The rank-order stability of cognitive performance across time in a longitudinal twin design

The non-shared environmental contributions to cognitive stability in later life was investigated in two twin longitudinal samples (SATSA and LSADT) representing two different countries (Denmark and Sweden). The contributing... [ view full abstract ]

The non-shared environmental contributions to cognitive stability in later life was investigated in two twin longitudinal samples (SATSA and LSADT) representing two different countries (Denmark and Sweden). The contributing twin samples were taken from the Interplay of Genes and Environment across Multiple Studies (IGEMS) consortium. A general cognitive factor score was created within in each sample and two individual cognitive test performances (Digits Span and Symbol Digit) were examined between twin pairs across the samples and waves. The initial intake average differed by about 10 years between the samples: SATSA (64) and LSADT (75). Twins’ cognitive performances were compared to their co-twin to determine twin order for each wave. The twin ranking on each test was then used to find the percentage of consistency for the higher scoring twin for each wave with intake being the baseline. The dizygotic twins had higher percentages of twin order consistency across the two samples for all the tests examined, but only modestly higher for Digit Span and Digit Symbol. Additionally, the younger sample (SATSA) had higher consistency on twin order for the general cognitive score than the older twin sample (LSADT). A multivariate Cholesky was used to assess the non-shared environmental contributions and correlations for each sample. The declining non-shared correlations pattern for general cognitive ability paralleled between samples across waves, but the magnitude was different, with SATSA reporting strong correlations and moderate for LSADT. Similar to general cognitive ability, the non-shared environmental correlation pattern across waves for Digit Span agreed between samples. The pattern and strength for the non-shared environmental correlations were stable across waves (~0.3). Unlike general cognitive ability and Digit Span, the Symbol Digit non-shared environmental correlation pattern were divergent after the first wave assessed and the pattern for SATSA and LSADT continued to mirror each other across waves. The observed differential patterns across samples and cognitive tests need further investigation to elucidate the meaning of these findings on the etiology of cognitive stability in later life.