Let's Inhibit Our Excitement: The Relationships Between Stroop, Behavioral Disinhibition, and the Frontal Lobes

Among the primary uses of neuropsychological assessment is measuring impairments that either reflect real-world skills and behaviors, or provide information about neurological structure. Although the latter use has been construed as the more inferential of the two, because it presupposes a relationship between task performance and neurological function, both uses are highly inferential. The supposition that a neuropsychological test measures real-world skills and behaviors is an assumption, and one whose merits are called into question by weak or inconsistent relationships between neuropsychological test scores and other measures of the skill or behavior (Shallice & Burgess, 1991; Ready et al., 2001; Chaytor et al., 2006) and poor construct differentiation (Dodrill, 1997; Duncan et al., 1997), particularly for executive functioning (Burgess, 1998). Put simply, the tests upon which neuropsychologists rely may not, in fact, measure the skills and behaviors they are assumed to measure. This assumption should be tested empirically. One of the most widely used neuropsychological tests is the Stroop Color-Word Interference Test (Stroop, 1935; Rabin, Barr, & Burton, 2005), which typically is used to measure inhibition of a prepotent response (Miyake et al., 2000; Vendrell et al., 1995). Although variations exist, it typically consists of three conditions: a color naming condition for which individuals are asked to name patches of colors on a page as quickly as possible, a color reading condition for which individuals are asked to read color names on a page as quickly as possible, and a color-word interference condition that contains color names written in incongruent colors (e.g., the word “blue” written in red ink) for which individuals are asked to name ink colors as quickly as possible. The increased time needed to name the colors on the color-word interference condition and the increased number of errors committed, relative to the simple color naming condition, are referred to as “the Stroop effect.” This effect is believed to be due to the difficulty of suppressing the natural tendency to read the words, which is thought to be a more practiced and thus automatic action (Stroop, 1935; Delis et al., 2001). Stroop (1935) originally conceptualized the task as one of interference or inhibition, noting that the terms had been used indiscriminately. Since his publication, researchers have been active in investigating the Stroop effect (see MacLeod, 1991 for a comprehensive review of the experimental literature), and in trying to understand the cognitive processes, largely from an experimental cognitive framework, responsible for producing the effect. The research from experimental psychology, however, appears almost entirely separate from the research and use in clinical neuropsychology, as Stuss and Levine (2002) have noted. Neuropsychologists may have adopted the Stroop, as Alvarez & Emory (2006) suggest, after one study found that patients with left frontal brain lesions showed significantly more slowing from word reading and color naming to Stroop inhibition than did patients with brain lesions in temporal, posterior, or right frontal regions (Perret, 1974). However, subsequent lesion studies with the Stroop have been inconsistent (Vendrell et al., 1995; Stuss et al., 2001), leading some to describe the Stroop as a “fundamentally ineffective test” for “evaluating frontal lobe functioning,” (Dodrill, 1999, p.563). Increased time taken to complete the Stroop interference typically is predictive of frontal lesions only when color naming speed is not taken into consideration (Foong et al., 1997; Stuss et al., 2001), suggesting that frontal lobe lesions or atrophy cause slowing in general, rather than a true exaggerated interference effect. When controlling for color naming, Stuss et al. (2001) found exaggeration of the Stroop interference only for patients with superior medial frontal lesions. A meta-analytic review that compared the sensitivity to fontal damage of the Wisconsin Card Sorting Test, verbal fluency, and the Stroop, showed the Stroop to be the least sensitive to frontal damage, with d=−0.30 (Alvarez & Emory, 2006). In adopting the Stroop for neuropsychological assessment, clinicians and researchers use an inverted interpretation of the interference effect: while Stroop's original study and subsequent experimental research found slowing in normal people on the interference trial, neuropsychologists tend to conclude that when individuals are abnormally slowed on the interference trial, it is due to impaired inhibition or concentration (Lezak et al., 2004). That is, when a normal process—interference—occurs at abnormally high levels, it is taken as evidence of a pathological process of disinhibition. Studies addressing the construct validity of the Stroop provide poor support for this interpretation, however, and suggest it may measure basic attention, working memory, and processing speed better than inhibition. Correlations with other neuropsychological tests have found that the Stroop shares a large amount of variance (e.g., correlations of .95) with measures of processing speed, leading the authors of one study to conclude that “interference scores from Stroop tasks are not simply measures of inhibition because they share most of their age-related effects with other measures of processing speed,” (Salthouse & Meinz, 1995). Indeed, processing speed appears to mediate the relationship between age and poorer Stroop performance (Earles et al., 1997). In contrast, substantial variation has been found between the Stroop interference task and other tests believed to measure inhibition (Shilling et al., 2002; Earles et al., 1997), suggesting the Stroop is a poor measure of inhibition. Factor analysis has shown the Stroop interference task to load on a measure of “sustained selective processing” along with Serial 7 and Serial 13 tasks, tasks that have little to do with inhibition (Shum et al., 1990). There thus appears to be little evidence from other neuropsychological tasks that Stroop interference is a measure of inhibition rather than simply attention and processing speed. More troublesome is that neuropsychologists' interpretation of the Stroop may go beyond cognitive interference to inferences about behavioral disinhibition. Such inferences occur despite the dearth of evidence that difficulty with Stroop inhibition has any relationship with behavioral disinhibition. Two studies using the Neuropsychiatric Inventory (NPI) have failed to find relationships between the Stroop and behavioral disinhibition in samples of people with probable Alzheimer's disease (AD) and Parkinson's disease (PD) (Aarsland et al., 1999; Back-Madruga et al., 2002), although it was significantly correlated with NPI total score (Back-Madruga et al., 2002), aggression (Back-Madruga et al., 2002), and apathy (Aarsland et al., 1999). In contrast, Chan (2001) found a borderline significant association (r=.18, p=.05) between an inhibition factor of the Disexecutive Questionnaire (DEX) and time taken to complete the Stroop among a sample of cognitively normal Hong Kong Chinese people aged 18 – 50. Despite the research that has cast doubt on the validity of Stroop as a measure of disinhibition, and the inconsistencies in its relationship with the frontal lobes, neuropsychologists continue to use it in these ways. The first aim of the present study is to directly examine how well the Stroop Interference test predicts behavioral disinhibition in daily life. The second aim is to examine the extent to which poor performance on the Stroop Interference test is associated with atrophy in frontal regions believed to be necessary for inhibition.