Accepted Manuscript Using principal component analysis to capture individual differences within a unified neuropsychological model of chronic post-stroke aphasia: Revealing the unique neural correlates of speech fluency, phonology and semantics Ajay D. Halai, PhD, Anna M. Woollams, Prof. Matthew A. Lambon Ralph, PhD PII: S0010-9452(16)30092-2 DOI: 10.1016/j.cortex.2016.04.016 Reference: CORTEX 1733 To appear in: Cortex Received Date: 14 January 2016 Revised Date: 15 April 2016 Accepted Date: 18 April 2016 Please cite this article as: Halai AD, Woollams AM, Lambon Ralph MA, Using principal component analysis to capture individual differences within a unified neuropsychological model of chronic poststroke aphasia: Revealing the unique neural correlates of speech fluency, phonology and semantics, CORTEX (2016), doi: 10.1016/j.cortex.2016.04.016. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT Using principal component analysis to capture individual differences within a unified neuropsychological model of chronic post-stroke aphasia: Revealing the unique neural correlates of RI PT speech fluency, phonology and semantics. Ajay D. HALAI1*, Anna M. WOOLLAMS1, & Matthew A. LAMBON RALPH1* Neuroscience and Aphasia Research Unit, University of Manchester, UK SC 1 Correspondence to: Ajay D. Halai, PhD ajay.halai@manchester.ac.uk M AN U Running head: PCA and individual differences in aphasia Neuroscience and Aphasia Research Unit TE D School of Psychological Sciences, Zochonis Building University of Manchester, Brunswick Street Manchester, M13 9PL, England or EP Ph: +44 161 275 7336 Prof. Matthew A. Lambon Ralph, PhD AC C matt.lambon-ralph@manchester.ac.uk Ph: +44 161 275 2551 ACCEPTED MANUSCRIPT Using principal component analysis to capture individual differences within a unified neuropsychological model of chronic post-stroke aphasia: Revealing the unique neural correlates of RI PT speech fluency, phonology and semantics. Ajay D. HALAI1*, Anna M. WOOLLAMS1, & Matthew A. LAMBON RALPH1* Neuroscience and Aphasia Research Unit, University of Manchester, UK SC 1 Correspondence to: Ajay D. Halai, PhD ajay.halai@manchester.ac.uk M AN U Running head: PCA and individual differences in aphasia Neuroscience and Aphasia Research Unit TE D School of Psychological Sciences, Zochonis Building University of Manchester, Brunswick Street Manchester, M13 9PL, England or EP Ph: +44 161 275 7336 Prof. Matthew A. Lambon Ralph, PhD AC C matt.lambon-ralph@manchester.ac.uk Ph: +44 161 275 2551 1 ACCEPTED MANUSCRIPT Abstract Individual differences in the performance profiles of neuropsychologically-impaired patients are pervasive yet there is still no resolution on the best way to model and account for the variation in their behavioural impairments and the associated neural correlates. To date, researchers have generally taken one of three different approaches: a single case-study RI PT methodology in which each case is considered separately; a case-series design in which all individual patients from a small coherent group are examined and directly compared; or, group studies, in which a sample of cases are investigated as one group with the assumption that they are drawn from a homogenous category and that performance differences are of no SC interest. In recent research, we have developed a complementary alternative through the use of principal component analysis of individual data from large patient cohorts. This datadriven approach not only generates a single unified model for the group as a whole M AN U (expressed in terms of the emergent principal components) but is also able to capture the individual differences between patients (in terms of their relative positions along the principal behavioural axes). We demonstrate the use of this approach by considering speech fluency, phonology and semantics in aphasia diagnosis and classification, as well as their unique neural correlates. Principal component analysis of the behavioural data from 31 patients with TE D chronic post-stroke aphasia resulted in four statistically-independent behavioural components reflecting phonological, semantic, executive-cognitive and fluency abilities. Even after accounting for lesion volume, entering the four behavioural components simultaneously into a voxel-based correlational methodology (VBCM) analysis revealed that speech fluency EP (speech quanta) was uniquely correlated with left motor cortex and underlying white matter (including the anterior section of the arcuate fasciculus and the frontal aslant tract), AC C phonological skills with regions in the superior temporal gryus and pars opercularis, and semantics with the anterior temporal stem. 2 ACCEPTED MANUSCRIPT Introduction For both clinical practice and basic research, it is important to understand the patterns and neural bases of impaired performance observed in both individual and groups of neuropsychological patients. Since the beginning of behavioural neurology and neuropsychology, contrastive impairments across patients have been used to make important RI PT inferences about the underlying cognitive computations and the neural structures that support them. In addition, stable models of variable patient performance are critical in accurate neurological differential diagnosis, clinical management and treatment. The generation of stable, reliable models of normal and impaired function rely on the SC ability to understand the nature and sources of individual differences across patients. This has always been a key challenge for the field (Shallice, 1988), leading to numerous debates and discussions (Caramazza & McCloskey, 1988; Lambon Ralph, Patterson, & Plaut, 2011; M AN U Schwartz & Dell, 2011; Shallice, 1988), and arguably still is. The kernel of the problem relates to the fact that there are multiple sources that underlie variable neuropsychological results. Previous formal considerations of the issue have set out a series of potential factors (Lambon Ralph, Moriarty, & Sage, 2002; Shallice, 1988) but, for brevity, we will note three types here: (A) the type of difference that neuropsychological dissociations and differential TE D diagnosis are based on, namely stable performance differences that arise when a task is supported by one or more computations underpinned by neuroanatomically-separated regions; (B) differences that relate, linearly or nonlinearly, to the severity of damage to the underlying neurocomputational component(s) – which can provide critical information about EP the nature and characteristics of the core neurocognitive computations, as well as the degree of recovery/deterioration in a patient’s disease or disorder; and (C) sources of individual AC C differences of no interest including random variations, measurement noise, and so forth. Over the years, researchers have adopted different approaches to this issue. The most classical and common approach (also utilised by experimental psychology and functional neuroimaging) is to recruit a coherent sample of patients and, through the power of central tendency, average the behavioural results or lesion distributions in order to remove noise and other individual variations of no interest. As noted previously (Shallice, 1988), this method relies on having or defining a ‘coherent’ group and, at worst, can suffer from two types of statistical error: (a) failing to uncover meaningful differences within the sample of patients which is lost through the averaging process; and (b) failing to detect or understand how systematic variation in patient/lesion severity changes task performance (which otherwise is falsely considered to reflect a behavioural dissociation). An alternative approach, single-case 3 ACCEPTED MANUSCRIPT study, possesses the opposite set of advantages and disadvantages. Of course, whilst single case-study avoids averaging away important behavioural dissociations, it means that it is hard/impossible to place each patient within a broader context or to generate a coherent, overarching cognitive and neuroanatomical model of the disorder. From a basic science perspective, this approach could lead to the logical absurdity of there being as many theories RI PT as there are patients; whilst, from a clinical perspective, it reduces our ability to generalise clinical knowledge from one patient to another – ultimately making it impossible to successfully diagnose, manage and treat the patient’s impairments. The single-case study makes it impossible to map the relationship between severity of damage and level of SC performance (which, like any mathematical function, requires more than one datum). Over the past decade or so, multiple researchers have adopted a hybrid of these two approaches in the form of case-series studies (Hodges, Patterson, Oxbury, & Funnell, 1992; Lambon Ralph M AN U et al., 2001; Schwartz et al., 2006; Woollams, Lambon Ralph, Plaut, & Patterson, 2007). These investigations recruit a sample of patients whom are all studied in detail at the individual level (i.e., often to the same precision as single-case studies), which can be used: to assess the consistency of performance, i.e., the coherence, in the sampled cases (Lambon Ralph et al., 2001; Schwartz et al., 2006; Woollams et al., 2007); to highlight TE D patterns/mechanisms/lesion correlates which generalise reliably across cases; to identify meaningful, consistent patterns of individual differences (rather than random noise: Woollams et al., 2007); as well as to relate impairment severity to task performance (leading to mathematical/computational models of the severity-based functions: Lambon Ralph et al., EP 2001; Schwartz et al., 2006; Woollams et al., 2007). Furthermore, ‘comparative’ case-series can be used to