Proceedings of The Physiological Society

Sleep Sleep and Circadian Rhythms (London, UK) (2018) Proc Physiol Soc 42, C27

Poster Communications

GRIA1 knockout mice show reduced global EEG sleep spindles, preserved local LFP spindles an retain long-term memory

C. Blanco Duque1, R. J. Purple2, T. Yamagata1, L. Krone1, L. McKillop1, D. M. Bannerman1, V. Vyazovskiy1

1. University of Oxford, Oxford, United Kingdom. 2. University of Bristol, Bristol, United Kingdom.

Sleep spindles have been implicated in cognitive functions and memory consolidation1,2. Deficits in spindles have been reported in brain disorders (e.g. schizophrenia) associated with polymorphisms of the GRIA1 gene3, which codes for the GluA1 AMPA receptor subunit. Here we investigated the dynamics of sleep spindles and their relationship with memory performance in GRIA1-/- and wild type (WT) mice. Chronic electroencephalogram (EEG) and the electromyogram (EMG) were recorded during spontaneous sleep in n=14 mice. Multichannel recordings of local field potentials (LFP) were also collected in a subset of mice from layer-V somatosensory cortex (SCx). Surgical procedures were performed under isoflurane anaesthesia. EEG and LFP power spectra were calculated with a Fast Fourier Transform using 4-second epochs. For individual spindle event detection, an automated algorithm based on autoregressive modelling was applied to the LFP and EEG signals. Spatial reference memory was assessed in an additional group of mice using a plus maze task. Frontal EEG spectral power during NREM sleep was significantly reduced in the spindle-frequency range (10-15 Hz) in GRIA1-/- relative to WT mice. Furthermore, individual EEG spindle events were readily detected in WT mice with the automated algorithm, while they were absent in GRIA1-/- mice. Interestingly, despite the absence of EEG spindles in GRIA1-/- mice, preliminary analyses of LFP signals revealed an occurrence of local spindle events in the SCx in both genotypes. A repeated measures analysis revealed no significant differences between GRIA1-/- and WT in memory performance (main effect of genotype and interaction by day; F< 1;p>0.20). This is consistent with previous evidence indicating that long-term memory formation is preserved in GRIA1-/- mice. The deletion of GluA1 in mice is associated with a profound reduction of EEG sleep spindling activity; yet local cortical sleep spindles may be preserved. Global EEG spindles do not seem necessary for memory consolidation, although a role for local LFP spindles cannot be excluded. These results suggest an important role of the GRIA1 gene in mediating the link between sleep and cognitive function.

Where applicable, experiments conform with Society ethical requirements