Actions seniors can take to improve brain function

What Actually Improves Memory and Thinking After 60: An Evidence-Ranked Guide

Question to Claude AI:

Nov, 2026 What are the best actions that seniors can take to improve memory and brain function? Possible actions might include reading, socializing, using brain exercise programs like Posit Science, taking supplements, visiting new locations, and learning new skills. The target does not have to be the reduced risk of Alzheimer's or dementia

The best-supported things an older adult can do now are regular aerobic and resistance exercise, a structured lifestyle program like U.S. POINTER that combines exercise, diet, cognitive training and health checks, and adaptive speed-of-processing training of the kind sold as BrainHQ. Fixing problems that drag cognition down also matters: high blood pressure, untreated hearing loss, cataracts, sleep apnea, and sedating or anticholinergic medicines. Most of the activities on your list are pleasant and worth doing, but reading, socializing, travel and most supplements have little or no randomized evidence that they improve memory itself.

TL;DR

Strongest evidence: exercise (aerobic, resistance, multimodal and tai chi-type programs) has the most consistent RCT support for better cognition in older adults. Meta-analyses find small-to-moderate gains at roughly 150+ minutes a week. A structured multidomain program (U.S. POINTER, JAMA 2025) beat a self-guided one by 0.029 SD a year, a small effect. Speed-of-processing training (ACTIVE/BrainHQ) gives large, long-lasting gains in processing speed, but transfer to memory and daily life is modest and still debated.
Moderate or conditional: these work mainly if you have the underlying problem or follow a specific regimen.
- Hearing aids if you have hearing loss (ACHIEVE showed 48% slower decline, but only in the higher-risk subgroup).
- Blood pressure control to a systolic target below 130 (SPRINT MIND and CRHCP reduced MCI and dementia).
- Cataract surgery (observational evidence only).
- CPAP for sleep apnea.
- A daily multivitamin (COSMOS pooled analysis: +0.07 SD global cognition, +0.06 SD episodic memory; small but consistent).
- B vitamins only if homocysteine is high (VITACOG).
- Learning demanding new skills (Synapse Project: memory gains at about 16 hours a week).
Weak, null or unproven: omega-3, vitamin D, cocoa flavanols (except in people with poor diets), magnesium, the MIND diet as a memory booster, reading, socializing alone, travel or novel places, and generic brain games. Several large RCTs were null, and the FTC fined Lumosity $2 million in 2016 over unsupported claims.[1] Creatine is an intriguing exception, but its evidence rests on small trials.[2]

Key Findings: Tiered Ranking

Rank	Action	Evidence strength	Effect on memory/cognition	Practical recommendation
1	Aerobic + resistance exercise	Strong (many RCTs, network meta-analyses)	SMD ~0.4–0.7 global; aerobic best for memory	≥150 min/wk moderate aerobic + 2 strength days
2	Structured multidomain program	Strong (U.S. POINTER, FINGER)	+0.029 SD/yr vs self-guided; mainly executive function	Join a structured, coached program if available
3	Speed-of-processing training (BrainHQ-type)	Moderate–strong for trained skill; contested for transfer	Large speed gains (ES ~0.6–0.8); modest IADL; dementia HR 0.75 with boosters	~10–15 h initial training plus boosters
4	Blood pressure control	Strong for MCI/dementia outcomes	MCI risk ↓19% (SPRINT MIND)	Systolic <130 if tolerated; discuss with doctor
5	Treat hearing loss	Moderate (ACHIEVE, subgroup benefit)	48% slower 3-yr decline in higher-risk group	Get hearing tested; wear fitted aids
6	Learn demanding new skills	Moderate (one good RCT)	Episodic memory gains	Sustained, effortful learning (hours/week)
7	Daily multivitamin	Moderate (3 COSMOS substudies)	Global +0.07 SD; memory +0.06 SD	Reasonable, low-cost option
8	Treat sleep apnea (CPAP)	Moderate–weak	Small gains (attention, speed) if adherent	Screen if snoring/sleepy; use CPAP nightly
9	Correct vision / cataract surgery	Observational	29% lower dementia risk (HR 0.71)	Don't delay indicated surgery
10	Medication review	Strong harm data; weak deprescribing RCTs	Avoids drug-induced impairment	Ask for anticholinergic/sedative review
11	B vitamins (if high homocysteine)	Moderate–weak (one RCT, subgroups)	Slowed atrophy; memory benefit only high tHcy	Test homocysteine first
12	Volunteering / purposeful activity	Weak–moderate	Brain volume gains in men (small trial)	Good for wellbeing; modest cognitive data
13	Creatine	Weak (small RCTs)	Memory SMD 0.88 in older subgroup	Optional; evidence immature
14	Music / language learning	Weak, mixed	Mostly null; some executive gains (piano)	Do it for enjoyment and challenge
15	MIND/Mediterranean diet	Observational strong; RCT null	No 3-yr difference vs control diet	Good for overall health; not a memory booster
16	Omega-3, vitamin D, cocoa, magnesium	Null in large RCTs (magnesium untested)	No overall effect	Not recommended for cognition
17	Reading, socializing, travel	Observational only	No RCT showing memory gains	Enjoy them; don't rely on them alone

