Word Generation Research Articles

The FinSwed study: Using verbal NEPSY-II subtests with a cultural minority group

While it has been established that cultural differences in test performance are common, few studies have compared groups from different countries that speak the same language. The aim of this study was to investigate the generalizability of selected linguistic and verbal memory subtests from the Swedish NEPSY-II to the Finland-Swedish minority group. The participants were 275 Swedish-speaking children from Finland aged 5–16-years. The performance was compared to the U.S. norms used in the test. The Finland-Swedish children scored around scaled score 12 and significantly higher than the normative mean on the subtests Comprehension of Instructions, Phonological Processing, Word Generation Semantic, List Memory, and Sentence Repetition, whereas no significant differences to the norms emerged for the subtests Word Generation Initial Letter, Narrative Memory, and Word List Interference. There was a significant age effect for two subtests, with scaled scores increasing with age. Lower parental education and male sex were associated with lower test scores on some subtests, whereas bilingualism was not. The findings were viewed in relation to previous comparison studies involving the NEPSY-II/NEPSY. The differences were suggested to relate to cultural, educational, and test-related differences between Finland, Sweden, and the U.S. The results add to the clinical utility of the NEPSY-II.

A - 74 Comparing HVLT-R Performance in Monolingual Vs. Spanish/English Bilingual Veterans

Abstract Objective Norms developed in monolingual populations may not be appropriate for use with bilingual patients. This study examined the impact of Spanish-English bilingualism on rote verbal memory using the Hopkins Verbal Learning Test-Revised (HVLT-R). Methods Clinically-referred veterans (86 monolingual; 30 bilingual) completed HVLT-R (English) during a comprehensive neuropsychological evaluation. Chi square analyses and independent samples t-tests were used to examine group demographic differences. Verbal fluency scores and HVLT-R scores were compared between groups using independent samples t-tests. Results Groups did not differ significantly in age (monolingual-M = 62.62; bilingual-M = 65.67) or education (monolingual-M = 13.9; bilingual-M = 13.2). Self-reported race/ethnicity was mostly White for the monolingual sample, while the bilingual sample was predominantly Hispanic. Groups also differed in gender composition (bilingual: 100% male; monolingual: 84% male). Groups did not differ in their speeded English phonemic word generation, t(108) = 0.311, p = 0.757, or speeded semantic word generation, t(109) = 1.931, p = 0.056, suggesting the current bilingual sample had high English proficiency. Cognitively unimpaired monolingual and bilingual veterans did not differ in their raw total recall, t(114) = 1.228, p = 0.222, delayed free recall, t(114) = 0.331, p = 0.741, percent retention, t(114) = 0.240, p = 0.811, or recognition discrimination, t(114) = 0.521, p = 0.603, for target words. Groups did not differ in their age-corrected T-scores in any of the above scores. Conclusions Results support the use of existing HVLT-R normative data for assessment of rote verbal memory in Spanish/English bilinguals with high English proficiency.

Intrahemispheric White Matter Asymmetries and Interhemispheric Connections Underlying the Lateralization of Language Production and Spatial Attention in Left-Handers

ABSTRACT Leftward language production and rightward spatial attention are salient features of functional organization in most humans, but their anatomical basis remains unclear. Interhemispheric connections and intrahemispheric white matter asymmetries have been proposed as important factors underlying functional lateralization. To investigate the role of white matter connectivity in functional lateralization, we first identified 96 left-handers using visual half field naming tasks. They were then divided into atypical and typical functional dominance based on the lateralization of brain activation in a word generation task (for language production) and a landmark task (for spatial attention). Using a novel fixel-based framework, we obtained fiber-specific properties of white matter pathways. Results showed, first, that differences between two language dominance groups occurred in the asymmetry of the superior longitudinal fasciculus-III (SLF-III), whereas differences between two spatial attention dominance groups occurred in the rostrum and rostral body of the corpus callosum. However, the directions of functional lateralization were not associated with the directions of white matter asymmetries. Second, the degree of language lateralization was predicted by SLF-III asymmetry and the rostral body of the corpus callosum, whereas the degree of spatial attention lateralization was predicted by the rostrum of the corpus callosum. Notably, the degree of each functional lateralization was positively correlated with the anterior and middle callosal connections, supporting the excitatory model of the corpus callosum. The results suggest that language lateralization is shaped by a combined effect of intra- and interhemispheric connections, whereas spatial attention lateralization relies more on interhemispheric connections.

