Functional alterations in visual networks underlie differences in the phenomenology of visual hallucinations in Lewy body disease

Abstract

AbstractBackgroundThe phenomenology of visual hallucinations (VH) in Lewy Body Disease (LBD) is heterogenous including minor phenomena (passage, presence, illusions) and complex hallucinations. Underlying mechanisms are largely unknown and here we investigated whether distinct neural networks underpin minor (MVH) and complex VH (CVH).MethodLBD patients (n=17 DLB, n=10 PDD; mean age=76.5±6.5; mean MMSE=20.2±5.7, male=16; 59.2%) with VH underwent resting state fMRI and the North East Visual Hallucinations Interview (NEVHI) to assess the phenomenology, duration and frequency of VH. Using multiple regression models, we performed a seed‐to‐seed functional connectivity (FC) analysis to identify specific nodes associated with i) duration, ii) frequency and iii) severity of MVH (controlling for CVH and MMSE) or CVH (controlling for MVH and MMSE).ResultThe severity, frequency and duration of MVH was negatively associated with FC between early visual areas (EVA) and regions of the ventral visual stream, and positively associated with FC between EVA and regions of the dorsal visual stream that may reflect the phenomenology of passage and presence. Furthermore, the duration of MVH was negatively associated with FC between the brainstem and EVA that may be linked to LBD brainstem neuropathology and positively associated with FC between areas of the dorsal visual stream (Table 1). The severity and duration of CVH was negatively associated with FC between areas of the ventral visual stream and positively associated with FC between lingual gyrus and supramarginal gyrus, within the salience network, that may relate to the complexity of the hallucination content. The frequency of CVH was negatively associated with FC between areas of the default mode network and the salience network (i.e. supramarginal gyrus) that may reflect attentional dysregulation (Table 2).ConclusionFunctional alterations in distinct visual and attentional networks underpin VH in LBD according to their phenomenology, showing a greater involvement of the brainstem and dorsal visual stream for MVH, and of the ventral visual stream and attentional networks for CVH. These findings can be interpreted according to the hodotopic hypothesis of VH (ffytche et al. 2008) and may reflect a link between VH phenomenology and the neuropathological progression of Braak stage (ffytche et al. 2017).