|Home | About | Journals | Submit | Contact Us | Français|
Acute abstinence from cigarette smoking by nicotine-dependent smokers has been linked with cognitive deficits, but the role of nicotine dependence per se in these effects is not known. We therefore tested the relationships of nicotine dependence and smoking history with performance in perceptual-motor, timed tests of attention. Nicotine-dependent smokers (n=37) and nonsmokers (n=48), 18–55 years old, took both the d2 Test of Attention and the Digit Symbol Test on each of 2 test days. For smokers, testing on one day began after ad libitum smoking (<45 min since last cigarette); and on the other day, it began after overnight abstinence (>13 hr since last cigarette). On each test day, there were two test blocks with an intervening break, when only the smokers each smoked one cigarette. There were no significant effects of abstinence or of smoking one cigarette on the performance of smokers; however, across conditions, the smokers’ performance on both tests correlated negatively with severity of nicotine dependence but not lifetime cigarette consumption or cigarette craving. Smokers with high nicotine dependence performed more slowly on both tests than less dependent smokers or nonsmokers. The findings suggest that severity of nicotine dependence and slowness in perceptual-motor tasks of attention share an underlying basis.
Chronic cigarette smoking and acute abstinence both influence a range of cognitive functions, notably attentional processes (see reviews by Heishman, Taylor, & Henningfield, 1994; Jorenby et al., 1999; Mansvelder, van Aerde, Couey, & Brussaard, 2006; Newhouse & Hughes, 1991; Newhouse, Potter, & Singh, 2004; Sherwood, 1993). In fact, the assertion that cigarette smoking improves concentration and other cognitive processes has been cited as a motive to continue smoking (Robinson & Pritchard, 1992; Wesnes & Warburton, 1983). In line with self-reports, research confirms that abrupt abstinence from smoking can impair attention and memory in nicotine-dependent individuals (Hirshman, Rhodes, Zinser, & Merritt, 2004; Mendrek et al., 2006), and that smoking (Parrott, Garnham, Wesnes, & Pincock, 1996) or nicotine administration (Ernst et al., 2001) can reverse withdrawal-related deficits. Although many factors contribute to nicotine dependence, studies suggest that cognitive processes are central to the initiation and maintenance of smoking behaviors (Newhouse et al., 2004; Shiffman et al., 2000). Understanding the interactions of cigarette smoking with cognitive functions, therefore, may help in the development of effective treatments for nicotine dependence.
One of the most robust effects of cigarette smoking is improvement of performance on attention tasks (see reviews by Kassel, 1997; Levin, McClernon, & Rezvani, 2006). Such improvement, as measured by reduced response times and increased accuracy, has been demonstrated using the Eye Movement Task (Mancuso, Lejeune, & Ansseau, 2001), the Digit Symbol Substitution Test (Petrie & Deary, 1989), the Letter Cancellation Test (Snyder, Davis, & Henningfield, 1989), the Rapid Visual Information Processing (RVIP) task (Wesnes & Warburton, 1983), the Sternberg Working Memory Paradigm (Pineda, Herrera, Kang, & Sandler, 1998), and the Stroop Color-Word Interference Test (Domier et al., 2007). While most findings support the view that smoking a cigarette improves attention, other studies indicate no effect (Herbert, Foulds, & Fife-Schaw, 2001) or disturbances in performance after cigarette smoking (see review by Heishman et al., 1994). For instance, some of these inconsistencies may reflect the fact that task performance in some studies required simple repetitive actions (O’Connor, 1986), while in other studies effortful and complex processes were involved (Sherwood, 1993; Spilich, June, & Renner, 1992).
Letter cancellation tasks use paper and pencil to test perceptual-motor scanning abilities and vigilance. The basic procedure includes scanning a row of letters or numbers and crossing the designated target item with a pencil stroke. Williams (1980) assessed the effects of smoking, using a letter cancellation test that involved crossing out each letter “E” randomly interspersed among other letters. Smoking a cigarette improved performance on this task more than sham smoking (with an unlit cigarette). In line with this finding, Snyder et al. (1989) reported improved performance in letter cancellation speed and accuracy following smoking a cigarette. Comparable improvement in speed was achieved with nicotine administration (Parrott et al., 1996). Performance on a letter cancellation test also was significantly impaired after 2–6 hr of abstinence (Parrott et al., 1996). Taken together, these findings suggest that both cigarette smoking and nicotine administration facilitate the cognitive functions that underlie attentional and perceptual-motor processes.
