Environmental sound processing in primary progressive aphasia: behavioral deficits and their neural correlates.

Introduction. Impaired language comprehension is frequently observed in primary progressive aphasia (PPA). Word comprehension deficits are characteristic of the semantic variant (PPA-S) whereas sentence comprehension deficits are more prevalent in the agrammatic (PPA-G) and logopenic (PPA-L) variants (Amici et al., 2007; Gorno-Tempini et al., 2011; Thompson et al., 2013). Word and sentence comprehension deficits have also been shown to have distinct neural substrates in PPA (Mesulam, Thompson, Weintraub, & Rogalski, in press). However, little is known about the relationship between word and sentence comprehension processes in PPA, specifically how words are accessed, combined, and used to predict upcoming elements within a sentence. A previous study demonstrated that listeners with stroke-induced agrammatic aphasia rapidly access verb meanings and use them to semantically integrate verb-arguments; however, they show deficits in using verb meanings predictively (Mack, Ji, & Thompson, 2013). The present study tested whether listeners with PPA are able to access verb meanings and to use this information to integrate and predict verb-arguments. Methods. Fifteen adults with PPA (8 with PPA-G, 3 with PPA-L, and 4 with PPA-S) and ten age-matched controls participated in two eyetracking experiments. In both experiments, participants heard sentences with restrictive verbs that were semantically compatible with only one object in a four-picture visual array (e.g., eat when the array included a cake and three non-edible objects) and unrestrictive verbs (e.g., move) that were compatible with all four objects. The verb-based integration experiment tested access to verb meaning and its effects on integration of the direct object (e.g., Susan will eat/move the cake); the verb-based prediction experiment examined prediction of the direct object (e.g., Susan will eat/move the …). The dependent variable was the rate of fixations on the target picture (e.g., the cake) in the first 500 ms after the offset of the verb. Logistic regression was used to compare the rate of target fixations between PPA and controls and to test for differences between PPA subtypes. Results. In the verb-based integration experiment (Fig. 1a), PPA listeners as well as controls showed rapid access to verb meaning, making more target fixations in the restrictive than unrestrictive conditions in the first 500 ms after verb offset. No significant differences were found between participant groups. In the verb-based prediction experiment (Fig. 1b), control listeners exhibited a greater difference between the restrictive and unrestrictive conditions compared to PPA listeners. In both experiments, no significant differences were found between PPA subtypes. Conclusion. The results of this study suggest that access to verb meaning is relatively preserved in PPA and can facilitate integration of verb-arguments. However, prediction of verb-arguments is impaired. These findings are in line with stroke-induced agrammatic aphasia, in which prediction is markedly impaired (Mack et al., 2013). The similar pattern of results across PPA subtypes should be interpreted cautiously due to small sample sizes. However, these findings suggest that – despite well-established differences in word and sentence comprehension impairments – there may also be shared deficits across PPA subtypes that affect the ability to use lexical information predictively during sentence comprehension.

BackgroundAdvancements in artificial intelligence and natural language processing (NLP) have made it a valuable tool for diagnostic and clinical applications. Building on Rezaii et al. (2024), who used AI‐driven NLP tools to identify linguistic features discriminating Primary Progressive Aphasia (PPA) subtypes from a picture description task, the current study aimed to replicate and extend these findings using an ecologically‐valid language sample often derived from the clinical question, “What did you do for work?” (WDYDFW?). We hypothesized that features of sentence length, word frequency, and content word usage, would discriminate among PPA subtypes.MethodSpeech samples from the WDYDFW? prompt were collected from 58 PPA participants (Mean age = 69.62 (SD = 7.38); 50.0% female) and 16 healthy controls (Mean age = 65.56 (SD = 4.95); 62.5% female). Digital language markers were extracted from transcribed samples using Quantitext, an NLP‐based toolbox our team has developed. Sentence length (average # words per sentence), word frequency (commonality in the English language of words used), content word usage (words that provide meaning, i.e., nouns, verbs, adjectives, adverbs), and syntax frequency (simple vs complex sentences structure) previously identified as key distinguishing linguistic features (identified by Rezaii et al. as key discriminating language features) were quantified for each sample. One‐way ANOVA with cohort as the dependent variable and linguistic features as the independent variables was conducted to detect differences in language markers among groups.ResultPreliminary analyses indicated individuals with nfvPPA produced significantly shorter sentences (F(3,70) = 12.583, p < .001) and lower‐frequency words compared to controls (F(3,70) =17.395, p = .01,), lvPPA (p < .001), and svPPA (p < 0.001). Controls used lower frequency words than lvPPA (F(3,70) =17.395, p = 0.02). lvPPA participants produced fewer content words than nfvPPA (F(3,70) = 3.77, p < 0.025) and control individuals (p < 0.05). Syntax frequency did not differ across groups.ConclusionConsistent with previous findings, NLP has the potential to discriminate among PPA subtypes using unstructured language samples. This study underscores the potential of AI tools to classify atypical cognitive‐behavioral syndromes. Analyses are ongoing to assess impact of symptom severity (CDR) and to validate findings against structured language samples.

