Causal interactions between fronto-parietal central executive and default-mode networks in humans. Proc Natl Acad Sci USA 110(49):19944–19949 Cheng B, Schlemm E, Schulz R, Boenstrup M, Messé A, Hilgetag C, Gerloff C (2019) & Tz Thomalla, G. Altered topology of large-scale structural brain networks in chronic stroke. Brain Commun, 1(1) Conroy P, Sage K, Ralph ML (2009) Improved vocabulary production after naming therapy in aphasia: can gains in picture naming generalise to connected speech? Int J Lang Communication Disorders 44(6):1036–1062 Dabul BP-E (2000) ABA-2: Apraxia Battery for Adults Fedorenko E, Thompson-Schill SL (2014) Reworking the language network. Trends Cogn Sci 18(3):120–126 Fedorenko E, Varley R (2016) Language and thought are not the same thing: evidence from neuroimaging and neurological patients. Ann N Y Acad Sci 1369(1):132–153 Fleming V, Brownsett S, Krason A, Maegli MA, Coley-Fisher H, Ong YH, Nardo D, Leach R, Howard D, Robson H, Warburton E, Ashburner J, Price CJ, Crinion JT, Leff (2020) A. P. Efficacy of spoken word comprehension therapy in patients with chronic aphasia: a cross-over randomised controlled trial with structural imaging. J Neurol Neurosurg Psychiatry 92(4):418–424 Foulon C, Cerliani L, Kinkingnéhun S, Levy R, Rosso C, Urbanski M, Volle E (2018) Schotten, M. T. Advanced lesion symptom mapping analyses and implementation as BCBtoolkit. GigaScience 7(3):1–17de Fridriksson J, den Ouden DB, Hillis AE, Hickok G, Rorden C, Basilakos A, Yourganov G, Bonilha L (2018) Anatomy of aphasia revisited. Brain 141(3):848–862 Friston KJ, Frith CD, Frackowiak RSJ, Turner R (1995) Characterizing Dynamic Brain Responses with fMRI: A Multivariate Approach. NeuroImage 2(2):166–172 Gow DW (2012) The cortical organization of lexical knowledge: A dual lexicon model of spoken language processing. Brain Lang 121(3):273–288 Griffis JC, Metcalf NV, Corbetta M, Shulman GL (2020) Damage to the shortest structural paths between brain regions is associated with disruptions of resting-state functional connectivity after stroke. NeuroImage 210:116589 Griffis JC, Metcalf NV, Corbetta M, Shulman GL (2021) Lesion Quantification Toolkit: A MATLAB software tool for estimating grey matter damage and white matter disconnections in patients with focal brain lesions. NeuroImage: Clin 30:102639 Guggisberg AG, Koch PJ, Hummel FC, Buetefisch CM (2019) Brain networks and their relevance for stroke rehabilitation. Clin Neurophysiol 130(7):1098–1124 Hartwigsen G, Henseler I, Stockert A, Wawrzyniak M, Wendt C, Klingbeil J, Baumgaertner A, Saur D (2017) Integration demands modulate effective connectivity in a fronto-temporal network for contextual sentence integration. NeuroImage 147:812–824 Hartwigsen G, Lim J-S, Bae H-J, Yu K-H, Kuijf HJ, Weaver NA, Biesbroek JM, Kopal J, Bzdok D (2024) Bayesian modelling disentangles language versus executive control disruption in stroke. Brain Commun, 6(3) Hartwigsen G, Saur D (2019) Neuroimaging of stroke recovery from aphasia – Insights into plasticity of the human language network. NeuroImage , 190, 14–31 Herbert R, Hickin J, Howard D, Osborne F, Best W (2008) Do picture-naming tests provide a valid assessment of lexical retrieval in conversation in aphasia? Aphasiology , 22 (2), 184–203. Heuer S, Ivanova MV, Hallowell B (2017) More Than the Verbal Stimulus Matters: Visual Attention in Language Assessment for People With Aphasia Using Multiple-Choice Image Displays. J Speech Lang Hear Res 60(5):1348–1361 Hickok G, Poeppel D (2007) The cortical organization of speech processing. Nat Rev Neurosci 8(5):393–402 Hope TMH, Leff AP, Price CJ (2018) Predicting language outcomes after stroke: Is structural disconnection a useful predictor? NeuroImage: Clin 19:22–29 Kaiser M (2011) A tutorial in connectome analysis: Topological and spatial features of brain networks. Neuroimage , 57 (3), 892–907 Kalm K, Davis MH, Norris D (2012) Neural mechanisms underlying the grouping effect in short-term memory. Hum Brain Mapp 33(7):1634–1647 Kalm K, Norris D (2014) The Representation of Order Information in Auditory-Verbal Short-Term Memory. J Neurosci 34(20):6879–6886 Kearney E, Guenther FH (2019) Articulating: the neural mechanisms of speech production. Lang Cognition Neurosci 34(9):1214–1229 Keil K, Kaszniak AW (2010) Examining executive function in individuals with brain injury: A review. Aphasiology 16(3):305–335 Lawrence AJ, Chung AW, Morris RG, Markus HS, Barrick TR (2014) Structural network efficiency is associated with cognitive impairment in small-vessel disease. Neurology 83(4):304–311 Leff AP, Schofield TM, Crinion JT, Seghier ML, Grogan A, Green DW, Price CJ (2009) The left superior temporal gyrus is a shared substrate for auditory short-term memory and speech comprehension: evidence from 210 patients with stroke. Brain 132(12):3401–3410 Liu T, Thiebaut de Schotten M, Altarelli I, Ramus F, Zhao J (2021) Maladaptive compensation of right fusiform gyrus in developmental dyslexia: A hub-based white matter network analysis. Cortex 145:57–66 Majerus S (2018) Working memory treatment in aphasia: A theoretical and quantitative review. J Neurolinguistics 48:157–175 Marebwa BK, Fridriksson J, Yourganov G, Feenaughty L, Rorden C, Bonilha L (2017) Chronic post-stroke aphasia severity is determined by fragmentation of residual white matter networks. Sci Rep, 7(1) Marini A, Galetto V, Zampieri E, Vorano L, Zettin M, Carlomagno S (2011) Narrative language in traumatic brain injury. Neuropsychologia 49(10):2904–2910 Martin N, Reilly J (2012) Short-term/working memory impairments in aphasia: Data, models, and their application to aphasia rehabilitation. Aphasiology 26(3–4):253–257 Murray LL (2012) Direct and indirect treatment approaches for addressing short-term or working memory deficits in aphasia. Aphasiology 26(3–4):317–337 Perkins L (1995) Applying conversation analysis to aphasia: Clinical implications and analytic issues. Eur J Disord Commun 30(3):372–383 Połczyńska M, Curtiss S, Walshaw P, Siddarth P, Benjamin C, Moseley BD, Vigil C, Jones M, Eliashiv D, Bookheimer S (2014) Grammar tests increase the ability to lateralize language function in the Wada test. Epilepsy Res 108(10):1864–1873 Price CJ, Hope TM, Seghier ML (2017) Ten problems and solutions when predicting individual outcome from lesion site after stroke. NeuroImage 145:200–208 Rorden C, Karnath H-O, Bonilha L (2007) Improving Lesion-Symptom Mapping. J Cogn Neurosci 19(7):1081–1088 Rubinov M, Sporns O (2010) Complex network measures of brain connectivity: Uses and interpretations. NeuroImage 52(3):1059–1069 Salis C, Kelly H, Code C (2015) Assessment and treatment of short-term and working memory impairments in stroke aphasia: A practical tutorial. Int J Lang Communication Disorders 50(6):721–736 Schumacher R, Halai AD, Lambon Ralph MA (2019) Assessing and mapping language, attention and executive multidimensional deficits in stroke aphasia. Brain 142(10):3202–3216 Seghier ML, Ramlackhansingh A, Crinion J, Leff AP, Price CJ (2008) Lesion identification using unified segmentation-normalisation models and fuzzy clustering. NeuroImage 41(4):1253–1266 Shain C, Blank IA, Fedorenko E, Gibson E, Schuler W (2022) Robust Effects of Working Memory Demand during Naturalistic Language Comprehension in Language-Selective Cortex. J Neurosci 42(39):7412–7430 Siegel, J. S., Ramsey, L. E., Snyder, A. Z., Metcalf, N. V., Chacko, R. V., Weinberger,K., … Corbetta, M. Disruptions of network connectivity predict impairment in multiple behavioral domains after stroke. Proceedings of the National Academy of Sciences, 113(30), E4367-E4376 (2016). Stefaniak JD, Alyahya RSW (2021) Lambon Ralph, M. A. Language networks in aphasia and health: A 1000 participant activation likelihood estimation meta-analysis. NeuroImage 233:117960 Swinburn K (2004) Comprehensive aphasia test. CAT, Psychology Termenon M, Achard S, Jaillard A, Delon-Martin C (2016) The hub disruption index, a reliable index sensitive to the brain networks reorganization. A study of the contralesional hemisphere in stroke. Front Comput Neurosci 10:84 Turkeltaub PE, Messing S, Norise C, Hamilton RH (2011) Are networks for residual language function and recovery consistent across aphasic patients? Neurology 76(20):1726 Tyler LK, Marslen-Wilson WD, Stamatakis EA, Tyler LK, Marslen-Wilson W (2005) Dissociating neuro-cognitive component processes: Voxel-based correlational methodology Age-Related Delay in Visual and Auditory Evoked Responses is Mediated by White-and Gray-matter Differences View project Ventral stream patients View project Dissociating neuro-cognitive component processes: voxel-based correlational methodology. Neuropsychologia 43:771–778 Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, Joliot M (2002) Automated Anatomical Labeling of Activations in SPM Using a Macroscopic Anatomical Parcellation of the MNI MRI Single-Subject Brain. NeuroImage 15(1):273–289 Van Sidtis L, D., Postman WA (2006) Formulaic expressions in spontaneous speech of left- and right‐hemisphere‐damaged subjects. Aphasiology 20(5):411–426 Wang J, Wang X, Xia M, Liao X, Evans A, He Y (2015) GRETNA: A graph theoretical network analysis toolbox for imaging connectomics. Frontiers in Human Neuroscience , 9 (JUNE), 147509 Wray A (2011) Formulaic Language as a Barrier to Effective Communication with People with Alzheimer’s Disease. Can Mod Lang Rev 67(4):429–458 Yeh FC (2022) Population-based tract-to-region connectome of the human brain and its hierarchical topology. Nat Commun 13(1):4933 Yeh FC, Panesar S, Fernandes D, Meola A, Yoshino M, Fernandez-Miranda JC, Vettel JM, Verstynen T (2018) Population-averaged atlas of the macroscale human structural connectome and its network topology. NeuroImage 178:57–68 Zalesky A, Fornito A, Bullmore ET (2010) Network-based statistic: Identifying differences in brain networks. NeuroImage 53(4):1197–1207 Zimmerer VC, Hardy CJD, Eastman J, Dutta S, Varnet L, Bond RL, Russell L, Rohrer JD, Warren JD, Varley RA (2020) Automated profiling of spontaneous speech in primary progressive aphasia and behavioral-variant frontotemporal dementia: An approach based on usage-frequency. Cortex 133:103–119 Zimmerer VC, Newman L, Thomson R, Coleman M, Varley RA (2018) Automated analysis of language production in aphasia and right-hemisphere damage: frequency and collocation strength. Aphasiology 32(11):1267–1283 Zimmerer VC, Wibrow M, Varley RA (2016) Formulaic Language in People with Probable Alzheimer’s Disease: A Frequency-Based Approach. J Alzheimer’s Disease 53(3):1145–1160 Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryNC.docx Supplementary Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5587181","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":390096891,"identity":"62f27fd0-8519-4f6f-9058-b62f23b15350","order_by":0,"name":"Ping-Jung Duh","email":"data:image/png;base64,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","orcid":"","institution":"Institute of Cognitive Neuroscience, University College London","correspondingAuthor":true,"prefix":"","firstName":"Ping-Jung","middleName":"","lastName":"Duh","suffix":""},{"id":390096892,"identity":"39442903-e0ec-4c07-b372-f119ad170c1b","order_by":1,"name":"Ioana Sederias","email":"","orcid":"","institution":"Clinical Neurosciences, University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Ioana","middleName":"","lastName":"Sederias","suffix":""},{"id":390096893,"identity":"2b69ddf7-a85d-4540-a0e0-9953214b8723","order_by":2,"name":"Vitor Zimmerer","email":"","orcid":"","institution":"Language and Cognition, University College London","correspondingAuthor":false,"prefix":"","firstName":"Vitor","middleName":"","lastName":"Zimmerer","suffix":""},{"id":390096894,"identity":"89bcab0e-c7d4-4803-a330-01b49a342a2d","order_by":3,"name":"Haya Akkad","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Haya","middleName":"","lastName":"Akkad","suffix":""},{"id":390096895,"identity":"e4df0ae5-52fe-485e-8ca1-397e65a76b53","order_by":4,"name":"Alex Leff","email":"","orcid":"","institution":"Institute of Cognitive Neuroscience, University College London","correspondingAuthor":false,"prefix":"","firstName":"Alex","middleName":"","lastName":"Leff","suffix":""},{"id":390096896,"identity":"d8001660-b2c7-4529-93c8-9ae476741cda","order_by":5,"name":"Thomas Hope","email":"","orcid":"https://orcid.org/0000-0003-0714-8545","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Hope","suffix":""},{"id":390096897,"identity":"b1f874cf-0f30-433d-b960-f34b830d5e0e","order_by":6,"name":"Jenny Crinion","email":"","orcid":"https://orcid.org/0000-0001-8080-6562","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Jenny","middleName":"","lastName":"Crinion","suffix":""}],"badges":[],"createdAt":"2024-12-05 13:05:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5587181/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5587181/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":71624906,"identity":"d46bd548-a90a-4557-a20a-75cab1dd5980","added_by":"auto","created_at":"2024-12-17 08:42:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":107331,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAn example illustrating the difference between streamline-based and voxel-based tract disconnection methods.