strengthen important behavioural dissociations by demonstrating that, despite AC C variations in severity, (i) the two case-series are systematically different in the expected way, yet (ii) are internally coherent (Jefferies & Lambon Ralph, 2006; Lambon Ralph, Lowe, & Rogers, 2007). Whilst the case-series methodology clearly has a number of strengths which benefit basic science and clinical practice, like group studies, the approach still relies heavily both on the ability to recruit reasonably coherent patient groups and also to know, a priori, what the relevant groupings of patient are. Indeed, we note here that one of the most common uses and development platforms of the case-series methodology were made through investigations of semantic dementia (Hodges et al., 1992; Lambon Ralph et al., 2001; Woollams et al., 2007), which is a highly consistent group of patients and thus makes this methodology particularly successful and powerful. It is in this context, therefore, that we consider a related approach – 4 ACCEPTED MANUSCRIPT the use of principal component analysis (PCA) as a data-driven method which uses the patterns of individual differences in order both to reveal the statistically-reliable distinctions within a patient dataset and also to place individual cases, relative to each other, in the resultant multidimensional model (Butler, Lambon Ralph, & Woollams, 2014; Lambon Ralph et al., 2002; Lambon Ralph et al., 2003). In effect, the PCA method spans group and RI PT individual levels of analysis because the emergent set of principal components provide a coherent, generalizable set of factors or model of the underlying systems whilst, unlike group studies, variations in individual patient performance can be placed systematically along one or more of the principal axes. If a group is relatively homogenous then PCA only generates SC one principal component along which all patients can be mapped. If, however, there are systematic differences within the cohort, then one or more statistically-independent (orthogonal) factors emerge. In contrast unsystematic sources of variation, such as M AN U measurement error and other random fluctuations are left as unexplained variance (or loaded onto many additional components with little explanatory power, i.e., with eigenvalues less than 1). In addition, in more recent work, we have demonstrated that the PCA factors can also be used to explore behaviour-brain correlates in a new fashion; rather than traditional lesionoverlapping methods (which rely on coherent group studies: (Caramazza & McCloskey, TE D 1988; Shallice, 1988) or correlating raw test scores with brain voxel status (as per lesionsymptom mapping: Bates et al., 2003), we have found that clear results emerge when the behavioural principal components are related to the patients’ lesion distributions (Butler et al., 2014; Chase, 2014). EP In this study, therefore, we demonstrate and use the PCA approach in order to explore the nature of post-stroke aphasia. Aphasia is an interesting test case/problem for multiple AC C reasons. Whilst it is possible to distinguish between patients with and without acquired language impairments, within the aphasic group there are multiple sources of individual differences. Traditionally, category-based classification schemes have been developed in order to assist with different diagnosis of aphasic subgroups, clinical management and interventions, as well as for basic science studies that have tried to relate aphasic profiles to underlying lesion distributions. Whilst these category-based systems may provide a useful ‘broad brushstroke’ summary for individual patients, there are multiple limitations which follow from the fact that the underlying distribution of patient data reflects continuous variations along multiple dimensions rather than coherent, mutually-exclusive categories. Instead, there are both very fuzzy boundaries between ‘categories’ and considerable variation of profile within each ‘category’. If this conceptualisation of post-stroke aphasia is correct 5 ACCEPTED MANUSCRIPT then we might be able to use PCA in order to define the core underlying dimensions and the relative positions of each individual patient along these axes. Secondly, in turn, this multidimensional behavioural description can then be compared directly to the variation in the patients’ lesion distributions. Before considering the PCA findings in this study, we first briefly consider some of RI PT the key clinical characteristics of chronic post-stroke aphasia. Speech production, phonological and semantic impairments are common following left hemisphere stroke and the nature of these deficits forms the principal behavioural divisions in most aphasia classification systems. Typically, patients are first divided into fluent (relatively effortless SC speech output) or non-fluent types (effortful, slow or reduced complexity and length) and then additionally by the status of their phonological and semantic skills (Goodglass, 1993; Goodglass & Kaplan, 1983; Kertesz, 1982). Despite their prominence in aphasia diagnostics M AN U and importance for understanding intact language function, the unique neural correlates of speech dys-/fluency, semantic and phonological deficits remain unclear. Furthermore, when these have been considered, they have been investigated separately and thus the interrelationship between them is unclear (e.g., is speech fluency, in part, a reflection of patients’ phonological or semantic skills?). TE D Classically, damage to left frontal areas was associated with poor fluency (Broca, 1861); however, studies using high-resolution MRI techniques have found that left frontal lesions do not always result in Broca’s aphasia (Basso, Lecours, Moraschini, & Vanier, 1985; Willmes & Poeck, 1993) and production deficits can occur following lesions outside of the EP frontal lobe, including white matter tracts and the anterior insula (aINS) (Bates et al., 2003; Blank et al., 2002; Damasio, 1992; Dronkers, 1996; Mohr et al., 1978; Wise, Greene, Büchel, AC C & Scott, 1999). Functional neuroimaging studies have identified a broader neural network during connected-speech production. For example, by comparing propositional with automatic speech (e.g., counting), a PET imaging study found extensive left hemisphere activation including the frontal lobe, pre supplementary motor area (pre-SMA), angular gyrus (AG), fusiform gyrus (FG), and perisylvian areas (Blank et al., 2002). A large-scale left hemisphere speech production network was also identified in a study that measured fluency, complexity and variety of post-stroke aphasic speech (Borovsky, Saygin, Bates, & Dronkers, 2007). These critical sub-components could not be separated when they were related to the patients’ neural damage, which hints at the multi-faceted nature of speech fluency. Specifically, considerable anatomical overlap was found across the measures: the number of speech tokens (T) correlated with aINS, pre-motor cortex (PMC); mean length of utterance 6 ACCEPTED MANUSCRIPT (MLU) with aINS, PMC and anterior superior temporal gyrus (aSTG); and type-token ratio (TTR) with AG, pMTG and aSTG. This outcome probably reflects the inter-correlations between both the behavioural measures and lesion distributions (middle cerebral artery MCA) lesions tend to sample the same subset of brain regions) (Phan, Donnan, Wright, & Reutens, 2005), and because severity-related variance is shared across individual RI PT measurements. In addition, it is likely that fluency itself is a multi-faceted aphasic feature (Basilakos et al., 2014), which may benefit from decomposition into unique elements before the relationship with lesions can be satisfactorily explored. Plus, to identify the neural correlates of the full range of aphasic features, it is critical to consider fluency simultaneously SC alongside phonological and semantic skills. To achieve this aim, the current study utilised the PCA methodology (Butler et al., 2014; Chase, 2014) to reveal, for the first time, the unique neural correlates of speech M AN U fluency, semantic and phonological abilities in chronic post-stroke aphasia. This extends previous findings by taking into account speech fluency measures as part of the wider behavioural profile (phonological and semantic deficits), which is a key dimension on which aphasia patients are typically classified. Specifically, two approaches for tackling the analytical challenges associated with brain-behaviour mapping in chronic aphasia are to TE D derive statistically-independent behavioural factors/dimension from a rotated principal component analysis (PCA), and lesion volume as a covariate for anatomical severity. Previous studies using PCA on neuropsychological data have also shown that it offers various benefits. First, there is additional statistical reliability by combining results from multiple EP tests (Lambon Ralph et al., 2002; Lambon Ralph et al., 2003). Secondly, by maintaining orthogonal components, it is possible avoid problems of collinearity when undertaking AC C lesion-symptom correlations. Thirdly, the use of varimax rotation promotes a clear cognitive interpretation of each of the principal components. Finally, when the rotated components are combined with lesion-symptom mapping, it is possible to establish the neural correlates of unique components of post-stroke aphasic behaviour (Butler et al., 2014; Mirman et al., 2015a; Mirman et al., 2015b). In summary, this study investigated the multi-dimensionality of post stroke aphasic deficits and fluency classification using this novel PCA-lesion mapping approach (Butler et al., 2014). The investigation was conducted in three stages: (i) confirmation