Abstract

Individuals with primary progressive aphasia (PPA) show selective breakdown in regions within the proposed dorsal (articulatory–phonological) and ventral (lexical–semantic) pathways involved in language processing. Phonological STM impairment, which has been attributed to selective damage to dorsal pathway structures, is considered to be a distinctive feature of the logopenic variant of PPA. By contrast, phonological abilities are considered to be relatively spared in the semantic variant and are largely unexplored in the nonfluent/agrammatic variant. Comprehensive assessment of phonological ability in the three variants of PPA has not been undertaken. We investigated phonological processing skills in a group of participants with PPA as well as healthy controls, with the goal of identifying whether patterns of performance support the dorsal versus ventral functional–anatomical framework and to discern whether phonological ability differs among PPA subtypes. We also explored the neural bases of phonological performance using voxel-based morphometry. Phonological performance was impaired in patients with damage to dorsal pathway structures (nonfluent/agrammatic and logopenic variants), with logopenic participants demonstrating particular difficulty on tasks involving nonwords. Binary logistic regression revealed that select phonological tasks predicted diagnostic group membership in the less fluent variants of PPA with a high degree of accuracy, particularly in conjunction with a motor speech measure. Brain–behavior correlations indicated a significant association between the integrity of gray matter in frontal and temporoparietal regions of the left hemisphere and phonological skill. Findings confirm the critical role of dorsal stream structures in phonological processing and demonstrate unique patterns of impaired phonological processing in logopenic and nonfluent/agrammatic variants of PPA.

INTRODUCTION

Primary progressive aphasia (PPA) is a debilitating condition wherein speech and language deteriorate as a result of neurodegenerative disease. Epidemiological data for PPA specifically are not available. However, incidence and prevalence of frontotemporal dementia (of which nonfluent and semantic variants of PPA are subtypes) are thought to be equal or greater than Alzheimer disease among early-onset (less than 60 years of age) dementias (Kirshner, 2014) with language variants accounting for approximately 43% of cases (Johnson et al., 2005). Three variants of PPA are now recognized, each of which shows a unique constellation of speech–language deficits and pattern of underlying atrophy in the brain (Gorno-Tempini et al., 2011). The nonfluent/agrammatic type presents with syntactic and motor speech deficits and frontoinsular atrophy in the left hemisphere (Gorno-Tempini et al., 2004; Grossman et al., 1996). The semantic variant is characterized by degradation of semantic knowledge in the context of anterior and inferior temporal lobe atrophy (typically left hemisphere greater than right; Gorno-Tempini et al., 2004; Hodges, Patterson, Oxbury, & Funnell, 1992). Finally, the more recently characterized logopenic variant demonstrates impairments in naming and repetition (Rohrer et al., 2010; Gorno-Tempini et al., 2008). This variant, associated with atrophy of temporoparietal regions in the left hemisphere, has also been referred to as the “phonological” variant of PPA due to deficits on tasks that require short-term phonological storage (i.e., the “phonological loop”) and to the presence of phonological paraphasias in connected speech (Leyton, Ballard, Piguet, & Hodges, 2014; Henry & Gorno-Tempini, 2010; Wilson et al., 2010; Gorno-Tempini et al., 2008).

Current models of the functional neuroanatomy of language propose two pathways by which speech and language are processed in the brain (Ueno, Saito, Rogers, & Lambon Ralph, 2011; Saur et al., 2008; Hickok & Poeppel, 2007). A dorsal pathway involving left temporoparietal and frontal peri-sylvian structures is thought to be involved, broadly, in articulatory–phonological processing, and a ventral language pathway, in the left middle and inferior temporal lobes, is considered crucial for lexical–semantic processing (Lambon Ralph, Sage, Jones, & Mayberry, 2010; Hickok & Poeppel, 2007).

Within the dorsal network, phonological processes are distributed over anterior and posterior peri-sylvian regions, which have been implicated in unique but complementary and highly interactive roles. The classic “phonological loop” literature implicates posterior temporoparietal cortex in short-term phonological storage, whereas frontal cortices are thought to support subvocal phonological rehearsal that serves to refresh the store (Baddeley, 2003). A more recent and elaborated model of the dorsal articulatory–phonological pathway, the State Feedback Control Model (and its extension, the Hierarchical State Feedback Control Model), proposes the existence of motor–phonological and auditory–phonological representations of speech, supported by left inferior frontal and superior temporal cortices, respectively (Hickok, 2012, 2014). Transcoding between these motor and auditory phonological codes occurs in left temporoparietal cortex (in an area referred to as Spt). This model, which focuses on volitional speech production, does not address phonological storage or manipulation, per se; however, functional neuroimaging research supports the idea that the aforementioned left frontal and temporoparietal regions are cooperatively and interactively engaged during tasks that require maintenance and manipulation of phonological information (Marvel & Desmond, 2012; Peschke, Ziegler, Eisenberger, & Baumgaertner, 2012; Champod & Petrides, 2010; Buchsbaum & D'Esposito, 2008). Previous work in individuals with aphasia and phonological impairment caused by stroke has not revealed significant differences in performance in those with anterior versus posterior left peri-sylvian lesions on phonological processing tasks (Rapcsak et al., 2009), suggesting that damage anywhere within the dorsal phonological–articulatory stream may produce phonological deficits. Nonetheless, it may be the case that individuals with anterior (frontal) damage within this network versus those with posterior (temporoparietal) damage have difficulty with phonological tasks for different underlying reasons.

In a previous study, we demonstrated that performance in specific language domains correlates with the structural integrity of dorsal (phonological) and ventral (semantic) systems in a mixed group of individuals with PPA (Henry, Beeson, Alexander, & Rapcsak, 2012). We found that gray matter volumes in left frontal and temporoparietal cortices correlated with performance on phonological tasks. Conversely, volumes in left angular and anterior temporal cortices correlated with performance on semantic measures. While confirming the general functional–anatomical separation of the dual processing streams, the study did not allow examination of relative performance across the three diagnostic variants of PPA.

Although semantic processing has been extensively examined in all three PPA variants, little research has explored phonological processing ability by PPA subtype. A phonological short-term storage deficit has been considered a distinctive feature of logopenic patients, who demonstrate impaired comprehension and repetition of sentences as well as reduced digit, letter, and word span (Rohrer et al., 2010; Gorno-Tempini et al; 2008). Logopenic patients also fail to show the classic “phonological similarity effect,” wherein strings of phonologically dissimilar letters are more accurately repeated than phonologically similar strings, a pattern thought to reflect degradation of the “store” component of the phonological loop (Gorno-Tempini et al., 2008). There has been relatively little investigation of phonological ability in nonfluent/agrammatic patients; however, they have demonstrated abnormal phonological facilitation effects during naming trials (Mack et al., 2013). Mack et al. showed that agrammatic patients exhibited naming facilitation with presentation of phonologically related words at a later time window than logopenic or control participants, suggesting impaired or delayed phonological encoding as an underlying cause for naming difficulty. Finally, individuals with semantic variant PPA have shown relative preservation of phonological processing, at least early in the disease course when damage is largely confined to anterior and inferior temporal regions (Jefferies, Jones, Bateman, & Lambon Ralph, 2005). Whereas this body of work suggests that phonological processing is impaired in logopenic and perhaps nonfluent/agrammatic patients, to our knowledge, no study has directly compared performance across the three variants on tasks designed to assess perception, storage, manipulation, and production of phonological information.

