Eye-Tracking Technologies for Cognitive Assessment After Acquired Brain Injury: Systematic Review.

Background: Acquired brain injury (ABI) is a heterogeneous umbrella term encompassing traumatic and nontraumatic etiologies and is frequently associated with persistent cognitive dysfunction. Conventional neuropsychological assessment remains central to clinical evaluation, but feasibility and measurement precision may be limited in individuals with motor impairment, aphasia, reduced stamina, or fluctuating arousal. Eye tracking offers an objective, low-burden approach that can quantify gaze behavior during task engagement and may provide complementary process-level markers of cognition. Objective: This study aimed to systematically synthesize the evidence on eye-tracking paradigms used as a primary approach for cognitive assessment in ABI and to summarize findings by cognitive domain, paradigm, and clinical interpretability. Methods: We conducted a PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020-compliant systematic review and registered the protocol in PROSPERO (CRD420251038768). PubMed, Web of Science, the Cochrane Library, Embase, EBSCOhost, PsycINFO, and Scopus were searched from inception to April 10, 2025. We included peer-reviewed English-language studies enrolling children or adults with ABI in which eye tracking was the primary assessment modality used to quantify at least one cognitive domain or clinically relevant cognitive-communication process. Two reviewers independently screened studies, extracted data, and assessed methodological quality using design-appropriate tools (Risk of Bias 2, Risk of Bias in Non-Randomized Studies of Interventions, Quality Assessment of Diagnostic Accuracy Studies 2, and the Newcastle-Ottawa Scale). A structured narrative synthesis was performed because of heterogeneity in paradigms and outcome definitions. Results: Twenty-seven studies met the inclusion criteria (N=872 participants; females: n=354 and males: n=518), with most evidence derived from mild traumatic brain injury cohorts, and fewer studies involving stroke, mixed etiologies, and disorders of consciousness. Across domains, antisaccade and related paradigms were commonly associated with differences in inhibitory control and executive function, while predictive tracking, smooth pursuit, and target-blanking paradigms frequently captured alterations in attentional prediction and timing. Virtual reality (VR) free-viewing paradigms identified visuospatial exploration asymmetries in stroke-related neglect, and gaze-based human-computer interface approaches demonstrated above-chance task performance in a subset of patients with disorders of consciousness. Evidence for incremental validity beyond conventional assessment was mixed and often indirect, and safety reporting was uncommon. Overall certainty of evidence was generally low and limited by small sample sizes, cross-sectional designs, and heterogeneity in acquisition procedures, metrics, and analytic pipelines. Conclusions: Eye tracking shows potential as an adjunctive, process-level approach for quantifying specific cognition-relevant behaviors after ABI, particularly within paradigms targeting inhibitory control and predictive attention. Current evidence is insufficient to support broad diagnostic claims or the routine replacement of conventional neuropsychological assessment. Future research should prioritize harmonized paradigms and reporting standards, external validation of classification models, longitudinal designs, and explicit feasibility and safety reporting to clarify when eye tracking provides incremental clinical value for precision neurorehabilitation.