Impulsivity as a determinant and consequence of drug use: a review of underlying processes

Abstract

Impulsive behaviors are closely linked to drug use and abuse, both as contributors to use and as consequences of use. Trait impulsivity is an important determinant of drug use during development, and in adults momentary 'state' increases in impulsive behavior may increase the likelihood of drug use, especially in individuals attempting to abstain. Conversely, acute and chronic effects of drug use may increase impulsive behaviors, which may in turn facilitate further drug use. However, these effects depend on the behavioral measure used to assess impulsivity. This article reviews data from controlled studies investigating different measures of impulsive behaviors, including delay discounting, behavioral inhibition and a newly proposed measure of inattention. Our findings support the hypothesis that drugs of abuse alter performance across independent behavioral measures of impulsivity. The findings lay the groundwork for studying the cognitive and neurobiological substrates of impulsivity, and for future studies on the role of impulsive behavior as both facilitator and a result of drug use.

Figure 1

Schematic of relationships between observable ‘impulsive’ behaviors, processes that may lead to these behaviors and laboratory tasks that measure the processes. ‘Difficulty awaiting turn’ refers to a common symptom of ADHD, ‘difficulty resisting drink’ refers to a drug-related behavior that involves impulsivity. Any of the three processes, behavioral inhibition, delay discounting and inattention, may contribute to these maladaptive behaviors

Figure 2

Schematic of delay discounting. The graph shows the indifference points, or points of equivalence, from choices between immediate (smaller) rewards and delayed (larger) rewards. More impulsive individuals ‘discount’ delayed rewards more steeply

Figure 3

Schematic of the Stop Task procedure. Subjects are presented with either Go trials (75% of trials; top panel) or Stop trials (25% of trials; bottom panel). A simple Go reaction time is calculated from the Go trials. On Stop trials, subject is give a Stop signal shortly after the Go signal, and instructed to inhibit their ‘Go’ response on these trials. The time between the Go and Stop signal is varied systematically until the delay is reached at which the subject is able to inhibit the response on 50% of trials. This provides an estimate of the time needed to stop a response. Longer stop times are indices of poorer inhibitory control

Figure 4

Schematic of lapses of attention. This figure shows two hypothetical distributions of simple reaction times, including the mean and the mode. The left panel shows the distribution without long reaction times, the right panel shows the separation of the mode and the mean when there are long reaction times or ‘lapses in attention’. It shows that long reaction times change the mean while leaving the mode relatively unaffected. The difference between the mean and the mode provides a measure of the skew, and the deviation from the mode (DevMod) is considered a measure of inattention. The inset shows the symbols to be used for mode, mean and DevMod in Figs 5, 6 and 7

Figure 5

Effects of sleep deprivation on lapses of attention. This figure shows the mean of the mode (circles), deviation from the mode (DevMod; horizontal line) and mean (squares) of reaction time (RT) distributions after 24 hours of sleep deprivation and after a normal night’s sleep. Healthy volunteers performed simple RT tasks at regular intervals on the day after sleep and no sleep. The mean RT = mode +DevMod. Therefore, the horizontal line connecting the mode to the mean RT corresponds to the DevMod. Filled symbols indicate significant differences between the normal sleep and deprivation conditions, at the corresponding timepoints (P < 0.05). A heavy horizontal line indicates that the DevMod is significantly different in the two conditions. The mean RT was increased at all four timepoints. Although there were small increases in modal RT, the plots indicate that most of the increase in mean RT was because of the increases in the DevMod measure, indicating more frequent long RTs

Figure 6

Effects of bupropion (BUP) and amphetamine (AMPH) on lapses of attention. This figure shows the means of the mode (circle), deviation from the mode (DevMod; horizontal line) and mean (square) of reaction time (RT) distributions after placebo, bupropion (150 mg) and d-amphetamine (20 mg). The mean RT = mode +DevMod. Filled symbols indicate significant differences between the drug and placebo conditions, at the corresponding timepoints (P <0.05). A heavy horizontal line indicates that the DevMod is significantly different in the two conditions. Both bupropion and d-amphetamine decreased the mean RT. However, bupropion selectively decreased the DevMod measure, whereas d-amphetamine decreased both the modal RT and the DevMod measure

Figure 7

Effects of caffeine on lapses of attention in fatigued subjects. Healthy volunteers were tested on two sessions from 5 p.m. to 5 a.m. At 3 a.m. they ingested capsules with either caffeine (200 mg) or placebo. The mean reaction time (RT) = mode +deviation from the mode (DevMod). Therefore, the horizontal line connecting the mode to the mean RT corresponds to the DevMod. Filled symbols indicate significant differences between the placebo and caffeine conditions, at the corresponding timepoints (P <0.05). A heavy horizontal line indicates that the DevMod is significantly different in the two conditions. This figure shows that reaction times increased over the course of both sessions. Caffeine significantly decreased both the mode and the mean reaction time, relative to the same timepoint in the placebo condition. Caffeine did not affect the DevMod, indicating that it did not specifically decreases lapses in attention