Primary objective: To examine the relationship between CT abnormalities and early neuropsychological outcome following traumatic brain injury (TBI) using quantitative CT analyses, data reduction methods for neuropsychological results and specific hypotheses based on literature review. Research design: Observational, prospective cohort study using acute (emergency) CT data and neuropsychological test data from 89 participants with TBI who were hospitalized for rehabilitation. Methodsand procedures: Principal components analysis with varimax rotation was used to reduce data from a standard battery of eight neuropsychological tests administered after clearance of post-traumatic amnesia (1 month post-TBI on average). Bivariate correlations were used to examine relationships of three factors (verbal memory, cognitive processing speed and verbal working memory) to quantitative volumetric analysis of CT scan abnormalities (size, number and location). Specific hypotheses as to CT predictors of poor performance on each factor were tested using multivariable linear regression that included injury severity and demographic variables. Main results: Eighty-nine per cent of participants had some pathology on initial CT. Age, education and time to follow commands (TFC), an index of overall injury severity, were significantly associated with the neuropsychological factors. However, none of the specific hypotheses about CT scan variables and cognitive outcome were strongly supported by the data. There was a trend for any CT abnormality to predict slower speed of processing and for higher number of brain lesions to predict worse memory performance. Conclusions: Despite the precision added by quantitative CT analysis, CT findings did not improve on demographic factors and TFC in predicting early cognitive outcome of TBI. Imaging methods that are more sensitive to white matter integrity may be needed to develop pathophysiologic predictors of TBI outcome.