\u003c/strong\u003e How a lesion is spatially localized across (intersects with) a tract will result in different outcomes from the two measurement approaches. The same five voxels out of 60 in the tract were lesioned, but the severity measures differed greatly. The streamline-based disconnection method showed 83% severity (left), while the voxel-based lesion load method only showed 8.3% severity (right).\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/ef6391bbceed5cd3d4b43fb7.png"},{"id":71627084,"identity":"e206f8a6-20fa-4b96-9700-e24650d63004","added_by":"auto","created_at":"2024-12-17 08:58:04","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":364627,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelations between structural network efficiency (SNE) and speech production performance in PWA. \u003c/strong\u003e(a):\u003cstrong\u003e \u003c/strong\u003eHeatmap shows Spearman's correlation coefficients (r) between structural network efficiency measures and behavioral scores. Color indicates correlation strength and direction (red = positive, blue = negative) for bilateral, left hemisphere, and right hemisphere language and multiple demand (MD) networks, with darker shades representing stronger correlations. Values marked with * are significant at corrected p \u0026lt; 0.05, with ** are significant at corrected p \u0026lt; 0.01. (b): Visualization of brain-behaviour correlations overlaid on cortical surface using SurfIce software. Right: functional networks of interest Brain rendering: Visualization of overlaid correlations in SurfIce (green- bilateral MD networks; orange- left language network; yellow\u0026amp;orange- bilateral language networks); Left: a visual representation of how SNE in each functional network relates to distinct measures of connected speech. WordComp = word comprehension; SenComp = sentence comprehension; PicDes = picture description; BigramFre = bigram frequency; BigramCol = bigram collocation.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/b952ff490cc4e7609a497b6a.png"},{"id":71623184,"identity":"739826dc-620c-4249-a4fb-723315d4795e","added_by":"auto","created_at":"2024-12-17 08:34:05","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":276048,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNeural correlates of word-level speech production deficits identified by VBCM. \u003c/strong\u003eResults are shown at p ≤ 0.001 voxel-level, p \u0026lt; 0.05 FWE) corrected at cluster-level. Colors indicate regions where gray matter damage correlates with impairments in different word-level speech production processes and sentence repetition: blue = naming deficits, yellow = spoken word comprehension deficits. Results are overlaid on a rendered brain template (MRIcro-GL).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/45790041d1445608cf4d3829.png"},{"id":71623163,"identity":"f0ccedd3-85f2-4a10-8189-4c23ccdaf2fa","added_by":"auto","created_at":"2024-12-17 08:34:04","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":198498,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCVA results: SNE and voxels with behaviour variables at different levels. \u003c/strong\u003eThis figure presents CVA results for language behavior at three levels: all, high, and word level. Each row represents a behavior level, with columns showing: (a) p-values from chi-squared tests for canonical variates (CVs), indicating significance (* p \u0026lt; 0.05, ** p \u0026lt; 0.01). The number of CVs is determined by the smaller of brain factors or behavior variables.; (b) brain factor loadings on the first CV, which accounts for the most variance.; and (c) bar charts illustrating the contribution of each behavior variable to the first canonical contrast. Abbreviations: CV- canonical variate; VF- voxels factor; SNE L- SNE language; WC- word comprehension; SC- sentence comprehension; Nam- naming; PicD- picture description; BiF- bigram frequency; BiC- bigram collocation; Con- connectivity\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/9e6123844c087d98b3181853.png"},{"id":71623171,"identity":"a7714b9d-6014-4c92-8d27-490f61f54290","added_by":"auto","created_at":"2024-12-17 08:34:04","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":171314,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLesion overlap map for 36 PWA. \u003c/strong\u003eColour-scale represents frequency of regional brain damage (hot-body scale with red indicating most frequently damaged brain regions i.e., \u0026gt;18 patients, while dark blue \u0026lt; 6 patients with damage to these regions). Results are shown overlaid on the MNI template brain, created in MRIcro-GL.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/8e7604d90e83869e24cac98f.png"},{"id":71623182,"identity":"ac0e9195-ef04-493a-a769-1d5548d26045","added_by":"auto","created_at":"2024-12-17 08:34:05","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":143466,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePipeline to estimate residual SNE in functionally-defined brain networks for each individual. \u003c/strong\u003eTo estimate residual structural connectivity matrices (c) each PWA’s lesion mask (a) is entered as an input to the lesion quantification toolkit (Griffis et al., 2021). Given a functional network mask (b) as an input, the quantification toolbox calculates the overlapping brain regions in the structural parcellation (d). A SNE measurement is then calculated by building a graph (e) combining the connectivity matrix and selected ROIs together. The edges were binarized at 70% threshold to form the graph.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/b83ffe329e90635a7cda0014.png"},{"id":71623170,"identity":"706fee94-e09d-457f-a747-11da097218c0","added_by":"auto","created_at":"2024-12-17 08:34:04","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":81285,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA three-node network examplar: A, B, C.\u003c/strong\u003e Structural efficiency between nodes: A and B and between nodes: B and C are both 1 because there are direct paths between them. For node A and node C, there is no direct path between them. The shortest path from node A to node C is to go through node B first therefore, the total path value is 2. The efficiency between them is the inverted value of this path, which is 0.5. The overall SNE is the average of all efficiency values in the network, which in this exemplar is 0.83.\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/be4d079f0b1c5405a9493a68.png"},{"id":71627266,"identity":"72ca7362-1bb4-4d77-9244-c3e132782f42","added_by":"auto","created_at":"2024-12-17 09:06:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2528104,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/ca4b9db3-3ce2-4ad1-b734-5cd408d73bc5.pdf"},{"id":71623159,"identity":"cb098622-1e8c-43b3-84ee-cca7a6a59dfc","added_by":"auto","created_at":"2024-12-17 08:34:04","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2611139,"visible":true,"origin":"","legend":"Supplementary","description":"","filename":"SupplementaryNC.docx","url":"https://assets-eu.researchsquare.com/files/rs-5587181/v1/8bcd3675ac7d1fb43dd4e285.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Uncovering Hidden Pathways: Structural Brain Networks Underpinning Connected Speech in Post-Stroke Aphasia","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAphasia is an acquired language disorder most commonly caused by a stroke affecting the dominant, typically left, hemisphere. While brain lesions causing aphasia primarily affect language functions, a growing body of research indicates that other cognitive domains crucial for daily communication, including working memory and cognitive flexibility are also impacted in post-stroke patients with aphasia (PWA) (Butler et al., 2014; Schumacher et al., 2019). Language and non-language cognitive functions are interrelated, with evidence suggesting bidirectional influences on speech communication abilities with language deficits disrupting non-language cognition performance (Keil \u0026amp; Kaszniak, 2010), and cognitive impairments impacting on functional language abilities (Heuer et al., 2017; Schumacher et al., 2019). For instance, working memory deficits frequently persist in PWA after recovery on standard language tests (Martin \u0026amp; Reilly, 2012), negatively impacting their ‘real-world’ daily abilities, and in some instances limiting treatment outcomes that rely on serial order memory for word re-learning (Murray, 2012; Salis et al., 2015).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCurrent standardised measures of language ability, especially those used in routine clinical practice, primarily assess linguistic impairments in isolation and often fail to fully capture the integrated cognitive-linguistic deficits that contribute to real-world communication difficulties (Marini et al., 2011). For instance, preserved performance on a naming task may not capture a PWA’s semantic retrieval and/or verbal working memory impairments that could affect their ability to communicate more complex ideas or sequence information in conversation. Performance on more naturalistic, cognitively demanding verbal tasks such as composite picture description tasks or conversational data (Perkins, 1995) may better reflect this complexity. Composite picture description tasks are often preferred in clinical assessment of PWA as they allow for control of the target spoken language and enable comparisons across individuals and time points (Conroy et al., 2009). However, measuring PWA’s performance on this task by, for example, primarily counting the number of information carrying words produced, often fails to capture the richness of the speech data and reduces the sensitivity of the task to connected speech function i.e., it is akin to another (single-word) naming test. A proposed alternative method for quantitative speech production analysis is the Frequency in Language Analysis Tool (FLAT) (Zimmerer et al., 2018, 2020), an automated toolbox for quantification of language-based frequentist-use of spoken language features. By characterising connected speech samples at multiple linguistic levels,\u0026nbsp;including\u0026nbsp;frequency, novelty and connectivity of word combinations used, it aims to\u0026nbsp;provide novel insights into more complex spoken language performance.\u003c/p\u003e\n\u003cp\u003eIt appears that familiarity of expressions is an important factor for organization of language in the brain. Surprisingly, analyzing connected speech from a usage-based or frequentist perspective has rarely been explored in the context of PWA. A few studies in PWA and those with Alzheimer's disease, have reported reduced linguistic creativity during speech, relying heavily on familiar phrases and fixed expressions (Van Lancker Sidtis \u0026amp; Postman, 2006; Zimmerer et al., 2018, 2020). For example, in non-fluent aphasia populations, bigram (i.e., two-word combination) frequency and collocation scores show significantly high values relative to controls, which can be associated with reduced lexical-semantic, grammatical, or working memory capacity, resulting in less flexible language processing (Zimmerer et al., 2018). Overreliance on familiar combinations has been regarded a compensatory mechanism to mask linguistic impairments (Bridges \u0026amp; Van Lancker Sidtis, 2013; Wray, 2011; Zimmerer et al., 2016).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWith this interplay between language and cognitive control underpinning everyday communication behaviour, speech production likely involves coordinated activity across multiple brain regions i.e., both language-specific as well as multiple-demand regions in both hemispheres (Hickok \u0026amp; Poeppel, 2007; Kearney \u0026amp; Guenther, 2019). For most people, the left hemisphere plays a crucial role in language processing. Sentence retrieval relies on left hemisphere language areas, including the inferior frontal gyrus and superior temporal gyrus (Hartwigsen et al., 2017; Stefaniak et al., 2021). However, additional cognitive functions including holding the conversation topic in mind, organizing ideas sequentially, and flexibly shifting between themes likely engage bilateral regions classically implicated in non-verbal executive functions and working memory, such as the dorsolateral prefrontal and parietal cortices (Chen et al., 2013; Kalm et al., 2012). With right and bilateral temporo-parietal functional activation supporting residual language capacity after aphasia stroke (Turkeltaub et al., 2011).