This study was conducted to determine whether provision of preformed dietary docosapentaenoic acid (DPAn-6) can replace docosahexaenoic acid (DHA) for brain function as assessed by spatial task performance. A newly modified artificial rearing method was employed to generate n-3 fatty acid-deficient rats. Newborn pups were separated from their mothers at 2 days of age and given artificial rat milk containing linoleic acid (LA), or LA supplemented with 1% DHA (DHA), 1% DPAn-6 (DPA) or 1% DHA plus 0.4% DPAn-6 (DHA/DPA). The animals were then weaned onto similar pelleted diets. At adulthood, behavioural tasks were administered and then the brains were collected for fatty acid analysis. The LA and DPA groups showed a lower (63–65%) brain DHA than the dam-reared, DHA and DHA/DPA groups and this loss was largely compensated for by an increase in brain DPAn-6. The brain fatty acid composition in the DPA group was the same as that in the LA group at adulthood. In the Morris water maze, the LA and DPA groups exhibited a longer escape latency than the dam-reared and DHA groups and had a defect in spatial retention. In conclusion, DPAn-6 could not replace DHA for brain function, indicating a highly specific structural requirement for DHA.