This article began with a hypothetical example of physicians having to make a critical decision in a stressful situation. We asked if it is possible to explain variations in the effectiveness with which different individuals reach correct decisions in emotion-laden conditions in such work environments—whether in medicine, the military, or in other everyday situations. Studies using behavioral, fMRI, ERP, and molecular genetic methods—when used to examine the cognitive functions of attention, working memory, and decision-making and the affective processes of anxiety, sensation seeking, and boredom—can systematically describe individual differences in these functions, both in isolation and in interaction with each other. Importantly, inter-individual variability must be examined from multiple levels of analysis—gene, protein, neurotransmitter, brain networks, cognition, and performance, and cannot be reduced to any one of these levels (Parasuraman, 2003
). Examples from each of these domains of cognition and emotion, and using all of these classes of methods, clearly show that analyses at the group level often mask important findings associated with sub-groups of individuals. Psychometric and molecular genetic methods can be used to identify such sub-groups. There is also growing interest in combining these two approaches. Studies using these diverse approaches show that group-level models of affect and cognition do not apply to all individuals, thus posing challenges for theory. A next step in theory development will involve the integration of the new psychometric and genetic information with neuroimaging data into more inclusive theories of cognition and affect.
We considered first individual differences in working memory and decision-making. Molecular genetic studies of individual variation in these functions point to variability in dopaminergic and noradrenergic genes. Given that working memory capacity is a key factor contributing to decision-making efficiency under time pressure, both may be mediated by overlapping molecular pathways. The experimental results indicate that a gene that regulates dopamine and norepinephrine availability, namely the DBH gene, plays a role in both functions. Specifically, variants of the DBH gene producing low levels of the DβH enzyme that converts dopamine to norpepinephrine (Cubells and Zabetian, 2004
) were associated with both high working memory capacity and superior decision making in a simulated battlefield command and control task, as well as altered subjective stress response (Parasuraman, 2009
). More importantly, the DBH gene also influenced the degree to which decision-making performance was adversely affected by imperfect decision aiding, with individuals with the low DβH enzyme variant showing less susceptibility. The modulation of the effects of a decision aid by genotype represents an example of a gene by task interaction. This represents a specific case of a gene-environment interaction. As Szalma (2009)
pointed out, inter-individual variability in cognitive performance often reflects the interaction between person and task factors, rather than these factors in isolation. Future research on the cognitive effects of the DBH and other genes should systematically examine such interactive effects.
In terms of practical applications of these findings, since decision automation is increasingly used in many work domains (Parasuraman et al., 2000
), these results carry implications for the design of these aids and for training individuals to use them effectively. Other implications for research and practice in human factors include providing practitioners with information on individual variation in decision-making so as to guide work design (Furham, 1992
) or for adapting interfaces to individual workers (Szalma, 2009
). Molecular genetic methods can also inform selection and training procedures aimed at developing teams of human operators who can make speedy, unbiased decisions in semi-automated systems (Parasuraman, 2009
). However, additional work needs to be done using neuroergonomic methods in more complex tasks representative of naturalistic decision-making, such as decisions made by consumers, jury members and judges, physicians, pilots, and military commanders.
Several affective processes play integral roles in decision-making. We discussed three areas of individual differences in affective processing in relation to decision-making and cognitive performance: affective evaluation, anxiety, and boredom susceptibility. Firstly, the influence of affective processes on decision-making can occur both consciously and automatically (Bechara, 2007
). The automatic aspect of affective processing is apparent in the phenomenon of affective priming. Results from affective priming studies indicate that evaluative decisions about words are influenced by automatically processed affective information presented between 150–250 ms before the decision stimulus, as reflected in both the LPP and N400 components of the ERP (Zhang et al., 2006
). Affective priming can also occur across stimulus domains (pictures and words), but the impact on evaluative decision-making differs as a function of processing style between individuals. A psychometrically validated scale (Childers et al., 1985
) indicates that some individuals are “visual” while others are “verbal”. In visual individuals, visual affective images are processed faster than words (Childers and Jiang, 2008
). Such differences in processing style could be used to assess everyday decision-making, as in consumers making choices of items to purchase. In a recent study, ERPs were measured as consumers made buying decisions for items presented as “on-sale” (e.g., a 15% reduction) or simply at regular price. Individuals with low-math anxiety had larger P300 responses than high-anxiety persons when making their non-buy purchasing decisions (Jones et al, 2011
). P300 had a more fronto-central distribution for high-math anxiety individuals whereas it was more centro-posterior for low-math anxietypersons. Thus, it is possible to identify neural correlates of individual differences in affect and cognition that influence performance in tasks that capture some of the characteristics of everyday decision-making.
