Introduction: Serotonin affects several brain functions including cognition. The serotonin transporter (SERT) regulates brain serotonin levels through reuptake into neurons. The gene encoding this transporter, the SERT gene, has several functional polymorphisms affecting the number of transporters and thereby the serotonin levels. SERT gene expression may be important for cognition and selective serotonin reuptake inhibitors (SSRI) may improve cognition post stroke. We therefore examined the association between SERT genotypes, cognitive function and early treatment with the SSRI citalopram among non-depressed Caucasian stroke patients. Patients and Methods: SERT gene polymorphisms in 270 non-depressed first-ever acute ischemic stroke patients randomized to citalopram, n = 130, or placebo, n = 140, were investigated. Patients were genotyped for a length polymorphism (L = long and S = short allele) and a single nucleotide polymorphism (A/G substitution) dividing the L-allele into LA and LG. According to these genotypes, patients were further grouped according to low (S/S, LG/S and LG/LG), medium (S/LA and LG/LA), or high functional gene expression (LALA). Cognition was measured by the Symbol Digit Modalities Test (SDMT) at 1 and 6 months. Mean SDMT scores according to genotype and randomization groups were compared using multiple logistic regression adjusting for age, stroke severity, premorbid functional status, and vascular risk factors including smoking, hypertension, and diabetes. Results: Stratified by genotype groups, there were no statistically significant differences in SDMT scores between randomization groups. Placebo-treated patients with low SERT expression genotypes, however, tended to have lower mean SDMT scores (at 1 month: 30.2, SD 10.8) compared to citalopram-treated patients (33.6, SD 13.7). Within the placebo group, the low genotype expression patients had significantly lower adjusted mean SDMT scores at 1 month compared to the high genotype expression patients (adjusted mean difference of –6 points, CI –12.0 to –0.05). We found similar results at 6 months, although not statistically significant. The genotype expression was not associated with SDMT scores among citalopram-treated patients. Conclusion: There was no difference in cognition between citalopram and placebo-treated patients according to the genotype group. Our results indicate, however, that low expression SERT genotype may contribute to reduced cognitive function post stroke as placebo-treated patients with low SERT expression tended to score lower on the SDMT. The significant difference in SDMT scores between low and high expression patients was present only in the placebo-treated group, thereby warranting further exploration of the potential effect of early citalopram treatment on cognitive functioning. Our results are preliminary and need replication in larger-scale studies.