The d2 Test of Attention, another cancellation test, has been proposed as a measure of selective attention and concentration (Brickenkamp & Zillmer, 1998; Rose et al., 2007). Unlike other cancellation tests, the d2 Test entails canceling targets that are interspersed among perceptually similar distractors. In this situation, the distractors compete for processing and the participant must consciously inhibit canceling them while selecting the targets. The d2 Test of Attention is more complex than other cancellation tests (such as the C’s and E’s Test) because its performance requires overcoming competing tendencies. No previous study has examined the effect of smoking on performance of the d2 Test of Attention.
The Digit Symbol Test assesses perceptual-motor speed during copying, visual scanning, and incidental learning (Joy, Fein, & Kaplan, 2003; Joy, Kaplan, & Fein, 2004). Similar to the d2 Test of Attention, the Digit Symbol Test requires selecting relevant items while ignoring distracting information in a designated 90-s period. Performance on the task is highly correlated with that of the d2 Test of Attention, suggesting that similar or overlapping functions underlie performance of the two tasks (Rose et al., 2007). These tests extend beyond sensory-motor integration and require allocation of cognitive resources such as attention and working memory (needed for matching symbols and discriminating targets from similar items that are non-targets) without confounding abilities in conceptual knowledge or abstract reasoning.
We have identified three studies that examined the relationship between performance on the Digit Symbol Test and cigarette smoking. One study found improved performance related to smoking after over-night deprivation (Petrie & Deary, 1989). Another study found no differences between the performance of heavy smokers and light-to-moderate smokers; but as 70% of the participants were at least 60 years old, the findings may be specific to the elderly (Razani, Boone, Lesser, & Weiss, 2004). In a third investigation, smokers exhibited slower copying speed than ex-smokers and nonsmokers (Grant et al., 1996).
In the present study, we tested for differences in performance between nonsmokers and smokers on the d2 Test of Attention and the Digit Symbol Test, and assessed the effects of abrupt cessation of smoking and the resumption of smoking after either overnight or brief abstinence on performance. We also explored whether performance among nicotine-dependent individuals varied as a function of cigarette craving, severity of nicotine dependence, and history of chronic smoking.
Thirty-seven cigarette smokers and 48 nonsmokers were studied. Potential research participants were recruited through flyers and newspaper advertisements. Smokers were included if they reported smoking between 15 and 40 cigarettes per day, and had been smoking continuously for at least 2 years. Exclusion criteria included: current use of medications that might affect cognition, any medical or psychiatric condition (other than nicotine dependence), history of head trauma, smoking marijuana more than once per week, drinking more than 10 alcoholic drinks per week, or regular substance abuse other than alcohol or marijuana. In addition, nonsmoker participants were excluded if they had a lifetime history of smoking more than five cigarettes.
Those that passed a telephone screening received an extensive description of the study in person and provided written informed consent, as approved by the UCLA Institutional Review Board. At this intake visit, carbon monoxide (CO) in expired air (Micro Smokerlyzer II, Bedfont Scientific Instruments) was taken as an index of recent smoking, with inclusion criteria of ≤5 ppm for nonsmokers and ≥10 ppm for smokers. Participants also completed questionnaires on demographic information, medical history, smoking history, handedness (Oldfield, 1971), childhood attention deficit hyperactivity disorder (Wender Utah Rating Scale; WURS) (Ward, Wender, & Reimherr, 1993), and depressive symptoms (Beck Depression Inventory; BDI) (Beck & Beamesderfer, 1974). A participant was excluded for having a score of ≥47 on the WURS and/or a score of ≥10 on the BDI. Intelligence was assessed through the Shipley Institute for Living Scale (Shipley Institute for Living Scale was administered only to a subset of research participants, 29 nonsmokers and 24 smokers), and those participants with scores ≤85 were excluded (Zachary, 1986). The Verbal Fluency Test was administered to non-native English speakers and those participants with scores ≤10 were excluded. All participants were free of current or past psychiatric illness, and had no history of alcohol or substance abuse. Urine drug screens were performed on both study days, and research participants positive for amphetamines, cocaine, marijuana, and/or opiates were excluded from further study. Of the 303 individuals who provided informed consent to participate, 68 smokers and 67 nonsmokers met the inclusion criteria. Twenty-seven smokers and 19 nonsmokers withdrew from participating. Data from four nicotine-dependent individuals were excluded because of errors in test administration or other technical problems.