Word Generation Interventions for Individuals with Aphasia: A Systematic Review

Word Generation Interventions for Individuals with Aphasia: A Systematic Review

SMART: Syntax-Calibrated Multi-Aspect Relation Transformer for Change Captioning.

Change captioning aims to describe the semantic change between two similar images. In this process, as the most typical distractor, viewpoint change leads to the pseudo changes about appearance and position of objects, thereby overwhelming the real change. Besides, since the visual signal of change appears in a local region with weak feature, it is difficult for the model to directly translate the learned change features into the sentence. In this paper, we propose a syntax-calibrated multi-aspect relation transformer to learn effective change features under different scenes, and build reliable cross-modal alignment between the change features and linguistic words during caption generation. Specifically, a multi-aspect relation learning network is designed to 1) explore the fine-grained changes under irrelevant distractors (e.g., viewpoint change) by embedding the relations of semantics and relative position into the features of each image; 2) learn two view-invariant image representations by strengthening their global contrastive alignment relation, so as to help capture a stable difference representation; 3) provide the model with the prior knowledge about whether and where the semantic change happened by measuring the relation between the representations of captured difference and the image pair. Through the above manner, the model can learn effective change features for caption generation. Further, we introduce the syntax knowledge of Part-of-Speech (POS) and devise a POS-based visual switch to calibrate the transformer decoder. The POS-based visual switch dynamically utilizes visual information during different word generation based on the POS of words. This enables the decoder to build reliable cross-modal alignment, so as to generate a high-level linguistic sentence about change. Extensive experiments show that the proposed method achieves the state-of-the-art performance on the three public datasets.

Artificial Intelligence Tools and Bias in Journalism-related Content Generation: Comparison Between Chat GPT-3.5, GPT-4 and Bing

This study explores the biases present in artificial intelligence (AI) tools, focusing on GPT-3.5, GPT-4, and Bing. The performance of the tools has been compared with a group of experts in linguistics, and journalists specialized in breaking news and international affairs. It reveals that GPT-3.5, widely accessible and free, exhibits a higher tendency rate in its word generation, suggesting an intrinsic bias within the tool itself rather than in the input data. Comparatively, GPT-4 and Bing demonstrate differing patterns in term generation and subjectivity, with GPT-4 aligning more closely with expert opinions and producing fewer opinative words. The research highlights the extensive use of generative AI in media and among the general populace, emphasizing the need for careful reliance on AI-generated content. The findings stress the risks of misinformation and biased reporting inherent in unexamined AI outputs. The challenge for journalists and information professionals is to ensure accuracy and ethical judgment in content creation to maintain the quality and diversity of content in journalistic practices.

QOL-24. TREATMENT- AND TIME-RELATED IMPACT ON LONG-TERM NEUROCOGNITIVE OUTCOMES IN ADULT SURVIVORS OF A PEDIATRIC BRAIN TUMOR

