OBJECTIVE: To examine the independent association between heart rate variability (HRV) and cognitive performance, in a nationally representative population study of older adults. METHODS: Cross-sectional analysis of wave 1 data from the Irish longitudinal study on ageing (TILDA) was performed. A subset of 4,763 participants who underwent ECG recording during resting and paced breathing periods were used for the analysis. HRV indices were divided into quintiles for comparison of values and cognitive performance was defined using the Montreal cognitive assessment (MOCA) score. Multivariate linear regression was used to model the association between cognition and different quintiles of each HRV index, after adjustment for covariates. RESULTS: The mean age was 61.7 +/- 8.3 years and 2,618 (55 %) were female. Lower quintiles of SDNN (P = 0.01-paced), LF (P = 0.001-paced), and LF:HF ratio (P = 0.049-paced) were significantly associated with lower MOCA scores (during both recording periods), independent of confounders. Sub-domains of MOCA responsible for the relationship were predominantly memory recall and language. INTERPRETATION: Reduced HRV is significantly associated with lower cognitive performance at a population level in people aged 50 and older. This further strengthens the relationship between autonomic dysfunction and cognitive disorders.