\u0026nbsp;Consistently with this, our group found increased bilateral temporal lobe grey matter tissue density associated with improved auditory comprehension abilities in chronic PWA following behavioral training (Fleming et al., 2020). Suggesting that preserved functional, and by association anatomical connectivity between hemispheres may underpin language recovery after brain damage.\u003c/p\u003e\n\u003cp\u003eThe complex interactions between multiple brain regions that support language recovery leads us to consider a broader network perspective in language processing. Cognitive impairments, including those involved in speech production, are often described in terms of network disruptions (Fedorenko \u0026amp; Varley, 2016; Hartwigsen \u0026amp; Saur, 2019). Deficits may arise from anatomically intact but disconnected brain regions within a network (Billot et al., 2022a; Fridriksson et al., 2018; Marebwa et al., 2017; Price et al., 2017). For example, structurally intact regions may be functionally impaired due to white matter disconnection from key network nodes. Residual anatomical connections of these regions could provide a neural basis for compensatory cognitive functions to support speech production (Bassi et al., 2019; Guggisberg et al., 2019).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo quantify anatomical damage associated with speech deficits in aphasia, researchers have traditionally used mass univariate approaches such as voxel-based lesion analyses. This method correlates variance in structural brain damage, e.g., from a T1 volumetric MRI brain scan, with variance in PWAs’ behavioural performance (Bates et al., 2003). However, a limitation of the voxel-based lesion method is that it treats each voxel independently and only considers regions with direct damage, which may not fully capture the distributed network disruptions underlying complex behaviors (Billot et al., 2022a; Carrera \u0026amp; Tononi, 2014). A proposed alternative and complementary approach is the analysis of brain structural connectivity, allowing for the examination of white matter pathways connecting various brain regions. Yet, this approach might overlook the complex interactions between different connectivity patterns (Foulon et al., 2018).\u003c/p\u003e\n\u003cp\u003eTo understand brain and behaviour function primarily in the healthy brain, many groups have used functional regions of interest (fROIs) defined by task-based fMRI activation patterns to parcellate language and multiple cognitive demand (MD) networks (Fedorenko et al., 2012; 2014). How residual structural connectivity between these functionally-defined nodes relates to post-stroke behavioral capacity remains unclear. Calculating the structural network efficiency (SNE) of fROIs using graph theory (Kaiser, 2011) may provide greater sensitivity and help bridge this gap by quantifying the efficiency of residual structural connections in the lesioned brain. The lesion quantification (Griffis et al., 2021) approach employs streamline-based tract disconnection severity, rather than voxel-wise tract lesion load (Hope et al., 2018). Tract lesion load estimates the proportion of damaged voxels within each tract, but it does not account for spatial distribution of lesions across streamlines. Thus, a small lesion volume could still substantially disrupt connectivity if distributed across streamlines (see Figure 1). The approach thereby directly captures tract disconnection and may better detect network disruptions relevant for brain function (Griffis et al., 2021).\u003c/p\u003e\n\u003cp\u003eBuilding on this, in this paper we examine whether the SNE approach can effectively measure the integrity of the disrupted structural network underlying speech production abilities in a group of 36 PWA. Our approach diverges from previous studies (Cheng et al., 2019; Lawrence et al., 2014;\u0026nbsp;Termenon et al., 2016) that looked at either whole-brain efficiency or local efficiency around a single region. To address the challenge of overfitted issues with our sample size, we reduced our brain variable dimensions, by focusing on two functional regions of interest networks(fROIs). These two fROIs were defined using the extensive database of normal healthy subjects’ functional activation patterns for language and multiple demand (MD) cognitive tasks available from Fedorenko and colleagues (Fedorenko, 2014; Fedorenko et al., 2010, 2012). These fROIs were chosen to capture the integrity of networks supporting language-specific and multiple-demand cognitive processes, which are often impaired in aphasia (Barbieri et al., 2019; Murray, 2012; Schumacher et al., 2019). The language fROI includes regions consistently activated during language tasks in healthy individuals, while the cognitive fROI comprises regions engaged in various demanding cognitive tasks, such as working memory and cognitive control. We then used SNE to quantify ipsilesional (left hemisphere), contralesional (right hemisphere) and bilateral connectivity with these same two functional brain networks. This enabled us to examine the impact of the left hemisphere lesions in our PWA sample (compared to the normative reference) on damaged dominant language and non-dominant (contralesional) language networks and bilateral multiple demand networks.\u003c/p\u003e\n\u003cp\u003eWe hypothesized that SNE captures brain-behaviour relationships at the network level. We predicted higher SNE values of bilateral language and MD brain networks would be associated with better spoken language abilities within the PWA group, complementing the localised lesion-deficit mappings from voxel-based method. We tested associations between variance in SNE and behavioral scores with bilateral, left hemisphere, and right hemisphere language and MD network measures, while controlling for lesion volume. To increase our sensitivity to connected speech production abilities, our analyses included 1) frequentist measures using FLAT analysis of PWAs’ speech samples from custom picture description tasks; and 2) anatomical measures of distributed brain network using SNE. \u0026nbsp;\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eBrain network correlations with connected (narrative) speech production\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this group of PWA, better performance on connected speech measures was associated with greater structural efficiency in the left language network. Specifically, FLAT connectivity measure (r = 0.46, p = 0.013) and CAT standardised picture description scores (r = 0.37, p =0.029) showed positive correlations with left hemisphere language network SNE. Complementing these findings, FLAT measures of bigram frequency and bigram collocation \u0026ndash; where lower scores reflects more creative, less formulaic language \u0026ndash; were correlated with higher SNE in both left and bilateral language networks (r = -0.49, p = 0.003 and r = -0.54, p = 0.001 respectively). The FLAT bigram collocation performance measure was the only one negatively associated with bilateral SNE measures of the MD network (r = -0.42, p = 0.013). See Figure 2 for full reported results.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBrain network correlations with sentence tasks but not\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eword-level speech production\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur analysis revealed a significant relationship between SNE and language behaviours at high-level but not at word-level language behaviours. There was a significant positive correlation between SNE of the bilateral language network and spoken sentence comprehension (r = 0.40, p = 0.018) in PWAs. The increased efficiency within residual bilateral hemisphere structural language network was associated with better spoken sentence comprehension. In contrast, word-level language processing including naming and spoken word comprehension, did not show significant correlations with SNE in any of the assessed networks.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePost-hoc 1: Voxel-based lesion symptom mapping\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate whether word-level language behaviours could be associated with focal lesions in specific brain regions, we conducted a post-hoc voxel-based correlational methodology (VBCM) (Akkad et al., 2023; Tyler et al., 2005), a variant of voxel-based lesion symptom mapping (VLSM; Bates et al., 2003). A voxel-wise F-test was used to compare the fuzzy lesion status between patients with their performance on each neuropsychological assessment. Total lesion volume was included as a covariate in the general linear model. Results are reported at p \u0026le; 0.001 voxel-level and p \u0026lt; 0.05 FWE corrected at cluster-level. All anatomical labels were based on the AAL atlas in MNI space. The analyses were run using SPM 12 with MATLAB 2022b.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe results revealed neural correlates with word-level speech production deficits (Figure 3 and Table 1). Damage to left middle temporal gyrus was correlated with impairments in naming. Lesions in the left superior parietal gyrus were associated with deficits in spoken word comprehension. No significant neural correlates were found for sentence comprehension and connected speech measures. The results complement our main findings by demonstrating that while word-level language behaviours do not show a significant relationship with SNE measures, they are associated with focal lesions in specific brain regions.\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"624\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\" style=\"width: 624px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 1. Neural correlates of word-level speech production deficits.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 161px;\"\u003e\n \u003cp\u003eBehaviour Assessments\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 170px;\"\u003e\n \u003cp\u003eBrain Location\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 111px;\"\u003e\n \u003cp\u003eExtend Voxels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 44px;\"\u003e\n \u003cp\u003eZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 139px;\"\u003e\n \u003cp\u003eMNI Coordinate\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 54px;\"\u003e\n \u003cp\u003ex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 39px;\"\u003e\n \u003cp\u003ey\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003ez\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003eCAT Naming\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003eLeft Middle\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eTemporal Gyrus\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 111px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e149\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 44px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e3.31\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 54px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e-58\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 39px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e-8\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e-20\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003eCAT Word Comprehension\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eLeft Superior \u0026nbsp;\u003c/p\u003e\n \u003cp\u003eParietal Gyrus\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 111px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e240\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 44px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e3.14\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 54px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e-26\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 39px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e-56\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003e48\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\" style=\"width: 624px;\"\u003e\n \u003cp\u003e*Only clusters with cluster-level FWE p \u0026lt; 0.05 are shown. Coordinates are in MNI space.