Social ties linked to better memory and thinking after 90
An observational study of 677 adults aged 90 and older linked social ties to better cognitive performance, but not decline or death over 1.63 years.
Our read
Populations- Studied:
- Cohort study of 677 adults aged 90 years and older, average 1.63 years
- What changed:
- Memory scores were higher among older adults with larger social networks
- The catch:
- Better thinking skills can help preserve friendships, rather than friendships improving thinking
- In the record:
- One of 9 papers on social network characteristics on record: 3 with results in people.
Editor of record: Boris Djordjevic · AgingResearch.News
The 677 adults aged 90 and older in the LifeAfter90 observational cohort lived in Northern California. Their average age was 92.9, and 61.2% were women. None had a recorded dementia diagnosis at enrollment. Participants completed thinking and memory tests every six months from 2019 to 2022.
Researchers assessed the size of participants’ close social networks, the share of friends versus family, and perceived emotional support. Larger networks, a greater share of friends and stronger support were associated with better executive function—the skills used to manage information and tasks—and memory for words. Findings were generally consistent in men and women. None of these social measures predicted cognitive decline or death during an average 1.63 years of follow-up. During that period, 163 participants died.
Why it matters
Social networks often change in advanced old age, making it important to distinguish links with current thinking skills from possible links to slower cognitive aging.
Caveats
Social ties and initial cognitive performance were assessed together, so better thinking abilities may have helped people maintain relationships. Follow-up averaged only 1.63 years, and enrollment excluded people with diagnosed dementia or other neurodegenerative diseases.
From the paper
- Participants
Social network information was first collected during the second wave, which included 757 participants and served as the baseline for this analysis.
- Follow-up
During an average follow‐up period of 1.63 years (SD = 1.18, maximum: 3.72), 163 (24.1%) participants died, and 45 (6.6%) participants were lost to follow‐up for other reasons.
- Limitation
As with many other studies on late‐life risk factors for cognitive health, we cannot rule out the possibility of reverse causation between social network characteristics and cognitive function.
Quoted word for word from the full text on PubMed Central.
The paper
University of California, Los Angeles; Boston University
Alzheimer's & Dementia, Jul 2026, Open access



Comments
Be the first to comment.