Objective: The Reynolds Intellectual Assessment Scales (RIAS) is a measure of intelligence that provides estimates of verbal and nonverbal intelligence, and verbal and nonverbal memory. Little is known about the psychometric properties of the RIAS memory scales. This study examined the construct validity of the RIAS memory scales index scores. Method: The sample included 78 children and adolescents (mean age = 14.6 years, SD = 3.0; 64.0% male) with various forms of brain injury (BI). A battery of neuropsychological tests, including the RIAS, was administered for rehabilitation purposes. To examine validity of RIAS memory scale scores, the BI group was compared to the standardization sample (SS) to determine overall level of impairment, and correlations were calculated between the RIAS Memory Indexes and the RIAS Intelligence Indexes, as well as other tests of verbal and perceptual abilities. Analyses examined the complete BI group, and just those children with TBI (n = 50). Results: As anticipated, single sample t-tests indicated the BI and TBI groups' Memory Indexes were significantly lower than the SS (p < .001). Patterns of associations among the various tests indicated stronger correlations were present between the Verbal Memory Index and verbal tests compared to perceptual tests. The opposite pattern was noted for the Nonverbal Memory Index. Conclusion: Results provide preliminary support for validity of RIAS Memory Indexes for assessment of children with BI. Similar investigations with larger, more homogeneous, and adult clinical groups are warranted, although results were substantially similar between our BI and TBI groups suggesting generalizability of findings to TBI and possibly other clinical samples.