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003ePost-hoc 2: Canonical Variate Analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate whether SNE was primarily sensitive to high-level language behaviors and provided additional brain-behaviour information beyond distributed voxels per se, we conducted a post-hoc Canonical Variate Analysis (CVA) in SPM12 (Friston et al., 1995). CVA is a multivariate method that identifies linear combinations of variables that best explain variance in dependent variables. Here, CVA was used to quantify how much language behaviour variance could be accounted for by brain measures (SNE and voxels). The resulting canonical variates, ranked by most explained shared variance, represent dimensions of maximal covariation between the brain measures and language behaviours. Chi-square tests indicate the statistical significance of the amount of shared variance that can be explained.\u003c/p\u003e\n\u003cp\u003eFirst, Singular Value Decomposition (SVD) was employed to reduce all voxels in language and MD ROIs to two factors, making them comparable to the two bilateral SNE measures. All variables were standardized to address scale differences. We performed three CVAs to examine reduced voxel factors plus SNE and language behaviours at different levels (Figure 4):\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe first CVA (Figure 4 first row) comprised of SNE and voxel factors with all language behaviours, which reached statistical significance (\u0026chi;\u0026sup2;=45.1, df=28, p = 0.021), was predominantly loaded on SNE and high-level language behaviours. The second CVA (Figure 4 second row) comprised SNE and voxel factors with high-level language behaviours. The result revealed a significant relationship (\u0026chi;\u0026sup2;=43, df=20, p = 0.0021), with SNE showing higher loading than voxel factors. The third CVA (Figure 4 third row) with SNE and voxel factors together with low-level language behaviours did not yield significant results (\u0026chi;\u0026sup2;=5.5, df=8, p = 0.708).\u003c/p\u003e\n\u003cp\u003eSNE analyses revealed significant relationships between bilateral language and MD networks and spoken language abilities. Two post-hoc analyses utilizing (i) VBCM, identified focal left hemisphere lesion correlates of single-word processing deficits; and (ii) CVA, demonstrated that SNE bilateral network measures were more sensitive in capturing higher-level connected language behaviours than word-level behaviours.\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur study aimed to investigate the relationship between SNE measures in bilateral language and MD networks and connected speech in PWA. Complementing localized lesion-deficit mappings derived from voxel-based methods we predicted a correlation between SNE measures in bilateral language and MD brain networks and spoken language abilities in PWA. The connected speech measure- bigram collocation, was associated with bilateral language and MD SNE brain measures. This suggests a complex interplay between linguistic and cognitive processes underlying speech function in PWA (Hartwigsen et al., 2024).\u0026nbsp;In line with this result, sentence-level speech comprehension abilities and connected speech bigram frequency correlated with bilateral language SNE measures. These ‘higher-level’ speech production and comprehension abilities were demonstrated using SNE, as opposed to classic voxel-based methods, suggesting that network-based approaches may have a greater sensitivity when mapping more complex language and cognitive processes which may be more widely distributed in the damaged brain. This provides a valuable complement to the more focal, often word-level behavioral associations, established using VLSM to date.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConnected speech production in language and multiple-demand networks\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCloser inspection of PWA's residual SNE brain measures and their speech performance found negative correlations between bilateral SNE measures and FLAT measures of bigram frequency and collocation. The bigram frequency negatively correlated with SNE measures in bilateral language networks suggests that generating less used lexical sequences (lower bigram frequency score) places greater demands on lexical processing, strengthening connections in bilateral language networks (Gow, 2012), while the bigram collocation measure also correlated with bilateral MD networks. Producing less common word combinations (lower bigram collocation score) likely recruits broader cognitive resources, engaging both language and MD cognitive processes.\u003c/p\u003e\n\u003cp\u003eThat the right hemisphere has a role to play in spoken language production is not new. Previous groups have proposed its role in maintaining connections between less semantically overlapping concepts and in processing less formulaic language (Van Lancker Sidtis \u0026amp; Postman, 2006). For PWA, given their language impairments and left hemisphere damage, they likely rely more heavily on residual bilateral cognitive control mechanisms to compensate for their deficits. The association of MD networks (Fedorenko \u0026amp; Thompson-Schill, 2014) with producing novel word combinations (as indexed by bigram collocation) suggests that this sample of PWA does likely recruit additional cognitive resources/networks to support flexible and creative spoken language use (Zimmerer et al., 2020).\u003c/p\u003e\n\u003cp\u003eNotably, the CAT picture description scores and FLAT speech connectivity measure correlated exclusively with solely left, not bilateral, language network SNE measures. This is consistent with the core aspects of generating precise words (information carrying units) and grammatically correct speech still relying heavily on left-lateralized language networks (Bradshaw et al., 2017; Połczyńska et al., 2014), and why lesions here result in aphasia. Reflecting on those results, when describing complex scenes, which is likely harder for PWA due to left damaged networks may recruit bilateral and MD resources to support coherent speech production.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSentence speech comprehension in bilateral language networks\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn our PWA sample, efficient residual structural network connections between bilateral language regions correlated with better sentence-level comprehension as opposed to word comprehension abilities. This correlation likely reflects the increased cognitive demands associated with processing more complex linguistic structures, in line with the integration of language and working memory processes (Siegel et al., 2016). Sentence comprehension places significant demands on working memory components, such as maintaining multiple items and serial order information (Majerus, 2018). These working memory components are consistently engaged in demanding sentence-level language assessments, which require online maintenance of increasing semantic and syntactic information for comprehension (Leff et al., 2009; Majerus, 2018). For PWA, to support the maintenance and manipulation of linguistic information during sentence processing, efficient working memory integration may be particularly important, to compensate for their language deficits\u0026nbsp;(Murray, 2012; Salis et al., 2015).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe bilateral language network templates (fROIs) (Fedorenko et al., 2011) that we used here are large and arguably incorporate regions also associated with verbal working memory processes (Kalm \u0026amp; Norris, 2014; Majerus, 2018). This may be due to the nature of the language localizer tasks used to define the templates (Fedorenko et al., 2011, 2012) or the inherent integration of verbal working memory in language processing. Supporting the latter interpretation, a recent fMRI study showed working memory demands during naturalistic speech comprehension primarily engaged the same bilateral language-template regions (Shain et al., 2022). This bilateral engagement may provide a mechanism for compensation from acquired brain damage in PWA, allowing the less affected hemisphere to support speech comprehension\u0026nbsp;(Fleming et al., 2020). Efficient bilateral connections potentially facilitate the exchange and integration of semantic and syntactic information, enhancing PWAs’ ability to understand meanings from complex inputs despite damage to specific language areas.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDistributed and focal findings: SNE and Voxel-based analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePost-hoc analyses revealed distinct neural substrates for basic and complex language skills. Through a series of VLSM and CVA analyses, we identified differing patterns of association between brain integrity and word-level and high-level language processes.\u003c/p\u003e\n\u003cp\u003eSNE analyses were sensitive to high-level language behaviors compared to word-level ones, as evidenced by the significant high-level language CVA (p = 0.0021) versus the non-significant word-level language CVA (p = 0.708). SNE method’s particular utility appears to be in capturing distributed neural substrates associated with more complex language processes. This aligns with the notion that complex cognitive functions rely on distributed neural networks rather than isolated brain regions, echoing recent findings by Billot et al. (2022)\u0026nbsp;on the effects of white matter disconnections in chronic PWA.\u003c/p\u003e\n\u003cp\u003eThe VLSM results identified specific regions in word comprehension and naming, which align with previous literature on the neuroanatomical basis of semantic and phonological processes (Butler et al., 2014; Schumacher et al., 2019). Unlike the higher-level language functions associated with our SNE factors, those word-level language abilities appear more closely tied to localized brain areas and focal damage.\u003c/p\u003e\n\u003cp\u003eTaken together, these findings underscore the complementary nature of SNE and voxel-based approaches. While VLSM may be sensitive to focal lesions affecting more fundamental language functions, SNE appears to be a valuable measure for capturing the distributed nature of higher-level language processing. This dissociation highlights the importance of employing multiple analytical approaches in studying the neural basis of language function in the damaged brain.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethodological considerations\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn terms of the structural organization of post-stroke brain networks, our data suggest that the efficiency of residual path connections (possibly an index of plasticity) may be critical for cognitive function. This highlights the importance of considering structural network-level properties in understanding language performance after stroke. Our analytical framework is based on T1 volumetric brain images, which are widely available and routinely collected in clinical settings, making this approach especially useful and easy to apply. We used Fedorenko's functional ROIs, as they are currently from the largest openly available participant cognitive, including language, localizer dataset. \u0026nbsp;Importantly, our framework can be tailored to incorporate any anatomical parcellation schemes and functional ROIs (we acknowledge there are many alternatives available), based on researcher’s specific needs. Future work could also investigate the impact of different parcellation schemes and thresholding approaches on SNE analyses.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eClinically, our findings also emphasize the importance of assessing spoken language abilities using broader linguistic tools, such as FLAT, that integrate multiple cognitive functions and may be more sensitive to capturing more complex speech use and difficulties following aphasic stroke. We selected bigram patterns and syntactic coherence from FLAT; however, future studies incorporating additional linguistic measures will be instrumental in advancing our understanding of connected speech and its relationship with brain network metrics. Evaluating both language-specific and broader cognitive brain networks and behaviour will provide a more comprehensive understanding of an individual's language capabilities post-stroke.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn this study of chronic PWA, good speech performance was associated with efficient residual structural brain connections across bilateral language and multiple cognitive demand networks. Post-hoc analyses highlighted how SNE measures capture higher-level speech and cognitive integration, with analyses of focal lesion patterns most sensitive to word-level deficits. These results\u0026nbsp;illustrate how spoken language research can be grounded in a broader approach that integrates cognitive behaviours and structural brain network theories. This\u0026nbsp;framework\u0026nbsp;signals a new perspective to long-term speech performance after aphasic stroke- it depends on efficient bilateral network structure and\u0026nbsp;could extend beyond language to include other cognitive domains.