Participants in both groups were tested in two blocks on each of two separate days. The smokers were tested after 13–16 hr abstinence on one day, and after 30–45 min abstinence on the other. On both days, the smokers smoked one cigarette (usual brand) during a break between the two blocks, and nonsmokers had a similar break in testing with no specific activity between blocks. Testing resumed within 5 min of the break. The order of sessions (overnight vs. brief abstinence) was counterbalanced for the smokers. Before each testing session, measurement of expired CO at a level of ≤10 ppm at the start of the ≥13-hr abstinence session verified compliance with the protocol. All testing sessions took place between 14:00 and 17:00 hr.
The d2 Test of Attention is a paper-pencil cancellation test, consisting of 14 lines, each with 47 letters, presented a total of 658 items (Brickenkamp & Zillmer, 1998). The participant is instructed to scan across each line with a time limit of 20 s per line, and to cross out all the target letters with a pencil. The target stimulus is the letter d accompanied with two dashes configured individually or in pairs above and/or below it, and the nontarget stimuli consist of a d with one or more than two dashes or a p with any number of dashes. Scores represent the total number of characters processed (errors excluded) within the designated time.
The Digit Symbol Test is a subset of the Wechsler Adult Intelligence Scale-Revised (WAIS-R), administered using paper and pencil (Wechsler, 1981). It involves copying abstract letter-like characters that are paired with the numbers 0 to 9. According to the instructions in the WAIS manual, four rows of empty boxes, labeled by a digit were presented on a sheet. A two-row code (with 9 digits in the upper row and 9 symbols in the lower row) was displayed at the top of the sheet. The instruction was to copy a symbol quickly into each empty box on the basis of the code at the top of the page. The score was computed as the number of correctly copied symbols within the 90-s time limit. Errors were extremely rare (<3%) and were excluded from the total number of copied symbols.
Cigarette craving was assessed using the self-report Urge to Smoke (UTS) scale (Jarvik et al., 2000). The UTS scale is a 10-item questionnaire and includes items such as “Do you have an urge for a cigarette right now?” and “Nothing would be better than smoking a cigarette right now.” The score of each item ranged from 1 to 7, with scores 1, 4, and 7 corresponding to definitely no craving symptom, possibly craving symptom, and definitely craving symptom, respectively. Severity of nicotine dependence was assessed using the Fagerström Test for Nicotine Dependence (FTND) (Heatherton, Kozlowski, Frecker, & Fagerström, 1991), with a maximum score possible of 10, and higher scores indicating greater severities of dependence. Cumulative smoking history was measured as pack-years, the product of packs of cigarettes smoked per day and the number of years smoked.
Demographic data were compared between groups using independent t-tests and chi-square. Task performance was analyzed using two-way repeated measures analyses of covariance (ANCOVA) to evaluate the main effects and interactions of group (smokers, nonsmokers), session (overnight abstinence, brief abstinence), and block (before vs. after the cigarette break), with age and gender as covariates. We used age and sex as covariates because of literature showing effects of both on performance on both tasks, and imbalance in the groups related to these two factors. For example, age is associated with slowness in performance of both the Digit Symbol Test and the d2 Test of Attention (Brickenkamp & Zillmer, 1998; Joy, Fein, Kaplan, & Freedman, 2000; Read et al., 2006). As the smokers in the present sample were older than the nonsmokers, we used age as a covariate to neutralize the effect of this variable. In light of evidence for better performance of women than men on both the Digit Symbol Test and the d2 Test of Attention (Rose et al., 2007; Royer, 1978), we used gender as a covariate to control for potential sex differences that may influence group effects. Since there was no experimental manipulation between the two sessions among the nonsmokers, the first session for half of the nonsmokers was nominally designated as corresponding to overnight abstinence as was the second session for the other half of nonsmokers. This was done so that we could include nonsmokers with smokers in repeated measures analyses of performance. Smoking status was used as a between-subjects factor and test session as a within-subjects factor.
