Are Prescription Stimulants “Smart Pills”?

Large U.S. Government Surveys

Most epidemiological research on nonmedical stimulant use has been focused on issues relevant to traditional problems of drug abuse and addiction, and so, stimulant use for cognitive enhancement is not generally distinguished from use for other purposes, such as staying awake or getting high. As Boyd and McCabe (2008) pointed out, the large national surveys of nonmedical prescription drug use have so far failed to distinguish the ways and reasons that people use the drugs, and this is certainly true where prescription stimulants are concerned. The largest survey to investigate prescription stimulant use in a nationally representative sample of Americans, the National Survey on Drug Use and Health (NSDUH), phrases the question about nonmedical use as follows: “Have you ever, even once, used any of these stimulants when they were not prescribed for you or that you took only for the experience or feeling they caused?” (Snodgrass & LeBaron 2007). This phrasing does not strictly exclude use for cognitive enhancement, but it emphasizes the noncognitive effects of the drugs. In 2008, the NSDUH found a prevalence of 8.5% for lifetime nonmedical stimulant use by Americans over the age of 12 years and a prevalence of 12.3% for Americans between 21 and 25 (Substance Abuse and Mental Health Services Administration, 2009).

In addition, large national surveys, including the NSDUH, have generally classified prescription stimulants with other stimulants including street drugs such as methamphetamine. For example, since 1975, the National Institute on Drug Abuse–sponsored Monitoring the Future (MTF) survey has gathered data on drug use by young people in the United States (Johnston, O’Malley, Bachman, & Schulenberg, 2009a, 2009b). Originally, MTF grouped prescription stimulants under a broader class of stimulants so that respondents were asked specifically about MPH only after they had indicated use of some drug in the category of AMPs. As rates of MPH prescriptions increased and anecdotal reports of nonmedical use grew, the 2001 version of the survey was changed to include a separate standalone question about MPH use. This resulted in more than a doubling of estimated annual use among 12th graders, from 2.4% to 5.1%. More recent data from the MTF suggests Ritalin use has declined (3.4% in 2008). However, this may still underestimate use of MPH, as the question refers specifically to Ritalin and does not include other brand names such as Concerta (an extended release formulation of MPH).

Turning to analyses related specifically to the drugs that are the subject of this article, reanalysis of the 2002 NSDUH data by Kroutil and colleagues (2006) found past-year nonmedical use of stimulants other than methamphetamine by 2% of individuals between the ages of 18 and 25 and by 0.3% of individuals 26 years of age and older. For ADHD medications in particular, these rates were 1.3% and 0.1%, respectively. Finally, Novak, Kroutil, Williams, and Van Brunt (2007) surveyed a sample of over four thousand individuals from the Harris Poll Online Panel and found that 4.3% of those surveyed between the ages of 18 and 25 had used prescription stimulants nonmedically in the past year, compared with only 1.3% between the ages of 26 and 49.

Surveys of Student Nonmedical Stimulant Use

We reviewed recent studies concerning prescription stimulant use specifically among students in the United States and Canada, using the method illustrated in Figure 1. Although less informative about the general population, these studies included questions about students’ specific reasons for using the drugs, as well as frequency of use and means of obtaining them. These studies typically found rates of use greater than those reported by the nationwide NSDUH or the MTF surveys. This probably reflects a true difference in rates of usage among the different populations. In support of that conclusion, the NSDUH data for college age Americans showed that college students were considerably more likely than nonstudents of the same age to use prescription stimulants nonmedically (odds ratio: 2.76; Herman-Stahl, Krebs, Kroutil, & Heller, 2007).

Table 1 shows all of the studies of middle school, secondary school, and college students that we identified. As indicated in the table, the studies are heterogeneous, with varying populations sampled, sample sizes, and year of data collection, and they focused on different subsets of the epidemiological questions addressed here, including prevalence and frequency of use, motivations for use, and method of obtaining the medication.

Do Prescription Stimulants Enhance Learning?

Most people would describe school as a place where they go to learn, so learning is an especially relevant cognitive process for students to enhance. Even outside of school, however, learning plays a role in most activities, and the ability to enhance the retention of information would be of value in many different occupational and recreational contexts.

Table 3 lists the results of 24 tasks from 22 articles on the effects of d-AMP or MPH on learning, assessed by a variety of declarative and nondeclarative memory tasks. Results for the 24 tasks are evenly split between enhanced learning and null results, but they yield a clearer pattern when the nature of the learning task and the retention interval are taken into account. In general, with single exposures of verbal material, no benefits are seen immediately following learning, but later recall and recognition are enhanced. Of the six articles reporting on memory performance (Camp-Bruno & Herting, 1994; Fleming, Bigelow, Weinberger, & Goldberg, 1995; Rapoport, Busbaum, & Weingartner, 1980; Soetens, D’Hooge, & Hueting, 1993; Unrug, Coenen, & van Luijtelaar, 1997; Zeeuws & Soetens 2007), encompassing eight separate experiments, only one of the experiments yielded significant memory enhancement at short delays (Rapoport et al., 1980). In contrast, retention was reliably enhanced by d-AMP when subjects were tested after longer delays, with recall improved after 1 hr through 1 week (Soetens, Casaer, D’Hooge, & Hueting, 1995; Soetens et al., 1993; Zeeuws & Soetens, 2007). Recognition improved after 1 week in one study (Soetens et al., 1995), while another found recognition improved after 2 hr (Mintzer & Griffiths, 2007). The one long-term memory study to examine the effects of MPH found a borderline-significant reduction in errors when subjects answered questions about a story (accompanied by slides) presented 1 week before (Brignell, Rosenthal, & Curran, 2007).