\u0026nbsp;This approach could potentially serve as a tool for assessing PWAs’ brain integrity to (a) predict individual behaviors using common SNE parameters, (b) explain individual differences via identified network sources (i.e. different functional regions), and (c) clarify recovery patterns and treatment response. Overall, the network perspective proposed here advances our understanding of PWA, shifting emphasis from regional damage to efficient residual circuits, providing insights to future brain research in understanding language recovery mechanisms.\u0026nbsp;\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eParticipants\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 36 PWA in the chronic post-stroke phase took part. All were native English speakers with a single left-hemisphere stroke at least 12 months prior to taking part in the study. The key inclusion criteria were that each participant had 1) anomia as assessed by the naming subtest of the Comprehensive Aphasia Test (CAT; Swinburn, 2004), 2) relatively good single word repetition and comprehension as per the subtests of the CAT (Swinburn et al., 2004) and 3) no evidence of speech apraxia as assessed by the Apraxia Battery for Adults (Dabul, 2000). All had good functional hearing and visual acuity, no prior neurological or psychiatric disorders and no contraindications for MRI. The lesion overlap map illustrates the PWAs\u0026rsquo; pattern of left hemisphere brain damage (Figure 5). The PWAs\u0026rsquo; demographics are described in Table 2. Participants were recruited at University College London and the study was approved by the Central London Research Ethics Committee, UK. The PWA in the current sample are the same participants reported in the study of Akkad and colleagues (Akkad et al., 2023) that used voxel-based correlational methodology (VBCM) to investigate the neural correlates of language and domain-general cognitive deficits in PWA.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBehavioural data\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll PWA completed a comprehensive assessment of spoken language abilities. This included a standardized assessment using subtests of the comprehensive aphasia test (CAT; Swinburn et al., 2004), as well as custom-made spoken picture description tasks. The custom tasks included a greater number of items and more complex scenes, compared to the CAT, allowing for larger speech samples to enable a comprehensive analysis of connected speech performance. Details of custom picture description tasks can be found in Supplementary Figure 1 \u0026ndash; Figure 4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConnected speech - FLAT frequency, collocation and connectivity score\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo quantify PWAs\u0026rsquo; connected speech performance, linguistic measures were computed utilizing the Formulaic Language Analysis Tool (FLAT, Zimmerer et al., 2017). FLAT is a software program that quantifies usage-frequency based language variables for word combinations. It extracts frequencies by looking up each unit within transcriptions from spoken language in the British National Corpus (BNC, 2007), which represents everyday communication. FLAT values cover several distinct linguistic dimensions such as: frequency, novelty, and connectivity. Here, we focused on the frequency of unique bigrams (i.e. 2 words combinations which only appear once per sample \u0026ndash; no repetition). Lower frequency word combinations generally indicate greater lexical capacity (Wray, 2011).\u003c/p\u003e\n\u003cp\u003eThe novelty dimension reflects the linguistic creativity within the sample, measured by collocation strength. Collocation strength refers to the occurrence of two words together relative to how often each individual word appears. Collocation strength is computed as bigram t-scores. Stronger collocations are likely processed in a holistic, formulaic manner (i.e. retrieved as one unit; Wray, 2011). Formulaic language poses fewer demands to lexical and syntactic processes (Zimmerer et al., 2018). Overreliance on formulaic language can therefore suggest a diminished capacity to generate rare or novel expressions (Bello-Lepe et al., 2024), which hinders communication.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFLAT measures connectivity of language production by computing the proportion of words occurring in grammatical trigrams (rather than single word or two-word chunks). The ability to produce higher proportions of grammatically meaningful word sequences reflects stronger skills in sentence formulation and syntactic processing (Zimmerer et al., 2016, 2020). The performance of PWA on standardised tests of CAT subtests and FLAT scores of our custom picture description task can be found in Table 3.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"709\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"10\" valign=\"bottom\" style=\"width: 709px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 2. \u0026nbsp;PWA demographic and clinical data\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003eID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003cp\u003e(years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 123px;\"\u003e\n \u003cp\u003eEducation\u003c/p\u003e\n \u003cp\u003e(years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003eTime Post Stroke\u003c/p\u003e\n \u003cp\u003e(years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 163px;\"\u003e\n \u003cp\u003eLesion Volume\u003c/p\u003e\n \u003cp\u003e()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e61.53\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e56\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e160.80\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e42.76\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e57.24\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e78.36\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e43.42\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e161.81\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e83.18\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e63.56\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e38.14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e29.49\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e8.93\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e117.55\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e171.71\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e71.31\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e63.53\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e22.25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e1.51\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e163.75\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e308.18\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e65.80\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e164.53\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e156.91\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e348.23\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e94.32\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e223.29\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e112.55\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e130.60\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e240.39\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e78.61\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e99.38\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e403.11\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e387.17\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e152.61\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e239.29\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003e36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 126px;\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 101px;\"\u003e\n \u003cp\u003e39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 113px;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 163px;\"\u003e\n \u003cp\u003e95.70\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 63px;\"\u003e\n \u003cp\u003eAvg(SD)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 142px;\"\u003e\n \u003cp\u003e26 M/10 F\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e59 (12.51)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e13.8 (2.38)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" style=\"width: 170px;\"\u003e\n \u003cp\u003e10.36 (8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\" style=\"width: 145px;\"\u003e\n \u003cp\u003e131.7 (79.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"10\" valign=\"bottom\" style=\"width: 709px;\"\u003e\n \u003cp\u003eThirty six participants include 26 males(M) and 10 females (F): mean age = 59 years [standard deviation (SD) = 12.51], range = 29-82 years; mean years of education = 13.8 years (SD= 2.38) , range = 10-17 years. Lesion volume were estimated by the automated lesion identification method (Seghier et al., 2008) and mean time post stroke = 10.36 years (SD=8), range =1-34 years; and average lesion volume was 131.7.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"756\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"18\" style=\"width: 747px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 3. Standardised assessment and FLAT connected speech scores of custom picture description task.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" style=\"width: 66px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 180px;\"\u003e\n \u003cp\u003eCAT Spoken Comprehension\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 189px;\"\u003e\n \u003cp\u003eCAT Speech Production\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"8\" valign=\"top\" style=\"width: 321px;\"\u003e\n \u003cp\u003eFLAT scores of custom picture description\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 9px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 38px;\"\u003e\n \u003cp\u003eID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" style=\"width: 113px;\"\u003e\n \u003cp\u003eWords\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003eSentences\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003eNaming\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 104px;\"\u003e\n \u003cp\u003ePicture Des\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 113px;\"\u003e\n \u003cp\u003eBigram Fre\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003eBigram Col\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 104px;\"\u003e\n \u003cp\u003eConnectivity\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e17\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e25\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e41\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e20\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e1973.69\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e15.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.47\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e19\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e14\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2205.73\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e15.96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e22\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e8\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1058.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e9.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.51\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e18\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e26\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e8\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e1932.78\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e12.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e42\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1269.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e10.96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.68\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1545.17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e11.62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.58\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e30\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e7\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2375.66\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e13.