Frequent Night-Time Micro-Awakenings Linked to Genetic Alzheimer's Risk in Middle Age

Brief night-time awakenings may point to a genetic vulnerability to Alzheimer's disease years before any symptoms appear, University of Liège researchers in Belgium report in the journal Sleep. The team at the university's GIGA Neurosciences unit analysed the sleep of more than 500 healthy people: among participants aged 50 to 69, more frequent nocturnal micro-awakenings went with a higher genetic risk of Alzheimer's — a link absent in young adults aged 18 to 31.

The team calculated each participant's polygenic risk — a single figure summarising the combined influence of genes on disease risk — and compared it with features of their sleep. The clearest link involved micro-awakenings: short bursts of brain activity that interrupt the sleep cycle without fully waking a person. "These micro-awakenings are therefore not insignificant," said Puneet Talwar, a researcher at the GIGA ULiège laboratory; certain sleep profiles, he said, could promote the accumulation of proteins involved in Alzheimer's disease.

The work also points to the locus coeruleus, a brainstem region about the size of a grain of rice that helps regulate wakefulness, attention and sleep. It is among the first places where abnormal protein deposits appear, sometimes beginning in adolescence. In a preliminary 2025 study using a 7-Tesla MRI scanner, an especially powerful brain imager, the Liège team found that sleep quality was related to the condition of this region from a young age, and that its proper functioning depended on the quality of REM (rapid eye movement) sleep, a stage important for memory.

"Sleep could become an accessible marker for the early identification of vulnerable individuals," said Gilles Vandewalle, co-director of the GIGA CRC In Vivo Imaging platform. The researchers are also examining whether improving sleep could help slow the disease in genetically predisposed people. Alzheimer's affects more than 220,000 people in Belgium; the team cautions the results are statistical associations needing confirmation and cannot predict who will develop the disease.