guest ::
| Home > Publications database > Regional activation patterns underlyingsemantic verbal fluency strategies |
| Home > Publications database > Regional activation patterns underlyingsemantic verbal fluency strategies |
| Poster (After Call) | FZJ-2026-04136 |
; ; ;
2026
This record in other databases:
Please use a persistent id in citations: doi:10.34734/FZJ-2026-04136
Abstract: Regional activation patterns underlying semantic verbal fluency strategies - An event-related fMRI study of productive speechGianna Kuhles1,2, Julia A. Camilleri1,2, Simon B. Eickhoff1,2, & Susanne Weis1,21 Institute of Systems Neuroscience, Heinrich Heine University Düsseldorf, Düsseldorf, Germany2 Institute of Neuroscience and Medicine (INM-7: Brain and Behaviour), Research Centre Jülich, Jülich, GermanyIntroductionEfficient retrieval of words related to a specific semantic category relies on dynamic navigation through semantic space via two complementary strategies: clustering, reflecting local exploitation within subcategories, and switching, reflecting exploratory transitions between subcategories [1]. Behavioural studies suggest that these verbal fluency strategies rely on partially dissociable cognitive functions, including lexical processing and executive control [2]. Systematic differences in clustering and switching have been associated with age, sex, and neuropsychiatric conditions, indicating that semantic search strategies constitute a sensitive behavioural phenotype [3, 4, 5]. Despite their clinical relevance, the neural implementation of clustering and switching during overt speech production remains poorly characterised. Critically, most prior neuroimaging studies have relied on covert speech or highly constrained lexical retrieval tasks, limiting ecological validity, due to technical challenges in acquiring and analysing speech in the MRI environment. Consequently, it remains unknown how clustering and switching during natural speech production are implemented at the level of regional brain activation and whether distinct cortical and subcortical systems preferentially support these strategy components.Here, we address this gap by leveraging an overt semantic verbal fluency paradigm to investigate the neural correlates of clustering and switching. This work is embedded within SpExNeuro – A behavioural and brain imaging data collection of speech and executive functions, designed to link productive speech with executive control and neurobiological mechanisms.MethodsThirty neurologically healthy, German-speaking adults from an ongoing cohort were examined (n = 30; 18 f, 12 m; mean age = 31,80 years, SD = 8,6). Participants performed an overt semantic verbal fluency task during fMRI, naming as many animals as possible over 120 s of continuous speech production.Functional MRI data were acquired on a 3 T Siemens Prisma scanner with a 64-channel head coil (64 slices; 2.2 × 2.2 × 2.0 mm³; TR = 980 ms; TE = 30 ms; flip angle = 70°; multiband factor = 4). Data were converted to BIDS format and preprocessed using fMRIPrep, including motion correction, susceptibility distortion correction, spatial normalisation, and nuisance regression. Audio recordings were obtained using the FOMRI-III+ MRI-compatible microphone, denoised using the deep learning–based source separation Ultimate Vocal Remover. This processingpipeline yielded sufficient signal quality to allow reliable word-level transcription and temporal alignment with the fMRI time series. Transcription was performed using the large-v3 faster-whisper model and manually quality-controlled. Each spoken word was time-stamped and classified as either part of a cluster or a switch according to established criteria [1]. Clusters were defined as sequences of at least two consecutive correct words belonging to the same semantic subcategory (e.g., eagle, duck... = birds), whereas switches were defined as transitions between semantic subcategories, including shifts between clusters and single unclustered words (e.g., eagle, duck, tuna, salmon, cod... = birds → fish). Voxel-wise first-level event-related GLMs were implemented in FSL FEAT to estimate task-evoked regional BOLD responses associated with verbal fluency strategies. Event regressors were convolved with a double-gamma haemodynamic response function, optimised for transient event-related designs. Motion parameters were included as nuisance regressors. Second-level mixed-effects group analyses were performed using FLAME 1 to assess population-level effects.ResultsDuring the verbal fluency task, participants generated an average of 42 words over 120 s, organising responses into 11 clusters on average with a mean cluster size of 3 words, and a mean number of switches of 19, indicating robust engagement of both clustering and switching strategies during overt task performance.Although overt speech production was accompanied by task-related head motion, the inclusion of motion parameters as nuisance regressors ensured that movement-related variance was modelled separately, allowing the reported activation