Primary progressive aphasia (PPA) is a neurodegenerative syndrome characterized by progressive decline in speech and/or language. There are three PPA subtypes with distinct speech-language profiles. Early diagnosis is essential for optimal provision of care but differential diagnosis by PPA subtype can be difficult and time consuming. We investigated the diagnostic utility of a novel electroencephalography (EEG)-based biomarker in conjunction with machine learning. Individuals with semantic, logopenic, or nonfluent/agrammatic variant PPA and healthy controls (n = 10 per group) listened to a continuous narrative while EEG responses were recorded. The speech envelope and linguistic features representing core language processes were extracted from the narrative speech and temporal response function (TRF) modeling was used to estimate the neural responses to these features. Although TRF modeling has shown promise for clinical applications, research is lacking regarding its diagnostic utility in populations like PPA. This study sought to provide preliminary evidence to address this gap. The resulting TRFs for channel Cz were used as input to machine learning algorithms for classification of PPA vs. healthy controls, three-way classification by PPA subtype, classification of a single PPA subtype relative to the other two (e.g., semantic vs. logopenic/nonfluent variant), and pairwise classification by PPA subtype. F1 scores were highest for the latter tasks (F1’s from 0.73 to 0.74), with better-than-chance classification in all tasks. Additional analyses determined that the TRF beta weights significantly improved classification over preprocessed EEG waveforms alone for all but one task (PPA vs. healthy controls). Our preliminary findings demonstrate the potential utility of this approach for differential diagnosis of PPA, warranting further investigation.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-13000-8.

Impacts of dementia syndromes on caregivers are well established, but research specific to Primary Progressive Aphasia (PPA) populations is scant. In particular, little is known about the impacts of non-language symptoms (e.g., emotion recognition and behavioural disturbance) on caregiver outcomes in PPA. The present study sought to investigate the interrelationships between non-language symptom profiles, caregiver coping behaviours, and caregiver outcomes (psychological wellbeing and burden) among PPA subtypes. Ninety-six PPA person-caregiver dyads (30 with logopenic variant [lvPPA], 26 with non-fluent variant [nfvPPA], and 40 with semantic variant [svPPA]) and 122 healthy controls were included in this cross-sectional study. Caregiver outcomes were assessed using the Zarit Burden Interview (ZBI) and the 21-item depression, anxiety, and stress scale (DASS-21). The study investigated whether a person with PPA-focused variables (psychological wellbeing, emotion recognition ability, behavioural disturbance) and caregiver coping style predicted caregiver outcomes. Differential caregiver coping styles were also examined. Overall, caregivers most commonly used adaptive coping styles (problem-focused, emotion-focused). Symptom profiles and use of dysfunctional coping correlated negatively with caregiver psychological wellbeing and positively with burden. Regression models indicated that caregiver psychological wellbeing was most strongly predicted by the use of dysfunctional coping strategies, and caregiver burden was most strongly predicted by reduced emotion recognition and presence of behavioural disturbance symptoms in persons with PPA. This study highlights the importance of considering non-language symptoms in persons diagnosed with PPA and their impact on caregiver outcomes. These findings may inform the development of psychoeducation materials and interventions for PPA caregivers. Further research is needed to identify the predictors of PPA caregiver outcomes with disease progression. Studies utilising qualitative approaches and considering caregiver gain are warranted to understand the experience of PPA caregivers further. What is already known on this subject Impacts on dementia caregivers are well established, including reduced psychological wellbeing and high perceived caregiver burden. However, little is known about the effects of increasingly documented non-language symptoms (behavioural disturbance, emotion recognition impairment) in persons with Primary Progressive Aphasia (PPA), and whether caregiver coping behaviour influences these outcomes. What this paper adds to the existing knowledge This study is the first to investigate the interrelationships between symptom presentation in the person with PPA, caregiver coping behaviours, and caregiver outcomes in PPA subtypes. Our investigations identified the combination of risk factors most strongly predictive of impacted caregiver outcomes, which varied according to PPA subtype. Overall, caregiver burden was most strongly predicted by the person with PPA-focused variables (emotion recognition, behavioural disturbance), while caregiver psychological wellbeing was most strongly predicted by caregivers' use of dysfunctional coping. What are the potential or actual clinical implications for this work? Our findings inform intervention design, identify targets for psychoeducation, and suggest directions for future research. To best preserve PPA caregivers' psychological wellbeing, supporting the development of adaptive coping skills appears crucial. Levels of caregiver burden should be monitored especially closely when persons with PPA showcase non-language symptoms, with protective measures such as respite care implemented pre-emptively.