Details

1. Exercise: the single best-supported action

Evidence. RCT-based network meta-analyses consistently find that exercise improves cognition in cognitively healthy older adults. - 2026 Bayesian network meta-analysis (38 RCTs, 4,047 community-dwelling older adults): - Gains versus control: aerobic SMD 0.58, resistance SMD 0.62, multimodal SMD 0.68.[3] - Meaningful benefits started at about 600 MET-minutes a week, with the best effects at 700–1,200 MET-minutes.[3] - 2022 meta-analysis (44 studies, 4,793 participants): found no lower threshold for benefit. The minimal clinically relevant dose was 724 MET-minutes a week, and resistance training performed best.[4] - 2025 network meta-analysis in healthy older adults, by domain: - Aerobic exercise was best for memory (SMD 0.42).[5] - Mind-body exercise such as tai chi was best for executive function and task-switching.[6] - Mind-body protocols of about 3 sessions a week, 60 minutes each, for 12 weeks performed well.[5]

Dose. 600–1,200 MET-minutes works out to roughly 150–300 minutes a week of moderate activity such as brisk walking (about 4 METs). Add 2 or more resistance sessions a week. Tai chi or dance count as well and particularly help executive function.

Caveats. Pooled SMDs of 0.6 or more are probably inflated by small, unblinded trials. Real-world effects are likely smaller. Benefits fade if exercise stops. Even so, exercise is the most reliable lever across memory, executive function and global cognition, and it carries large non-cognitive benefits.

2. Structured multidomain programs (U.S. POINTER, FINGER)

Evidence. U.S. POINTER (JAMA, 2025) randomized 2,111 sedentary adults aged 60–79 at elevated risk to one of two programs for 2 years.[7][8] - Structured arm: 38 facilitated peer-team meetings, supervised aerobic and resistance exercise, the MIND diet, BrainHQ cognitive training, social and cognitive challenge, and monthly blood pressure monitoring.[9][10] - Self-guided arm: the same goals with much less support.[7]

Results. - Global cognition rose 0.243 SD a year in the structured arm versus 0.213 SD a year in the self-guided arm. The difference was 0.029 SD a year (95% CI 0.008–0.050; P=.008).[11] - The edge was mainly in executive function. Memory improved in both groups with no significant difference between them.[12] - The benefit was consistent in APOE ε4 carriers and non-carriers, and was larger in people with lower baseline cognition.[13] - U.S. POINTER principal investigator Laura Baker (Wake Forest) said at AAIC 2025 that structured-arm participants "performed at a level comparable to adults 1 to 2 years younger," and the Alzheimer's Association's July 2025 release described the structured arm as "protecting cognition from normal age-related decline for up to two years."

Caveats. There was no untreated control group, so part of the gain in both arms may be practice effects. A JAMA editorial called the clinical relevance of a roughly 14% relative difference "uncertain." It also noted that "the more striking finding is perhaps the similarity of the cognitive benefits across both groups."[14] The practical takeaway: adopting the package matters most, and structure and accountability add a little more.