Anxiety clearly impairs performance on many other cognitive tasks, although the effects tend to be most robust for tasks with working memory or multitasking components (Eysenck, 1992
). Trait anxiety also impairs economic decision-making, and some of the underlying mechanisms can be revealed using ERPs. For example, the FRN component of the ERP, which reflects outcome evaluation in the Iowa Gambling task, differentiates between individuals low and high in trait anxiety (Gu et al., 2010
). This finding suggests that anxiety involves a conscious strategy of evaluation of decisions, and is particularly debilitating when outcomes are ambiguous. Of course, this does not rule out the view that anxiety also impairs decision-making automatically, as studies using masked stimuli in the affective priming (Li et al., 2008
) and “emotional Stroop” paradigms (Gotlib et al., 1984
) have shown.
Individual differences in the interactive effects between anxiety and the cognitive processes of working memory and decision-making may be influenced by genetic factors in dopaminergic molecular pathways. A “worrier vs. warrior” model has been proposed to account for such individual differences between cognition and affect (Goldman et al., 2005
). Whereas “worriers” with the COMT (Met158) variation may have high working memory (Goldberg and Weinberger, 2004
), they also tend to exhibit high trait anxiety. In contrast, those with the COMT (Val158) variation have the opposite pattern, not high in working memory performance but with little or no anxiety. This “warrior” genotype has better stress resiliency but also modest diminution of executive cognitive performance.
Thirdly, individual differences in sensation seeking and its sub-component boredom susceptibility also play a role in cognitive performance. Brain activation patterns during a matching task differentiate between individuals low and high sensation seeking (Jiang et al., 2009
). Boredom susceptibility is also associated with poor performance on repetitive tasks. Not surprisingly, participants with trait boredom do poorly on long-duration vigilance tasks (Davies et al., 1983
; Thackray et al., 1973
). Such tasks typically induce stress and fatigue (Szalma et al., 2004
), which is exacerbated in individuals with high boredom susceptibility. Given that many modern work environments, such as airport security screening (McCarley et al., 2004
; Wolfe et al., 2005
) and military sonar monitoring (Arabito et al., 2007
) involve vigilance, boredom susceptibility can negatively impact performance in many industrial tasks (Smith, 1981
). Trait boredom is also associated with job dissatisfaction and absenteeism in manufacturing (Kass et al., 2001
) and healthcare (Watt and Hargis, 2009
Finally, some limitations of the neuroergonomic approach we have outlined must be acknowledged. Candidate gene studies have been criticized because of failure to replicate reported associations, although this has been primarily in the context of neuropsychiatric disorders (Iannonidis et al., 2001
). The purported solution—studies examining variation across the entire genome—is expensive due to the need for very large samples, and the results to date have not offered significant improvement. A compromise in which both GWAS and candidate gene studies with multiple phenotypes are conducted in a theory-driven manner appears to be a measured approach (Reinvang et al., 2010
). Another limitation of the present review is that much—but not all—of the evidence discussed has not involved examination of brain function and performance in ecologically valid simulations of work or everyday environments, which is one goal of neuroergonomics (Parasuraman and Rizzo, 2008
). But some data from such studies were discussed, and the methods described here could be extended to more complex situations representative of real-world performance.