of the basic results from a previous focussed-investigation of post-stroke fluency (Borovsky et al., 2007) – which represents a more “standard” approach to brain-behaviour correlation; (ii) decomposition of speech production measures using rotated PCA and identification of the 7 ACCEPTED MANUSCRIPT neural correlates; and then, most importantly, (iii) simultaneous decomposition of a large behavioural dataset that included measures of speech production, phonology, semantics and executive abilities using an omnibus PCA, and subsequent identification of their unique neural correlates. RI PT Methods Participants Thirty one chronic stroke patients (either ischaemic or haemorrhagic) were recruited (the same patients as reported by Butler et al., 2014), who had impairment in producing SC and/or understanding spoken language and their aphasia was classified using the BDAE. No restrictions were placed according to aphasia type or severity (spanning global to minimal aphasia). All patients were at least 12 months post-stroke at the time of scanning and M AN U assessment, were native English speakers with normal or corrected-to-normal hearing and vision (see Table 1 for a summary of demographic details). Our selection criteria excluded participants if they had any contraindications for scanning, were pre-morbidly left handed, had more than one stroke or had any other significant neurological conditions. Informed consent was obtained from all participants prior to participation under approval from the local TE D ethics committee. Structural imaging data from a healthy age and education matched control group (8 female, 11 male) were used to determine the lesion outline in the patients using the Seghier et al., (2008) lesion identification toolbox. EP Neuropsychological Assessments and Analysis In order to assess speech production deficits, participants were asked to undertake a AC C picture description task (‘Cookie theft’ from the Boston Diagnostic Aphasia Examination) (Goodglass & Kaplan, 1983), which was audio recorded. Participants were instructed to “tell me everything you see going on in this picture”. In addition to the fluency measures, we utilised an extensive battery of neuropsychological tests to assess participants’ language and cognitive abilities (described in Butler et al., 2014), enabling us to understand how speech production measures relate to the patients’ input and output phonological, semantic and general executive abilities. These included subtests from the Psycholinguistic Assessments of Language Processing in Aphasia (PALPA) battery (Kay, Lesser, & Coltheart, 1992): auditory discrimination using non-word (PALPA 1) and word minimal pairs (PALPA 2); and immediate and delayed repetition of non-words (PALPA 8) and words (PALPA 9). Tests 8 ACCEPTED MANUSCRIPT from the 64-item Cambridge Semantic Battery (Bozeat et al., 2000) were included: spoken and written versions of the word-to-picture matching task; Camel and Cactus Test (picture); and the picture naming test. To increase the sensitivity to mild naming and semantic deficits we used The Boston Naming Test (BNT) (Kaplan, Goodglass, & Weintraub, 1983) and a written 96-trial synonym judgement test (Jefferies, Patterson, Jones, & Lambon Ralph, 2009). RI PT The spoken sentence comprehension task from the Comprehensive Aphasia Test (CAT) (Swinburn, Baker, & Howard, 2005) was used to assess sentential receptive skills. The additional cognitive tests included forward and backward digit span (Wechsler, 1987), the Brixton Spatial Rule Anticipation Task (Burgess & Shallice, 1997), and Raven’s Coloured SC Progressive Matrices (Raven, 1962). All scores were converted into percentage. Assessments were conducted with participants over several testing sessions, with the pace and number M AN U determined by the participant. The ‘Cookie Theft’ description was recorded and then transcribed. The coding procedure followed the method used by Borovsky et al., (2007) Each utterance was marked and bound morphemes, repetitions, false starts, retraces, unintelligible material and interruptions were coded separately. Repeated and retraced utterances were excluded from analysis and only correct full utterances were coded. When the boundary of an utterance was TE D unclear, or quite lengthy, transcribers applied the rule from Lee (1974) that only one “and” conjunction per sentence was allowed when the “and” connected two independent clauses. From these transcriptions, we extracted the number of word tokens (T) and types, type/token ratio (TTR), number of morphemes and mean length of utterance in morphemes (MLU). In EP order to control for the length of response given by each participant, we also computed words-per-minute (WPM). We considered these measures as indices of speech fluency as AC C they are among the simplest to quantify and mirror those commonly used in other studies and clinical practice. There are a number of additional measures that could be used to elaborate the model of fluency, including syntactical features (i.e. see Thompson et al., 2012). These, however, require a more open-ended speech sample and are relatively time-consuming to code. We believe that between the four measures used here, we are capturing the amount of speech output within a confined context. All scores were converted into percentage, where the max score across participants was used. Principal Component Analysis 9 ACCEPTED MANUSCRIPT This analysis was split into two parts. First, only the speech production measures were entered into a PCA with varimax rotation (SPSS 16.0). Factors with an eigenvalue exceeding 1.0 were extracted and then rotated. Following orthogonal rotation, the loadings of each test allowed a clear behavioural interpretation of each factor. Individual participants’ scores on 0each extracted factor were then used as behavioural covariates in the neuroimaging analysis. RI PT A second PCA was performed on a larger dataset which combined the background neuropsychological data with speech production measures. This allowed us to determine how the speech production factors related to the three previously-extracted (Butler et al., 2014) SC factors relating to phonology, semantic and executive abilities. Acquisition of Neuroimaging Data High resolution structural T1-weighted Magnetic Resonance Imaging (MRI) scans M AN U were acquired on a 3.0 Tesla Philips Achieva scanner (Philips Healthcare, Best, The Netherlands) using an 8-element SENSE head coil. A T1-weighted inversion recovery sequence with 3D acquisition was employed, with the following parameters: TR (repetition time) = 9.0 ms, TE (echo time) = 3.93 ms, flip angle = 8°, 150 contiguous slices, slice thickness = 1 mm, acquired voxel size 1.0 x 1.0 x 1.0 mm3, matrix size 256 x 256, FOV = TE D 256 mm, TI (inversion time) = 1150 ms, SENSE acceleration factor 2.5, total scan acquisition time = 575 s. Analysis of Neuroimaging Data software EP Structural MRI scans were pre-processed with Statistical Parametric Mapping (SPM8: Wellcome Trust Centre for Neuroimaging, AC C http://www.fil.ion.ucl.ac.uk/spm/). The images were normalised into standard Montreal Neurological Institute (MNI) space using a modified unified segmentation-normalisation procedure optimised for focal lesioned brains (Seghier et al., 2008). Data from all participants with stroke aphasia and all healthy controls were entered into the segmentationnormalisation. This procedure combines segmentation, bias correction and spatial normalisation through the inversion of a single unified model (see Ashburner & Friston, 2005 for more details). In brief, the unified model combines tissue class (with an additional tissue class for abnormal voxels), intensity bias and non-linear warping into the same probabilistic models that are assumed to generate subject-specific images. Images were then smoothed with an 8mm full-width-half-maximum (FWHM) Gaussian kernel and used in the lesion analyses described below. The lesion of each patient was automatically identified using an 10 ACCEPTED MANUSCRIPT outlier detection algorithm, compared to healthy controls, based on fuzzy clustering. The default parameters were used apart from the lesion definition ‘U-threshold’, which was set to 0.5 to create a binary lesion image. We modified the U-threshold from 0.3 to 0.5 after comparing the results obtained from a sample of patients to what would be nominated as lesioned tissue by an expert neurologist. The images generated for each patient were RI PT individually checked and visually inspected with respect to the original scan, and were used to create the lesion overlap map in Fig.1 and the individual lesion outlines in Fig.3. We selected the Seghier et al. (2008) method as it is objective and efficient for a large sample of patients (Wilke, de Haan, Juenger, & Karnath, 2011), in comparison to a labour intensive SC hand-traced lesion mask. The method has been shown to have a DICE overlap >0.64 with manual segmentation of the lesion and >0.7 with a simulated ‘real’ lesion (where real lesions are superimposed onto healthy brains; Seghier et al., 2008). We should note here, explicitly, M AN U that although commonly referred to as an automated ‘lesion’ segmentation method, the technique detects areas of unexpected tissue class – and, thus, identifies missing grey and white matter but also areas of augmented CSF space. We used the T1-weighted images with continuous signal intensity values across the whole brain and correlated these with individual measures or PCA factor scores using a TE D voxel-based correlational methodology (VBCM: Tyler, Marslen-Wilson, & Stamatakis, 2005), a variant of voxel-lesion symptom mapping (VSLM: Bates et al., 2003). VBCM does not require a binary classification of the intact/lesioned brain to be marked, as in the case of VSLM, as both the behaviour and tissue concentration measures are treated as continuous EP variables (conducted in SPM8). Three analyses were conducted and all reported anatomical regions were located on the left hemisphere. First, each fluency measure was analysed AC C separately, in an attempt to confirm the basic results from Borovsky et al. (2007). Secondly, the factor scores from the first speech production PCA were simultaneously entered to investigate how variation in tissue concentration corresponded to the unique components of speech dys-/fluency. Finally, the participants’ factor scores from the omnibus PCA of the entire neuropsychological test battery and speech measures were entered into a VBCM analysis. In order to ensure that the results were not merely attributable to lesion size, each participants’ lesion volume was calculated from the lesion identified by the automated lesion identification method (Seghier et al., 2008) and this was entered as a covariate in each VBCM. Hence, all analyses were performed with and without a correction for lesion volume. All anatomical labels were based on the Harvard-Oxford atlas in MNI space. 