3. Computerized brain training: Posit Science / BrainHQ

ACTIVE trial (N=2,802, mean age 73.6). Participants did up to 10 sessions of 60–75 minutes over 5–6 weeks, training in memory, reasoning or speed of processing.[15][16][17] Some later received booster sessions. - At 10 years: - Speed training kept a medium-to-large effect on processing speed (ES 0.66).[18] - Reasoning training kept a small effect on reasoning (ES 0.23).[18] - Memory training's effect on memory was not maintained.[19] - All three arms reported less difficulty with instrumental activities of daily living (ES 0.36–0.48).[20] These were self-reported outcomes. - At 20 years (Medicare claims; Alzheimer's & Dementia: TRCI, February 2026): - Speed training with at least one booster was associated with lower risk of a dementia diagnosis (HR 0.75).[16][21] - Speed training without boosters showed no benefit (HR 1.01), and neither did memory or reasoning training.[21][22] - 48.7% of controls were diagnosed during follow-up.[17] - A 2026 reanalysis (Dickens et al.) challenged the analysis, arguing that splitting by booster status breaks randomization.[23] Booster completers may simply be healthier people.

IMPACT study (Smith et al., JAGS 2009; n=487, age 65+). About 40 hours of Posit Science auditory training (1 hour a day, 5 days a week, for 8 weeks) was compared with an active control.[24] - The primary outcome, auditory memory and attention, improved (ES about 0.23).[24] - Secondary effects were roughly 0.20–0.33.[25] - About 0.09 SD remained at 3 months, so the gains faded quickly without continued practice.[24]

Broader evidence and controversy. - Cochrane review (2020; 8 trials, 1,183 participants): computerized cognitive training "may lead to slightly better overall cognitive function at the end of 12 weeks of training," but found "no evidence that the effect persists a year later." Overall evidence quality was low.[26][27] - Far transfer: gains on trained tasks, especially speed, are large. Transfer to everyday memory is small.[28] - FTC action: in January 2016 the FTC settled with Lumos Labs for $2 million (a $50 million judgment, suspended) over unsupported claims that Lumosity improved everyday performance and delayed cognitive decline.[29][30][31][32]

Recommendation. If you want brain training, use the specific adaptive speed-of-processing exercises (the ACTIVE "useful field of view" task, sold as BrainHQ's Double Decision). Do about 10–15 hours initially, add periodic boosters, and pair it with exercise. Treat generic brain games as entertainment.

4. Blood pressure control

Evidence. - SPRINT MIND: a systolic target below 120 versus below 140 mmHg reduced mild cognitive impairment by 19% and MCI-or-dementia by 15%. The effect on probable dementia alone was not significant (HR 0.83; 0.67–1.04), partly because the trial stopped early.[33][34] Extended follow-up confirmed a lower MCI rate (HR 0.87).[35] - China Rural Hypertension Control Project (33,995 people; Nature Medicine 2025): a 22 mmHg net systolic reduction over 48 months cut all-cause dementia by 15% (RR 0.85; 0.76–0.95) and cognitive impairment without dementia by 16%.[36] - Guidelines: the 2025 AHA/ACC guideline recommends a systolic target below 130 mmHg partly to prevent MCI and dementia.[37]

Caveat. This mainly prevents decline rather than sharpening memory within months. It is still one of the highest-yield actions for anyone with hypertension.

5. Hearing and vision correction

Hearing. ACHIEVE (Lancet 2023; 977 adults aged 70–84 with untreated hearing loss) tested audiologist-fitted hearing aids plus counseling for 3 years.[38][39] - In the full trial, there was no significant effect on 3-year cognitive decline.[40] - In the prespecified higher-risk subgroup recruited from the ARIC cohort (n=238), decline was 48% slower.[40] - A secondary analysis found 58% slower decline in the top quartile of risk.[41] - Blood biomarkers of neurodegeneration (GFAP, NfL) improved in parallel.[38] - The 2024 Lancet Commission calls the evidence that treating hearing loss reduces dementia risk "now stronger."[42]

Vision. The 2024 Lancet Commission added untreated vision loss to its list of 14 modifiable risk factors, which together account for an estimated 45% of dementia.[43][44] In the Adult Changes in Thought cohort of 3,038 adults aged 65+ (Lee et al., JAMA Internal Medicine 2022), cataract surgery was associated with 29% lower dementia risk (adjusted HR 0.71; 95% CI 0.62–0.83) for at least 10 years. Glaucoma surgery, which does not restore vision, showed no such link (HR 1.08; 0.75–1.56). This is observational, so healthy-patient bias is possible, and there is no RCT.[45]