Abstract Treatment for pediatric brain tumors has been associated with the risk of developing long-term neurocognitive impairment. However, it is not well understood which neurocognitive profiles exist within this population. We investigated neurocognitive functioning (NF) in survivors in early adulthood and aimed to assess the impact of radiotherapy dose (RTD) and time since treatment (TxT) on NF. Forty-six survivors of pediatric brain tumors (mean age 23.99±4.71, mean age at diagnosis 9.29±4.54) performed an extensive neurocognitive test battery resulting in intelligence (n=5), memory (n=2), and language (n=3) scores. NF in long-term survivors was compared to sex-, age-, and education-matched controls, and between the different treatment strategies (surgery-only n=19, chemotherapy n=8, radiotherapy n=11, or chemo-radiotherapy n=8). Furthermore, relation between max-RTD (range 40.28 – 61.61 Gy) as well as TxT and NF was assessed. Three ANCOVA models were tested, and false discovery rate (FDR) was applied. For all scales and domain scores, survivors scored lower compared to controls. Scores were significant lower on full scale intelligence (pfdr< 0.001), verbal comprehension (pfdr< 0.001), working memory (pfdr= 0.003), perceptual reasoning (pfdr< 0.001), processing speed (pfdr< 0.001), spoken word comprehension (pfdr= 0.001), and semantic word generation (pfdr= 0.036) compared to controls. NF was not significantly different between the different treatments, nor did the association with max-RTD reach significance. TxT was associated with verbal comprehension and spoken word comprehension; albeit significance was not retained after FDR correction. In this study, the overall survivor cohort scored lower on multiple cognitive outcomes. More detailed analysis of various risk factors might further elucidate the different neurocognitive profiles for this population. Furthermore, more detailed analysis of RTD distributions and affected brain structures may provide greater insight on the impact of radiotherapy on NF.

Probabilistic Presurgical Language fMRI Atlas of Patients with Brain Tumors.

Patients with brain tumors have high intersubject variation in putative language regions, which may limit the utility of straightforward application of healthy subject brain atlases in clinical scenarios. The purpose of this study was to develop a probabilistic functional brain atlas that consolidates language functional activations of sentence completion and Silent Word Generation language paradigms using a large sample of patients with brain tumors. The atlas was developed using retrospectively collected fMRI data from patients with brain tumors who underwent their first standard-of-care presurgical language fMRI scan at our institution between July 18, 2015, and May 13, 2022. Three hundred seventeen patients (861 fMRI scans) were used to develop the language functional atlas. An independent presurgical language fMRI data set of 39 patients with brain tumors from a previous study was used to evaluate our atlas. Family-wise error-corrected binary functional activation maps from sentence completion, letter fluency, and category fluency presurgical fMRI were used to create probability overlap maps and pooled probabilistic overlap maps in Montreal Neurological Institute standard space. The Wilcoxon signed-rank test was used to determine a significant difference in the maximum Dice coefficient for our atlas compared with a meta-analysis-based template with respect to expert-delineated primary language area activations. Probabilities of activating the left anterior primary language area and left posterior primary language area in the temporal lobe were 87.9% and 91.5%, respectively, for sentence completion, 88.5% and 74.2%, respectively, for letter fluency, and 83.6% and 67.6%, respectively, for category fluency. Maximum Dice coefficients for templates derived from our language atlas were significantly higher than the meta-analysis-based template in the left anterior primary language area (0.351 and 0.326, respectively, P < .05) and the left posterior primary language area in the temporal lobe (0.274 and 0.244, respectively, P < .005). Brain tumor patient- and paradigm-specific probabilistic language atlases were developed. These atlases had superior spatial agreement with fMRI activations in individual patients compared with the meta-analysis-based template.

Functional and Effective Connectivity Underlying Semantic Verbal Fluency.