In this study, we investigated phonological processing in a large cohort of individuals with PPA, with the following aims: to further assess the validity of the dorsal–ventral pathway distinction in language processing, to compare phonological processing ability across PPA variants, and to determine the clinical utility of phonological tasks in differential diagnosis by PPA variant. We used a brief battery of tasks involving both spoken and written language. On the basis of previous work, we selected three subtests from the Arizona Phonological Battery that have proven particularly sensitive to phonological impairment in individuals with aphasia caused by stroke as well as individuals with PPA (Henry et al., 2012; Beeson, Rising, Kim, & Rapcsak, 2010; Rapcsak et al., 2009). Written language measures included reading and spelling of nonwords (Rapcsak et al., 2009). Given the well-established loci of anatomical damage in the subtypes of PPA, we hypothesized that patients with damage to the dorsal pathway (nonfluent/agrammatic and logopenic variants) would show greater impairment on phonological processing tasks, whereas patients with damage to the ventral pathway (semantic variant) would show relative preservation of phonological abilities. Furthermore, given that individuals with logopenic PPA have damage to posterior peri-sylvian structures implicated in primary storage and transcoding of phonological information as well as a pattern of language features consistent with a phonological deficit (e.g., repetition, span, and nonword reading deficits; Brambati, Ogar, Neuhaus, Miller, & Gorno-Tempini, 2009; Gorno-Tempini et al., 2008), we predicted that this variant would show greatest impairment on phonological measures. Finally, we directly examined cortical regions involved in phonological processing using voxel-based morphometry (VBM). We predicted that these analyses would confirm significant involvement of left peri-sylvian (frontal and temporoparietal) cortical regions in phonological processing ability.

METHODS

Participants

Thirty-six individuals with PPA (12 individuals with semantic, 12 with logopenic, and 12 with nonfluent/agrammatic variant PPA) and 13 healthy controls were included in the study. All participants gave written informed consent, and the University of California San Francisco (UCSF) Committee on Human Research approved all study procedures. To be included in the study, participants with language impairment met current diagnostic criteria for PPA (Gorno-Tempini et al., 2011). A PPA diagnosis required progressive deterioration of speech–language functions, with deficits largely restricted to these domains during early stages of the disease. Patients were diagnosed with nonfluent/agrammatic, semantic, or logopenic variant of PPA based on current guidelines (Gorno-Tempini et al., 2011). Diagnosis by variant was reached by consensus, following a multidisciplinary evaluation comprising language and neuropsychological testing (see Table 1 for assessments) and neurological examination. Structural neuroimaging (MRI) was conducted for all participants but was not considered for diagnosis. At the time of the study, participants with PPA were required to score 10 or greater on the Mini-Mental State Examination (MMSE). In addition, they had to demonstrate adequate motor speech ability to perform spoken phonological tasks, including correct repetition of up to three syllables without error (as measured by the Western Aphasia Battery [WAB] repetition subtest; Kertesz, 1982) and apraxia of speech (AOS) and dysarthria ratings of 4 (moderate) or less on the Motor Speech Evaluation (MSE; Wertz, LaPointe, & Rosenbek, 1984). To ensure adequate perceptual abilities, potential participants could report no unaided hearing impairment and were required to perform at 85% or better on a test of 20 auditory minimal pair judgments in words and nonwords (adapted from PALPA Subtests 1 and 2; Kay, Lesser, & Coltheart, 1992). Sixty-six individuals were evaluated for participation in this study, and 17 were excluded. Of those, 12 were individuals with the nonfluent variant of PPA (seven did not meet the motor speech criteria, one individual scored less than 10 on the MMSE, and four had unaided hearing loss). Five individuals with the logopenic variant of PPA were excluded (three with a score of less than 10 on the MMSE and two with hearing loss).

Table 1. 

Demographic Information and Cognitive Scores (with Standard Deviation) for Groups with PPA and Healthy Controls

LV (n = 12)NFV (n = 12)SV (n = 12)Control (n = 13)
Demographic 
 Age 62.75 (8.9) 70.08 (7.4) 63.00 (6.4) 66.46 (3.9) 
 Sex 2M:10F 5M:7F 6M:6F 4M:9F 
 Education (years) 17.8 (1.1) 18.5 (1.8) 16.50 (2.7) 18.0 (3.8) 
 Handedness 9R;2L;1Amb 11R;1L 11R;1L 9R;4L 
 MMSE (30) 24.1 (5.5)d 27.4 (1.90) 27.4 (2.8) 29.6 (0.7) 
Language and speech 
 BNT (15) 9.8 (3.7)b 13.5 (1.6) 5.9 (3.0)a,b  
 Fluency rating (WAB; 10) 8.8 (0.9) 7.6 (1.8)c 9.0 (0.4)  
 Repetition (WAB; 100) 78.6 (12.9)b,c 91.9 (5.8) 95.7 (4.6)  
 Sequential commands (WAB; 80) 72.6 (8.6)c 73.5 (8.3) 80.0 (0.0)  
 AOS rating (0–7) 0.3 (1.2) 1.9 (1.3)a,c 0.0  
 Dysarthria rating (0–7) 0.0 2.1 (1.9)a,c 0.0  
 PPVT-short (16)e 14.0 (2.1) 15.6 (0.5) 10.5 (3.8)a,b,d 15.5 (0.7) 
Working memory 
 Digit span forward 4.3 (1.4)c,d 4.6 (1.0)c,d 7.1 (1.3) 7.5 (1.5) 
Visuospatial function 
 Benson figure copy (17)e 14.7 (2.6) 15.2 (1.2) 15.8 (0.8) 15.4 (0.5) 
Visual memory 
 Benson figure recall (17)e 6.0 (3.4)b,d 11.3 (2.5) 7.7 (4.8) 12.1 (2.9) 
Verbal memory 
 CVLT-SF Trials 1–4 (40)e 17.6 (8.5) 24.6 (5.8) 18.8 (5.9)  
 CVLT-SF 30-sec free recall (10)e 4.9 (2.9) 6.3 (1.8) 3.8 (2.4)  
 CVLT-SF 10-min free recall (10)e 3.5 (3.6) 5.8 (2.0) 2.4 (1.9)b  
Executive functions 
 Digit span backward 2.9 (0.9)c,d 3.9 (1.0)c,d 6.2 (1.8) 5.9 (1.5) 
 Modified trails (lines per minute)e 15.8 (11.9)d 12.7 (8.8)d 25.3 (10.1)d 44.6 (16.0) 
LV (n = 12)NFV (n = 12)SV (n = 12)Control (n = 13)
Demographic 
 Age 62.75 (8.9) 70.08 (7.4) 63.00 (6.4) 66.46 (3.9) 
 Sex 2M:10F 5M:7F 6M:6F 4M:9F 
 Education (years) 17.8 (1.1) 18.5 (1.8) 16.50 (2.7) 18.0 (3.8) 
 Handedness 9R;2L;1Amb 11R;1L 11R;1L 9R;4L 
 MMSE (30) 24.1 (5.5)d 27.4 (1.90) 27.4 (2.8) 29.6 (0.7) 
Language and speech 
 BNT (15) 9.8 (3.7)b 13.5 (1.6) 5.9 (3.0)a,b  
 Fluency rating (WAB; 10) 8.8 (0.9) 7.6 (1.8)c 9.0 (0.4)  
 Repetition (WAB; 100) 78.6 (12.9)b,c 91.9 (5.8) 95.7 (4.6)  
 Sequential commands (WAB; 80) 72.6 (8.6)c 73.5 (8.3) 80.0 (0.0)  
 AOS rating (0–7) 0.3 (1.2) 1.9 (1.3)a,c 0.0  
 Dysarthria rating (0–7) 0.0 2.1 (1.9)a,c 0.0  
 PPVT-short (16)e 14.0 (2.1) 15.6 (0.5) 10.5 (3.8)a,b,d 15.5 (0.7) 
Working memory 
 Digit span forward 4.3 (1.4)c,d 4.6 (1.0)c,d 7.1 (1.3) 7.5 (1.5) 
Visuospatial function 
 Benson figure copy (17)e 14.7 (2.6) 15.2 (1.2) 15.8 (0.8) 15.4 (0.5) 
Visual memory 
 Benson figure recall (17)e 6.0 (3.4)b,d 11.3 (2.5) 7.7 (4.8) 12.1 (2.9) 
Verbal memory 
 CVLT-SF Trials 1–4 (40)e 17.6 (8.5) 24.6 (5.8) 18.8 (5.9)  
 CVLT-SF 30-sec free recall (10)e 4.9 (2.9) 6.3 (1.8) 3.8 (2.4)  
 CVLT-SF 10-min free recall (10)e 3.5 (3.6) 5.8 (2.0) 2.4 (1.9)b  
Executive functions 
 Digit span backward 2.9 (0.9)c,d 3.9 (1.0)c,d 6.2 (1.8) 5.9 (1.5) 
 Modified trails (lines per minute)e 15.8 (11.9)d 12.7 (8.8)d 25.3 (10.1)d 44.6 (16.0) 

Group differences: significantly impaired relative to alogopenic, bnonfluent/agrammatic, csemantic and dcontrol groups (p < .05 with Sidak correction).

eFrom Kramer et al. (2003).

