Impaired consolidation of spatial memory during sleep in patients with leucine-rich glioma-inactivated 1-associated limbic encephalitis.

Sleep promotes the systems consolidation of hippocampus (HC)-dependent spatial memories by reprocessing of previously encoded hippocampal representations. Hippocampal reprocessing involves pattern separation and pattern completion as central hippocampal functions performed by the dentate gyrus (DG) and cornu ammonis region 3 (CA3), respectively. The leucine-rich, glioma inactivated 1 (LGI1)-associated limbic encephalitis (LE) is an autoimmune brain disorder particularly affecting the DG and CA3 regions, thereby impairing hippocampal function. We studied 15 LGI1 patients (and matched healthy controls) to examine hippocampal contributions to the sleep-associated consolidation of spatial memory. Spatial memory was assessed using the virtual Morris water maze (VWM) during learning before nocturnal sleep. Spatial retrieval of target locations (as indicated by dwell time in target area) was tested in the next morning, with separate trials testing pattern separation and pattern completion functions, as well as place memory precision and reversal learning capabilities. Leucine-rich, glioma inactivated 1-associated limbic encephalitis (LGI1-LE) patients were able to learn and retrieve spatial locations, albeit to a lesser extent than controls. Recall of place memories was decreased in LGI1-LE patients in comparison with learning performance before sleep and with healthy controls, especially in trials assessing pattern separation. Moreover, at recall, LGI1 patients showed a less flexible adaptation to the reversal learning task, in comparison with the controls. Sleep quality, macro-sleep architecture and EEG slow oscillations (SOs) and spindles were comparable in both groups. However, in LGI1-LE patients, phase-amplitude coupling of SO-spindle events appeared diminished although the group difference did not remain significant after correction for multiple comparisons. In addition, a negative correlation between spindle density and retrieval of target locations was observed. Magnetic resonance imaging confirmed smaller volumes of the HC and its subfields (subiculum, CA1, CA3, DG) in the patients. Divergent structure-function relationships emerged between patients with LGI1-associated encephalitis and healthy controls: In patients, larger volumes of DG and CA3 were associated with weaker sleep-dependent consolidation but greater stability under cue deprivation. In controls, larger hippocampal, CA1, and subicular volumes correlated with better memory retrieval and reversal learning performance. Our results show an impaired sleep-associated consolidation of spatial memory in LGI1-LE patients highlighting the involvement of DG and CA3 areas in sleep-associated spatial memory formation and cognitive flexibility.