Age-related changes and longitudinal stability of individual differences in ABCD Neurocognition measures. (April 2022)

• Record Type: Journal Article
• Authors: Anokhin, Andrey P.; Luciana, Monica; Banich, Marie; Barch, Deanna; Bjork, James M.; Gonzalez, Marybel R.; Gonzalez, Raul; Haist, Frank; Jacobus, Joanna; Lisdahl, Krista; McGlade, Erin; McCandliss, Bruce; Nagel, Bonnie; Nixon, Sara Jo; Tapert, Susan; Kennedy, James T.; Thompson, Wesley
• Is Part Of: Developmental cognitive neuroscience. Volume 54(2022)
• Publication Date: 2022-04
• Subjects: Neurocognition -- Longitudinal -- Development -- Test-retest reliability; Cognitive neuroscience -- Periodicals; Developmental neurobiology -- Periodicals; Neuropsychology -- Periodicals; Neuropsychiatry -- Periodicals; 612.8233
• DOI: 10.1016/j.dcn.2022.101078
• Languages: English
• ISSNs: 1878-9293

Abstract

Abstract: Temporal stability of individual differences is an important prerequisite for accurate tracking of prospective relationships between neurocognition and real-world behavioral outcomes such as substance abuse and psychopathology. Here we report age-related changes and longitudinal test-retest stability (TRS) for the Neurocognition battery of the Adolescent Brain and Cognitive Development (ABCD) study, which included the NIH Toolbox (TB) Cognitive Domain and additional memory and visuospatial processing tests administered at baseline (ages 9–11) and two-year follow-up. As expected, performance improved significantly with age, but the effect size varied broadly, with Pattern Comparison and the Crystallized Cognition Composite showing the largest age-related gain (Cohen's d:.99 and.97, respectively). TRS ranged from fair (Flanker test: r = 0.44) to excellent (Crystallized Cognition Composite: r = 0.82). A comparison of longitudinal changes and cross-sectional age-related differences within baseline and follow-up assessments suggested that, for some measures, longitudinal changes may be confounded by practice effects and differences in task stimuli or procedure between baseline and follow-up. In conclusion, a subset of measures showed good stability of individual differences despite significant age-related changes, warranting their use as prospective predictors. However, caution is needed in the interpretation of observed longitudinal changes as indicators of neurocognitive development. Highlights: Neurocognitive performance improved with age for all measures except RAVLT. Longitudinal test-retest stability of test performance ranged from fair to excellent. For some tests, developmental changes may be confounded with practice effects. Ceiling effects (score compression with increasing age) were evident for some metrics.