Amb = ambidextrous; F = female; M = male.

LV = logopenic variant, NFV = nonfluent/agrammatic variant, SV = semantic variant; WAB = Western Aphasia Battery; PPVT = Peabody Picture Vocabulary Test; CVLT-SF = California Verbal Learning Test-UCSF version.

Healthy controls were screened for history of neurological or psychological illness and scored within normal limits on a neuropsychological screen (≥27 on the MMSE). Demographic characteristics for participants and relevant cognitive and speech–language measures are reported in Table 1.

Speech–Language Measures

In addition to the standard UCSF Memory and Aging Center speech–language assessments detailed elsewhere (Gorno-Tempini et al., 2004), each individual was administered three phonological tasks from the Arizona Phonological Battery (Henry et al., 2012; Beeson et al., 2010; Rapcsak et al., 2009). Tasks included phoneme deletion, phoneme blending, and phoneme replacement, in both words and phonologically plausible nonwords (see Table 2 for sample items). A speech–language pathologist (M. H. or M. B.) conducted all testing with patients, and research assistants tested healthy controls. Each session was videotaped. Responses were transcribed and scored online, and videotapes were reviewed for accuracy of coding. For each of the phonological tasks, three practice items were given before administration of experimental items. Scores were calculated for each of the deletion, blending, and replacement tests as well as a summary score that equaled total percent correct on all three phonological tasks. As in previous studies, motor speech was characterized using the MSE, designed to detect characteristics of AOS and dysarthria (Wertz et al., 1984) via tasks of increasing difficulty (from simple phonation and production of monosyllables to repetition of multisyllabic words and sentences).

Table 2. 

Sample Tasks from the Arizona Phonological Battery (Beeson et al., 2010)

TaskExample
Phoneme deletion (n = 10 words, 10 pseudowords) Say “fat”…now take away “f” → “at” 
Say “zane”…now take away “z” → “ane” 
Phoneme blending (n = 10 words, 10 pseudowords) Blend these sounds together 
 /b/ /oi/ /l/ → “boil” 
 /z/ /aI/ /p/ → “zipe” 
Phoneme replacement (n = 15 words, 15 pseudowords) Say “mouth”…now change /th/ to /s/ → “mouse” 
Say “bazz”…now change /b/ to /d/ → “dazz” 
TaskExample
Phoneme deletion (n = 10 words, 10 pseudowords) Say “fat”…now take away “f” → “at” 
Say “zane”…now take away “z” → “ane” 
Phoneme blending (n = 10 words, 10 pseudowords) Blend these sounds together 
 /b/ /oi/ /l/ → “boil” 
 /z/ /aI/ /p/ → “zipe” 
Phoneme replacement (n = 15 words, 15 pseudowords) Say “mouth”…now change /th/ to /s/ → “mouse” 
Say “bazz”…now change /b/ to /d/ → “dazz” 

Written language measures consisted of oral reading and writing to dictation of 80 real words (40 regularly spelled and 40 irregularly spelled) and 20 phonologically plausible nonwords, whose linguistic characteristics have been described previously (Henry et al., 2012; Rapcsak et al., 2009). Performance on written language measures was available for a subset of participants (reading = 10 logopenic PPA, 11 nonfluent/agrammatic PPA, 12 semantic PPA, 12 healthy controls; spelling = 9 logopenic PPA, nine nonfluent/agrammatic PPA, 12 semantic PPA, 12 healthy controls). Nonword reading/spelling was of particular importance, as pronunciations and written renditions must be derived phonologically. For this reason, we were also interested in the relative performance of our participants on nonword stimuli relative to real words across spoken and written tasks. To examine the effect of stimulus type on performance, we calculated lexicality effect sizes by subtracting nonword score from real word score. An effect of lexicality, or word versus nonword status, is observed when letter combinations with a lexical representation are read/spelled/repeated correctly, whereas unfamiliar but pronounceable letter combinations are produced less accurately. This derived measure serves as an index of phonological impairment. We compared the size of this effect across groups and tasks (phonological battery and written language). Scores on the phonological tasks and the reading and spelling trials were examined across clinical subgroups using ANOVA. Sidak's procedure was used to adjust alpha for follow-up contrasts. We also directly compared real word with nonword performance on tasks within each diagnostic group using t tests.

To determine whether any of our spoken or written phonological measures might be a positive predictor of PPA variant diagnosis, independently of other speech–language measures, we entered all phonological test scores and reading and spelling of nonwords into a forward, stepwise binary logistic regression. This analysis used the likelihood ratio method and a p < .05 criterion for inclusion of variables in the model. Multicollinearity was assessed using variance inflation factor (<10 considered acceptable; Field, 2009). Logopenic and nonfluent/agrammatic patients were the focus of this analysis, as these groups may be difficult to distinguish behaviorally (both groups present with speech sound errors and reduced fluency of language output). Group membership (logopenic vs. nonfluent/agrammatic) served as a dependent variable for this analysis. We followed this with an analysis in which other standard speech–language measures were added to the previous model (in addition to the phonological tasks) in a stepwise fashion (scores reported in Table 1; including Boston Naming Test, WAB fluency rating, WAB repetition, WAB sequential commands, and AOS/dysarthria rating from the MSE). This analysis was intended to determine how phonological measures might contribute to differential diagnosis in conjunction with standard speech–language measures. Significance of changes in the model's predictive accuracy was assessed by comparing the log-likelihood of each new model with that of the previous model (this statistic has a chi-square distribution).

Neuroimaging and Brain–Behavior Correlations

Patients underwent high-resolution structural MRI scanning on a Siemens 3-T Trio scanner using a T1-weighted 3-D magnetization prepared rapid acquisition gradient-echo sequence with the following parameters: slice thickness = 1 mm, field of view = 256 × 256 mm, matrix = 256 × 256, repetition time = 2300 msec, echo time = 2.98 msec, flip angle = 9°. All image processing and analysis were performed using SPM8 (www.fil.ion.ucl.ac.uk/spm/software/spm8) and MATLAB Version 7.10 (The MathWorks, Inc., Natick, MA). All T1 structural images were segmented and bias corrected and then registered to Montreal Neurological Institute space through an affine and a nonlinear deformation. The nonlinear deformation parameters were calculated with the high dimensional diffeomorphic anatomical registration through exponentiated lie (DARTEL) algorithm and the predefined templates within the SPM DARTEL toolbox (Ashburner, 2007). Voxel-based morphometry (VBM) analysis was conducted using modulated gray matter images, with voxel values multiplied by Jacobian determinants derived from the spatial normalization to preserve the total amount of gray matter from the original images. Modulated images were smoothed with a Gaussian kernel (10-mm FWHM).