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.34\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e22\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e16\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e30\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e19\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1651.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e12.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.34\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e20\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e17\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e5\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2409.99\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e21.43\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.44\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e18\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e34\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e20\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1571.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e14.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.43\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e16\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e32\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1600.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e13.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.52\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e25\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e25\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e35\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e26\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1152.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e10.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.48\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e21\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e41\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2277.86\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e17.99\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.57\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e29\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1182.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e9.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.50\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e40\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1773.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e15.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.51\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e23\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e41\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e22\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1371.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e12.96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.61\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e14\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e35\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1620.55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e13.10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.47\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e39\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e32\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1333.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e11.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.55\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e36\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e25\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1228.18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e11.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.55\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e21\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e29\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e11\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1107.79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e5.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.16\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e37\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e31\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1386.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e10.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.57\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e22\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2020.29\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e14.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.41\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e40\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1264.18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e10.41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.65\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e18\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e40\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e18\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e562.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e1.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.51\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e40\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e13\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e204.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e3.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.43\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e23\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e43\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1496.96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e13.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.60\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e41\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1344.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e10.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.69\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e25\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e39\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e20\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1131.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e10.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.53\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e30\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e4\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2847.93\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e20.82\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.37\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e22\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e18\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e33\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e32\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1090.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e10.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.59\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e21\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e37\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e22\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1641.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e12.41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.56\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e20\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e21\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e1\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e3620.38\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e21.58\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.05\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e22\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e39\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e8\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e1810.27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e13.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.15\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e43\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eN/A\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1744.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e15.94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.55\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e29\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eN/A\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2061.85\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e14.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.26\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" style=\"width: 57px;\"\u003e\n \u003cp\u003e36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 104px;\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 85px;\"\u003e\n \u003cp\u003e48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e11\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1610.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e11.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.50\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" style=\"width: 66px;\"\u003e\n \u003cp\u003eAvg(SD)range\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e\u0026nbsp; 26.75 (3.08)\u003c/p\u003e\n \u003cp\u003e17-30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e22.5 (5.34)\u003c/p\u003e\n \u003cp\u003e13-32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e35.11 (7.47)\u003c/p\u003e\n \u003cp\u003e13-48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e19.56 (10.21)\u003c/p\u003e\n \u003cp\u003e1-40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1624.5 (620.19)\u003c/p\u003e\n \u003cp\u003e204-3620\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e12.65 (4.19)\u003c/p\u003e\n \u003cp\u003e1-21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.46 (0.15)\u003c/p\u003e\n \u003cp\u003e0.05-0.69\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\" style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNorm\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"5\" valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 66px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 132px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\" style=\"width: 76px;\"\u003e\n \u003cp\u003eAvg(SD)\u003c/p\u003e\n \u003cp\u003erange\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e29.15 (1.35) \u0026nbsp; \u0026nbsp; \u0026nbsp;25-30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e30.17 (1.85) 26-32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e46.37 (1.6)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;42-48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e52.20 (18.25) 33-87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1092.94 (356.10)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e691 - 2293\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e10.02 (3.25)\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e6 - 21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.79 (0.04)\u003c/p\u003e\n \u003cp\u003e0.7-0.88\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\" style=\"width: 76px;\"\u003e\n \u003cp\u003ecut off\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1804\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e16.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e0.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"19\" style=\"width: 756px;\"\u003e\n \u003cp\u003eThe table\u0026apos;s bottom two rows present patient data and normative data for healthy individuals: average score (Avg), standard deviation (SD), and range. Normative data for CAT is sourced from the CAT manual (Swinburn et al., 2004), while FLAT normative data was obtained from 18 healthy participants completing identical composite picture tasks (see Supplementary S2). Bold values indicate scores below cutoff: for CAT, these are as suggested by the manual; for FLAT, bold values in frequency and collocation are above average plus two standard deviations, while for connectivity, they are below average minus two standard deviations. Abbreviation: Picture Des- picture description; Bigram Fre- bigram frequency; Bigram Col- bigram collocation; FLAT- Formulaic Language Analysis Tool (Zimmerer et al., 2017). N/A values were replaced with the mean for subsequent statistical analyses.