patterns to reflect strategy-specific neural responses rather than residual motion effects.At the first level, the GLMs successfully captured trial-by-trial BOLD modulation associated with individual word onsets classified as clustering or switching events at the single-subject level, demonstrating robust model fit for overt speech production in verbal fluency. Group-level whole-brain analyses using voxel-wise FWE–corrected inference (p < .05) revealed significantly greater activation for clustering compared to switching, among others in frontal regions as well as the central and cingulate gyri, indicating neural processes specifically associated with semantic clustering. Conversely, switching relative to clustering elicited significantly increased activation, among others in the frontal pole, insular cortex, and opercular cortex, reflecting neural mechanisms uniquely related to switching processes.Fig 1. Semantic Verbal Fluency Task Outcomes.Fig 2. Activation Patterns underlying Semantic Verbal Fluency Strategies.Group-level voxel-wise activation maps illustrate brain regions associated with clustering (blue) and switching (red) events during overt semantic verbal fluency.DiscussionMethodologically, this study demonstrates that overt verbal fluency can be analysed at the level of individual strategy events using fMRI, provided that high-quality audio acquisition is available.The findings suggest evidence at the level of regional task-evoked BOLD activity that clustering and switching during overt semantic fluency rely on partially dissociable neural substrates. Clustering within semantic categories is typically associated with more automatic retrieval processes supported by posterior parietal and sensorimotor regions, whereas switching between clusters places higher demands on executive control and engages broader fronto-parietal networks, particularly in the left hemisphere. Consistent with this framework, the present results revealed greater activation for clustering in frontal regions as well as central and cingulate gyri, suggesting engagement of networks supporting efficient retrieval within already activated semantic fields. In contrast, switching recruited a more extended frontal network, including the frontal pole, insular cortex, and opercular regions. This pattern is consistent with previous findings linking switching in verbal fluency tasks to increased demands on cognitive control, set-shifting, and controlled retrieval processes.Importantly, verbal fluency tasks are among the most widely used tools in clinical neuropsychological assessment for the screening of executive dysfunction, semantic memory impairment, and (early) cognitive decline. The present findings indicate that specific clinical impairments in verbal fluency may be traceable to distinct underlying neural mechanisms, rather than reflecting a global deficit. Thus, the present approach holds promise for improving the neurofunctional specificity of verbal fluency as a clinical marker, particularly in conditions characterised by early executive or semantic decline. The present findings represent results from an ongoing data collection and will be validated in larger samples. Future work will extend this framework to additional semantic categories (e.g., jobs, emotions) and phonemic fluency conditions to test the generalisability and domain specificity of the observed strategy-dependent activation patterns.Data Resource StatementSpExNeuro will be released as an open multimodal dataset of speech and executive function integrating a comprehensive range of variables across behavioral, linguistic, neuroimaging, and neuroendocrine domains to facilitate investigations of individual differences in speech–cognition–brain relationships.References1.Troyer, A. K., & Moscovitch, M. (2006). Cognitive processes of verbal fluency tasks. In A. M. Poreh (Ed.), The quantified process approach to neuropsychological assessment (pp. 143–160). Psychology Press.2.Haugrud, N., Crossley, M., & Vrbancic, M. (2011). Clustering and switching strategies during verbal fluency performance differentiate Alzheimer’s disease and healthy aging. Journal of the International Neuropsychological Society, 17(6), 1153–1157. https://doi.org/10.1017/S13556177110011963.Lanting, S., Haugrud, N., & Crossley, M. (2009). The effect of age and sex on clustering and switching during speeded verbal fluency tasks. Journal of the International Neuropsychological Society, 15(2), 196–204. https://doi.org/10.1017/S13556177090902374.Weiss, E. M., Ragland, J. D., Brensinger, C. M., Bilker, W. B., Deisenhammer, E. A., & Delazer, M. (2006).Sex differences in clustering and switching in verbal fluency. Journal of the International Neuropsychological Society, 12(4), 502–509. https://doi.org/10.1017/S13556177060606565.Fossati, P., Guillaume, L. B., Ergis, A. M., & Allilaire, J. F. (2003). Qualitative analysis of verbal fluency in depression. Psychiatry Research, 117(1), 17–24. https://doi.org/10.1016/S0165-1781(02)00300-1
|
The record appears in these collections: |
PDF