11 ACCEPTED MANUSCRIPT Results Neuropsychological and lesion profiles Table 2 summarises the participants’ scores on the speech production measures and a large neuropsychological battery (Butler et al., 2014) and is ordered according to patients’ scores on the Boston Naming Test. A lesion overlap map for stroke aphasic participants is RI PT provided in Fig.1A, and primarily covers the left hemisphere area supplied by the middle cerebral artery (Phan et al., 2005). The maximum number of participants who had a lesion in any one voxel was 26 (-45, -23, 27; ventral portion of the postcentral gyrus). Confirmation of previous basic findings on speech fluency alone SC A bivariate correlation analysis on the speech production measures showed a high degree of inter-correlation. The token measure was correlated with MLU and WPM (r=0.633, M AN U p<0.001 and r=0.513, p<0.001 respectively). The MLU measure was correlated with TTR and WPM (r=0.505, p=0.004 and r=0.674, p<0.001 respectively). No other correlations reached significance. As noted by Borovsky et al., (2007), it is not unexpected to find intercorrelations between these measures. Indeed, the shared variance could reflect common factors underlying speech, including semantics, phonology and general aphasia severity. As highlighted in the Introduction, these high intercorrelations provide a challenge when below). TE D attempting to relate the behavioural measures to unique aspects of the patients’ lesions (see The first VBCM analysis examined the neural correlates of fluency (T), complexity EP (MLU), semantic variety (TTR) and words-per-minute (WPM) each entered separately (Fig.2A-D). The results for T revealed involvement of the superior-posterior insula and portions of the frontal lobe including the pre-central gyrus and middle frontal gyrus. The AC C status of underlying white matter was also correlated. The analysis for MLU revealed an overlapping albeit broader region, extending throughout the left IFG, insula and putamen, along the SMA, motor cortex and the underlying white matter. The results for TTR, contrasted with those for T and MLU, in revealing a region that extended from SMG and pMTG/STG through to aSTG as well as a portion of the anterior/orbito frontal gyrus. Peak voxels were in the IFG/frontal pole, SMG and pSTG. Finally, the analysis for WPM revealed no significant correlations with lesion areas, although at an uncorrected voxel height of p=0.01 (cluster extent 50 voxels) the results were very similar to those obtained for T. 12 ACCEPTED MANUSCRIPT Effect of lesion size Given that some regions are more likely than others to be damaged after MCA stroke (Phan et al., 2005), we explored the intrinsic relationship between lesion location and size, and then controlled for this factor in the subsequent lesion-symptom analyses. Each participant’s volume was calculated from the lesion identified by the modified segmentation- RI PT normalisation procedure. Lesion volume itself was correlated with the integrity of various regions with the MCA territory (Fig.1B), representing the outer belt of the lesion overlap map (Fig.1A). Crucially, including lesion size in the VBCM analysis for each speech measure reduced the significance across all measures and the underlying neural correlates (see Fig.2A- SC D). The TTR and T measures did not correlate with any regions after accounting for lesion volume (even at a reduced image threshold of p<0.01, FWE-cluster corrected p<0.05). MLU was the only measure to reveal significant relationships with the left superior insula, putamen M AN U and underlying white matter after accounting for lesion volume. Identifying principal speech-production components The speech production measures were subjected to a rotated PCA and produced a two independent factor solution, which accounted for 84.65% of variance in participants’ TE D performance (F1=59.28%; F2=25.35%). The factor loading of each production measure is given in Table 3. The measures that tapped into overall quantity of speech (e.g., number of words and words-per-minute) loaded heavily on Factor 1; hence we refer to this factor as ‘speech quanta’. Factor 2 was interpreted as ‘semantic variety’, as the TTR measurement EP loaded heavily on it (e.g., proportion of unique words spoken). The MLU loaded heavily on Factor 1, while loading to a lesser degree on Factor 2. Lesion analysis for each factor, entered AC C simultaneously, is displayed in Fig.2E. The speech quanta measure (F1) was uniquely correlated with the left precentral gyrus, superior insula and putamen, and the underlying white matter (including the superior longitudinal fasciculus, caudate-premotor tracts and the frontal ‘aslant’ tract that connects the medial-superior portion of the frontal lobe to the inferior-lateral frontal region) (Basilakos et al., 2014; Catani et al., 2012; Rojkova et al., 2015). The semantic variety factor (F2) was uniquely correlated with the SMG and pMTG/STG plus the anterior/orbito frontal gyrus. Weaker correlation is observed along the STG to anterior lateral portions. After adding a covariate for lesion volume, no significant clusters remained for F1 or F2. Identifying principal speech, language and executive behavioural factors 13 ACCEPTED MANUSCRIPT In order to determine the relationship between speech production measures and a range of language and cognitive skills, the factor scores from the speech PCA were correlated with the factor scores for phonology, semantics and executive abilities, computed in our previous study (Butler et al., 2014). The speech quanta factor only correlated with the first (phonology) factor (r=0.404, p=0.01). The semantic variety speech factor only correlated RI PT with the second (semantics) factor (r=0.452, p<0.001). To understand the relationship between speech fluency and various language and cognitive measures more thoroughly, an omnibus PCA was conducted. This resulted in a four factor solution including phonological skills (54.8% variance), semantic ability (11.6% variance) and cognitive-executive function SC (8.3% variance). We note here that a variety of different tasks over-and-above the executive tests loaded on this third factor. These include measures associated with auditoryphonological processing (e.g., minimal pairs) or semantic memory (e.g., CCT), although in M AN U both cases the loading on the component one would typically associate the tests with are generally high. This outcome probably reflects the fact that the PCA process ‘decomposes’ the components of each task (reflecting the fact that no task is a pure measure of a single underpinning cognitive function). In this case, all the tasks that load onto the third factor require a high degree of additional ‘problem-solving’ and executively-loaded working TE D memory. In this omnibus PCA, the ‘semantic variety’ characteristics of speech production were subsumed into the general semantics factor and a new additional fourth factor emerged based entirely on ‘speech quanta’ (6.3% variance) (see Table 4 for factor loadings and Table 5 for EP the patients’ factors scores). The VBCM analysis for each independent factor, entered simultaneously, is shown in Fig.3A and peak MNI co-ordinates are shown in Table 6. Each AC C cluster shows where tissue concentration covaries uniquely with a given factor score, after accounting for lesion volume (the results without the lesion volume covariate were very similar). Results are thresholded at p<0.005 voxel-level, p≤0.001 family-wise error (FWE)corrected cluster-level. Performance on the phonological factor was uniquely correlated with voxels across a number of left hemisphere regions: primary auditory cortex (BA41 and 42); mid to posterior middle and superior temporal gyri (MTG and STG); superior temporal sulcus (STS); and posterior portions of the insula, Heschl’s gyrus and the planum temporale. The phonological cluster also overlapped with white matter regions, encompassing part of the arcuate fasciculus, a key aspect of the dorsal language pathway (Catani & ffytche, 2005; Catani, Jones, & ffytche, 2005; Duffau et al., 2005; Parker et al., 2005; Saur et al., 2008). 