Rapid automatized naming (RAN) assesses pre-literacy and could be useful for assessing the effects of HIV infection on the development of literacy. Many children with HIV live in sub-Saharan Africa, but the RAN has not been well-studied there. Cultural differences in when children learn color names and in how they prioritize between accuracy and speed can affect RAN results. Typically, RAN and other cognitive functions (e.g., general intelligence, processing speed, working memory, attention) are strongly related. Our goal is to examine how RAN performance relates to neurocognitive function in sub-Saharan Africa and whether RAN performance could reveal the effect of HIV on pre-literacy. The current study examines the relationship between RAN performance and cognitive variables in both children living with HIV (CLWH) and healthy controls in Tanzania.

486 children (ages 3-8) were administered RAN Color and Object tasks as part of a larger longitudinal study in Dar es Salaam, Tanzania. All participants were also administered the Leiter International Performance Scale - 3rd Edition (Leiter-3), a test of nonverbal intelligence and general cognitive function. Binomial logistic regression examined the likelihood of completing a RAN task and included age, Leiter-3 composite standard scores, and HIV status. Multiple linear regression using the same predictors assessed factors associated with RAN completion time in children who completed the tests. SES and education were not included in the model specifications because they were not correlated with any RAN measures.

Only 40% of children could complete the RAN Color task, while 88% completed the Object task. Logistic regression models showed that age (p<0.001), Leiter-3 Nonverbal IQ (p<0.01), and Processing Speed (p<0.001) composite scores were all unique predictors of whether children would complete the RAN Color task. Age (p <0.001) and Nonverbal IQ (p<0.001) were predictors of completing of the RAN Object task. Of those who could complete the RAN Color task, multiple linear regression demonstrated that age (p<0.01), Leiter-3 Nonverbal IQ (p=0.01) and Processing Speed (p=0.001) composites predicted completion time, with the model accounting for 25% of the variance. For the RAN Object task, multiple linear regression indicated age (p<0.001), Leiter-3 Processing Speed (p=0.01) and Nonverbal Memory (p=0.01) composites, and living with HIV (p=0.01), predicted completion time, with the model accounting for 42% of the variance.

Completion rates for RAN Color and Object tasks were low but improved with age. Consistent with brain maturation, increasing age and processing speed improved completion time, regardless of the RAN task. General cognitive ability predicted RAN Color performance, and Nonverbal Memory (encompassing attention, working memory, and retrieval) and HIV status additionally predicted RAN Object performance. Results extend research indicating RAN is distinct yet multifactorial, relying on various neurocognitive functions working together. Additionally, the relationship of HIV to RAN Object performance implies an overlap between the neurocognitive functions inherent in RAN and the neurocognitive weaknesses often reported in CLWH. These findings suggest cognitive vulnerabilities in CLWH may extend to literacy skills in sub-Saharan Africa, which requires further study.