Among nicotine-dependent individuals, scores on the UTS (craving) were analyzed using two-way repeated measures analyses of variance (ANOVA) to evaluate the main effect and interactions of session (overnight abstinence, brief abstinence), and block (before vs. after the cigarette break). Pearson’s correlations were then computed to assess whether cigarette craving scores were correlated with performance scores in the corresponding sessions and blocks. To test for association with level of nicotine dependence (FTND) and lifetime cigarette consumption (pack-years), partial correlations (adjusted for age) between these variables and task performance were computed as well.
An exploratory analysis based on the severity of nicotine dependence was also conducted, with low-moderate levels of nicotine dependence defined as scores ≤5 on the FTND and high levels having scores ≥6. Two-way ANCOVA (age and gender as a covariates) was used to evaluate the interactions and main effects of group (nonsmokers, low-to-moderate smokers, and high smokers), session (overnight abstinence, brief abstinence), and block (before vs. after the cigarette break). For all analyses, the criterion for statistical significance was p<.05 (two-tailed). Effect sizes are reported as partial eta-squares. Data were analyzed using SPSS 13 (SPSS Inc., Chicago, IL).
The two groups did not differ significantly in education, estimated intelligence, WURS or BDI scores (Table 1). They were unbalanced, however, with respect to age (nonsmokers, M=30.27, SD=9.9; smokers, M=35.08, SD=9.1; t(83)=2.3, p=.02) and gender (nonsmokers, 16 males, 32 females; smokers, 20 males, 17 females; χ2=3.68, p=.04). The smokers reported a current mean (SD) smoking level of 20.58 (5.33) cigarettes per day for an average of 14.7 (8.47) years, and the mean FTND score for the group was 5.28 (1.89).
ANCOVAs (age and gender as covariates), with performance score as the dependent variable, session and block as within-subject independent variables, and group as a between-subject independent variable, revealed no significant difference between smokers and nonsmokers on the d2 Test of Attention (F[1,81]=.01, p=.97) or the Digit Symbol Test (F[1,81]=.90, p=.34) (Figure 1).
In both tasks, there was no main effect of session (d2 Test of Attention, F[1,81]=.08, p=.77; Digit Symbol Test, F[1,81]=.03, p=.85) or interaction between session and group (d2 Test of Attention F[1,81]=.09, p=.92; Digit Symbol Test F[1,81]=.04, p=.83). There was a main effect for block, reflecting enhanced performance after the break (d2 Test of Attention, F[1,81]=13.85, p=.001, ηp2=.23; Digit Symbol Test, F[1,81]=11.28, p=.001, ηp2=.12), but no interaction between group and block (d2 Test of Attention, F[1,82]=2.50, p=.11; Digit Symbol Test, F[1,81]=1.54, p=.21), and no interaction of group by session by block (d2 Test of Attention, F[1,81]=.02, p=.89; Digit Symbol Test, F[1,81]=.54, p=.46).
Given that there was no main effect of session, and no interaction between session and block, performance scores were collapsed across both sessions and blocks to reflect a composite score. Partial correlation adjusted for age, revealed significant negative correlations of composite performance scores with the severity of nicotine dependence on both the d2 Test of Attention (r=−.44, p=.008) and the Digit Symbol Test (r=−.53, p=.001). Partial correlations corresponding to each session (overnight abstinence, relative satiety) and each block (before-break, after-break) showed significant negative associations as well. Performance scores were not correlated with pack-years for the d2 Test of Attention (r=−.21, p=.21) and the Digit Symbol Test (r=.002, p=.9).