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003eMRI data processing \u0026ndash; lesion quantification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStructural T1-weighted whole brain MRI scans were acquired on a 3T Siemens TIM-Trio system at the Wellcome Centre for Human Neuroimaging. Images were registered into standard Montreal Neurological Institute (MNI) space using a modified unified segmentation\u0026ndash;normalisation procedure (Seghier et al., 2008). Images were smoothed with an 8mm full-width at half-maximum (FWHM) Gaussian kernel. To match anatomical parcellations for later lesion quantification measurements, the lesion masks were binarised from these images, re-sliced to 1 mm isotropic voxel dimension and re-sampled to image dimensions of 181x217x181 in nifti file format. All pre-processed procedures were computed within Statistical Parametric Mapping software (SPM 12) running under MATLAB 2022a. The segmented lesion masks were then used as inputs in the Lesion Quantification Toolkit (Griffis et al., 2021).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMRI data-\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003estructural network efficiency (SNE)\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo investigate the relationship between the lesioned brain\u0026rsquo;s residual structural connectivity and PWAs\u0026rsquo; speech production performance, we used SNE to quantify the connectivity between regions within two functional brain networks supporting language-specific and multiple-demand cognitive processes. The two functional brain networks were selected based on the work by Fedorenko and colleagues (Fedorenko, 2014; Fedorenko et al., 2010, 2012). The data processing pipeline is outlined in Figure 6. First, the structural connectivity matrices were derived from lesion and functional regions of interest (fROI) masks (see Figure 6 a-b) using Automated Anatomical Labeling parcellations (AAL; Tzourio-Mazoyer et al., 2002) and HCP-842 tractography atlas (Yeh et al., 2018) in the Lesion Quantification Toolkit (Griffis et al., 2021). The AAL parcellation offers a well-established and widely adopted anatomical parcellation scheme (Lawrence et al., 2014; Liu et al., 2021; Rubinov \u0026amp; Sporns, 2010; Yeh et al., 2018), facilitating comparison across studies and promoting reproducibility. With 116 cortical regions, it provides a reasonable balance between anatomical detail and computational feasibility. The HCP-842 tractography atlas was chosen for its high-resolution and extensive coverage of white matter pathways (Griffis et al., 2021; Yeh et al., 2022). To maintain consistency with Fedorenko\u0026apos;s lab\u0026apos;s fROI definitions while working in AAL space, we quantified the overlap between their fROIs and AAL parcels, incorporating all overlapping regions except those with minimal overlap (less than 10% of the total area). Details of AAL fROI nodes and their corresponding proportions can be found in Supplementary Figure 5 and Supplementary Table 7.\u003c/p\u003e\n\u003cp\u003eThe residual structural connectivity matrix comprised intact white matter connections obtained from the residual tract disconnection between each AAL brain parcel (Figure 6c). After constructing the residual structural connectivity matrices, brain structural parcellations were designated as nodes, with white matter connections exceeding a 70% threshold binarised to 1 and utilised as edges. The threshold was selected based on prior studies (Griffis et al., 2021; Wang et al., 2015; Zalesky et al., 2010) as it balances network sparsity, reducing weak or spurious links while preserving the brain network\u0026apos;s core structure. Nodes and edges within functional regions of interest (fROIs) formed network graphs (Figure 6d and 6e). The shortest path is the minimum structural connection(s) needing traversal from one region to another. Graph theory defines efficiency as the inverted shortest path length between two nodes (Kaiser, 2011). Thus, efficiency ranges from 0 to 1, with higher values indicating greater efficiency. We implemented the breadth-first search algorithm (Rubinov \u0026amp; Sporns, 2010) to calculate shortest path. Structural network efficiency, SNE, was defined as the mean efficiency across all nodes within the fROI. Figure 7 illustrates a graph including 3 nodes and its SNE.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe residual structural connectivity matrix was calculated by Lesion Quantification Toolkit (Griffis et al., 2021) running under MATLAB 2019b and the SNE was measured by Networkx package running under Python 3.9 (see Supplementary S4).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSpearman correlation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate whether higher residual SNE within fROIs correlated with better preserved behavioural functions in PWA at the group level, we examined the correlation between individual behavioural scores and SNE in both lateralised and bilateral fROIs. Given the PWA brain data did not always satisfy the assumption of normal distribution we used two-tailed Spearman\u0026apos;s rank correlation coefficients consistent with prior brain-behaviour mapping studies (Bertoux et al., 2020; Schumacher et al., 2019). For greater lesion-behavior statistical power (Rorden et al., 2007) we used p-value correction based on 10,000 random permutation tests and reported at p \u0026lt; 0.05 and p \u0026lt; 0.01 alpha threshold. Lesion volume, estimated by the automated lesion identification method (Seghier et al., 2008), was included as a covariate in the statistical models (Price et al., 2017). Analyses were run with SciPy under Python 3.9.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data described in this study is available to accredited researchers from J. C. - on request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe are grateful to Karl Friston and Peter Zeidman for their valuable guidance during the SPM Methods Clinic meeting, with regard to the Canonical Variates Analysis (CVA) methodologies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eP. J. D. and J. C. performed the research; P. J. D. conducted brain image quantification and analysis of SNE; I. S. conducted FLAT connected speech quantification; P. J. D. and J. C. wrote the original draft; V. C. Z., T. H., and A. L. reviewed methods and results; H. A. preprocessed brain image data; \u0026nbsp;all authors reviewed and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was funded in part, by the Wellcome Trust [203147/Z/16/Z and106161/Z/14/Z J.C]. For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. The funders had no participation in the design and results of this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of Competing Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAkkad H, Hope TMH, Howland C, Ondobaka S, Pappa K, Nardo D, Duncan J, Leff AP, Crinion J (2023) Mapping spoken language and cognitive deficits in post-stroke aphasia. NeuroImage: Clin 39:103452\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarbieri E, Mack J, Chiappetta B, Europa E, Thompson CK (2019) Recovery of offline and online sentence processing in aphasia: Language and domain-general network neuroplasticity. Cortex 120:394\u0026ndash;418\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBassi MS, Iezzi E, Gilio L, Centonze D, Buttari F (2019) Synaptic Plasticity Shapes Brain Connectivity: Implications for Network Topology. Int J Mol Sci 20(24):6193\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBates E, Wilson SM, Saygin AP, Dick F, Sereno MI, Knight RT, Dronker NF (2003) Voxel-based lesion\u0026ndash;symptom mapping. Nat Neurosci 6(5):448\u0026ndash;450\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBello-Lepe S, Mahmood S, Varley R, Zimmerer V (2024) Speech pauses in speakers with and without aphasia: A usage-based approach. Cortex 178:287\u0026ndash;298\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBertoux M, Duclos H, Caillaud M, Segobin S, Merck C, De La Sayette V, Belliard S, Desgranges B, Eustache F, Laisney M (2020) When affect overlaps with concept: emotion recognition in semantic variant of primary progressive aphasia. Brain 143(12):3850\u0026ndash;3864\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBillot A, Thiebaut de Schotten M, Parrish TB, Thompson CK, Rapp B, Caplan D, Kiran S (2022) Structural disconnections associated with language impairments in chronic post-stroke aphasia using disconnectome maps. Cortex 155:90\u0026ndash;106\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBradshaw AR, Thompson PA, Wilson AC, Bishop DVM, Woodhead ZV (2017) J. Measuring language lateralisation with different language tasks: A systematic review. PeerJ 10:e3929\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBridges KA, Van Sidtis L (2013) Formulaic language in Alzheimer\u0026rsquo;s disease. Aphasiology 27(7):799\u0026ndash;810\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eButler RA, Ralph MAL, Woollams AM (2014) Capturing multidimensionality in stroke aphasia: mapping principal behavioural components to neural structures. Brain 137(12):3248\u0026ndash;3266\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarrera E, Tononi G (2014) Diaschisis: past, present, future. Brain 137(9):2408\u0026ndash;2422\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen AC, Oathes DJ, Chang C, Bradley T, Zhou ZW, Williamsa LM, Glover GH, Deisseroth K, Etkin A (2013) Causal interactions between fronto-parietal central executive and default-mode networks in humans. Proc Natl Acad Sci USA 110(49):19944\u0026ndash;19949\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheng B, Schlemm E, Schulz R, Boenstrup M, Mess\u0026eacute; A, Hilgetag C, Gerloff C (2019) \u0026amp; Tz Thomalla, G. Altered topology of large-scale structural brain networks in chronic stroke. Brain Commun, 1(1)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eConroy P, Sage K, Ralph ML (2009) Improved vocabulary production after naming therapy in aphasia: can gains in picture naming generalise to connected speech? Int J Lang Communication Disorders 44(6):1036\u0026ndash;1062\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDabul BP-E (2000) ABA-2: Apraxia Battery for Adults\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFedorenko E, Thompson-Schill SL (2014) Reworking the language network. Trends Cogn Sci 18(3):120\u0026ndash;126\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFedorenko E, Varley R (2016) Language and thought are not the same thing: evidence from neuroimaging and neurological patients. Ann N Y Acad Sci 1369(1):132\u0026ndash;153\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFleming V, Brownsett S, Krason A, Maegli MA, Coley-Fisher H, Ong YH, Nardo D, Leach R, Howard D, Robson H, Warburton E, Ashburner J, Price CJ, Crinion JT, Leff (2020) A. P. Efficacy of spoken word comprehension therapy in patients with chronic aphasia: a cross-over randomised controlled trial with structural imaging. J Neurol Neurosurg Psychiatry 92(4):418\u0026ndash;424\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFoulon C, Cerliani L, Kinkingn\u0026eacute;hun S, Levy R, Rosso C, Urbanski M, Volle E (2018) Schotten, M. T. Advanced lesion symptom mapping analyses and implementation as BCBtoolkit. GigaScience 7(3):1\u0026ndash;17de\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFridriksson J, den Ouden DB, Hillis AE, Hickok G, Rorden C, Basilakos A, Yourganov G, Bonilha L (2018) Anatomy of aphasia revisited. Brain 141(3):848\u0026ndash;862\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFriston KJ, Frith CD, Frackowiak RSJ, Turner R (1995) Characterizing Dynamic Brain Responses with fMRI: A Multivariate Approach. NeuroImage 2(2):166\u0026ndash;172\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGow DW (2012) The cortical organization of lexical knowledge: A dual lexicon model of spoken language processing. Brain Lang 121(3):273\u0026ndash;288\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGriffis JC, Metcalf NV, Corbetta M, Shulman GL (2020) Damage to the shortest structural paths between brain regions is associated with disruptions of resting-state functional connectivity after stroke. NeuroImage 210:116589\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGriffis JC, Metcalf NV, Corbetta M, Shulman GL (2021) Lesion Quantification Toolkit: A MATLAB software tool for estimating grey matter damage and white matter disconnections in patients with focal brain lesions. NeuroImage: Clin 30:102639\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuggisberg AG, Koch PJ, Hummel FC, Buetefisch CM (2019) Brain networks and their relevance for stroke rehabilitation. Clin Neurophysiol 130(7):1098\u0026ndash;1124\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHartwigsen G, Henseler I, Stockert A, Wawrzyniak M, Wendt C, Klingbeil J, Baumgaertner A, Saur D (2017) Integration demands modulate effective connectivity in a fronto-temporal network for contextual sentence integration. NeuroImage 147:812\u0026ndash;824\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHartwigsen G, Lim J-S, Bae H-J, Yu K-H, Kuijf HJ, Weaver NA, Biesbroek JM, Kopal J, Bzdok D (2024) Bayesian modelling disentangles language versus executive control disruption in stroke. Brain Commun, 6(3)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHartwigsen G, Saur D (2019) Neuroimaging of stroke recovery from aphasia \u0026ndash; Insights into plasticity of the human language network. \u003cem\u003eNeuroImage\u003c/em\u003e, 190, \u003cem\u003e14\u0026ndash;31\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHerbert R, Hickin J, Howard D, Osborne F, Best W (2008) Do picture-naming tests provide a valid assessment of lexical retrieval in conversation in aphasia? \u003cem\u003eAphasiology\u003c/em\u003e, 22\u003cem\u003e(2), 184\u0026ndash;203.