Clinical assessment of frontotemporal lobar degeneration (FTLD)/primary progressive aphasia (PPA) patients is challenging, given that common cognitive assessments rely extensively on language. Since asymmetry in neuroimaging biomarkers is often described as a central finding in these patients, our study evaluated [18F]-fluorodeoxyglucose (FDG) uptake patterns in patients meeting clinical and imaging criteria for FTLD, with emphasis on PPA. Fifty-one subjects underwent brain [18F]-FDG positron-emission tomography/magnetic resonance imaging (PET/MRI) as part of their routine clinical workup for dementia and neurodegenerative disease. Images were obtained using a Siemens Biograph mMR integrated 3T PET/MRI scanner. PET surface maps and fusion fluid-attenuated inversion recovery-PET images were generated utilizing MIMneuro software. Two board-certified neuroradiologists and one nuclear medicine physician blinded to patient history classified each FTLD/PPA subtype and assessed for left- versus right-side dominant hypometabolism. Qualitative and semiquantitative assessment demonstrated 18 cases of PPA, 16 behavioral variant frontotemporal dementia (bvFTD), 12 corticobasal degeneration, and 5 progressive supranuclear palsy. Among the 18 PPA subjects (11 svPPA, 5 lvPPA, and 2 agPPA), 12 (67%) demonstrated left-dominant hypometabolism and 6 (33%) right-dominant hypometabolism. While existing literature stresses left-dominant hypometabolism as a key imaging feature in the PPA subtypes, a third of our cases demonstrated right-dominant hypometabolism, suggesting that emphasis should be placed on the functionality of specific brain regions affected, rather than left versus right sidedness of hypometabolism patterns.

Brain total creatine differs between primary progressive aphasia (PPA) subtypes and correlates with disease severity

Convolutional neural networks (CNNs) with log-mel audio representation and CNN-based end-to-end learning have both been used for environmental event sound recognition (ESC). However, log-mel features can be complemented by features learned from the raw audio waveform with an effective fusion method. In this paper, we first propose a novel stacked CNN model with multiple convolutional layers of decreasing filter sizes to improve the performance of CNN models with either log-mel feature input or raw waveform input. These two models are then combined using the Dempster–Shafer (DS) evidence theory to build the ensemble DS-CNN model for ESC. Our experiments over three public datasets showed that our method could achieve much higher performance in environmental sound recognition than other CNN models with the same types of input features. This is achieved by exploiting the complementarity of the model based on log-mel feature input and the model based on learning features directly from raw waveforms.

