Differentiating adolescents with severe internalizing symptoms using a brief computerized cognitive task: a LASSO regression approach

BACKGROUND: Adolescent depression and anxiety are major global public health concerns; however, current diagnostic approaches rely primarily on subjective self‑report and lack objective cognitive metrics. METHODS: This matched case-control study (controls matched by age and sex) recruited 400 adolescents aged 13–19 years with severe internalizing symptoms, defined by Patient Health Questionnaire-9 (PHQ-9 ≥ 20) and Generalized Anxiety Disorder-7 (GAD-7 ≥ 15), along with 400 healthy controls (PHQ-9 = 0, GAD-7 = 0). The participants completed self-report questionnaires followed by a 2-min computerized interactive game assessing sustained attention, inhibitory control, and visuomotor precision. Least absolute shrinkage and selection operator (LASSO) regression with five-fold cross-validation was performed to identify the most discriminative features. RESULTS: Adolescents with severe internalizing symptoms showed a longer reaction time (median 0.85 s [IQR 0.76–0.94] vs. 0.83 s [IQR 0.77–0.92]; p = 0.38) and a lower correct response rate (98 [88, 100]% vs. 100 [98, 100]%, p < 0.001) with greater spatial deviation (0.21 [0.18, 0.27] cm vs. 0.19 [0.16, 0.23] cm, p < 0.001). LASSO retained three predictors, namely reaction time, correct response rate, and response deviation, as optimal (λ = 0.0017; three non-zero coefficients). Validation yielded an area under the curve (AUC) value of 0.71 (95% CI 0.64–0.77), with sensitivity of 0.50 (95% CI 0.41–0.59) and specificity of 0.78 (95% CI 0.70–0.85). DeLong’s test showed no overfitting (p = 0.18). CONCLUSIONS: The brief, gamified cognitive task demonstrated moderate ability to differentiate adolescents with severe internalizing symptoms from healthy peers. Results indicate cognitive performance differences in sustained attention, inhibitory control, and response precision, which may serve as objective, scalable complementary indicators for early identification in youth mental health screening.