Semantic verbal fluency (SVF) impairment is present in several neurological disorders. Although activation in SVF-related areas has been reported, how these regions are connected and their functional roles in the network remain divergent. We assessed SVF static and dynamic functional connectivity (FC) and effective connectivity in healthy participants using functional magnetic resonance imaging. We observed activation in the inferior frontal (IFG), middle temporal (pMTG) and angular gyri (AG), anterior cingulate (AC), insular cortex, and regions of the superior, middle, and medial frontal gyri (SFG, MFG, MidFG). Our static FC analysis showed a highly interconnected task and resting state network. Increased connectivity of AC with the pMTG and AG was observed for the task. The dynamic FC analysis provided circuits with connections similarly modulated across time and regions related to category identification, language comprehension, word selection and recovery, word generation, inhibition of speaking, speech planning, and articulatory planning of orofacial movements. Finally, the effective connectivity analysis provided a network that best explained our data, starting at the AG and going to the pMTG, from which there was a division between the ventral and dorsal streams. The SFG and MFG regions were connected and modulated by the MidFG, while the inferior regions formed the ventral stream. Therefore, we successfully assessed the SVF network, exploring regions associated with the entire processing, from category identification to word generation. The methodological approach can be helpful for further investigation of the SVF network in neurological disorders.

Approaches to Measuring Language Lateralisation: An Exploratory Study Comparing Two fMRI Methods and Functional Transcranial Doppler Ultrasound.

In this exploratory study we compare and contrast two methods for deriving a laterality index (LI) from functional magnetic resonance imaging (fMRI) data: the weighted bootstrapped mean from the LI Toolbox (toolbox method), and a novel method that uses subtraction of activations from homologous regions in left and right hemispheres to give an array of difference scores (mirror method). Data came from 31 individuals who had been selected to include a high proportion of people with atypical laterality when tested with functional transcranial Doppler ultrasound (fTCD). On two tasks, word generation and semantic matching, the mirror method generally gave better agreement with fTCD laterality than the toolbox method, both for individual regions of interest, and for a large region corresponding to the middle cerebral artery. LI estimates from this method had much smaller confidence intervals (CIs) than those from the toolbox method; with the mirror method, most participants were reliably lateralised to left or right, whereas with the toolbox method, a higher proportion were categorised as bilateral (i.e., the CI for the LI spanned zero). Reasons for discrepancies between fMRI methods are discussed: one issue is that the toolbox method averages the LI across a wide range of thresholds. Furthermore, examination of task-related t-statistic maps from the two hemispheres showed that language lateralisation is evident in regions characterised by deactivation, and so key information may be lost by ignoring voxel activations below zero, as is done with conventional estimates of the LI.

Comparative analysis of brain language templates with primary language areas detected from presurgical fMRI of brain tumor patients.

Functional brain templates are often used in the analysis of clinical functional MRI (fMRI) studies. However, these templates are mostly built based on anatomy or fMRI of healthy subjects, which have not been fully vetted in clinical cohorts. Our aim was to evaluate language templates by comparing with primary language areas (PLAs) detected from presurgical fMRI of brain tumor patients. Four language templates (A-D) based on anatomy, task-based fMRI, resting-state fMRI, and meta-analysis, respectively, were compared with PLAs detected by fMRI with word generation and sentence completion paradigms. For each template, the fraction of PLA activations enclosed by the template (positive inclusion fraction, [PIF]), the fraction of activations within the template but that did not belong to PLAs (false inclusion fraction, [FIF]), and their Dice similarity coefficient (DSC) with PLA activations were calculated. For anterior PLAs, Template A had the greatest PIF (median, 0.95), whereas Template D had both the lowest FIF (median, 0.074), and the highest DSC (median, 0.30), which were all significant compared to other templates. For posterior PLAs, Templates B and D had similar PIF (median, 0.91 and 0.90, respectively) and DSC (both medians, 0.059), which were all significantly higher than that of Template C. Templates B and C had significantly lower FIF (median, 0.061 and 0.054, respectively) compared to Template D. This study demonstrated significant differences between language templates in their inclusiveness of and spatial agreement with the PLAs detected in the presurgical fMRI of the patient cohort. These findings may help guide the selection of language templates tailored to their applications in clinical fMRI studies.

Detecting language network alterations in mild cognitive impairment using task-based fMRI and resting-state fMRI: A comparative study.