14 ACCEPTED MANUSCRIPT Performance on the semantic factor was uniquely related to a cluster in the left hemisphere anterior temporal lobe (ATL; see Fig.3A). The cluster overlapped with the anterior MTG and the temporal stem (including the dorsal edges of the inferior temporal gyrus (ITG) and fusiform gyrus). Thus, with regards to white matter, the cluster included part of the ventral language route, overlapping with the inferior longitudinal fasciculus, inferior RI PT fronto-occipital fasciculus and uncinate fasciculus (Duffau, Gatignol, Moritz-Gasser, & Mandonnet, 2009; Mummery, Ashburner, Scott, & Wise, 1999; Parker et al., 2005; Saur et al., 2008). Performance on the speech quanta factor was uniquely related to a cluster in the left SC hemisphere precentral gyrus, superior insula and putamen, extending medially to the caudate nucleus. With regards to white matter, the cluster included an area corresponding to the anterior part of the dorsal language route (superior longitudinal fasciculus) as well as the M AN U caudate-premotor tracts and the frontal ‘aslant’ tract (Rojkova et al., 2015). It is notable that the effect of lesion size on the speech quanta factor when calculated using all neuropsychological scores is negligible which was not the case when using the speech production measures alone (see above, Fig.2E and Fig.3A). In contrast to the phonological, semantic and fluency factors, there were no clusters TE D that correlated uniquely with the executive factor score which survived correction for multiple comparisons. This factor did, however, correlate with lesion volume in a simple ability. Individual Cases EP correlation (r(31) = -.373, p=.039), suggesting that larger lesions result in poorer executive AC C The relationship of individual patients to the group-level analyses was explored for two reasons. First, to test clinical utility, it is important to explore how individual results relate to the group-level maps for each language-cognitive factor. Secondly, exemplar cases help interpretation of the behavioural factors and their neural correlates (given that PCA, by design, generates scores that are one step removed from raw clinical measures: see Lambon Ralph et al., 2003). To achieve this aim, the group was divided into two using a median split based on executive factor scores to cover a range of patient types. Two participants were selected from each split to provide contrasting pairs who scored high or low on the speech quanta factor (see Fig. 3B and 3C). In the high executive group, patient KW (Broca’s aphasia) had the lowest speech quanta score (orange outline) and EB (anomic) had the highest (yellow outline). In the low executive group, patient ESb (global) had the lowest 15 ACCEPTED MANUSCRIPT speech quanta score (purple outline) and KS (TSA) the highest (pink outline). It is clear and encouraging from these examples that the individual patient lesions and associated behavioural profiles both align with the group-level VBCM results and are stable regardless of overall executive functioning (as should be expected given that the PCA approach extracts RI PT statistically-independent factors). Discussion Mirroring the principal features of natural language, speech fluency, phonological and semantic deficits are common following left MCA stroke. In most aphasia classification SC systems, patients are first divided by speech fluency followed by the relative balance of the patients’ phonological and semantic abilities (Goodglass & Kaplan, 1983; Kertesz, 1982). By measuring and applying novel analyses to the overall rate, complexity and variety of M AN U naturalistic speech simultaneously with other measures of language performance, we were able to provide insights into their neural substrates. The results provide novel insights not only about the core features and neural bases of naturalistic language, but also how these underpin the varieties and nature of post-stroke aphasia. In the first analysis, we largely confirmed the basic speech fluency results from TE D Borovsky et al., (2007), by showing that the rate of tokens (T) and mean length of utterance (MLU) in the patients’ speech correlated with the insula, SMA and underlying white matter. In contrast, type-to-token ratio (TTR) correlated with the SMG, pMTG, aSTG, and anterior/orbito frontal gyrus. Whilst these results are interesting and support previous EP observations that speech fluency is a multi-faceted aphasic feature (Ackermann & Riecker, 2004; Basilakos et al., 2014), their interpretation is complicated by the fact that these AC C behaviour-lesion correlations were not significant when lesion volume was controlled and no account is taken of the patients’ other aphasic features – thus the picture of the lesionsymptom mappings is incomplete. We addressed these limitations by utilising the PCA-VBCM methodology in the two subsequent stages. We initially undertook a PCA of the speech production measures alone. This PCA suggested that the clinical concept of ‘fluency’ reflects two more general and independent factors: ‘speech quanta’ (the quantity of speech produced) and ‘variety’ (the range of different words/token irrespective of the quantity produced). Scores from the speech quanta factor correlated uniquely with tissue damage in the left precentral gyrus, superior insula and putamen, and the neighbouring white matter (including the superior longitudinal fasciculus, caudate-premotor tracts and the frontal ‘aslant’ tract that connects the medial16 ACCEPTED MANUSCRIPT superior portion of the frontal lobe to the inferior-lateral frontal region) (Rojkova et al., 2015). In contrast, the speech variety factor was uniquely related to the SMG and pMTG/STG plus the anterior/orbito frontal gyrus. However, again the results for these speech production-only data were largely confounded with lesion volume. These twin challenges of intercorrelation and severity were overcome in the final step RI PT in which we completed an omnibus PCA containing all the detailed behavioural data and the speech production measures, which also allowed us to explore the relationship between speech fluency and other aphasic features (Basilakos et al., 2014). This analysis revealed four unique factors (phonology, executive processing, semantic function which included the SC speech variety measure, and speech quanta). The phonological factor uniquely correlated with damage to left perisylvian areas, including left mid to posterior STG, MTG, STS and HG, as well as the underlying white matter (arcuate fasciculus portion of the dorsal language M AN U pathway). The semantic factor correlated uniquely with left anterior-MTG and the underlying temporal stem. The speech quanta factor correlated uniquely with left hemisphere precentral gyrus, superior insula and putamen, extending medially to the caudate nucleus. With regards to white matter, the cluster included an area corresponding to the anterior part of the dorsal language route (superior longitudinal fasciculus) as well as the caudate-premotor tracts and TE D the frontal ‘aslant’ tract. The executive factor did not covary, uniquely, with any brain regions, as found previously (Butler et al., 2014). The speech quanta-related region included key cortical and subcortical areas associated with deficits of articulatory planning (Bates et al., 2003; Borovsky et al., 2007; EP Dronkers, 1996) and motor coordination of speech related movements (Ackermann & Riecker, 2004). Likewise, direct stimulation of ventral premotor cortex and pars opercularis AC C leads to speech inhibition (Kinoshita et al., 2014). The underlying white matter damage might also be crucial and, indeed, the core of the identified speech quanta-related region overlaps with the frontal ‘aslant’ tract (FAT: Rojkova et al., 2015; see Figure 4). The integrity of this tract has been shown to correlate with non-fluent speech in primary progressive aphasia (Catani et al., 2013; Mandelli et al., 2014) and, when combined with the integrity of the anterior section of the arcuate fasciculus, to correlate with clinician-rated fluency in poststroke aphasia (Basilakos et al., 2014). In keeping with our finding of a separate principal component for speech quanta over and above phonology and semantic skills, Catani et al. (2013) did not find a relationship between FAT integrity and repetition or syntactic abilities in primary progressive aphasia. Additional convergent evidence has been derived from neurosurgical investigations: direct electrical stimulation of FAT inhibits speech and 17 ACCEPTED MANUSCRIPT subsequent surgical resection leads to transient postoperative inhibition (Kinoshita et al., 2014). The results for the phonological, semantic and executive factors reproduced those observed in our previous study and related investigations (Butler et al., 2014; Mirman et al., 2015a; Mirman et al., 2015b). It is interesting to note that speech ‘variety’ seems to be RI PT inherently linked to the patients’ degree of remaining semantic abilities (factor 2) presumably because the use of a wide vocabulary is underpinned by fine semantic distinctions. More generally, the observed neural correlates of the phonological and semantic factors are highly consistent with data from other methods. Phonological processing was associated with left SC HG, mid to posterior MTG, STG and STS, which have all been identified in various previous reviews of phonological processing (Hickok & Poeppel, 2004, 2007; Price, 2010, 2012; Vigneau et al., 2006; Wise et al., 2001) and rTMS to pSTG has been shown to increase error M AN U rates in the speech production of neurologically-intact participants (Acheson, Hamidi, Binder, & Postle, 2011). The neural substrate for semantic performance (left anterior temporal regions) again converges with findings from large scale fMRI reviews (Binder, Desai, Graves, & Conant, 2009; Price, 2010; Vigneau et al., 2006; Visser, Jefferies, & Lambon Ralph, 2009), data from semantic dementia (Mion et al., 2010), rTMS (Pobric, Jefferies, & TE D Lambon Ralph, 2007), direct