We examined the relation between regional gray matter volumes and phonological summary score, as well as reading and spelling of nonwords, with age, sex, and total intracranial volume included in the model as covariates. Motor speech rating from the MSE (AOS rating plus dysarthria rating) was included as an additional covariate in the phonological summary score analysis. Threshold for significance was set at p < .001, and correction for multiple comparisons was achieved by permutation analysis, as described previously (Wilson et al., 2010). Subsequently, we examined the unique requirements of specific phonological subtests (deletion, blending, and replacement tasks) within the broader phonological network by examining the independent contribution of frontal and temporoparietal ROIs to phonological subtest performance. A frontal ROI was defined as regions where t > 4 in the phonology map in the inferior frontal gyrus, middle frontal gyrus, or precentral gyrus. A temporoparietal ROI was defined as regions where t > 4 in the phonology map in the superior temporal gyrus, supramarginal gyrus, or inferior parietal lobule. Anatomical labels were from Tzourio-Mazoyer et al. (2002). We carried out three ANOVAs: one with each phonology measure as the dependent variable. In each ANOVA, the independent variables were Mean gray matter volume in each of these two ROIs, along with covariates of age, sex, and total intracranial volume.

RESULTS

Participant groups did not differ with regard to age or education (Table 1). Consistent with current criteria for PPA subtyping, individuals in the semantic group were impaired on confrontation naming on the BNT (Kaplan, Goodglass, & Weintraub, 2001) and showed impaired single-word comprehension on a shortened (16-item) version of the Peabody Picture Vocabulary Test-Revised (Kramer et al., 2003; Dunn & Dunn, 1981) relative to other participants. Individuals in the nonfluent/agrammatic group had significantly more severe ratings for AOS and dysarthria, and individuals with logopenic variant were impaired relative to the other groups on the WAB (Kertesz, 1982) repetition subtest.

Phonological Performance

Phonological task performance in participants with PPA and healthy controls is presented in Figure 1 and Table 3. We conducted separate ANOVAs examining phonological test scores across the diagnostic groups, with follow-up contrasts conducted if the omnibus test was significant. The Sidak procedure was used to correct for multiple comparisons (adjusted alpha: p < .009). The ANOVA comparing overall phonological score was significant (F(3, 45) = 13.99, p < .001), and follow-up contrasts revealed significantly worse performance in logopenic patients relative to controls and semantic patients as well as worse performance in nonfluent/agrammatic patients relative to controls and semantic patients. Nonfluent/agrammatic and logopenic patients were not significantly different.

Figure 1. 

Phonology scores in each PPA variant on tasks using nonword (NW) and word stimuli. Error bars represent standard errors. Blend = blending; Repl = replacement.

Figure 1. 

Phonology scores in each PPA variant on tasks using nonword (NW) and word stimuli. Error bars represent standard errors. Blend = blending; Repl = replacement.

Table 3. 

Mean Scores (and Standard Deviations) for Individual Phonological Subtests and Written Language Tasks

LVNFVSVControl
Deletion NW 67.50 (33.61)*b,***c,d 88.33 (15.86) 99.17 (2.87) 100.00 (0.00) 
Deletion words 79.17 (24.66)**b,***c,d 96.67 (6.51) 100.00 (0.00) 99.23 (2.77) 
Blend NW 40.83 (38.95)***c,d 53.33 (31.14)*c,**d 84.17 (13.11) 88.46 (11.44) 
Blend words 53.33 (32.29)***c,d 63.33 (23.48)*c,**d 88.33 (15.86) 91.54 (13.45) 
Repl NW 36.67 (32.10)***c,d 47.78 (30.86)***c,d 88.89 (14.86) 89.74 (7.99) 
Repl words 50.01 (32.10)***c,d 59.03 (32.38)**c,d 88.89 (15.13) 90.77 (7.95) 
Total phonology score 52.98 (28.20)***c,d 66.24 (21.56)**c,d 91.19 (9.11) 92.86 (4.16) 
 
Read NW 84.5 (19.92) 92.73 (12.12) 96.67 (7.18) 97.92 (3.34) 
Read words 95.13 (5.25) 97.05 (5.76) 97.60 (2.53) 99.90 (0.36) 
Spell NW 63.89 (26.07)***b,c,d 90.00 (10.35) 99.17 (1.95) 96.67 (6.85) 
Spell words 85.42 (9.88)***d 94.84 (5.53) 86.67 (10.74)***d 99.79 (.49) 
LVNFVSVControl
Deletion NW 67.50 (33.61)*b,***c,d 88.33 (15.86) 99.17 (2.87) 100.00 (0.00) 
Deletion words 79.17 (24.66)**b,***c,d 96.67 (6.51) 100.00 (0.00) 99.23 (2.77) 
Blend NW 40.83 (38.95)***c,d 53.33 (31.14)*c,**d 84.17 (13.11) 88.46 (11.44) 
Blend words 53.33 (32.29)***c,d 63.33 (23.48)*c,**d 88.33 (15.86) 91.54 (13.45) 
Repl NW 36.67 (32.10)***c,d 47.78 (30.86)***c,d 88.89 (14.86) 89.74 (7.99) 
Repl words 50.01 (32.10)***c,d 59.03 (32.38)**c,d 88.89 (15.13) 90.77 (7.95) 
Total phonology score 52.98 (28.20)***c,d 66.24 (21.56)**c,d 91.19 (9.11) 92.86 (4.16) 
 
Read NW 84.5 (19.92) 92.73 (12.12) 96.67 (7.18) 97.92 (3.34) 
Read words 95.13 (5.25) 97.05 (5.76) 97.60 (2.53) 99.90 (0.36) 
Spell NW 63.89 (26.07)***b,c,d 90.00 (10.35) 99.17 (1.95) 96.67 (6.85) 
Spell words 85.42 (9.88)***d 94.84 (5.53) 86.67 (10.74)***d 99.79 (.49) 

*p < .009, **p < .005, ***p < .001 significantly impaired relative to alogopenic, bnonfluent/agrammatic, csemantic, and dcontrol group (Sidak-adjusted alpha = .009).

LV = logopenic variant; NFV = nonfluent/agrammatic variant; SV = semantic variant.

An effect of Diagnostic group was revealed by the ANOVAs examining phoneme deletion in nonwords (F(3, 45) = 8.21, p < .001) and words (F(3, 45) = 7.27, p < .001), and follow-up contrasts identified worse performance in logopenic participants relative to all other groups on these tasks. Group differences were also observed on the phoneme blending in nonword (F(3, 45) = 9.48, p < .001) and word (F(3, 45) = 9.22, p < .001) tests and phoneme replacement in nonword (F(3, 45) = 16.75, p < .001) and word (F(3, 45) = 9.07, p < .001) tests. Follow-up contrasts revealed that logopenic and nonfluent participants performed worse than semantic patients and controls but did not differ significantly.

Written Language Battery Performance

Reading and spelling scores by word type are presented in Figure 2 and Table 3. Nonword performance was of particular importance because of high phonological processing demands. ANOVAs revealed a significant Group effect for nonword spelling but not reading (F(3, 39) = 16.08, p < .001), and follow-up contrasts revealed that logopenic patients were significantly impaired relative to nonfluent/agrammatic, semantic, and control groups on this measure.

Figure 2. 

Reading and spelling performance in patients with PPA and healthy controls on word and nonword (NW) stimuli. Error bars represent standard errors.

Figure 2. 

Reading and spelling performance in patients with PPA and healthy controls on word and nonword (NW) stimuli. Error bars represent standard errors.

Word versus Nonword Performance across Tasks

Lexicality effect sizes (word score minus nonword score) were calculated for reading/spelling (averaged) and phonological tasks (Figure 3). Semantic variant participants showed a reverse lexicality effect, with nonword spelling performance superior to real word performance (because of impaired irregular word reading/spelling or surface dyslexia/dysgraphia). As such, this group was excluded from the analysis to avoid biasing the results. The reading/spelling lexicality effect size analysis did not reveal a significant effect of group (F(2, 25) = 3.05, p = .065). By contrast, the ANOVA for the lexicality effect from the spoken phonological tasks revealed a significant effect of group (F(3, 45) = 4.74, p = .006), with follow-up contrasts (adjusted alpha: p < .009) showing a significantly larger effect in logopenic participants relative to the semantic variant group (p = .006) and healthy controls (p = .004). Direct comparisons of real word versus nonword performance were conducted for each task using paired t tests, revealing a significant decrement in nonword compared with real word performance on the following tasks in the logopenic group: spelling (t(8) = 2.38, p = .045), phoneme deletion (t(11) = 2.55, p = .027), and phoneme replacement (t(11) = 3.19, p = .009). Significantly worse performance on nonwords relative to real words was not observed on any tasks in any of the other groups.

