Can Crossword Puzzles Improve Memory Function in Patients With MCI?

Crossword puzzles have an advantage over computerized memory games for improving memory function in adults with mild cognitive impairment.

Completing computerized crossword puzzles is more effective at improving cognitive scores than computerized games among patients with mild cognitive impairment (MCI), according to study findings published in NEJM Evidence.

Risk for dementia is increased among adults with MCI. Current evidence suggests that complex mental activities, such as reading, playing checkers, and completing crosswords or puzzles, can decrease risk for dementia by up to 46%.

Researchers conducted a 2-site, single-blinded, 78-week trial, COG-IT (Cognitive Training and Neuroplasticity in Mild Cognitive Impairment; Identifier: NCT03205709) to compare cognitive outcomes of different types of home-based computerized games. Patients (N=107) with MCI were recruited at Columbia University and Duke University between 2017 and 2020. Patients were randomized in a 1:1 ratio to play games (n=51) or complete crosswords (n=56). The interventions were provided by Lumos Labs.

During a session of the game intervention, 6 random modules were selected from 18 games that included memory, matching, spatial recognition, and processing speed tasks and the difficulty was scaled by the Lumosity Performance Index. The crossword intervention presented a crossword puzzle equivalent to The New York Times’ Thursday puzzles, in which participants were given 30 minutes complete a puzzle and if they completed in in 15 minutes, they were given a second puzzle.

…[C]rossword puzzle training could become a home-based, scalable, cognitive enhancement tool for individuals with MCI.

The primary outcome was the change in the 11-item Alzheimer’s Disease Assessment Scale-Cognitive (ADAS-Cog) total score at 78 weeks compared with baseline.

The games and crosswords cohorts included patients mean age, 71.1±8.5 and 71.3±9.1 years, 33.3% and 50.0% were men, 76.5% and 75.0% were White, they completed 16.6±3.0 and 16.8±3.3 years of education, and ADAS-Cog score was 9.5±3.5 and 9.6±3.5 points, respectively.

The target number of game or crossword sessions was 72. The median number of sessions played was 72 for patients enrolled at Duke and 68 for participants enrolled at Columbia. Some patients participated in more sessions than intended (maximum, 116 sessions).

At week 78, the ADAS-Cog score declined by an average -0.89±0.51 points among patients in the games group and improved by 0.55±0.48 points among those in the crosswords group (P =.04).

For functional outcomes, Functional Assessment Questionnaire (FAQ) scores decreased among individuals in both groups, but to a smaller extent among patients in the crosswords group compared with those the games cohort (mean, -0.89 vs -2.07), respectively.

Progression to dementia was observed among 10.7% of patients in the crosswords group and 15.7% of patients in the games group, favoring crosswords (odds ratio [OR], 0.65; 95% CI, 0.17-2.32). Reversion of MCI to normal cognition occurred among 30.4% of the crosswords and 23.5% of the games cohorts, favoring neither group (OR, 1.41; 95% CI, 0.55-3.71).

At magnetic resonance imaging (MRI), individuals in the games group had a greater decline in hippocampal volume (LSMD, 34.07) and cortical thickness (LSMD, 0.02) compared with those in the crosswords group, after adjusting for covariates.

No intervention-related adverse events were reported.

A potential limitation for this study was that the follow-up assessments were delayed by ³60 days for 6.3% of participants due to the COVID-19 pandemic.

The researchers concluded that “If these effects are replicated and expanded in future trials with the inclusion of a control group that does not receive cognitive training, crossword puzzle training could become a home-based, scalable, cognitive enhancement tool for individuals with MCI.”

Disclosure: Multiple authors declared affiliations with industry. Please refer to the original article for a full list of disclosures.


Devanand DP, Goldberg TE, Qian M, et al. Computerized games versus crosswords training in mild cognitive impairment. NEJM Evidence. Published online October 27, 2022. doi:10.1056/EVIDoa2200121