electrical stimulation of IFOF (Duffau et al., 2008) and neuroanatomically-constrained models (Ueno, Saito, Rogers, & Lambon Ralph, 2011). One thing that must be kept in mind when using PCA is that the factor solution is always dependent on the tests included in the decomposition. Thus, if an analysis fails to EP include behavioural measures that reflect a core feature of the target clinical populations then, by definition, the resultant PCA model will not derive a principal dimension for this AC C clinically-relevant feature. Our current PCA model contains measures of phonological, semantic, speech fluency and executive skill. Whilst this model captures some of the core aspects of post-stroke aphasia, future studies will need to augment the range of assessments in order to capture other key features of this clinical disorder including syntactic processing, speech errors, written language processing, etc. To conclude, in addition to revealing specific behavioural and neuroanatomical information about the nature of chronic post-stroke aphasia, the current study serves as a worked example for the utility of our PCA-VBCM method. Rather than adopting either a classical group-study or single-case investigation, the PCA data-driven approach not only generates a single unified model for the group as a whole (expressed in terms of the four emergent principal components) but is also able to capture the individual differences between 18 ACCEPTED MANUSCRIPT patients (in terms of their relative positions along the principal behavioural axes). This method not only preserves the individual-level data (as per single-case and case-series methods) but is also able to place it within the broader context of the whole group (akin to group studies). Thus rather than ignoring or averaging across individual variations, the methodology actively embraces individual differences and extracts coherent variations. We RI PT also note here that by utilising varimax rotation, the emergent factors become relatively straightforward to interpret from a cognitive perspective and the fact that they are statistically-independent makes them ideal for the neuroimaging analyses (which call for orthogonal predictors rather than the strong inter-correlations that are found between the raw SC test results). Whilst we have applied this PCA-VBCM approach to post-stroke chronic aphasia in the current investigation, this methodological approach should be applicable and AC C EP TE D M AN U beneficial across a range of acute and progressive neurological conditions. 19 ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT Acknowledgements: We are grateful to all the patients and carers for their continued support of our research programme. We thank Dr. Rebecca Butler for collection of the speech data analysed in this study and to Prof. Marco Catani and Dr. Michel Thiebaut de Schotten for providing the anatomical overlay for the frontal aslant tract. This research was support by funding from the Rosetrees Trust, an NIHR (UK) senior investigator award and an MRC (UK) programme grant to MALR (MR/J004146/1). 20 ACCEPTED MANUSCRIPT References Acheson, D. J., Hamidi, M., Binder, J. R., & Postle, B. R. (2011). A Common Neural Substrate for Language Production and Verbal Working Memory. Journal of Cognitive Neuroscience, 23(6), 1358-1367. doi: 10.1162/jocn.2010.21519 Ackermann, H., & Riecker, A. (2004). The contribution of the insula to motor aspects of http://dx.doi.org/10.1016/S0093-934X(03)00347-X RI PT speech production: A review and a hypothesis. Brain and Language, 89(2), 320-328. doi: Ashburner, J., & Friston, K. J. (2005). Unified segmentation. NeuroImage, 26(3), 839-851. doi: http://dx.doi.org/10.1016/j.neuroimage.2005.02.018 SC Basilakos, A., Fillmore, P. T., Rorden, C., Guo, D., Bonilha, L., & Fridriksson, J. (2014). Regional White Matter Damage Predicts Speech Fluency in Chronic Post-Stroke Aphasia. M AN U Frontiers in Human Neuroscience, 8, 845. doi: 10.3389/fnhum.2014.00845 Basso, A., Lecours, A. R., Moraschini, S., & Vanier, M. (1985). Anatomoclinical correlations of the aphasias as defined through computerized tomography: Exceptions. Brain and Language, 26(2), 201-229. doi: http://dx.doi.org/10.1016/0093-934X(85)90039-2 Bates, E., Wilson, S. M., Saygin, A. P., Dick, F., Sereno, M. I., Knight, R. T., & Dronkers, N. F. (2003). Voxel-based lesion-symptom mapping. [10.1038/nn1050]. Nature TE D Neuroscience, 6(5), 448-450. doi: http://dx.doi.org/10.1038/nn1050 Binder, J. R., Desai, R. H., Graves, W. W., & Conant, L. L. (2009). Where Is the Semantic System? A Critical Review and Meta-Analysis of 120 Functional Neuroimaging Studies. Cerebral Cortex, 19(12), 2767-2796. doi: http://dx.doi.org/10.1093/cercor/bhp055 production: EP Blank, C., Scott, S. K., Murphy, K., Warburton, E. A., & Wise, R. J. S. (2002). Speech Wernicke, Broca and beyond. Brain, 125(8), 1829-1838. doi: AC C http://dx.doi.org/10.1093/brain/awf191 Borovsky, A., Saygin, A. P., Bates, E., & Dronkers, N. (2007). Lesion correlates of conversational speech production deficits. Neuropsychologia, 45(11), 2525-2533. doi: http://dx.doi.org/10.1016/j.neuropsychologia.2007.03.023 Bozeat, S., Lambon Ralph, M. A., Patterson, K., Garrard, P., & Hodges, J. R. (2000). Nonverbal semantic impairment in semantic dementia. Neuropsychologia, 38(9), 1207-1215. doi: http://dx.doi.org/10.1016/S0028-3932(00)00034-8 Broca, P. (1861). Remarques sur le siege de la faculte du language articule suivies d'une observation d'aphemie. Bulletins Société anatomique de Paris, 6, 330-357. 21 ACCEPTED MANUSCRIPT Burgess, P. W., & Shallice, T. (1997). The Hayling and Brixton Tests. Thurston, Suffolk: Thames Valley Test Company. Butler, R. A., Lambon Ralph, M. A., & Woollams, A. M. (2014). Capturing multidimensionality in stroke aphasia: mapping principal behavioural components to neural structures. [Journal Article]. Brain, 137(12), 3248-3266. doi: 10.1093/brain/awu286 RI PT Caramazza, A., & McCloskey, M. (1988). The case for single-patient studies. Special Issue: Methodological problems in cognitive neuropsychology. Cognitive Neuropsychology, 5(5), 517-527. Catani, M., Dell’Acqua, F., Vergani, F., Malik, F., Hodge, H., Roy, P., . . . Thiebaut de SC Schotten, M. (2012). Short frontal lobe connections of the human brain. Cortex, 48(2), 273-291. doi: http://dx.doi.org/10.1016/j.cortex.2011.12.001 128). M AN U Catani, M., & ffytche, D. H. (2005). The rises and falls of disconnection syndromes (Vol. Catani, M., Jones, D. K., & ffytche, D. H. (2005). Perisylvian language networks of the human brain. Annals of Neurology, 57(1), 8-16. doi: 10.1002/ana.20319 Catani, M., Mesulam, M. M., Jakobsen, E., Malik, F., Martersteck, A., Wieneke, C., . . . Rogalski, E. (2013). A novel frontal pathway underlies verbal fluency in primary aphasia. 10.1093/brain/awt163 [Journal Article]. Brain, 136(8), 2619-2628. doi: TE D progressive Chase, A. (2014). Stroke: Improved lesion-symptom mapping in poststroke aphasia. [Research Highlight]. Nat Rev Neurol, 10(12), 674-674. doi: 10.1038/nrneurol.2014.217 EP Damasio, A. R. (1992). Aphasia. New England Journal of Medicine, 326(8), 531-539. doi: doi:10.1056/NEJM199202203260806 AC C Dronkers, N. F. (1996). A new brain region for coordinating speech articulation. [10.1038/384159a0]. Nature, 384(6605), 159-161. Duffau, H., Gatignol, P., Mandonnet, E., Capelle, L., & Taillandier, L. (2008). Intraoperative subcortical stimulation mapping of language pathways in a consecutive series of 115 patients with Grade II glioma in the left dominant hemisphere. Journal of Neurosurgery, 109(3), 461-471. doi: doi:10.3171/JNS/2008/109/9/0461 Duffau, H., Gatignol, P., Mandonnet, E., Peruzzi, P., Tzourio-Mazoyer, N., & Capelle, L. (2005). New insights into the anatomo-functional connectivity of the semantic system: a study using cortico-subcortical electrostimulations (Vol. 128). 22 ACCEPTED MANUSCRIPT Duffau, H., Gatignol, P., Moritz-Gasser, S., & Mandonnet, E. (2009). Is the left uncinate fasciculus essential for language? Journal of Neurology, 256(3), 382-389. doi: 10.1007/s00415-009-0053-9 Goodglass, H. (1993). Understanding aphasia. San Diego: Academic Press. Goodglass, H., & Kaplan, E. (1983). The assessment of aphasia and related disorders (2nd RI PT ed.). Philadelphia: Lea & Febiger. Hickok, G., & Poeppel, D. (2004). Dorsal and ventral streams: a framework for understanding aspects of the functional anatomy of language. Cognition, 92(1-2), 67-99. Hickok, G., & Poeppel, D. (2007). The cortical organization of speech processing. Nature SC Reviews Neuroscience, 8(5), 393-402. Hodges, J. R., Patterson, K., Oxbury, S., & Funnell, E. (1992). Semantic dementia: Progressive fluent aphasia with temporal lobe atrophy. Brain, 115(Pt6), 1783-1806. M AN U Jefferies, E., & Lambon Ralph, M. A. (2006). Semantic impairment in stroke aphasia vs. semantic dementia: A case-series comparison. Brain, 129, 2132-2147. Jefferies, E., Patterson, K., Jones, R. W., & Lambon Ralph, M. A. (2009). Comprehension of concrete and abstract words in semantic dementia. Neuropsychology, 23(4), 492-499. doi: 10.1037/a0015452 Febiger. TE D Kaplan, E., Goodglass, H., & Weintraub, S. (1983). Boston Naming Test. Philadelphia: Lea & Kay, J., Lesser, R., & Coltheart, M. (1992). Psycholinguistic Assessments of Language Processing in Aphasia. Hove, UK: Lawrence Erlbaum Associates Ltd. EP Kertesz, A. (1982). Western aphasia battery. New York: Grune & Stratton. Kinoshita, M., de Champfleur, N., Deverdun, J., Moritz-Gasser, S., Herbet, G., & Duffau, H. AC C (2014). Role of fronto-striatal tract and frontal aslant tract in movement and speech: an axonal mapping study. Brain Structure and Function, 1-14. doi: 10.1007/s00429-0140863-0 Lambon Ralph, M. A., Lowe, C., & Rogers, T. T. (2007). Neural basis of category-specific semantic deficits for living things: evidence from semantic dementia, HSVE and a neural network model. Brain, 130, 1127-1137. Lambon Ralph, M. A., McClelland, J. L., Patterson, K., Galton, C. J., & Hodges, J. R. (2001). No right to speak? The relationship between object naming and semantic impairment: Neuropsychological evidence and a computational model. Journal of Cognitive Neuroscience, 13, 341-356. 