Figure 3. 

Lexicality effect sizes for written language and phonology tasks. Error bars represent standard errors.

Figure 3. 

Lexicality effect sizes for written language and phonology tasks. Error bars represent standard errors.

Phonology Measures as Predictors of Clinical Subtype: Nonfluent/Agrammatic and Logopenic Patients

Because nonfluent/agrammatic and logopenic patients can be difficult to discriminate behaviorally due to overlapping clinical features, we sought to determine whether any of our phonological tasks might hold utility for predicting to which diagnostic group an individual belongs. As such, we conducted stepwise, forward, binary logistic regression with all phonological task scores and reading and spelling of nonwords as predictor variables. The overall fit of the final model was significant (χ2(2) = 13.02, p = .001), and two variables were kept in the model as significant predictors (at criterion of p < .05): nonword spelling and sound blending in nonwords, which together predicted group membership with 83.3% accuracy (Table 4). However, multicollinearity diagnostics indicated a variance inflation factor value of 11.15, which is considered to be outside acceptable bounds for multicollinearity (Field, 2009). As such, we conducted the analysis again with only scores from the nonword versions of each task, as these were the more sensitive tasks across patient groups. This eliminated the problem of multicollinearity. The resulting model was virtually unchanged (χ2(2) = 13.02, p = .001; nonword spelling and sound blending in nonwords were kept in the model as significant predictors; the model showed 83.3% correct group assignment). In a separate logistic regression analysis, standard speech and language measures were added to this model in a stepwise manner to determine the predictive power of phonological scores in conjunction with widely used clinical measures. The addition of MSE (AOS plus dysarthria) rating to the model resulted in a significant improvement in the model's predictive power (χ2(1) = 11.94, p = .001), with the resulting model (χ2(3) = 24.95, p < .001) showing improved prediction at 100% accuracy. Other speech–language measures (repetition, fluency, and sequential commands subtests from the WAB as well as BNT score) did not significantly improve the model's predictive accuracy based on the model chi-square statistic.

Table 4. 

Classification Tables Showing Results of Binary Logistic Regression Predicting Logopenic versus Nonfluent/Agrammatic PPA Group Membership Based on (a) Sound Blending in Nonwords and Nonword Spelling and (b) Sound Blending in Nonwords, Nonword Spelling, and Motor Speech Rating (AOS Plus Dysarthria Ratings from the MSE)

APredicted Diagnostic GroupBPredicted Diagnostic Group
LVNFV% CorrectLVNFV% Correct
LV 88.9 LV 100 
NFV 77.8 NFV 100 
Total   83.3 Total   100 
APredicted Diagnostic GroupBPredicted Diagnostic Group
LVNFV% CorrectLVNFV% Correct
LV 88.9 LV 100 
NFV 77.8 NFV 100 
Total   83.3 Total   100 

Neuroimaging and Brain–Behavior Correlations

Imaging analysis using VBM revealed a significant relationship between gray matter volumes and total phonology score in left frontal and temporoparietal regions (Figure 4 and Table 5). Specifically, phonological score was associated with gray matter volumes in a large left hemisphere cluster including anterior regions (superior, middle, and inferior frontal gyri, insula, supplementary motor area, and precentral gyrus); posterior regions (postcentral gyrus, superior and inferior parietal lobe, temporo-parietal junction, and superior temporal gyrus); as well as the occipital lobe and cingulate gyrus. Additional, small clusters were observed in the right superior and middle temporal gyri, the right inferior parietal lobe, and the thalamus.

Figure 4. 

Results of VBM (n = 12 semantic variant, 12 logopenic variant, 12 nonfluent/agrammatic variant, 13 healthy controls) analysis examining relation between phonology score and gray matter volumes (p < .001 corrected for multiple comparisons based on cluster size; age, sex, total intracranial volume, and motor speech score included as covariates).

Figure 4. 

Results of VBM (n = 12 semantic variant, 12 logopenic variant, 12 nonfluent/agrammatic variant, 13 healthy controls) analysis examining relation between phonology score and gray matter volumes (p < .001 corrected for multiple comparisons based on cluster size; age, sex, total intracranial volume, and motor speech score included as covariates).

Table 5. 

Significant Clusters and Corresponding Anatomical Regions from Phonology and Written Language VBM Analyses

ClusterVolume (mm3)Center of MassMax tpAnatomical Regions
xyz
Phonology score 
172,200 −31 −18 37 5.37 .003 LH: superior/middle/inferior frontal gyri, insula, precentral/postcentral gyri, inferior parietal lobe, superior parietal lobe, superior temporal gyrus, cingulate, SMA, rolandic operculum, precuneus, occipital lobe 
29,352 58 −18 −4 4.92 .020 RH: superior/middle temporal gyri 
27,264 −18 4.89 .021 Thalamus 
13,696 55 −41 39 4.34 .041 RH: inferior parietal lobe 
 
Read nonwords 
48,504 −25 −51 49 5.45 .012 LH: superior/inferior parietal lobe, precuneus, postcentral gyrus 
25,624 −25 −81 23 4.89 .028 LH: occipital, cuneus, precuneus, middle temporal gyrus 
15,592 22 −70 27 5.09 .041 RH: parietal, occipital, cuneus 
 
Spell nonwords 
18,544 −44 −53 43 4.64 .031 LH: superior/inferior parietal lobe 
11,920 −18 −47 59 4.56 .047 LH: precentral/postcentral gyri, precuneus, superior parietal lobe 
ClusterVolume (mm3)Center of MassMax tpAnatomical Regions
xyz
Phonology score 
172,200 −31 −18 37 5.37 .003 LH: superior/middle/inferior frontal gyri, insula, precentral/postcentral gyri, inferior parietal lobe, superior parietal lobe, superior temporal gyrus, cingulate, SMA, rolandic operculum, precuneus, occipital lobe 
29,352 58 −18 −4 4.92 .020 RH: superior/middle temporal gyri 
27,264 −18 4.89 .021 Thalamus 
13,696 55 −41 39 4.34 .041 RH: inferior parietal lobe 
 
Read nonwords 
48,504 −25 −51 49 5.45 .012 LH: superior/inferior parietal lobe, precuneus, postcentral gyrus 
25,624 −25 −81 23 4.89 .028 LH: occipital, cuneus, precuneus, middle temporal gyrus 
15,592 22 −70 27 5.09 .041 RH: parietal, occipital, cuneus 
 
Spell nonwords 
18,544 −44 −53 43 4.64 .031 LH: superior/inferior parietal lobe 
11,920 −18 −47 59 4.56 .047 LH: precentral/postcentral gyri, precuneus, superior parietal lobe 

LH = left hemisphere; RH = right hemisphere.

VBM analyses for each of the three individual phonological tasks (not shown) revealed patterns highly similar to the overall phonological score, with involvement of left frontal as well as temporoparietal regions for each measure. To further investigate differential reliance of the three phonological measures on frontal and temporoparietal regions, we carried out a follow-up ROI analysis. Deletion scores were predicted by atrophy in the frontal ROI (F(1, 43) = 7.57, p = .009), and the parietal ROI made only a marginal independent contribution (F(1, 43) = 2.93, p = .094). Similarly, replacement scores were predicted by atrophy in the frontal ROI (F(1, 43) = 16.37, p < .001), but not independently by the temporoparietal ROI (F(1, 43) < 1). In contrast, blending scores were predicted independently by atrophy in the frontal ROI (F(1, 43) = 8.92, p = .005) as well as the temporoparietal ROI (F(1, 43) = 5.98, p = .019).

