Large‐Scale Network Dysfunction in Dementia with Lewy Bodies: Preliminary Results from an fMRI Connectomic Mapping Study

Background The diagnosis of dementia with Lewy bodies (DLB) is often delayed in clinical practice. The sensitivity and specificity of structural MRI is suboptimal: a multisite study with autopsy confirmation demonstrated 64% sensitivity and 68% specificity for differentiating DLB from Alzheimer's Disease (AD; Harper et.al, 2016). There is an urgent need to develop additional reliable biomarkers for DLB. We describe resting state functional MRI (rsfMRI) results using an automated connectome mapping program (Infinitome). Method Participants meeting criteria for probable DLB (McKeith et al, 2017) aged between 40 ‐90 years with a MMSE >15 were recruited from a multidisciplinary dementia clinic in this cross‐sectional study. MRI imaging was performed using a 3T Siemens scanner. DTI (diffusion tensor imaging), rsfMRI and structural images were gathered and analyzed through Infinitome (Rosenbloom et al, 2021), a program that creates a machine learning‐based, subject‐specific version of the Human Connectome Project‐Multimodal Parcellation atlas based upon diffusion tractography structural connectivity. Result Twenty‐six participants were contacted, of which 14 (54%) were consented and one withdrawn by the investigator. Of the 13 subjects that completed the study, 77% were male, 100% white and non‐Hispanic origin, with mean age 73.29 years (± 5.8), and average MMSE score 23.14 (±4.3). All participants demonstrated abnormal connectivity in regions involving visuospatial attention (areas 8Av, PGs, TE2a) and language processing (areas 55b, TGd, TGv; Table 1; Baker et al, 2018). Five parcellations were part of the extended default mode network (DMN) whereas one was part of the extended central executive network. Seven participants demonstrated changes in connectivity of area 46 (R & L; cognitive processing and mood) within the salience network (SN). Conclusion Preliminary findings suggest reduced connectivity in regions involved in visuospatial attention, language processing, and mood within in large‐scale networks such as the DMN and SN, consistent with previously published fMRI studies in DLB (Nicastro 2021; Schumacher 2018). The use of rsfMRI and connectome mapping may enable identification of a distinct radiological “signature” for DLB. Future research will evaluate rsfMRI features that distinguish DLB from AD.