23 ACCEPTED MANUSCRIPT Lambon Ralph, M. A., Moriarty, L., & Sage, K. (2002). Anomia is simply a reflection of semantic and phonological impairments: Evidence from a case-series study. Aphasiology, 16(1-2), 56-82. doi: 10.1080/02687040143000448 Lambon Ralph, M. A., Patterson, K., Graham, N., Dawson, K., & Hodges, J. R. (2003). Homogeneity and heterogeneity in mild cognitive impairment and Alzheimer’s disease: a RI PT cross sectional and longitudinal study of 55 cases (Vol. 126). Lambon Ralph, M. A., Patterson, K., & Plaut, D. C. (2011). Finite case series or infinite single-case studies? Comments on 'Case series investigations in cognitive neuropsychology' by Schwartz and Dell (2010). Cognitive Neuropsychology, 28(7), 466- SC 474. Lee, L. L. (1974). Developmental sentence analysis. Chicago, IL: Northwestern University Press. M AN U Mandelli, M. L., Caverzasi, E., Binney, R. J., Henry, M. L., Lobach, I., Block, N., . . . GornoTempini, M. L. (2014). Frontal White Matter Tracts Sustaining Speech Production in Primary Progressive Aphasia. The Journal of Neuroscience, 34(29), 9754-9767. doi: 10.1523/jneurosci.3464-13.2014 Mion, M., Patterson, K., Acosta-Cabronero, J., Pengas, G., Izquierdo-Garcia, D., Hong, Y. TE D T., . . . Nestor, P. J. (2010). What the left and right anterior fusiform gyri tell us about semantic memory (Vol. 133). Mirman, D., Chen, Q., Zhang, Y., Wang, Z., Faseyitan, O. K., Coslett, H. B., & Schwartz, M. F. (2015a). Neural organization of spoken language revealed by lesion-symptom mapping. EP [Article]. Nat Commun, 6. doi: 10.1038/ncomms7762 Mirman, D., Zhang, Y., Wang, Z., Coslett, H. B., & Schwartz, M. F. (2015b). The ins and AC C outs of meaning: Behavioral and neuroanatomical dissociation of semantically-driven word retrieval and multimodal semantic recognition in aphasia. Neuropsychologia(0). doi: http://dx.doi.org/10.1016/j.neuropsychologia.2015.02.014 Mohr, J., Pessin, M., Finkelstein, S., Funkenstein, H., Duncan, G., & Davis, K. (1978). Broca aphasia: Pathologic and clinical. Neurology, 28(4), 311-324. Mummery, C. J., Ashburner, J., Scott, S. K., & Wise, R. J. S. (1999). Functional neuroimaging of speech perception in six normal and two aphasic subjects. The Journal of the Acoustical Society of America, 106(1), 449-457. Parker, G. J. M., Luzzi, S., Alexander, D. C., Wheeler-Kingshott, C. A. M., Ciccarelli, O., & Lambon Ralph, M. A. (2005). Lateralization of ventral and dorsal auditory-language pathways in the human brain. NeuroImage, 24(3), 656-666. 24 ACCEPTED MANUSCRIPT Phan, T. G., Donnan, G. A., Wright, P. M., & Reutens, D. C. (2005). A Digital Map of Middle Cerebral Artery Infarcts Associated With Middle Cerebral Artery Trunk and Branch Occlusion. Stroke, 36(5), 986-991. doi: 10.1161/01.STR.0000163087.66828.e9 Pobric, G., Jefferies, E., & Lambon Ralph, M. A. (2007). Anterior temporal lobes mediate semantic representation: Mimicking semantic dementia by using rTMS in normal RI PT participants. Proceedings of the National Academy of Sciences, 104(50), 20137-20141. doi: http://dx.doi.org/10.1073/pnas.0707383104 Price, C. J. (2010). The anatomy of language: a review of 100 fMRI studies published in 2009. Annals of the New York Academy of Sciences, 1191(1), 62-88. doi: SC http://dx.doi.org/10.1111/j.1749-6632.2010.05444.x Price, C. J. (2012). A review and synthesis of the first 20 years of PET and fMRI studies of heard speech, spoken language and reading. NeuroImage, 62(2), 816-847. doi: M AN U http://dx.doi.org/10.1016/j.neuroimage.2012.04.062 Raven, J. C. (1962). Advanced Progressive Matrices, Set II. London: H. K. Lewis. Rojkova, K., Volle, E., Urbanski, M., Humbert, F., Dell’Acqua, F., & Thiebaut de Schotten, M. (2015). Atlasing the frontal lobe connections and their variability due to age and education: a spherical deconvolution tractography study. [journal article]. Brain Structure TE D and Function, 221(3), 1751-1766. doi: 10.1007/s00429-015-1001-3 Saur, D., Kreher, B. W., Schnell, S., Kummerer, D., Kellmeyer, P., Vry, M.-S., . . . Weiller, C. (2008). Ventral and dorsal pathways for language. Proceedings of the National Academy of Sciences, 105(46), 18035-18040. doi: EP http://dx.doi.org/10.1073/pnas.0805234105 Schwartz, M. F., & Dell, G. S. (2011). Case series investigations in cognitive AC C neuropsychology. Cognitive Neuropsychology, 27(6), 477-494. Schwartz, M. F., Dell, G. S., Martin, N., Gahl, S., & Sobel, P. (2006). A case-series test of the interactive two-step model of lexical access: Evidence from picture naming. Journal of Memory and Language, 54(2), 228-264. Seghier, M. L., Ramlackhansingh, A., Crinion, J. T., Leff, A. P., & Price, C. (2008). Lesion identification using unified segmentation-normalisation models and fuzzy clustering. NeuroImage, 41, 1253-1266. Shallice, T. (1988). From neuropsychology to mental structure. Cambridge: Cambridge University Press. Swinburn, K., Baker, G., & Howard, D. (2005). CAT: The Comprehensive Aphasia Test. New York: Psychology Press. 25 ACCEPTED MANUSCRIPT Thompson, C. K., Cho, S., Hsu, C.-J., Wieneke, C., Rademaker, A., Weitner, B. B., . . . Weintraub, S. (2012). Dissociations Between Fluency And Agrammatism In Primary Progressive Aphasia. Aphasiology, 26(1), 20-43. doi: 10.1080/02687038.2011.584691 Tyler, L. K., Marslen-Wilson, W., & Stamatakis, E. A. (2005). Dissociating neuro-cognitive component processes: voxel-based correlational methodology. Neuropsychologia, 43(5), RI PT 771-778. doi: http://dx.doi.org/10.1016/j.neuropsychologia.2004.07.020 Ueno, T., Saito, S., Rogers, Timothy T., & Lambon Ralph, Matthew A. (2011). Lichtheim 2: Synthesizing Aphasia and the Neural Basis of Language in a Neurocomputational Model of the Dual Dorsal-Ventral Language Pathways. Neuron, 72(2), 385-396. doi: SC 10.1016/j.neuron.2011.09.013 Vigneau, M., Beaucousin, V., Hervé, P. Y., Duffau, H., Crivello, F., Houdé, O., . . . TzourioMazoyer, N. (2006). Meta-analyzing left hemisphere language areas: Phonology, M AN U semantics, and sentence processing. NeuroImage, 30(4), 1414-1432. Visser, M., Jefferies, E., & Lambon Ralph, M. A. (2009). Semantic Processing in the Anterior Temporal Lobes: A Meta-analysis of the Functional Neuroimaging Literature. Journal of Cognitive Neuroscience, 22(6), 1083-1094. doi: http://dx.doi.org/10.1162/jocn.2009.21309 TE D Wechsler, D. A. (1987). Wechsler Memory Scale—Revised manual. New York: Psychological Corporation. Wilke, M., de Haan, B., Juenger, H., & Karnath, H.-O. (2011). Manual, semi-automated, and automated delineation of chronic brain lesions: A comparison of methods. NeuroImage, EP 56(4), 2038-2046. doi: http://dx.doi.org/10.1016/j.neuroimage.2011.04.014 Willmes, K., & Poeck, K. (1993). To what extent can aphasic syndromes be localized? (Vol. AC C 116). Wise, R. J. S., Greene, J., Büchel, C., & Scott, S. K. (1999). Brain regions involved in articulation. The Lancet, 353(9158), 1057-1061. Wise, R. J. S., Scott, S. K., Blank, C., Mummery, C. J., Murphy, K., & Warburton, E. A. (2001). Separate neural subsystems within `Wernicke's area'. Brain, 124(1), 83-95. doi: http://dx.doi.org/10.1093/brain/124.1.83 Woollams, A. M., Lambon Ralph, M. A., Plaut, D. C., & Patterson, K. (2007). SD-squared: On the association between semantic dementia and surface dyslexia. Psychological Review, 114(2), 316-339. 26 ACCEPTED MANUSCRIPT Figure 1: (A) Lesion overlap map across the 31 patients (threshold 1-26). (B) VBCM between regional integrity and lesion volume (t-scale 3-6). Figure 2: VBCM analysis using each speech production measure as the dependent measure [each t- RI PT map (scale 3) is thresholded at p<0.005, cluster corrected at familywise error of p<0.001; without lesion volume correction]. Note each result reflects the raw (non-unique) correlation with each speech fluency measure (see text). (A). Lesion correlates for tokens. (B) Lesion correlates for mean length per utterance, (C) type-token ratio. (D) Lesion correlates for words-per-minute (uncorrected threshold SC p=0.01, voxel-extent 50) (E) Lesion correlates of the unique PCA factors: F1 reflects speech quanta (yellow) and F2 reflects semantic variety (cyan). Note - no significant cluster remains for any factor M AN U after lesion volume correction (see text for further details & Figure 1B). Figure 3: (A) VBCM analysis showing significant and unique neural correlates to phonological (red), semantic (blue) and speech quanta (green) performance, including lesion volume covariate. Image TE D thresholded at p<0.005, cluster corrected at familywise error of p<0.01 (image scale (t) 3–5). Illustrative exemplar cases are shown in Panels B and C. Panel B (patients with good executive scores) – patient EB (yellow) with good fluency vs. KW (orange) with poor fluency. Panel C (patients EP with poor executive skills) – patient KS (pink) with good fluency vs. ESb (purple) with poor fluency. AC C Figure 4: Overlap (yellow) between the speech quanta-related lesion correlate (green) and the frontal aslant tract (red). The speech quanta image is thresholded at p<0.005, familywise error of p<0.01 (image scale (t) 3-5), and the frontal aslant tract is a probabilistic map (image scale 0.5-1). 