JMIR Publications

ABSTRACT

Background:

Academic performance depends not only on acquired knowledge but also on cognitive and oculomotor processes that support efficient engagement with learning-related information. Multidomain cognitive training may influence some of these processes, although evidence for generalized far transfer remains limited.

Objective:

To characterize longitudinal differences in cognitive and learning-related performance among secondary-school students who completed Mastermind Cognitive Training compared with students receiving treatment as usual.

Methods:

This multisite, parallel-group, nonrandomized study included 65 participants (Training, n = 38; Control, n = 27) across seven Fusion Academy locations. Pre-to-Post assessment intervals ranged from approximately 10 to 18 weeks across sites. Linear mixed-effects models tested Time × Program Type interactions across reading, oculomotor, cognitive, psychomotor, executive-attention, and exploratory academic-processing outcomes, with Benjamini-Hochberg false discovery rate (BH-FDR) correction applied within analytic families. Site, assessment-interval, training-exposure, delivery-modality, and baseline-adjusted sensitivity analyses evaluated robustness of the primary findings.

Results:

All four primary outcomes showed significant between-group differences in longitudinal change after BH-FDR correction. Reading Comprehension increased by 18.42 percentage points in Training versus 5.93 points in Control (q = .008; partial η² = .10), while Reading Rate increased by 38.37 words per minute in Training and decreased by 18.96 words per minute in Control (q = .004; partial η² = .12). Larger effects were observed for RightEye Oculomotor Function (q < .001; partial η² = .29) and the MM Metric (q < .001; partial η² = .35). Three domain-specific Mastermind indices and BRT mean response time also survived correction. Exploratory SEA findings were concentrated in Reading response-time variability and Math Fluency and Math Verification processing/response time, while accuracy did not differ significantly. Baseline-adjusted sensitivity analyses supported all four primary findings (all q ≤ .002).

Conclusions:

Students who participated in Mastermind Cognitive Training demonstrated greater longitudinal improvement than treatment-as-usual students across reading comprehension, reading rate, oculomotor function, global cognitive performance, and selected measures of processing efficiency. Particularly notable were differences in reading comprehension and reading rate, learning-related outcomes assessed outside the Mastermind platform. Although the nonrandomized design prevents causal attribution, the findings support further randomized evaluation of multidomain cognitive training across cognitive, reading, and other educationally relevant outcomes in secondary-school students.