Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Does microstructure of white matter association tracts in the temporal lobe predict neural activity and behaviour? An EEG-DWI study of picture naming
Poster Session A, Wednesday, September 30, 11:00 am - 1:00 pm, Wangari Maathai
Irina Chupina1, Natascha M. Roos1, Joanna Sierpowska2, Robert Oostenveld1,3, Anil M. Tuladhar4,1, Britta U. Westner4,1, Vitória Piai1; 1Donders Institute, Radboud University, Nijmegen, the Netherlands, 2The Institute of Neurosciences, University of Barcelona, Spain, 3Karolinska Institute, Stockholm, Sweden, 4Radboud University Medical Center, Nijmegen, the Netherlands
To date, multiple studies have established associations between picture naming performance and the physiological properties of white matter (WM) tracts. However, the predominant focus on accuracy measures in prior literature has left the role of WM in the timing of word retrieval relatively understudied. In addition, while electrophysiological studies of visual processing have demonstrated that WM properties can predict neural activity underlying behavioural responses, such investigations are lacking in the language field. To address these gaps, we investigated the structure-activity-behaviour relationship in picture naming. To this end, we performed electroencephalography (EEG) and diffusion-weighted imaging (DWI) in stroke-free individuals and individuals after left- and right-hemispheric stroke (total N = 31). We probabilistically reconstructed the bilateral ventral association tracts traversing the temporal lobe, namely inferior longitudinal fasciculus (ILF), inferior fronto-occipital fasciculus (IFOF), uncinate fasciculus (UF), and middle longitudinal fasciculus (MdLF). These tracts have been linked to visual and lexical-semantic processes. We also reconstructed the motor corticospinal tract (CST) as a control tract. Using diffusion tensor imaging, we obtained fractional anisotropy (FA) values for each tract. We then correlated these FA values with mean response times (RTs) and the coefficient of variation (CV) of RTs, capturing the average picture naming speed and naming speed variability, respectively. In addition, we correlated FA values with the onset latency of the electrophysiological P2 event-related component time locked to picture onset. The P2 shows an occipito-parietal scalp topography and has been previously linked to lexical access in picture naming. Based on its anatomy and function, we hypothesised that the P2 latency might be linked to the microstructure of ventral WM tracts. Finally, following the correlation results, we selected tracts that showed associations with both P2 latency and naming performance and conducted structural equation modelling (path) analyses to test the full structure-activity-behaviour relationship. All analyses were controlled for age and presence of stroke. Correlation analyses showed that higher FA in the bilateral IFOF and right UF was associated with faster naming, whereas higher FA in the right CST was linked to more consistent naming speed. FA in the bilateral ILF and MdLF was related to both mean RTs and variability. These findings suggest that ventral tracts may differentially contribute to overall speed (mean RT) and processing consistency (CV). Furthermore, higher FA in the right IFOF and left MdLF was also linked to earlier P2 onset. Path models for these two tracts revealed that right IFOF microstructure predicted the P2 latency which, in turn, predicted average naming speed. Some of the variance behind the WM and naming speed relationship was explained by factors beyond P2 latency. This right IFOF pathway likely reflects the shared involvement of all three observation levels in lexical-semantic processing. Overall, this study provides a proof of concept for modelling structure-activity-behaviour relationships in language research, while highlighting the need for further physiology- and neuroanatomy-driven elaboration of such models.
Topic Areas: Language Production,