Cognition is a complex genetic trait that may be influenced by numerous genes of small influence and the identification of these genes may be enhanced by the study of endophenotypes such as event-related potentials (ERPs). This thesis examined in detail whether genetic influence on the ERP slow wave (SW) potential varied according to recording site, recording interval, and trial type. The ultimate goal was to examine genetic associations among the SW (and P300 amplitude and latency) and more complex behavioural measures of cognitive function (i.e. processing speed (CRT), inspection time, working memory performance, full-scale IQ (MAB)).
ERPs were elicited in a delayed-response working-memory task and collected from 391 twin pairs (185 MZ, 206 DZ) aged 16. Data were collected at 15 sites and SW amplitude was computed for six intervals (1 x 500 ms, 3 x 1000 ms, 1 x 3000 ms, and 1 x 3500 ms) and four trial types (sensory and memory - 50% of each included a distracting stimulus). Fifty twin pairs were tested on two occasions to assess test-retest reliability (average r''s = 0.48 for SW, 0.67 for P300 amplitude, 0.55 for P300 latency).
Genetic influences on SW amplitude were regionally variable with magnitudes greater at posterior (h2 = 0.30 to 0.42) than anterior sites (h2 = 0.08 to 0.28). Common genes had considerable influence at adjacent sites (average rg = 0.80), but genetic influence at anterior sites appeared to be largely independent of that at posterior sites. This suggested prefrontal and parietal SW amplitude might provide unique information about working memory processes, which are enhanced in prefrontal and parietal brain regions during visuospatial working memory tasks.
Sources of genetic influence on four consecutive SW intervals varied over time with two sources being identified. For parietal SW, the first genetic factor was most influential on the first SW interval (38% of total variance) with decreasing influence on subsequent intervals (26% on the second to 8% on the fourth). Conversely, the second factor had increasing influence over the second to fourth intervals (3% to 17%). Thus, the first genetic factor appeared to capture working memory processes that decay over time, and the second, preparatory motor processes that increase in activity, as a motor response becomes imminent.
In contrast to the independent genetic influences for site (anterior vs. posterior) and interval (early vs. late), common genetic factors influenced SW recorded in trial types that varied in memory requirement and distraction (rg = 0.93 to 0.99). However, heritability increased with increasing trial type difficulty (0.29 to 0.48 at parietal). Therefore, while the same neural network may have been active during all trial types, the level of activity may have varied with task difficulty and affected heritability.
Unique environmental factors influenced all remaining variance. Analyses that included retest data (to isolate measurement error from unique environmental influence) indicated that for SW recorded in the first interval (all trial types), measurement error inflated unique environmental estimates by approximately 54%.
The final analyses examined genetic associations among ERP and behavioural measures of cognitive function. A latent genetic factor was found to influence parietal SW amplitude (h2 = 0.53), accounting for 18% of the variance, and IQ (h2 = 0.83), accounting for 4% of the variance. Thus genes influencing spatially related working memory processes may have a small effect on general cognition. In addition, a further latent factor was found to influence prefrontal P300 latency (h2 = 0.46), accounting for 46% of the variance, and working memory performance (h2 = 0.46), accounting for 12% of the variance, suggesting that genes influencing prefrontal neural efficiency may influence memory performance. This factor also had small influences (3% to 13%) on choice reaction time (h2 - 0.47), prefrontal SW amplitude (h2 = 0.32), parietal P300 amplitude (h2 = 0.58), and parietal P300 latency (h2 = 0.53), but did not influence IQ. No association was found between P300 amplitude and IQ, with remaining latent factors influencing either ERP amplitude (SW and P300) or the behavioural measures (choice reaction time, inspection time (h2 = 0.40), working memory performance, and IQ). Analyses showed that the heritability of selected ERP measures compared favourably with that of behavioural measures and that examining ERPs provided unique insights into processes underlying cognitive ability.