Predicting primary progressive aphasias with support vector machine approaches in structural MRI data

The three variants of primary progressive aphasia (PPA) differ in clinical presentation, underlying brain pathology, and clinical course, which stresses the need for early differentiation. However, brief cognitive tests that validly distinguish between all PPA variants are lacking. The Sydney Language Battery (SYDBAT) is a promising screening instrument that can be used as a first step in a comprehensive neuropsychological assessment to distinguish PPA subtypes, but evidence on its validity and reliability is to date limited. In the current study, the validation and diagnostic value of the SYDBAT are described for discriminating PPA subtypes as well as distinguishing PPA from mild cognitive impairment (MCI) or Alzheimer's dementia (AD). Forty-five patients with PPA (13 with semantic PPA, 20 with logopenic PPA, and 12 with nonfluent/agrammatic PPA), 25 MCI patients, 13 AD patients, and 50 cognitively unimpaired controls were included in this study. Both patients and controls completed the SYDBAT-NL (Dutch version). Performance on and predictive ability of the four subtests (i.e., Naming, Word Comprehension, Repetition, and Semantic Association) were assessed. In addition, construct validity and internal consistency were examined. Different SYDBAT performance patterns were found across PPA and non-PPA patient groups. While a discriminant function analysis based on SYDBAT subtest scores could predict PPA subtype with 78% accuracy, it was more difficult to disentangle PPA from non-PPA patients based on SYDBAT scores alone. For assisting in clinical interpretation, simple rules were set up and translated into a diagnostic decision tree for subtyping PPA, which was capable of diagnosing a large proportion of the cases. Satisfying validity and reliability measures were found. The SYDBAT is an easy-to-use and promising screen for assessing single-word language processes, which may contribute to the differential diagnostic process of PPA and the assessment of language impairment in MCI and AD. It can be easily implemented for initial screening of patients in a memory clinic.

Primary progressive aphasia (PPA) is a neurodegenerative syndrome of progressive language decline. PPA has 3 main subtypes: logopenic, semantic, and agrammatic. Observational studies suggested an association between language-related neurodevelopmental phenotypes and an increased risk of PPA. We sought to assess such relationships through Mendelian randomization (MR) approach, which can suggest potentially causal associations. Genome-wide significant single-nucleotide polymorphisms (SNPs) associated with dyslexia (42 SNPs), developmental speech disorders (29 SNPs), and left-handedness (41 SNPs) were used as genetic proxies for the exposures. Eighteen of 41 SNPs of left-handedness were associated with structural asymmetry of the cerebral cortex. Genome-wide association study summary statistics were obtained from publicly available databases for semantic (308 cases/616 controls) and agrammatic PPA (269 cases/538 controls). The logopenic PPA (324 cases/3,444 controls) was approximated by proxy through the rubric of clinically diagnosed Alzheimer disease with salient language impairment. Inverse-weighted variance MR was performed as the main analysis for testing the relationship between the exposures and outcomes. Sensitivity analyses were completed to test the robustness of the results. Dyslexia, developmental speech disorders, and left-handedness were not associated with any PPA subtype (p > 0.05). The genetic proxy of cortical asymmetry in left-handedness was significantly associated with agrammatic PPA (β = 4.3, p = 0.007), but not with other PPA subtypes. This association was driven by microtubule-related genes, primarily by a variant that is in complete linkage disequilibrium with MAPT gene. Sensitivity analyses were overall consistent with the primary analyses. Our results do not support a causal association between dyslexia, developmental speech disorders, and handedness with any of the PPA subtypes. Our data suggest a complex association between cortical asymmetry genes and agrammatic PPA. Whether the additional association with left-handedness is necessary remains to be determined but is unlikely, given the absence of association between left-handedness and PPA. Genetic proxy of brain asymmetry (regardless of handedness) was not tested as an exposure due to lack of suitable genetic proxy. Furthermore, the genes related to cortical asymmetry associated with agrammatic PPA are implicated in microtubule-related proteins (TUBA1B, TUBB, and MAPT), which is keeping with the association of tau-related neurodegeneration in this PPA variant.