27 ACCEPTED MANUSCRIPT Table 1. Participant background information. Cases are ordered according to BNT score BDAE classification RI PT Wernicke Global Global Broca Broca Mixed Nonfluent Global Wernicke Wernicke/Conduction TSA Broca Mixed Nonfluent Broca Broca Mixed Nonfluent Broca Mixed Nonfluent Anomia Anomia Anomia Mixed Nonfluent Anomia Mixed Nonfluent Anomia Broca TMA Broca Anomia Anomia Anomia Anomia SC M AN U TE D EP AC C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Time Lesion Age Years of postInitials Gender Volume (years) Education stroke (voxels) (months) DBb 66 M 12 59 42687 ES 69 M 11 39 28146 ESb 68 M 11 142 33193 KW 81 M 10 24 11393 BS 59 M 11 103 27242 KL 55 M 13 31 14625 LM 63 M 11 13 14990 DB 60 M 12 44 31644 PE 73 F 16 22 6959 KS 59 M 12 12 5822 KK 48 M 12 33 20043 WM 77 M 11 66 33282 GL 47 M 12 18 26319 DCS 45 F 12 12 5273 JSa 73 M 11 190 45875 JSc 78 M 12 76 18459 JA 65 M 11 128 26097 JJ 84 M 12 25 8951 JM 62 M 11 110 15492 JSb 72 M 11 23 1481 ER 64 F 14 181 26480 HN 81 M 10 56 25963 BH 64 M 11 26 8795 EB 61 M 17 12 4806 DM 49 M 17 42 11915 DS 72 M 11 106 11446 AG 55 M 11 131 21270 LH 65 M 11 81 10073 JMf 70 F 11 84 8921 AL 49 F 12 69 9767 TJ 60 M 12 23 19975 28 ACCEPTED MANUSCRIPT Table 2.Participant scores on the behavioural assessment battery and speech production measures. ID NW rep I NW rep d W rep I W rep d Picture naming Boston naming NW min pairs W min pairs SWPM WWPM CAT synonym CCT Brixton Ravens Digit F Digit B WPM TTR MLU T DBb 0 0 37.5 0 0 0 22.22 52.78 57.81 31.25 12.5 48.96 53.13 38.18 30.56 25 0 63.33 53.91 60.74 23.89 0 0 0 0 4.69 0 48.61 54.17 78.13 90.63 25 72.92 73.44 40 66.67 0 0 15.12 94.09 35.19 24.63 ESb 0 0 0 0 0 0 54.17 50 87.5 60.94 34.38 52.08 43.75 23.64 38.89 0 0 0 0 0 0 KW 0 0 3.75 0 1.56 0 75 65.28 95.31 92.19 84.38 82.29 89.06 50.91 80.56 50 42.86 2 86.25 12.35 2.99 BS 3.33 0 5 1.25 3.13 1.67 65.28 75 92.19 100 31.25 78.13 84.38 38.18 91.67 0 0 52.61 63.45 48.15 21.65 RI PT ES 0 0 6.25 0 4.69 1.67 75 77.78 92.19 98.44 28.13 68.75 78.13 61.82 88.89 0 0 11.87 44.72 21.16 13.44 13.33 3.33 27.5 0 1.56 1.67 43.06 54.17 67.19 53.13 28.13 57.29 68.75 32.73 61.11 0 0 14.51 46 18.52 7.47 DB 70 30 85 83.75 7.81 8.33 87.5 58.33 64.06 76.56 31.25 59.38 82.81 40 86.11 37.5 14.29 17.42 38.33 65.36 89.58 PE 13.33 3.33 45 41.25 20.31 11.67 77.78 86.11 96.88 100 50 79.17 84.38 41.82 80.56 25 28.57 49.1 60.62 93.7 100 KS 73.33 80 93.75 95 31.25 13.33 94.44 95.83 71.88 67.19 84.38 84.38 68.75 52.73 86.11 100 57.14 70.53 86.6 90.53 60.47 KK 33.33 3.33 56.25 26.25 42.19 15 72.22 95.83 93.75 95.31 46.88 81.25 84.38 76.36 100 0 0 14.75 62.21 50 45.54 WM 36.67 30 55 41.25 39.06 25 47.22 63.89 92.19 75 50 61.46 51.56 43.64 61.11 37.5 28.57 5.44 65.71 22.22 10.45 GL 93.33 63.33 100 81.25 68.75 31.67 98.61 97.22 96.88 95.31 65.63 75 73.44 58.18 91.67 37.5 28.57 9.24 55.86 29.97 52.26 40 56.67 72.5 68.75 67.19 43.33 97.22 97.22 100 98.44 93.75 91.67 95.31 81.82 100 62.5 57.14 14.83 92.74 70.37 23.14 30 3.33 75 65 62.5 46.67 75 77.78 92.19 98.44 59.38 81.25 76.56 67.27 83.33 50 28.57 30.84 74.41 49.38 25.38 JSc 36.67 63.33 90 91.25 71.88 53.33 75 86.11 98.44 98.44 75 76.04 82.81 43.64 77.78 62.5 42.86 23.9 90.36 61.11 20.9 JA 36.67 40 85 78.75 79.69 63.33 90.28 95.83 100 98.44 78.13 63.54 87.5 61.82 80.56 37.5 0 46.26 74.41 60.74 25.38 98.44 98.44 56.25 93.75 53.13 43.64 41.67 62.5 42.86 41.98 85.19 82.72 40.31 100 100 100 82.29 81.25 76.36 94.44 50 57.14 79.16 77.63 84.77 59.72 36.67 23.33 82.5 73.75 85.94 63.33 51.39 80.56 JM 83.33 83.33 100 98.75 81.25 63.33 93.06 95.83 JSb 63.33 36.67 86.25 81.25 75 63.33 76.39 88.89 93.75 90.63 84.38 75 78.13 60 86.11 62.5 28.57 41.18 92 47.62 26.13 ER 53.33 36.67 70 81.25 71.88 65 81.94 88.89 95.31 93.75 56.25 84.38 90.63 41.82 38.89 25 0 46.49 82.14 95.06 47.03 HN 36.67 23.33 83.75 80 65.63 65 77.78 76.39 93.75 93.75 37.5 85.42 85.94 25.45 75 50 42.86 25.4 91.27 91.36 47.03 BH 86.67 80 100 96.25 95.31 66.67 93.06 94.44 98.44 93.75 78.13 83.33 73.44 67.27 66.67 62.5 57.14 36.77 78.68 60.49 28.37 EB 83.33 53.33 100 100 81.25 66.67 94.44 98.61 98.44 100 71.88 94.79 90.63 80 100 75 57.14 78.87 67.16 98.52 93.31 DM 60 10 73.75 68.75 75 71.67 80.56 93.06 98.44 98.44 56.25 95.83 98.44 50.91 91.67 37.5 0 23.63 84.74 50.79 28.37 DS 56.67 33.33 88.75 91.25 84.38 73.33 79.17 77.78 100 100 87.5 93.75 89.06 72.73 72.22 50 28.57 34.53 100 100 17.17 AG 73.33 83.33 77.5 87.5 87.5 78.33 100 98.61 100 100 87.5 89.58 75 56.36 75 100 100 13.06 80.5 54.81 22.4 LH 56.67 50 82.5 88.75 81.25 78.33 95.83 97.22 96.88 100 90.63 92.71 87.5 76.36 88.89 87.5 57.14 66.11 74.33 52.91 61.21 JMf 93.33 66.67 96.25 98.75 96.88 80 90.28 95.83 100 100 71.88 91.67 93.75 50.91 83.33 62.5 57.14 50.53 69 73.15 48.52 AC C JJ EP TE D DCS JSa M AN U SC KL LM 29 ACCEPTED MANUSCRIPT 93.75 88.33 91.67 100 100 100 84.38 93.75 79.69 60 91.67 87.5 85.71 100 86.25 44.79 38.07 SC 98.75 M AN U 100 TE D 90 EP 90 AC C AL RI PT 87.65 TJ 93.33 83.33 98.75 92.5 95.31 95 87.5 98.61 98.44 100 68.75 88.54 70.31 52.73 50 75 28.57 38.28 76.67 81.48 Abbreviations: W – word; NW – nonword; rep I – immediate repetition; rep D – delayed repetition; S/WWPM – spoken/written word-picture matching; CAT – sentence comprehension; CCT – Camel & Cactus Test; Dig F – forward digit span; Dig B – backward digit span; WPM – words per minute; TTR – type-token ratio; MLU – mean length utterance; T - tokens 30 ACCEPTED MANUSCRIPT Table 3. Factor loadings for PCA of speech production measures. Measure TTR MLU T WPM F1 0.082 0.803 0.904 0.800 F2 0.973 0.488 -0.157 0.258 AC C EP TE D M AN U SC RI PT Footnote: factor loadings exceeding 0.5 are marked in bold and with an asterisk 31 ACCEPTED MANUSCRIPT Table 4. Factor loadings from the omnibus PCA 96-Synonym judgement Spoken word to picture matching Written word to picture matching 0.353 0.311 0.156 0.714 0.722 0.699 -0.078 0.091 0.391 0.235 0.310 0.294 0.154 TE D Type-Token Ratio Words-per-minute Mean Length per Utterance Tokens 4 0.326 0.175 0.430 0.411 0.198 0.070 0.167 0.237 -0.050 0.168 0.209 0.347 0.378 0.578 0.244 -0.190 0.022 0.461 0.034 0.232 0.702 0.918 0.629 0.276 0.079 0.018 0.749 0.103 0.397 -0.152 0.055 0.013 0.000 0.350 0.160 0.716 0.814 0.818 M AN U Camel and Cactus – pictures Ravens Coloured Progressive Brixton spatial anticipation test 3 0.220 0.126 0.122 0.168 0.058 0.166 0.683 0.505 0.374 0.096 0.013 RI PT 2 0.108 0.106 0.230 0.315 0.239 0.191 0.203 0.421 0.388 0.600 0.629 SC 1 Non word repetition Immediate 0.822 Non word repetition Delayed 0.912 Word repetition Immediate 0.780 Word repetition Delayed 0.783 Forward digit span 0.841 Backward digit span 0.774 Non word minimal pairs 0.584 Word minimal pairs 0.586 Comprehensive Aphasia Test – sentence 0.726 Cambridge 64 item naming 0.712 Boston Naming Test 0.659 AC C EP Footnote: factor loadings exceeding 0.5 are marked in bold and with an asterisk 32 ACCEPTED MANUSCRIPT Table 5: PCA factor scores from Butler et al., (2014), fluency-only PCA and the omnibus PCA. F1 -0.775 -1.842 -0.300 -0.832 -1.952 -1.391 -0.843 -0.149 -1.426 1.274 -1.129 0.363 0.971 0.346 -0.240 0.461 0.009 0.131 0.998 0.537 -0.622 -0.534 1.422 0.451 -0.727 -0.051 1.810 0.765 0.755 1.353 1.165 F2 -1.600 0.384 -1.773 0.387 0.249 -0.516 -1.445 -2.555 -0.090 -1.524 -0.158 -0.567 -1.148 0.564 0.358 0.657 0.377 1.420 -0.210 0.269 1.273 1.048 0.090 -0.204 1.239 1.653 0.244 0.101 0.391 0.365 0.718 SC F2-Var -0.499 0.797 -3.030 0.540 -0.217 -1.185 -1.072 -2.058 -0.980 0.616 -0.642 -0.297 -1.181 1.080 0.148 0.947 0.279 0.706 0.227 0.854 0.612 0.909 0.392 -0.526 0.525 1.760 0.451 -0.240 -0.131 0.842 0.373 M AN U F1-Flu 0.290 -1.024 -1.360 -1.853 -0.136 -1.046 -1.171 1.151 1.961 1.209 -0.164 -1.364 -0.284 -0.642 -0.453 -0.643 -0.099 0.307 1.328 -0.456 0.677 0.253 -0.214 2.256 -0.566 -0.199 -0.759 0.830 0.624 1.285 0.265 All scores combined F3 -2.606 -0.564 -1.290 0.700 0.604 1.193 -0.946 0.658 0.706 -0.015 1.504 -1.198 1.107 1.351 0.221 -0.344 0.417 -2.027 0.805 0.003 -0.957 -0.754 -0.198 1.215 0.551 -0.282 -0.018 0.885 0.294 0.000 -1.013 RI PT Speech PCA TE D EP AC C DBb ES ESb KW BS KL LM DB PE KS KK WM GL DCS JSa JSc JA JJ JM JSb ER HN BH EB DM DS AG LH JMf AL TJ Butler et al. (2014) F1F2F3Phon Sem Exe -0.332 -2.321 -2.018 -1.668 -0.266 -0.331 -1.078 -0.701 -1.911 -1.211 -0.202 0.972 -1.935 0.267 0.678 -1.810 -0.072 0.972 -0.873 -1.682 -0.765 0.315 -2.295 0.594 -1.067 0.358 0.459 1.726 -2.395 0.718 -1.207 0.161 1.319 -0.036 -0.525 -1.344 0.618 -0.361 0.583 0.208 0.045 1.680 -0.321 0.309 0.233 0.416 0.356 -0.256 -0.124 0.593 0.308 0.236 1.429 -2.197 0.993 -0.053 0.863 0.472 0.006 0.120 -0.241 1.337 -1.092 -0.089 0.712 -0.692 1.204 0.216 -0.296 0.760 0.174 1.150 -0.607 1.269 0.420 0.128 1.045 0.041 1.419 0.210 -0.064 0.704 0.354 0.807 0.811 0.737 0.110 1.446 0.262 0.228 1.141 1.029 -1.289 F4 1.043 -0.256 -1.923 -1.929 0.120 -1.175 -0.606 1.491 1.723 1.274 -0.149 -1.458 -0.715 -0.819 -0.398 -0.500 -0.170 0.348 0.683 -0.409 1.189 0.927 -0.522 1.739 -0.013 0.129 -1.321 0.234 0.467 0.834 0.161 33 ACCEPTED MANUSCRIPT Table 6. Neural correlates for omnibus PCA factors after accounting for lesion volume. MNI co-ordinates Principal component Factor 1 Location Left temporal lobe (Phonology) Planum Polare 4.24 -48 -16 0 Posterior supramarginal gyrus 4.23 -48 -50 14 4.06 -62 -18 -14 3.68 -60 -6 -26 3.63 -38 -6 -28 2.77 -70 -24 -30 3.85 -24 -4 32 3.80 -30 -4 28 3.58 -54 8 32 Posterior middle temporal gyrus Left temporal lobe (Semantic) Anterior middle temporal gyrus Anterior temporal fusiform cortex Posterior inferior temporal gyrus Left prefrontal lobe (Speech quanta) Cortico Spinal tract Precentral gyrus AC C EP TE D Precentral gyrus z 2164 M AN U Factor 4 y 1030 SC Factor 2 x RI PT Extent (voxels) Z 4387 34 AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT AC C EP TE D M AN U SC RI PT ACCEPTED MANUSCRIPT