The objective of this study was to investigate the functional changes associated with mild cognitive impairment (MCI) using independent component analysis (ICA) with the word generation task functional magnetic resonance imaging (fMRI) and resting-state fMRI. In this study 17 patients with MCI and age and education-matched 17 healthy individuals as control group are investigated. All participants underwent resting-state fMRI and task-based fMRI while performing the word generation task. ICA was used to identify the appropriate independent components (ICs) and their associated networks. The Dice Coefficient method was used to determine the relevance of the ICs to the networks of interest. IC-14 was found relevant to language network in both resting-state and task-based fMRI, IC-4 to visual, and IC-28 to dorsal attention network (DAN) in word generation task-based fMRI by Sorento-Dice Coefficient. ICA showed increased activation in language network, which had a larger voxel size in resting-state functional MRI than word generation task-based fMRI in the bilateral lingual gyrus. Right temporo-occipital fusiform cortex, right hippocampus, and right thalamus were also activated in the task-based fMRI. Decreased activation was found in DAN and visual network MCI patients in word generation task-based fMRI. Task-based fMRI and ICA are more sophisticated and reliable tools in evaluation cognitive impairments in language processing. Our findings support the neural mechanisms of the cognitive impairments in MCI.

Canadian Practice and Recommendations on Functional MRI to Lateralize Language in Epilepsy.

Identifying a patient's dominant language hemisphere is an important evaluation performed prior to epilepsy surgery and is commonly assessed using functional magnetic resonance imaging (fMRI). However, the lack of standardization and resultant heterogeneity of fMRI paradigms used in clinical practice limits the ability of cross-center comparisons to be made regarding language laterality results. Through surveying Canadian Epilepsy Centres in combination with reviewing supporting literature, current fMRI language lateralization practices for the clinical evaluation of patients with epilepsy were assessed. To encourage standardization of this practice, we outlined a two-part paradigm series that demonstrates widespread acceptance, reliability and accessibility in lateralizing various aspects of language functioning in individuals with average or near-average IQ and normal literacy skills. The collected data confirm a lack of standardization in fMRI laterality assessments leading to clinical heterogeneity in stimulation and control tasks, paradigm design and timing, laterality index calculations, thresholding values and analysis software and technique. We suggest a Sentence Completion (SC) and Word Generation (WG) paradigm series as it was most commonly employed across Canada, demonstrated reliability in lateralizing both receptive and expressive language areas in supporting literature, and could be readily intelligible to an inclusive population. Through providing recommendations for a two-part paradigm series, we hope to contribute to the standardization of this practice across Canada to reduce clinical heterogeneity, encourage communicability between institutions, and enhance methodologies for the surgical treatment of epilepsy for the benefit of all individuals living with epilepsy in Canada.

Noise-Aware Image Captioning with Progressively Exploring Mismatched Words

Image captioning aims to automatically generate captions for images by learning a cross-modal generator from vision to language. The large amount of image-text pairs required for training is usually sourced from the internet due to the manual cost, which brings the noise with mismatched relevance that affects the learning process. Unlike traditional noisy label learning, the key challenge in processing noisy image-text pairs is to finely identify the mismatched words to make the most use of trustworthy information in the text, rather than coarsely weighing the entire examples. To tackle this challenge, we propose a Noise-aware Image Captioning method (NIC) to adaptively mitigate the erroneous guidance from noise by progressively exploring mismatched words. Specifically, NIC first identifies mismatched words by quantifying word-label reliability from two aspects: 1) inter-modal representativeness, which measures the significance of the current word by assessing cross-modal correlation via prediction certainty; 2) intra-modal informativeness, which amplifies the effect of current prediction by combining the quality of subsequent word generation. During optimization, NIC constructs the pseudo-word-labels considering the reliability of the origin word-labels and model convergence to periodically coordinate mismatched words. As a result, NIC can effectively exploit both clean and noisy image-text pairs to learn a more robust mapping function. Extensive experiments conducted on the MS-COCO and Conceptual Caption datasets validate the effectiveness of our method in various noisy scenarios.