6. Learning new skills (Synapse Project, music, language)

Synapse Project (Park et al., Psychological Science 2014): adults aged 60–90 were randomized to learn digital photography, quilting or both, averaging 16.5 hours a week for 3 months.[46] - Episodic memory improved compared with the "receptive" groups.[47] - The receptive groups were a social group with no new learning and a placebo group doing low-challenge activities such as word puzzles. Neither improved.[47][48] - Photography had the clearest individual effect. Some effects were still significant a year later.[48][49] - The key ingredient appears to be sustained mental effort, not simply being busy.[50]

Music. - A piano RCT (Bugos 2022; n=155, ages 60–80, 16 weeks) found some executive-function and psychosocial gains.[51] - The "Keys to Staying Sharp" trial (2023) found no cognitive improvement from piano versus music listening.[52]

Language. A language-learning RCT in adults aged 65–75 (Berggren 2020; n=160, 11 weeks) found no memory or intelligence gains versus relaxation training.[53]

Takeaway. Learning works when it is demanding, sustained and high-volume. A few casual hours a week is unlikely to measurably improve memory.

7. Supplements

Multivitamin (COSMOS):
- In COSMOS-Clinic (n=573), global cognition was +0.06 SD (not significant) and episodic memory was +0.12 SD (significant) over 2 years.[54]
- Pooling 3 COSMOS substudies (about 5,200 people; Vyas et al., Am J Clin Nutr 2024), the gains were +0.07 SD for global cognition (95% CI 0.03–0.11) and +0.06 SD for episodic memory (95% CI 0.03–0.10). Mass General Brigham says the multivitamin "slowed global cognitive aging by the equivalent of two years compared to placebo."
- Caveats: all three are substudies of one trial, so this is not an independent meta-analysis. Partial funding came from Mars Edge, and the participants were mostly White.[55][56] The effect is small but cheap and low-risk.
Cocoa flavanols (500 mg/day):
- No effect on global cognition in COSMOS-Mind or COSMOS-Clinic.[57][58]
- In COSMOS-Web (Brickman, PNAS 2023), memory improved only in people with the poorest baseline diet or lowest flavanol intake (d≈0.09–0.14).[59]
- Eating more berries, tea and apples is a simpler route for people with poor diets.
Omega-3/DHA: null.
- VITAL-Cog (1 g/day fish oil; n=3,424 plus 794 tested in person) showed no difference in cognitive change.[60][61]
- The MAPT and DO-HEALTH trials and a Cochrane review were also null.[62][63][64][65]
Vitamin D: null in VITAL (2,000 IU/day), DO-HEALTH (2,000 IU/day; n=2,157) and D-Health (60,000 IU/month; 5 years).[66][67][68] Only correct an actual deficiency.
B vitamins (VITACOG):
- Population: 266 adults aged 70+ with MCI, given folic acid, B6 and B12 for 2 years.[69]
- Brain atrophy slowed from 1.08% to 0.76% a year. Memory and global cognition benefited only in those with high homocysteine, and more so when omega-3 status was good (atrophy slowed 40% in that group).[70][71][72][73]
- A pooled analysis with a second trial (n=545) confirmed less atrophy but no significant effect on memory or executive function.[74]
- Recommendation: test homocysteine first and supplement if it is elevated or B12 is low.
Creatine: a meta-analysis found a memory SMD of 0.88 in adults aged 66–76 versus 0.03 in younger adults. Trials are small and heterogeneous, so the evidence is promising but immature.[2][75]
Magnesium: the only cognition RCT of a branded magnesium L-threonate had 44 participants aged 50–70 and was industry-funded.[76][77] No independent trials exist in older adults.

8. Diet

The MIND diet trial (NEJM 2023; 604 adults aged 65+, 3 years) found no difference in cognition or brain MRI versus a control diet with mild calorie restriction (P=0.23). Both groups improved.[78][79][80] Post-hoc analyses link higher adherence to better cognition, but that is association.[81] A Mediterranean/MIND-style diet remains sensible for cardiovascular health and as part of a multidomain package (as in POINTER), but it is not proven as a stand-alone memory booster.