The exploratory ANCOVAs revealed that the severity of nicotine dependence was a predictor of group differences on both the d2 Test of Attention (F[1,34]=5.23, p=.007, ηp2=.14) and the Digit Symbol Test (F[1,34]=6.86, p=.002, ηp2=.28). Planned contrasts were calculated for each outcome measure, comparing nicotine-dependent individuals with a high level of nicotine dependence severity (M=7, SD=1.03) to those with low-moderate levels (M=4.05, SD=1.32), and nonsmokers, respectively. Results revealed that smokers with a high level of nicotine dependence performed more poorly than those with low-to-moderate dependence levels and nonsmokers on both tests (p<.01, for all comparisons) (Figure 2). No group differences were found between smokers with low-moderate levels and those with a high level of nicotine dependence. There were no other significant main effects or interactions.
Self-reports of craving indicated significant main effects of session (F[1,35]=32.34, p=.001) and block (F[1,35]=84.23, p=.001); and the interaction between session and block was significant (F[1,35]=22.14, p=.001). Smokers had higher scores on the UTS scale following the overnight abstinence (M=6.1, SD=1.64) than after brief abstinence (M=4.4, SD=1). Smokers also reported a greater decrease in craving after the break following the overnight abstinence (M=3.0, SD=1.4) than after brief abstinence (M=2.4, SD=1.5). Pearson’s correlations among cigarette craving scores and perceptual-motor speed in the overnight abstinence session before break (d2 Test of Attention, r=−.12, p=.47, Digit Symbol Test, r=−.22, p=.19) and after break (d2 Test of Attention, r=.00, p=.96, Digit Symbol Test, r=−.01, p=.56) for both tests were not significant. The correlations among cigarette craving scores and perceptual-motor speed in the brief abstinence session before break (d2 Test of Attention, r=−.14, p=.40, Digit Symbol Test, r=−.19, p=.26) and after break (d2 Test of Attention, r=−.05, p=.75, Digit Symbol Test, r=−.26, p=.13) for both tests were not significant as well.
In this study, performance of smokers on both the d2 Test of Attention and the Digit Symbol Test was correlated negatively with severity of nicotine dependence, but not with history of smoking or with craving. Although the FTND and the pack-years measures contain an overlapping component (e.g., number of cigarettes consumed per day), the FTND includes five other items that address other behavioral characteristics of dependence. For example, FTND includes items such as: “How soon after you wake up do you smoke your first cigarette?” and “Do you find it difficult not to smoke in places where you shouldn’t, such as in church or school, in a movie, at the library, on a bus, in court or in a hospital?”
Our findings support that severity of nicotine dependence is a stronger predictor of perceptual-motor speed than how much or how long a person smokes. Although the present data cannot identify the source of the observed slowness in performance, the findings can help to characterize some of the individual differences that contribute to perceptual-motor processes among nicotine-dependent individuals.
The group differences in perceptual-motor speed may reflect toxic effects of smoking or may pre-date nicotine dependence. Smokers differ from nonsmokers with respect to brain structure (Brody et al., 2004) and metabolite concentrations (Durazzo, Gazdzinski, Banys, & Meyerhoff, 2004; Gallinat et al., 2007), as well as in neural activation in response to smoking-related stimuli (David et al., 2005; Due, Huettel, Hall, & Rubin, 2002) and during performance of cognitive tasks (Lawrence, Ross, & Stein, 2002; Xu et al., 2007). While providing consistent findings of gray matter deficits in smokers (Brody et al., 2004; Gallinat et al., 2006), these studies lack longitudinal findings or assessments prior to the initiation of smoking. Therefore, they do not show whether the reported group differences in brain structure reflect risk factors, consequences of smoking, or a combination of factors.
Evidence from family, twin, and adoption studies implies a biological predisposition to nicotine dependence (Batra, Patkar, Berrettini, Weinstein, & Leone, 2003; Broms, Silventoinen, Madden, Heath, & Kaprio, 2006). Although we know of no evidence of specific genetic determinants of speed on perceptual-motor tasks of attention, it is plausible that the underlying processes have heritable components, and that genetic mediators of slowness overlap with mediators of severity of nicotine dependence.
Research indicates that cigarette craving disturbs cognition, notably in tests of sustained attention (Cepeda-Benito & Tiffany, 1996; Hendricks, Ditre, Drobes, & Brandon, 2006; Mogg, Field, & Bradley, 2005). In the present study, however, abstinence-induced craving was not related to task performance. This null finding is inconsistent with previous demonstrations that nicotine deprivation and craving impairs attention in smokers. One possibility may be that while 13 hr of abstinence may be detrimental under complex task conditions, such as those that required working memory (Mendrek et al., 2006), it does not disturb performance in perceptual-motor tasks.