VBM analyses for reading and spelling of nonwords are presented in Figure 5 and Table 5. The nonword reading score was associated with gray matter volumes in the left superior and inferior parietal lobe, postcentral gyrus, middle temporal gyrus and occipital lobe. A smaller cluster was observed in the right parietal lobe and occipital lobe. The score for spelling nonwords was associated with volumes in left superior and inferior parietal lobe as well as precentral and postcentral gyri.

Figure 5. 

Results of VBM analysis examining relation between (top) reading nonword score and (bottom) spelling nonword score and gray matter volumes (p < .001 corrected for multiple comparisons based on cluster size; age, sex, and total intracranial volume included as covariates).

Figure 5. 

Results of VBM analysis examining relation between (top) reading nonword score and (bottom) spelling nonword score and gray matter volumes (p < .001 corrected for multiple comparisons based on cluster size; age, sex, and total intracranial volume included as covariates).

DISCUSSION

Imaging studies in PPA have revealed that the three variants show selective degeneration of regions within the proposed dorsal (nonfluent and logopenic variants) and ventral (semantic variant) pathways involved in processing spoken language. Findings from our phonological tasks as well as from written language tasks confirm the dorsal (articulatory–phonological) versus ventral (lexical–semantic) stream distinction in a relatively large cohort of individuals with PPA. As predicted, patients with damage to the dorsal pathway showed impaired performance on phonologically mediated tasks relative to both controls and patients with ventral pathway damage. Findings are in accordance with an increased prevalence of phonological paraphasias in the connected speech of individuals with nonfluent/agrammatic and logopenic variants of PPA (Wilson et al., 2010) and with previous work indicating abnormal phonological processing in these patient groups (Mack et al., 2013; Gorno-Tempini et al., 2008).

Logopenic patients demonstrate a characteristic pattern of atrophy affecting posterior peri-sylvian regions implicated in storage and transcoding of phonological codes as well as a well-established speech–language profile characterized by difficulty with phonological processing (repetition, span, and nonword tasks; Rohrer et al., 2010; Gorno-Tempini et al., 2008). As such, we predicted that the logopenic group would show poorer performance across spoken and written phonological measures relative to the other variants, including the nonfluent/agrammatic group. Counter to our prediction, the phonological summary score was not significantly different in logopenic versus nonfluent/agrammatic groups, indicating that damage to anterior or posterior peri-sylvian regions may cause significant impairment of spoken phonological task performance. The logopenic subgroup did demonstrate lower scores on the phoneme deletion tasks relative to nonfluent/agrammatic patients (and other participants). Furthermore, nonword spelling was significantly worse in these patients relative to all other participants, and unlike other groups, they showed a significant decrement in nonword relative to real word performance for two of the three spoken language measures as well as spelling.

The two less fluent variants of PPA (logopenic and nonfluent/agrammatic), both of which present with segmental speech sound errors and reductions in speech fluency, can be difficult to dissociate behaviorally, particularly in mild, early stages. Acoustic measures of speech production and speech sound error typing have proven useful in differentiating between these clinical subtypes (Ballard et al., 2014; Croot, Ballard, Leyton, & Hodges, 2012; Wilson et al., 2010). We were interested to discern whether any of our phonological tasks might be additional useful tools for differential diagnosis in PPA. We found that the two most discriminating tasks in our battery (nonword spelling and sound blending in nonwords), which together take less than 10 min to administer, distinguished between logopenic and nonfluent/agrammatic patients with 83% accuracy. This was striking, given that neither of these measures is currently used in diagnosis of PPA by variant. In conjunction with a rating for motor speech impairment (a more established measure for differentiating between PPA types), we were able to predict group membership without error. Thus, phonological processing tasks may play a role in helping to assign individuals with PPA to appropriate diagnostic subtypes, particularly in mild-to-moderate stages of severity. Diagnosis by variant is particularly relevant given that PPA variant status is predictive of underlying pathology (Gorno-Tempini et al., 2011) and thus important for determining candidacy for pharmacological and other treatments that may be forthcoming. Future studies with larger cohorts need to be conducted to confirm the adequacy of these and other measures for discriminating between variants and to determine relevant cutoff scores for aiding in individual diagnosis.

Brain–behavior correlations conducted via VBM confirmed the involvement of a primarily left hemisphere fronto-temporoparietal network in phonological processing for spoken phonology tasks. Overall, these findings are consistent with current models of the functional neuroanatomy of language (Ueno et al., 2011; Saur et al., 2008; Hickok & Poeppel, 2004), previous functional imaging research (Price, 2012), prior work in PPA (Henry et al., 2012), and lesion data from individuals with phonological processing difficulties subsequent to stroke (Rapcsak et al., 2009). Lesion studies in patients, however, have not identified specific peri-sylvian subregions that are critical for different aspects of phonological processing. The phonological tasks employed in this study engage a number of procedures that may be collectively referred to as “phonological”; however, more specifically, they require processing of auditory input, assembly of phonological codes from sublexical information, short-term storage of phonological codes, manipulation of stored phonological representations, subvocal rehearsal, and ultimately, motor output. This study was not designed to explore these subprocesses or their neural bases via specific tasks. Nonetheless, the various tasks employed may place different demands on particular components within the phonological network. Consistent with this possibility were the unique patterns of impairment on individual tasks across PPA variants and partially distinct patterns of regional recruitment (revealed by VBM) within the phonological network for each task.

The TPJ/ventral inferior parietal lobe was implicated in all of our measures using whole-brain VBM, confirming the area's critical role in execution of spoken and written phonological tasks. This region, which is consistently damaged in logopenic PPA, is thought to support transcoding of auditory into articulatory–verbal information (Hickok & Poeppel, 2007) and storage/maintenance of phonological information (Buchsbaum & D'Esposito, 2008; Jacquemot & Scott, 2006; Paulesu, Frith, & Frackowiak, 1993), processes that are common to all of our phonological measures. Accordingly, only logopenic patients were impaired across spoken and written phonological tasks. Dorsal inferior parietal lobe/intraparietal sulcus, also damaged in logopenic PPA, is thought to play a role in analysis and construction of syllable order (Moser, Baker, Sanchez, Rorden, & Fridriksson, 2009), attention to temporal order during verbal working memory tasks (Owen, McMillan, Laird, & Bullmore, 2005), and serial sublexical (sound–letter conversion) processes invoked during written language processing (Wilson et al., 2009). Consistent with this role were the results of VBM analyses for reading and spelling of nonwords, which revealed correlations between the dorsal inferior parietal lobe/intraparietal sulcus and sublexical written language skills. In addition, ROI analyses revealed a significant independent contribution of temporoparietal cortex to phoneme blending performance, a task that requires not only maintenance of phonological information but also attention to serial ordering of phonemes before transcoding into articulatory gestures associated with whole words/pseudowords. The finding that spelling of nonwords and sound blending in nonwords were the most discriminating clinical measures for differentiation of logopenic relative to nonfluent/agrammatic patients also supports a critical contribution of temporoparietal cortex to these tasks.

Left frontal cortex was also implicated in all spoken phonological subtests on whole-brain VBM. Left middle and inferior frontal and premotor regions are thought to support overt and covert articulatory processes as well as monitoring and manipulation of phonological material (Marvel & Desmond, 2012; Peschke et al., 2012; Price, 2012; Champod & Petrides, 2010). To rule out the possibility that impaired performance on spoken phonological tasks was related solely to motor output difficulty, we screened participants to confirm adequate speech production ability for repetition of up to three syllables without error. This rather conservative criterion served to ensure that all individuals were more than capable of providing the monosyllabic responses required for completion of spoken phonological tasks.

