Eye-tracking and game-based digital tools for autism assessment in Latin America: a narrative review

Autism assessment in Latin America relies on clinician time that is scarce, and digital tools that record gaze or play behavior could widen access if they perform well in regional samples. This narrative review examined eye-tracking and game-based tools for autism screening or assessment developed or validated with Latin American participants, and the evidence of accuracy, feasibility and cultural adaptation they report. Literature published between 2018 and 2026 was searched. Nineteen primary studies came from Brazil, seven from Ecuador, three from Chile, two from Mexico and one from Peru. Eight eye-tracking or gaze-based studies reported classification metrics against a clinical reference, with the largest field sample (585 toddlers in community nurseries) yielding an area under the curve of 0.65. No game-based study reported accuracy against a reference standard; five described design, operability or usability. Cultural adaptation was documented almost exclusively for questionnaires, whereas the adaptation of eye-tracking stimuli was not described and only one study compared socioeconomic strata. Computational classifiers reached high internal metrics, and a single replication tested one of them on a new regional sample. The evidence supports eye tracking as a feasible adjunct in Latin American settings, leaves game-based assessment unvalidated, and indicates that multicenter, prospectively registered accuracy studies with locally adapted stimuli are the main gap.