There was no evidence of gain in performance achieved by smoking a cigarette similar to those reported in studies of attention and choice reaction time (Parrott et al., 1996). The improvement in performance between the test blocks reflected practice effects rather than an intervening effect of smoking, as improvement also was manifested by nonsmokers, who did not smoke between blocks. These findings suggest that the acute beneficial effects of smoking are selective to specific cognitive operations. This study revealed that the severity of nicotine dependence was a reliable indicator of perceptual-motor speed irrespective of chronic smoking, craving, and overnight or brief abstinence. The mechanism linking severity of nicotine dependence with cognitive performance is unknown and warrants further investigation. In a recent functional neuroimaging study, the severity of nicotine dependence was positively correlated with blood oxygen level-dependent response to smoking cues in the brain regions (dorsal anterior cingulate cortex, inferior parietal cortex, secondary visual areas, parahippocampal, fusiform gyri) that function in visuospatial attention (Smolka et al., 2006). In another study, the same brain regions were activated in tasks that require sustained attention and selecting rare targets (Kirino, Belger, Goldman-Rakic, & McCarthy, 2000). Other research suggests that faster-performing individuals show less neural activity than slower performers (Rypma et al., 2006). The incorporation of functional neuro-imaging into investigations of the relationship between perceptual-motor performance and severity of nicotine-dependence may help identify the neural links between slowness and nicotine-dependence.
This study had some limitations. One of these is the potential confound of improvement with repeated testing on neuropsychological instruments (Hinton-Bayre & Geffen, 2005). Another limitation is that perceptual-motor speed was measured using two tests. Further investigations with other neuropsychological instruments (e.g., Trail Making, Finger Tapping) that measure attentional and perceptual-motor processes are needed to confirm that the observed relationship between severity of nicotine dependence and speed is not specific to the particular tasks used.
In summary, the performance of smokers on the d2 Test of Attention and the Digit Symbol Test correlated negatively with severity of nicotine dependence but not lifetime consumption or cigarette craving. There were no significant effects of abstinence or of smoking one cigarette on the performance of smokers. The findings suggest that severity of nicotine-dependence that could explain some of the variance in perceptual-motor speed. The etiology of slowness remains unclear and may reflect the toxic consequences of smoking or factors that predispose a person to smoke. The extent to which smoking cessation can reverse these deficits merits further examination in longitudinal and smoking cessation studies.
Supported by NIH grants R01 DA014093 (EDL), R01 DA20872 (ALB); UC Tobacco-Related Disease Research Program Award 10RT-0091, UCLA GCRC MOI RR 00865; a Department of Veterans Affairs Merit Review Award (ALB); T32 DA07272 (AA); and Philip Morris USA and Philip Morris International (EDL).
Allen Azizian, Departments of Psychiatry and Biobehavioral Sciences, Laboratory of Molecular Neuroimaging, UCLA Semel Institute for Neuroscience and Human Behavior, University of California, Los Angeles, CA.
John R. Monterosso, Departments of Psychiatry and Biobehavioral Sciences, Laboratory of Molecular Neuroimaging, UCLA Semel Institute for Neuroscience and Human Behavior, University of California, Los Angeles, CA.
Arthur L. Brody, Departments of Psychiatry and Biobehavioral Sciences, Laboratory of Molecular Neuroimaging, UCLA Semel Institute for Neuroscience and Human Behavior, University of California, Los Angeles, CA.
Sara L. Simon, Departments of Psychiatry and Biobehavioral Sciences, Laboratory of Molecular Neuroimaging, UCLA Semel Institute for Neuroscience and Human Behavior, University of California, Los Angeles, CA.
Edythe D. London, Departments of Psychiatry and Biobehavioral Sciences, Laboratory of Molecular Neuroimaging, UCLA Semel Institute for Neuroscience and Human Behavior, and Molecular and Medical Pharmacology, David Geffen School of Medicine, Brain Research Institute, and Biomedical Engineering Interdepartmental Faculty, University of California, Los Angeles, CA.