\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeuer S, Ivanova MV, Hallowell B (2017) More Than the Verbal Stimulus Matters: Visual Attention in Language Assessment for People With Aphasia Using Multiple-Choice Image Displays. J Speech Lang Hear Res 60(5):1348\u0026ndash;1361\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHickok G, Poeppel D (2007) The cortical organization of speech processing. Nat Rev Neurosci 8(5):393\u0026ndash;402\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHope TMH, Leff AP, Price CJ (2018) Predicting language outcomes after stroke: Is structural disconnection a useful predictor? NeuroImage: Clin 19:22\u0026ndash;29\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKaiser M (2011) A tutorial in connectome analysis: Topological and spatial features of brain networks. \u003cem\u003eNeuroimage\u003c/em\u003e, 57\u003cem\u003e(3), 892\u0026ndash;907\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKalm K, Davis MH, Norris D (2012) Neural mechanisms underlying the grouping effect in short-term memory. Hum Brain Mapp 33(7):1634\u0026ndash;1647\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKalm K, Norris D (2014) The Representation of Order Information in Auditory-Verbal Short-Term Memory. J Neurosci 34(20):6879\u0026ndash;6886\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKearney E, Guenther FH (2019) Articulating: the neural mechanisms of speech production. Lang Cognition Neurosci 34(9):1214\u0026ndash;1229\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKeil K, Kaszniak AW (2010) Examining executive function in individuals with brain injury: A review. Aphasiology 16(3):305\u0026ndash;335\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLawrence AJ, Chung AW, Morris RG, Markus HS, Barrick TR (2014) Structural network efficiency is associated with cognitive impairment in small-vessel disease. Neurology 83(4):304\u0026ndash;311\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeff AP, Schofield TM, Crinion JT, Seghier ML, Grogan A, Green DW, Price CJ (2009) The left superior temporal gyrus is a shared substrate for auditory short-term memory and speech comprehension: evidence from 210 patients with stroke. Brain 132(12):3401\u0026ndash;3410\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu T, Thiebaut de Schotten M, Altarelli I, Ramus F, Zhao J (2021) Maladaptive compensation of right fusiform gyrus in developmental dyslexia: A hub-based white matter network analysis. Cortex 145:57\u0026ndash;66\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMajerus S (2018) Working memory treatment in aphasia: A theoretical and quantitative review. J Neurolinguistics 48:157\u0026ndash;175\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarebwa BK, Fridriksson J, Yourganov G, Feenaughty L, Rorden C, Bonilha L (2017) Chronic post-stroke aphasia severity is determined by fragmentation of residual white matter networks. Sci Rep, 7(1)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarini A, Galetto V, Zampieri E, Vorano L, Zettin M, Carlomagno S (2011) Narrative language in traumatic brain injury. Neuropsychologia 49(10):2904\u0026ndash;2910\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartin N, Reilly J (2012) Short-term/working memory impairments in aphasia: Data, models, and their application to aphasia rehabilitation. Aphasiology 26(3\u0026ndash;4):253\u0026ndash;257\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMurray LL (2012) Direct and indirect treatment approaches for addressing short-term or working memory deficits in aphasia. Aphasiology 26(3\u0026ndash;4):317\u0026ndash;337\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePerkins L (1995) Applying conversation analysis to aphasia: Clinical implications and analytic issues. Eur J Disord Commun 30(3):372\u0026ndash;383\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePołczyńska M, Curtiss S, Walshaw P, Siddarth P, Benjamin C, Moseley BD, Vigil C, Jones M, Eliashiv D, Bookheimer S (2014) Grammar tests increase the ability to lateralize language function in the Wada test. Epilepsy Res 108(10):1864\u0026ndash;1873\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrice CJ, Hope TM, Seghier ML (2017) Ten problems and solutions when predicting individual outcome from lesion site after stroke. NeuroImage 145:200\u0026ndash;208\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRorden C, Karnath H-O, Bonilha L (2007) Improving Lesion-Symptom Mapping. J Cogn Neurosci 19(7):1081\u0026ndash;1088\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRubinov M, Sporns O (2010) Complex network measures of brain connectivity: Uses and interpretations. NeuroImage 52(3):1059\u0026ndash;1069\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSalis C, Kelly H, Code C (2015) Assessment and treatment of short-term and working memory impairments in stroke aphasia: A practical tutorial. Int J Lang Communication Disorders 50(6):721\u0026ndash;736\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchumacher R, Halai AD, Lambon Ralph MA (2019) Assessing and mapping language, attention and executive multidimensional deficits in stroke aphasia. Brain 142(10):3202\u0026ndash;3216\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeghier ML, Ramlackhansingh A, Crinion J, Leff AP, Price CJ (2008) Lesion identification using unified segmentation-normalisation models and fuzzy clustering. NeuroImage 41(4):1253\u0026ndash;1266\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShain C, Blank IA, Fedorenko E, Gibson E, Schuler W (2022) Robust Effects of Working Memory Demand during Naturalistic Language Comprehension in Language-Selective Cortex. J Neurosci 42(39):7412\u0026ndash;7430\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSiegel, J. S., Ramsey, L. E., Snyder, A. Z., Metcalf, N. V., Chacko, R. V., Weinberger,K., \u0026hellip; Corbetta, M. Disruptions of network connectivity predict impairment in multiple behavioral domains after stroke. Proceedings of the National Academy of Sciences, 113(30), E4367-E4376 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStefaniak JD, Alyahya RSW (2021) Lambon Ralph, M. A. Language networks in aphasia and health: A 1000 participant activation likelihood estimation meta-analysis. NeuroImage 233:117960\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSwinburn K (2004) Comprehensive aphasia test. CAT, Psychology\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTermenon M, Achard S, Jaillard A, Delon-Martin C (2016) The hub disruption index, a reliable index sensitive to the brain networks reorganization. A study of the contralesional hemisphere in stroke. Front Comput Neurosci 10:84\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTurkeltaub PE, Messing S, Norise C, Hamilton RH (2011) Are networks for residual language function and recovery consistent across aphasic patients? Neurology 76(20):1726\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTyler LK, Marslen-Wilson WD, Stamatakis EA, Tyler LK, Marslen-Wilson W (2005) Dissociating neuro-cognitive component processes: Voxel-based correlational methodology Age-Related Delay in Visual and Auditory Evoked Responses is Mediated by White-and Gray-matter Differences View project Ventral stream patients View project Dissociating neuro-cognitive component processes: voxel-based correlational methodology. Neuropsychologia 43:771\u0026ndash;778\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, Joliot M (2002) Automated Anatomical Labeling of Activations in SPM Using a Macroscopic Anatomical Parcellation of the MNI MRI Single-Subject Brain. NeuroImage 15(1):273\u0026ndash;289\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan Sidtis L, D., Postman WA (2006) Formulaic expressions in spontaneous speech of left- and right‐hemisphere‐damaged subjects. Aphasiology 20(5):411\u0026ndash;426\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang J, Wang X, Xia M, Liao X, Evans A, He Y (2015) GRETNA: A graph theoretical network analysis toolbox for imaging connectomics. \u003cem\u003eFrontiers in Human Neuroscience\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(JUNE), 147509\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWray A (2011) Formulaic Language as a Barrier to Effective Communication with People with Alzheimer\u0026rsquo;s Disease. Can Mod Lang Rev 67(4):429\u0026ndash;458\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYeh FC (2022) Population-based tract-to-region connectome of the human brain and its hierarchical topology. Nat Commun 13(1):4933\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYeh FC, Panesar S, Fernandes D, Meola A, Yoshino M, Fernandez-Miranda JC, Vettel JM, Verstynen T (2018) Population-averaged atlas of the macroscale human structural connectome and its network topology. NeuroImage 178:57\u0026ndash;68\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZalesky A, Fornito A, Bullmore ET (2010) Network-based statistic: Identifying differences in brain networks. NeuroImage 53(4):1197\u0026ndash;1207\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZimmerer VC, Hardy CJD, Eastman J, Dutta S, Varnet L, Bond RL, Russell L, Rohrer JD, Warren JD, Varley RA (2020) Automated profiling of spontaneous speech in primary progressive aphasia and behavioral-variant frontotemporal dementia: An approach based on usage-frequency. Cortex 133:103\u0026ndash;119\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZimmerer VC, Newman L, Thomson R, Coleman M, Varley RA (2018) Automated analysis of language production in aphasia and right-hemisphere damage: frequency and collocation strength. Aphasiology 32(11):1267\u0026ndash;1283\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZimmerer VC, Wibrow M, Varley RA (2016) Formulaic Language in People with Probable Alzheimer\u0026rsquo;s Disease: A Frequency-Based Approach. J Alzheimer\u0026rsquo;s Disease 53(3):1145\u0026ndash;1160\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-5587181/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5587181/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIndividuals with post-stroke aphasia (PWA), an acquired language disorder, face significant communication challenges essential for daily life. Surprisingly, little is known about how brain damage following stroke disrupts the anatomical integration of language and multiple-demand brain networks required for speech production. To address this, we used a measure of brain structural network efficiency (SNE) to investigate the anatomical correlates of spoken language abilities in 36 chronic PWA. Connected speech abilities were correlated with higher SNE not only within the left (dominant) language network but also with bilateral language and multiple-demand networks. Post-hoc analyses found: 1) word-level language behaviours were associated solely with discrete left temporoparietal lesions, using voxel-based correlational methodology; 2) using canonical variate analysis, SNE was primarily sensitive to higher-level language behaviours, loading additional brain-behaviour variance beyond distributed voxels. Taken together, our findings illustrate how speech abilities in PWA rely on distributed bilateral networks, outlining higher-level language-cognition interplay.\u003c/p\u003e","manuscriptTitle":"Uncovering Hidden Pathways: Structural Brain Networks Underpinning Connected Speech in Post-Stroke Aphasia","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-17 08:33:59","doi":"10.21203/rs.3.rs-5587181/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4cfc2cf2-6cfc-4fc4-a798-39fc3d1c4f27","owner":[],"postedDate":"December 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":41584262,"name":"Health sciences/Diseases/Neurological disorders/Stroke"},{"id":41584263,"name":"Biological sciences/Neuroscience/Cognitive neuroscience/Language"}],"tags":[],"updatedAt":"2024-12-17T08:33:59+00:00","versionOfRecord":[],"versionCreatedAt":"2024-12-17 08:33:59","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5587181","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5587181","identity":"rs-5587181","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}