Importantly, however, impaired subvocal rehearsal and inner speech generation have been documented in patients with apraxia of speech caused by stroke (Waters, Rochon, & Caplan, 1992) and anarthria (Cubelli & Nichelli, 1992), indicating that frontal cortices traditionally viewed as “motoric” likely play a role in both overt speech production as well as subvocal speech required for prearticulatory maintenance and assembly of phonological codes. This can account for the significant correlations between phonological scores and frontal volumes even after controlling for motor speech scores (Figure 4) and for the results of ROI analyses indicating frontal involvement in all phonological subtests. In addition to subvocal rehearsal, each of our spoken phonological tasks requires fundamental alterations in phonological representations, involving both monitoring and manipulation of phonological strings, also considered to be frontally mediated processes (in conjunction with posterior peri-sylvian regions). In contrast to previous findings (Henry et al., 2012), we did not observe a significant frontal contribution to reading and spelling of nonwords, although nonword scores did correlate with frontal volumes at an uncorrected threshold (p < .005). It may be the case that left frontal cortex provides a greater contribution to spoken relative to written phonological tasks due to modality of output or to a greater reliance on frontally mediated functions such as subvocal rehearsal during execution of these tasks.

In conclusion, our results confirm the selective impairment of dorsal articulatory–phonological processes in individuals with nonfluent/agrammatic and logopenic PPA and the sparing of this network in semantic variant PPA. We observed a striking impairment of phonological abilities across spoken and written tasks in individuals with the logopenic variant, supporting the assertion that language deficits in this patient group are phonological in nature. The contrast between the central phonological deficit in logopenic PPA and the sparing of phonological processing in semantic variant patients has implications for language treatment in these groups. Whereas both groups exhibit anomia as a primary characteristic, the underlying basis (semantic vs. phonological) differs in each. As such, naming treatments may be designed to capitalize on the relative sparing of these core linguistic processes in each patient group (Henry et al., 2013). Our findings also document impaired phonological task performance in individuals with nonfluent/agrammatic variant and suggest that phonological tasks engage frontal regions independently of their role in overt speech production. Specifically, they likely support subvocal rehearsal as well as manipulation and monitoring of phonological information. Finally, we have shown that phonological tasks (particularly those involving nonwords) have the potential to contribute to differential diagnosis in PPA and, as such, should be considered for inclusion in clinical assessment protocols.

Acknowledgments

This work was supported by the following grants: NIH, F32DC010945 (M. L. H.), R03DC013403 (M. L. H.), and R01DC007646 (P. M. B.); NIA P01AG019724 (Bruce Miller); and NINDS R01NS050915 (M. L. G. T.). We thank our participants and their families for their commitment to research. We would also like to thank Minerva Meese, LMS, SLP, for contributing to data management and analysis that supported this research.

Reprint requests should be sent to Maya L. Henry, Department of Communication Sciences and Disorders, University of Texas, Austin, 2504A Whitis Ave., Austin, TX 78712, or via e-mail: maya.henry@austin.utexas.edu.