9. Sleep and sleep apnea

Severe OSA: a meta-analysis of 14 RCTs (1,926 participants) found that CPAP partially improves cognition, mainly attention and processing speed.[82]
2025 meta-analysis: found processing speed improved (WMD 0.25).[83]
Older adults specifically: a 12-month RCT in adults 65+ found no cognitive benefit despite less sleepiness. Adherence was low.[84]
Recommendation: get screened if you snore, stop breathing at night or are sleepy in the daytime, and use CPAP consistently if prescribed. Insomnia and sleep quality matter too, but sleeping pills are themselves a cognitive risk (see section 10).

10. Medication review

Anticholinergic burden: in a cohort of 3,434 adults aged 65+ (Gray, JAMA Internal Medicine 2015), the highest cumulative anticholinergic use (more than 1,095 total standardized daily doses, about 3 years of daily use) was associated with 54% higher dementia risk (adjusted HR 1.54; 95% CI 1.21–1.96) and 63% higher Alzheimer's risk (HR 1.63; 1.24–2.14), with a clear dose-response.
Deprescribing evidence: a 2023 Cochrane review found only low-certainty evidence that cutting anticholinergic burden improves cognition in the short term.[85]
Common culprits: sedating antihistamines, bladder antimuscarinics, tricyclics, and benzodiazepine or "Z-drug" sleep aids.
Recommendation: ask a pharmacist or physician for a structured review. The upside of removing an unneeded sedating drug is immediate and the risk is low.

11. Social engagement, volunteering, purpose, mood

Socializing alone: the Synapse social group did not improve memory.[48] Social isolation is still a Lancet Commission risk factor and is worth addressing for mood and wellbeing.[86]
Volunteering: in the Baltimore Experience Corps brain substudy (Carlson et al., Alzheimer's & Dementia 2015; 111 participants, mean age 67.2; volunteers asked to tutor children more than 15 hours a week), men in the intervention arm showed 0.7–1.6% gains in cortical and hippocampal volume over 2 years while control men declined (P ≤ .04; significant in men only). The cognitive data are limited.
Depression: a Lancet Commission risk factor.[86] Treating it is worthwhile in its own right and removes a known drag on concentration and memory.
Meditation and mindfulness: not systematically reviewed for this report. Treat as unproven for memory.

12. Reading and visiting new places

No RCTs show that reading or travel improves memory in older adults. The positive associations in cohort studies are vulnerable to reverse causation, because people with sharper cognition read and travel more. Both are worthwhile, and travel often adds walking, novelty and social contact. To get a cognitive "dose," make them effortful: read challenging non-fiction and discuss it, or plan and navigate trips yourself.

Recommendations (a practical weekly plan)

Move: about 150–300 minutes a week of moderate aerobic activity, plus 2 strength sessions. Add tai chi or dance if you enjoy them.
Get checked: blood pressure (aim for systolic below 130 with your doctor), a hearing test (and wear aids), eyes and cataracts, a sleep apnea screen, and a medication review for anticholinergics and sedatives. Consider checking homocysteine and B12.
Train and learn: for brain training, do 10–15 hours of adaptive speed-of-processing training such as BrainHQ Double Decision, with periodic boosters. Better still, take up a demanding new skill for several hours a week.
Structure it: join a coached, group-based program like POINTER if one exists locally. Accountability adds a measurable edge.
Supplements: a standard daily multivitamin is reasonable. Skip omega-3, vitamin D (unless deficient), cocoa extract and magnesium for cognition. Creatine is optional and experimental.
People in their 80s: the same principles apply, with lower-intensity, supervised or multimodal exercise. Fixing sensory, sleep and medication problems has proportionally larger payoffs, and multidomain trials suggest bigger gains for those starting from lower cognition.

Caveats

Most effects are small (about 0.03–0.3 SD) and hard to notice in daily life. The realistic goal is slower decline plus modest gains, not a dramatic reversal.
Many headline findings come from subgroups: ACHIEVE's ARIC cohort, VITACOG's high-homocysteine group, COSMOS-Web's poor-diet group, and ACTIVE's booster completers. Treat these as hypotheses for those groups, not general proof.
Dementia-risk findings for blood pressure, cataract surgery and anticholinergics are mostly secondary to the goal of improving memory now, and several are observational.
Industry ties exist. Posit Science sponsors BrainHQ research, and Mars Edge co-funded COSMOS.