Behavioral assessment has been investigated in frontotemporal lobar degeneration and Alzheimer’s disease, but has not been explored extensively in subtypes of primary progressive aphasia (PPA). We explored the ability of a modified version of the Frontal Behavioral Inventory (FBI-mod) to discriminate between patients with distinct subtypes of PPA and patients with mild cognitive impairment (MCI). We hypothesized that individuals with nonfluent agrammatic PPA (nfaPPA) would have higher negative behavior scores than other groups and that individuals with semantic variant PPA (svPPA) would have higher disinhibition scores than other groups. Family members and/or caregivers of 120 individuals with PPA and MCI (mean age 69.54+8.75 years; 65 (54%) female; education 16.06±2.68 years; disease duration 46.47±34.26 months) completed the FBI-mod [logopenic PPA (lvPPA) n = 40. nfaPPA n = 29, svPPA n = 27, MCI n = 24]. The groups were not significantly different in age, gender, education, or disease duration. There were no significant differences between the groups for negative behaviors (p = 0.72) and disinhibition scores (p = 0.14). When comparing negative and disinhibition scores (in percent), negative scores were significantly higher in all groups (p < 0.001). When comparing subtest items, there was a pairwise difference between lvPPA and svPPA for restlessness (lvPPA < svPPA, p = 0.02, after adjusting for multiple between-group comparisons). There was a significant difference in the proportion of severe neglect between the groups with lvPPA having a lower proportion than the other two variants (p = 0.05), and there was a significant difference in the proportion of severe poor judgment between the groups with lvPPA also having a lower proportion than nfaPPA (p = 0.04). This study reveals the greater negative behavioral disturbance than disinhibition in the PPA and MCI groups of similar age and duration since onset and identifies different profiles for some specific behaviors for the PPA groups. These findings may have clinical and practical implications.

Motor speech disorders in the nonfluent, semantic and logopenic variants of primary progressive aphasia

Two-thirds of the world’s languages, including Mandarin and Cantonese, employ pitch variation to convey meaning (lexical tone). Existing diagnostic frameworks for primary progressive aphasia (PPA) have been developed for English speakers, and have not considered the impact of salient language-specific variations, such as tone. This study investigates lexical tone processing in Mandarin- and Cantonese-speaking individuals with PPA and examines their neural signatures using structural neuroimaging.Methods:Seventy-eight native Chinese speakers (54 with PPA; 24 healthy controls) were assessed using the CLAP (Chinese Language Assessment for PPA) battery, a series of neuropsychological and linguistic tasks developed to characterize the linguistic features of Mandarin and Cantonese speakers with PPA. Lexical tone production was examined through repetition and reading of “tone-twister” phrases, as well as repetition of multicharacter phrases varying in articulatory features (place, manner, and tone). Tone perception and comprehension was assessed via identification, discrimination, and tone-word/picture matching tasks. Group differences were analyzed using nonparametric tests and generalized estimating equations, with ROC analyses determining diagnostic accuracy. Structural MRI data were acquired for 55 participants, and voxel-based morphometry (VBM) was used to examine the neural correlates of tone performance.Results:Participants with nonfluent/agrammatic variant PPA (nfvPPA) showed marked impairments in lexical tone production (all p<0.001), with a disproportionately high rate of tonal relative to syllabic errors (p<0.001). In contrast, semantic variant PPA (svPPA) exhibited prominent deficits in three tone perception tasks (all p<0.001). Patients with logopenic variant PPA (lvPPA) showed relatively preserved tone production but a predominance of syllabic errors (p<0.001), suggesting underlying phonological deficits. Lexical tone production tasks demonstrated strong discrimination of nfvPPA (AUC= 0.702–0.907). In contrast, three tone perception tasks exhibited high sensitivity for detecting svPPA (90.9–100%), though specificity was modest (37–63%). Neural correlate analyses revealed that tone production deficits were associated with reduced grey matter volume in the left inferior frontal gyrus, insula, and temporal cortex, whereas tone perception performance correlated with atrophy in the left superior and middle temporal gyri, temporal pole, and orbitofrontal regions.Discussion:Lexical tone processing is differentially impaired across PPA subtypes, with tone production and perception deficits mapping onto distinct neural substrates. These findings underscore the necessity of developing language-specific diagnostic approaches for tonal language speakers and call into question the cross-cultural applicability of current PPA diagnostic strategies, which have been largely shaped by Indo-European language frameworks.