REFERENCES

REFERENCES
Ashburner
,
J.
(
2007
).
A fast diffeomorphic image registration algorithm
.
Neuroimage
,
38
,
95
113
.
Baddeley
,
A.
(
2003
).
Working memory: Looking back and looking forward
.
Nature Reviews Neuroscience
,
4
,
829
839
.
Ballard
,
K. J.
,
Savage
,
S.
,
Leyton
,
C. E.
,
Vogel
,
A. P.
,
Hornberger
,
M.
, &
Hodges
,
J. R.
(
2014
).
Logopenic and nonfluent variants of primary progressive aphasia are differentiated by acoustic measures of speech production
.
PLoS One
,
9
,
e89864
.
Beeson
,
P. M.
,
Rising
,
K.
,
Kim
,
E. S.
, &
Rapcsak
,
S. Z.
(
2010
).
A treatment sequence for phonological alexia/agraphia
.
Journal of Speech, Language, and Hearing Research
,
53
,
450
468
.
Brambati
,
S. M.
,
Ogar
,
J.
,
Neuhaus
,
J.
,
Miller
,
B. L.
, &
Gorno-Tempini
,
M. L.
(
2009
).
Reading disorders in primary progressive aphasia: A behavioral and neuroimaging study
.
Neuropsychologia
,
47
,
1893
1900
.
Buchsbaum
,
B. R.
, &
D'Esposito
,
M.
(
2008
).
The search for the phonological store: From loop to convolution
.
Journal of Cognitive Neuroscience
,
20
,
762
778
.
Champod
,
A. S.
, &
Petrides
,
M.
(
2010
).
Dissociation within the frontoparietal network in verbal working memory: A parametric functional magnetic resonance imaging study
.
Journal of Neuroscience
,
30
,
3849
3856
.
Croot
,
K.
,
Ballard
,
K.
,
Leyton
,
C. E.
, &
Hodges
,
J. R.
(
2012
).
Apraxia of speech and phonological errors in the diagnosis of nonfluent/agrammatic and logopenic variants of primary progressive aphasia
.
Journal of Speech, Language, and Hearing Research
,
55
,
S1562
S1572
.
Cubelli
,
R.
, &
Nichelli
,
P.
(
1992
).
Inner speech in anarthria: Neuropsychological evidence of differential effects of cerebral lesions on subvocal articulation
.
Journal of Clinical and Experimental Neuropsychology
,
14
,
499
517
.
Dunn
,
L. M.
, &
Dunn
,
L. M.
(
1981
).
Manual for the Peabody Picture Vocabulary Test-Revised
.
Circle Pines, MN
:
American Guidance Service
.
Field
,
A.
(
2009
).
Discovering statistics using SPSS
.
London
:
Sage Publications
.
Gorno-Tempini
,
M. L.
,
Brambati
,
S. M.
,
Ginex
,
V.
,
Ogar
,
J.
,
Dronkers
,
N. F.
,
Marcone
,
A.
, et al
(
2008
).
The logopenic/phonological variant of primary progressive aphasia
.
Neurology
,
71
,
1227
1234
.
Gorno-Tempini
,
M. L.
,
Dronkers
,
N. F.
,
Rankin
,
K. P.
,
Ogar
,
J. M.
,
Phengrasamy
,
L.
,
Rosen
,
H. J.
, et al
(
2004
).
Cognition and anatomy in three variants of primary progressive aphasia
.
Annals of Neurology
,
55
,
335
346
.
Gorno-Tempini
,
M. L.
,
Hillis
,
A. E.
,
Weintraub
,
S.
,
Kertesz
,
A.
,
Mendez
,
M.
,
Cappa
,
S. F.
, et al
(
2011
).
Classification of primary progressive aphasia and its variants
.
Neurology
,
76
,
1006
1014
.
Grossman
,
M.
,
Mickanin
,
J.
,
Onishi
,
K.
,
Hughes
,
E.
,
D'Esposito
,
M.
,
Ding
,
X. S.
, et al
(
1996
).
Progressive nonfluent aphasia: Language, cognitive, and PET measures contrasted with probable Alzheimer's disease
.
Journal of Cognitive Neuroscience
,
8
,
135
154
.
Henry
,
M. L.
,
Beeson
,
P. M.
,
Alexander
,
G. E.
, &
Rapcsak
,
S. Z.
(
2012
).
Written language impairments in primary progressive aphasia: A reflection of damage to central semantic and phonological processes
.
Journal of Cognitive Neuroscience
,
24
,
261
275
.
Henry
,
M. L.
, &
Gorno-Tempini
,
M. L.
(
2010
).
The logopenic variant of primary progressive aphasia
.
Current Opinion in Neurology
,
23
,
633
637
.
Henry
,
M. L.
,
Rising
,
K.
,
DeMarco
,
A. T.
,
Miller
,
B. L.
,
Gorno-Tempini
,
M. L.
, &
Beeson
,
P. M.
(
2013
).
Examining the value of lexical retrieval treatment in primary progressive aphasia: Two positive cases
.
Brain and Language
,
127
,
145
156
.
Hickok
,
G.
(
2012
).
The cortical organization of speech processing: Feedback control and predictive coding the context of a dual-stream model
.
Journal of Communication Disorders
,
45
,
393
402
.
Hickok
,
G.
(
2014
).
The architecture of speech production and the role of the phoneme in speech processing
.
Language and Cognitive Processes
,
29
,
2
20
.
Hickok
,
G.
, &
Poeppel
,
D.
(
2004
).
Dorsal and ventral streams: A framework for understanding aspects of the functional anatomy of language
.
Cognition
,
92
,
67
99
.
Hickok
,
G.
, &
Poeppel
,
D.
(
2007
).
The cortical organization of speech processing
.
Nature Reviews Neuroscience
,
8
,
393
402
.
Hodges
,
J. R.
,
Patterson
,
K.
,
Oxbury
,
S.
, &
Funnell
,
E.
(
1992
).
Semantic dementia: Progressive fluent aphasia with temporal lobe atrophy
.
Brain
,
115
,
1783
1806
.
Jacquemot
,
C.
, &
Scott
,
S. K.
(
2006
).
What is the relationship between phonological short-term memory and speech processing?
Trends in Cognitive Sciences
,
10
,
480
486
.
Jefferies
,
E.
,
Jones
,
R. W.
,
Bateman
,
D.
, &
Lambon Ralph
,
M. A.
(
2005
).
A semantic contribution to nonword recall? Evidence for intact phonological processes in semantic dementia
.
Cognitive Neuropsychology
,
22
,
183
212
.
Johnson
,
J. K.
,
Diehl
,
J.
,
Mendez
,
M. F.
,
Neuhaus
,
J.
,
Shapira
,
J. S.
,
Forman
,
M.
, et al
(
2005
).
Frontotemporal lobar degeneration: Demographic characteristics of 353 patients
.
Archives of Neurology
,
62
,
925
930
.
Kaplan
,
E.
,
Goodglass
,
H.
, &
Weintraub
,
S.
(
2001
).
Boston Naming Test
.
Philadelphia
:
Lippincott, Williams and Wilkins
.
Kay
,
J.
,
Lesser
,
R.
, &
Coltheart
,
M.
(
1992
).
PALPA: Psycholinguistic Assessments of Language Processing in Aphasia
.
Hove, UK
:
Lawrence Erlbaum Associates Ltd.
Kertesz
,
A.
(
1982
).
The Western Aphasia Battery
.
New York
:
Grune & Stratton
.
Kirshner
,
H. S.
(
2014
).
Frontotemporal dementia and primary progressive aphasia, a review
.
Neuropsychiatric Disease and Treatment
,
10
,
1045
1055
.
Kramer
,
J. H.
,
Jurik
,
J.
,
Sha
,
S. J.
,
Rankin
,
K. P.
,
Rosen
,
H. J.
,
Johnson
,
J. K.
, et al
(
2003
).
Distinctive neuropsychological patterns in frontotemporal dementia, semantic dementia, and Alzheimer disease
.
Cognitive and Behavioral Neurology
,
16
,
211
218
.
Lambon Ralph
,
M. A.
,
Sage
,
K.
,
Jones
,
R. W.
, &
Mayberry
,
E. J.
(
2010
).
Coherent concepts are computed in the anterior temporal lobes
.
Proceedings of the National Academy of Sciences, U.S.A.
,
107
,
2717
2722
.
Leyton
,
C. E.
,
Ballard
,
K. J.
,
Piguet
,
O.
, &
Hodges
,
J. R.
(
2014
).
Phonologic errors as a clinical marker of the logopenic variant of PPA
.
Neurology
,
82
,
1620
1627
.
Mack
,
J. E.
,
Cho-Reyes
,
S.
,
Kloet
,
J. D.
,
Weintraub
,
S.
,
Mesulam
,
M.
, &
Thompson
,
C. K.
(
2013
).
Phonological facilitation of object naming in agrammatic and logopenic primary progressive aphasia (PPA)
.
Cognitive Neuropsychology
,
30
,
172
193
.
Marvel
,
C. L.
, &
Desmond
,
J. E.
(
2012
).
From storage to manipulation: How the neural correlates of verbal working memory reflect varying demands on inner speech
.
Brain and Language
,
120
,
42
51
.
Moser
,
D.
,
Baker
,
J. M.
,
Sanchez
,
C. E.
,
Rorden
,
C.
, &
Fridriksson
,
J.
(
2009
).
Temporal order processing of syllables in the left parietal lobe
.
Journal of Neuroscience
,
29
,
12568
12573
.
Owen
,
A. M.
,
McMillan
,
K. M.
,
Laird
,
A. R.
, &
Bullmore
,
E.
(
2005
).
N-back working memory paradigm: A meta-analysis of normative functional neuroimaging studies
.
Human Brain Mapping
,
25
,
46
59
.
Paulesu
,
E.
,
Frith
,
C. D.
, &
Frackowiak
,
R. S.
(
1993
).
The neural correlates of the verbal component of working memory
.
Nature
,
362
,
342
345
.
Peschke
,
C.
,
Ziegler
,
W.
,
Eisenberger
,
J.
, &
Baumgaertner
,
A.
(
2012
).
Phonological manipulation between speech perception and production activates a parieto-frontal circuit
.
Neuroimage
,
59
,
788
799
.
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
,
816
847
.
Rapcsak
,
S. Z.
,
Beeson
,
P. M.
,
Henry
,
M. L.
,
Leyden
,
A.
,
Kim
,
E.
,
Rising
,
K.
, et al
(
2009
).
Phonological dyslexia and dysgraphia: Cognitive mechanisms and neural substrates
.
Cortex
,
45
,
575
591
.
Rohrer
,
J. D.
,
Ridgway
,
G. R.
,
Crutch
,
S. J.
,
Hailstone
,
J.
,
Goll
,
J. C.
,
Clarkson
,
M. J.
, et al
(
2010
).
Progressive logopenic/phonological aphasia: Erosion of the language network
.
Neuroimage
,
49
,
984
993
.
Saur
,
D.
,
Kreher
,
B. W.
,
Schnell
,
S.
,
Kümmerer
,
D.
,
Kellmeyer
,
P.
,
Vry
,
M.
, et al
(
2008
).
Ventral and dorsal pathways for language
.
Proceedings of the National Academy of Sciences, U.S.A.
,
105
,
18035
18040
.
Tzourio-Mazoyer
,
N.
,
Landeau
,
B.
,
Papathanassiou
,
D.
,
Crivello
,
F.
,
Etard
,
O.
,
Delcroix
,
N.
, et al
(
2002
).
Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain
.
Neuroimage
,
15
,
273
289
.
Ueno
,
T.
,
Saito
,
S.
,
Rogers
,
T. T.
, &
Lambon Ralph
,
M. 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
,
385
396
.
Waters
,
G. S.
,
Rochon
,
E.
, &
Caplan
,
D.
(
1992
).
The role of high-level speech planning in rehearsal: Evidence from patients with apraxia of speech
.
Journal of Memory and Language
,
31
,
54
73
.
Wertz
,
R. T.
,
LaPointe
,
L. L.
, &
Rosenbek
,
J. C.
(
1984
).
Apraxia of speech in adults: The disorder and its management
.
New York
:
Grune and Stratton
.
Wilson
,
S. M.
,
Brambati
,
S. M.
,
Henry
,
R. G.
,
Handwerker
,
D. A.
,
Agosta
,
F.
,
Miller
,
B. L.
, et al
(
2009
).
The neural basis of surface dyslexia in semantic dementia
.
Brain
,
132
,
71
86
.
Wilson
,
S. M.
,
Henry
,
M. L.
,
Besbris
,
M.
,
Ogar
,
J. M.
,
Dronkers
,
N. F.
,
Jarrold
,
W.
, et al
(
2010
).
Connected speech production in three variants of primary progressive aphasia
.
Brain
,
133
,
2069
2088
.