EF-Net: Mental State Recognition by Analyzing Multimodal EEG-fNIRS via CNN.

Analysis of brain signals is essential to the study of mental states and various neurological conditions. The two most prevalent noninvasive signals for measuring brain activities are electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS). EEG, characterized by its higher sampling frequency, captures more temporal features, while fNIRS, with a greater number of channels, provides richer spatial information. Although a few previous studies have explored the use of multimodal deep-learning models to analyze brain activity for both EEG and fNIRS, subject-independent training-testing split analysis remains underexplored. The results of the subject-independent setting directly show the model's ability on unseen subjects, which is crucial for real-world applications. In this paper, we introduce EF-Net, a new CNN-based multimodal deep-learning model. We evaluate EF-Net on an EEG-fNIRS word generation (WG) dataset on the mental state recognition task, primarily focusing on the subject-independent setting. For completeness, we report results in the subject-dependent and subject-semidependent settings as well. We compare our model with five baseline approaches, including three traditional machine learning methods and two deep learning methods. EF-Net demonstrates superior performance in both accuracy and F1 score, surpassing these baselines. Our model achieves F1 scores of 99.36%, 98.31%, and 65.05% in the subject-dependent, subject-semidependent, and subject-independent settings, respectively, surpassing the best baseline F1 scores by 1.83%, 4.34%, and 2.13% These results highlight EF-Net's capability to effectively learn and interpret mental states and brain activity across different and unseen subjects.

Supporting social strengths amid emerging bilingualism: effects of Word Generation on social perspective taking in English learners’ writing

ABSTRACT Bilinguals often outperform monolinguals at apprehending the perspectives of others—an apparent consequence of their experiences moving across linguistic and sociocultural contexts. Whether English learners' (ELs') use of such skill in academic writing may be affected by literacy curriculum is the focus of this study. The study explored students' use of social perspective taking (SPT) in writing by examining their responses to a novel argumentative writing task, the iPad–Take A Stand assessment (iTAS). The iTAS was designed to assess the efficacy of the Word Generation curriculum (WG), which provides supports for discussing and deploying multiple perspectives, during its 2012–2014 randomized controlled trial. iTAS essays produced in that study by 4th–7th-grade ELs and their non-EL peers were coded along two dimensions of SPT, acknowledgment and articulation. Analyses revealed a positive impact of WG irrespective of language status on social perspective acknowledgment, the less sophisticated dimension; in contrast, a ‘bilingual boost' of the curriculum was detected in students' use of social perspective articulation, the more sophisticated dimension. These findings suggest ELs may draw on their SPT skills when constructing written arguments, at least in the presence of WG’s sociocognitive and linguistic supports.

Word-of-mouth generated influences of different prepared dishes via online consumer purchases: Preliminary text-based research findings from "Jingdong Mall" flagship shops.

With the rapid development of China's prepared vegetable product market, consumer demand for prepared vegetable products is increasing. The study adopts a qualitative research method to construct a model of factors influencing the generation of consumer word-of-mouth (IWOM) in the online consumption context, taking the real consumer word-of-mouth of Jingdong's own flagship shop of prepared dishes as the object of the study. The model states that the objective factors that promote the generation of word of mouth include specific consumption context (emotionality, initiality, scarcity, convenience and process) and specific product attributes (richness, safety, accessibility and emotionality). Meanwhile subjective factors design consumer satisfaction with the product (satisfaction with a single attribute and satisfaction with the brand as a whole) and emotions (positive self-conscious emotions, high arousal positive emotions and medium arousal positive emotions). Objective factors may contribute directly to the generation of word-of-mouth, or they may further contribute to word-of-mouth generation through subjective consumer factors. In addition, IWOM is often generated not by a single factor, but by a combination of factors.