The primary progressive aphasias are rare, language-led dementias, with three main variants: semantic, non-fluent/agrammatic and logopenic. Although the semantic variant has a clear neuroanatomical profile, the non-fluent/agrammatic and logopenic variants are difficult to discriminate from neuroimaging. Previous phenotype-driven studies have characterized neuroanatomical profiles of each variant on MRI. In this work, we used a machine learning algorithm known as SuStaIn to discover data-driven neuroanatomical 'subtype' progression profiles and performed an in-depth subtype-phenotype analysis to characterize the heterogeneity of primary progressive aphasia. Our study included 270 participants with primary progressive aphasia seen for research in the UCL Queen Square Institute of Neurology Dementia Research Centre, with follow-up scans available for 137 participants. This dataset included individuals diagnosed with all three main variants (semantic, n = 94; non-fluent/agrammatic, n = 109; logopenic, n = 51) and individuals with unspecified primary progressive aphasia (n = 16). A dataset of 66 patients (semantic, n = 37; non-fluent/agrammatic, n = 29) from the ARTFL LEFFTDS Longitudinal Frontotemporal Lobar Degeneration (ALLFTD) Research Study was used to validate our results. MRI scans were segmented, and SuStaIn was used on 19 regions of interest to identify neuroanatomical profiles independent of the diagnosis. We assessed the assignment of subtypes and stages, in addition to their longitudinal consistency. We discovered four neuroanatomical subtypes of primary progressive aphasia, labelled S1 (left temporal), S2 (insula), S3 (temporoparietal) and S4 (frontoparietal), exhibiting robustness to statistical scrutiny. S1 was correlated strongly with the semantic variant, whereas S2, S3 and S4 showed mixed associations with the logopenic and non-fluent/agrammatic variants. Notably, S3 displayed a neuroanatomical signature akin to a logopenic-only signature, yet a significant proportion of logopenic cases were allocated to S2. The non-fluent/agrammatic variant demonstrated diverse associations with S2, S3 and S4. No clear relationship emerged between any of the neuroanatomical subtypes and the unspecified cases. At first follow-up, subtype assignment was stable for 84% of patients, and stage assignment was stable for 91.9% of patients. We partially validated our findings in the ALLFTD dataset, finding comparable qualitative patterns. Our study, leveraging machine learning on a large primary progressive aphasia dataset, delineated four distinct neuroanatomical patterns. Our findings suggest that separable spatiotemporal neuroanatomical phenotypes do exist within the primary progressive aphasia spectrum, but that these are noisy, particularly for the non-fluent/agrammatic non-fluent/agrammatic and logopenic variants. Furthermore, these phenotypes do not always conform to standard formulations of clinico-anatomical correlation. Understanding the multifaceted profiles of the disease, encompassing neuroanatomical, molecular, clinical and cognitive dimensions, has potential implications for clinical decision support.

Background/aims Primary progressive aphasia (PPA) encompasses a group of neurodegenerative disorders characterized by progressive language decline. However, diagnosing mixed variant of PPA (mvPPA) can be challenging because symptoms overlap across PPA subtypes. This ambiguity is further compounded by cross-linguistic variations in language function and the limited availability of diagnostic tools for languages other than English. This study bridges a critical gap in the field by providing the first review regarding mvPPA from both neuropsychological and cross-linguistic perspectives. Methods/procedures This narrative review examined studies from four databases: MEDLINE, Web of Science, Google Scholar, and Scopus, spanning the period from 2000 to 2024. A summary of the core clinical and neuropsychological features of different PPA subtypes, along with their corresponding clinical characteristics, is presented for the selected studies. Results/outcomes Eighteen selected studies included patients with mixed PPA variant. These patients were heterogeneous, with complex neuropsychological profiles that combine features of established PPA variants. Neuropsychological findings showed heterogeneous profiles in mixed PPA. Naming impairments, grammatical errors, and phonological difficulties were common. This stage of disease evolves from the spread of atrophy and mixed pathology, but it can also emerge earlier. Neuroimaging findings, such as atrophy and hypoperfusion on MRI and PET scans, were sometimes inconsistent with clinical diagnoses. Conclusion Diagnosing mixed variant of PPA presents significant challenges. There are considerable gaps in our understanding of how language background, neuroimaging markers, and neuropsychological functions interact in clinical and cross-linguistic manifestations. Urgent cross-linguistic research is needed, especially for bilingual or multilingual individuals with PPA, to better characterize linguistic typologies and shared brain atrophy sites across language systems.