Depression in older adults with cognitive impairment increases progression to dementia. Microbiota is associated with current mood and cognition, but the extent to which it predicts future symptoms is unknown. In this work, we identified microbial features that reflect current and predict future cognitive and depressive symptoms. Clinical assessments and stool samples were collected from 268 participants with varying cognitive and depressive symptoms. Seventy participants underwent 2-year follow-up. Microbial community diversity, structure, and composition were assessed using high-resolution 16 S rRNA marker gene sequencing. We implemented linear regression to characterize the relationship between microbiome composition, current cognitive impairment, and depressive symptoms. We leveraged elastic net regression to discover features that reflect current or future cognitive function and depressive symptoms. Greater microbial community diversity associated with lower current cognition in the whole sample, and greater depression in participants not on antidepressants. Poor current cognitive function associated with lower relative abundance of Bifidobacterium, while greater GABA degradation associated with greater current depression severity. Future cognitive decline associated with lower cognitive function, lower relative abundance of Intestinibacter, lower glutamate degradation, and higher baseline histamine synthesis. Future increase in depressive symptoms associated with higher baseline depression and anxiety, lower cognitive function, diabetes, lower relative abundance of Bacteroidota, and lower glutamate degradation. Our results suggest cognitive dysfunction and depression are unique states with an overall biological effect detectable through gut microbiota. The microbiome may present a noninvasive readout and prognostic tool for cognitive and psychiatric states.

患有认知障碍的老年人的抑郁症会加速痴呆的进展。微生物群与当前情绪和认知相关，但其预测未来症状的程度尚不清楚。在这项工作中，我们确定了反映当前并预测未来认知和抑郁症状的微生物特征。从 268 名具有不同认知和抑郁症状的参与者中收集了临床评估和粪便样本。七十名参与者接受了两年的随访。使用高分辨率 16 S rRNA 标记基因测序评估微生物群落多样性、结构和组成。我们采用线性回归来表征微生物组组成、当前认知障碍和抑郁症状之间的关系。我们利用弹性网络回归来发现反映当前或未来认知功能和抑郁症状的特征。微生物群落多样性越高，整个样本中当前认知能力越低，并且未服用抗抑郁药物的参与者抑郁程度越高。当前认知功能较差与双歧杆菌相对丰度较低相关，而 GABA 降解程度较高与当前抑郁症严重程度相关。未来的认知能力下降与较低的认知功能、较低的肠杆菌相对丰度、较低的谷氨酸降解和较高的基线组胺合成有关。未来抑郁症状的增加与较高的基线抑郁和焦虑、较低的认知功能、糖尿病、较低的拟杆菌相对丰度和较低的谷氨酸降解有关。我们的结果表明认知功能障碍和抑郁症是独特的状态，其总体生物效应可通过肠道微生物群检测到。微生物组可能为认知和精神状态提供无创读数和预后工具。