Computational strategy for aphasia support

Aphasia, a language disorder resulting from brain lesions, impacts communication abilities, and is commonly diagnosed through formal tests assessing speech fluency, comprehension, naming, and repetition abilities. These tests aid in devising treatment plans and evaluating recovery potential. Technological advances have enabled the development of devices and techniques for recording brain activity. Despite their potential, the high cost, device robustness, and complexity often limit their application in rehabilitation centers outside hospitals. In this study, we propose a computational strategy that uses electroencephalography devices to support aphasia rehabilitation therapies. The research was conducted with aphasic participants undergoing rehabilitation at the Center for Prevention and Rehabilitation of People with Disabilities within the Unified Health System in Bahia-Brazil. Preliminary results obtained from the silent word generation task, as outlined in the adult language paradigm by the American Society for Functional Neuroradiology, suggest that increased electrical activation in the right hemisphere may indicate language migration to the non-dominant hemisphere, potentially leading to improved recovery.

Adaptive Semantic-Enhanced Transformer for Image Captioning.

In the research on image captioning, rich semantic information is very important for generating critical caption words as guiding information. However, semantic information from offline object detectors involves many semantic objects that do not appear in the caption, thereby bringing noise into the decoding process. To produce more accurate semantic guiding information and further optimize the decoding process, we propose an end-to-end adaptive semantic-enhanced transformer (AS-Transformer) model for image captioning. For semantic enhancement information extraction, we propose a constrained weaklysupervised learning (CWSL) module, which reconstructs the semantic object's probability distribution detected by the multiple instances learning (MIL) through a joint loss function. These strengthened semantic objects from the reconstructed probability distribution can better depict the semantic meaning of images. Also, for semantic enhancement decoding, we propose an adaptive gated mechanism (AGM) module to adjust the attention between visual and semantic information adaptively for the more accurate generation of caption words. Through the joint control of the CWSL module and AGM module, our proposed model constructs a complete adaptive enhancement mechanism from encoding to decoding and obtains visual context that is more suitable for captions. Experiments on the public Microsoft Common Objects in COntext (MSCOCO) and Flickr30K datasets illustrate that our proposed AS-Transformer can adaptively obtain effective semantic information and adjust the attention weights between semantic and visual information automatically, which achieves more accurate captions compared with semantic enhancement methods and outperforms state-of-the-art methods.

Patterns of performance on the animal fluency task in logopenic variant of primary progressive aphasia: A reflection of phonological and semantic skills

PurposeThis study aimed to characterize the quantitative (total number of correct words generated) and qualitative (psycholinguistic properties of correct words generated) performance patterns on the animal fluency task in individuals with the logopenic variant of primary progressive aphasia and to investigate the influence of phonological and semantic abilities to these patterns. MethodsFifteen participants with lvPPA and twenty neurotypical adults completed the animal fluency task and an assessment battery to characterize their phonological and semantic abilities. We recorded the total number of correct words produced and their psycholinguistic properties. Group differences were analyzed using independent samples t-tests and analysis of covariance. Stepwise and multiple linear regression analyses were implemented to investigate the contribution of psycholinguistic properties on word generation as well as the role of phonological and semantic abilities on performance. We also investigated the mediating role of phonological and semantic abilities on the relationship between relevant psycholinguistic properties and word generation output. ResultsCompared to neurotypical controls, participants with lvPPA produced fewer correct responses and more words with lower age of acquisition. The total number of correct words generated was predicted by the age of word acquisition, such that individuals who generated more responses, produced words acquired later in life. Phonology and semantics influenced the number of correct words generated and their frequency, age of acquisition, and semantic neighborhood density. Familiarity and arousal were driven by semantic abilities. Phonological abilities partially mediated the relationship between age of acquisition and word generation output. ConclusionsThis study provides valuable insights into the performance patterns of the animal fluency task in lvPPA. Individuals with lvPPA with more intact phonological and semantic abilities generated greater number of words with more complex psycholinguistic properties. Our findings contribute to the understanding of language processes underlying word retrieval in lvPPA.