Enhanced learning was also observed in two studies that involved multiple repeated encoding opportunities. Camp-Bruno and Herting (1994) found MPH enhanced summed recall in the Buschke Selective Reminding Test (Buschke, 1973; Buschke & Fuld, 1974) when 1-hr and 2-hr delays were combined, although individually only the 2-hr delay approached significance. Likewise, de Wit, Enggasser, and Richards (2002) found no effect of d-AMP on the Hopkins Verbal Learning Test (Brandt, 1991) after a 25-min delay. Willett (1962) tested rote learning of nonsense syllables with repeated presentations, and his results indicate that d-AMP decreased the number of trials needed to reach criterion.

In paired-associates learning, subjects are presented with pairs of stimuli and must learn to recall the second item of the pair when presented with the first. For these tasks, as with tasks involving memory for individual items, there is a trend for stimulants to enhance performance with longer delays. For immediate measures of learning, no effects of d-AMP or MPH were observed by Brumaghim and Klorman (1998); Fleming et al. (1995); Hurst, Radlow, and Weidner (1968); or Strauss et al. (1984). However, when Hurst et al.’s subjects were tested a week later, they recalled more if their initial learning had been carried out with d-AMP than with placebo. Weitzner (1965) assessed paired-associates learning with an immediate cued-recall test and found facilitation when the associate word was semantically related to the cue, provided it was not also related to other cue words. Finally, Burns, House, French, and Miller (1967) found a borderline-significant impairment of performance with d-AMP on a nonverbal associative learning task.

The information learned in the tasks reviewed so far was explicit, declarative, and consistent within each experiment. In contrast, probabilistic and procedural learning tasks require the subject to gradually extract a regularity in the associations among stimuli from multiple presentations in which the correct associations are only presented some of the time, with incorrect associations also presented. Findings are mixed in these tasks. Breitenstein and colleagues (2004, 2006) showed subjects drawings of common objects accompanied by nonsense word sounds in training sessions that extended over multiple days. They found faster learning of the to-be-learned, higher probability pairings between sessions (consistent with enhanced retention over longer delays). Breitenstein et al. (2004) found that this enhancement remained a year later. Schlösser et al. (2009) tested subjects’ probabilistic learning ability in the context of a functional magnetic resonance imaging (fMRI) study, comparing performance and brain activation with MPH and placebo. MPH did not affect learning performance as measured by accuracy. Although subjects were overall faster in responding on MPH, this difference was independent of the difficulty of the learning task, and the authors accordingly attributed it to response processes rather than learning.

Two additional studies assessed the effects of d-AMP on visual–motor sequence learning, a form of nondeclarative, procedural learning, and found no effect (Kumari et al., 1997; Makris, Rush, Frederich, Taylor, & Kelly, 2007). In a related experimental paradigm, Ward, Kelly, Foltin, and Fischman (1997) assessed the effect of d-AMP on the learning of motor sequences from immediate feedback and also failed to find an effect.

A provisional conclusion about the effects of stimulants on learning is that they do help with the consolidation of declarative learning, with effect sizes varying widely from small to large depending on the task and individual study. Indeed, as a practical matter, stimulants may be more helpful than many of the laboratory tasks indicate, given the apparent dependence of enhancement on length of delay before testing. Although, as a matter of convenience, experimenters tend to test memory for learned material soon after the learning, this method has not generally demonstrated stimulant-enhanced learning. However, when longer periods intervene between learning and test, a more robust enhancement effect can be seen. Note that the persistence of the enhancement effect well past the time of drug action implies that state-dependent learning is not responsible. In general, long-term effects on learning are of greater practical value to people. Even students cramming for exams need to retain information for more than an hour or two. We therefore conclude that stimulant medication does enhance learning in ways that may be useful in the real world.

Do Prescription Stimulants Enhance Working Memory?

In contrast to the types of memory discussed in the previous section, which are long-lasting and formed as a result of learning, working memory is a temporary store of information. Working memory has been studied extensively by cognitive psychologists and cognitive neuroscientists because of its role in executive function. It has been likened to an internal scratch pad; by holding information in working memory, one keeps it available to consult and manipulate in the service of performing tasks as diverse as parsing a sentence and planning a route through the environment. Presumably for this reason, working memory ability correlates with measures of general intelligence (Friedman et al., 2006). The possibility of enhancing working memory ability is therefore of potential real-world interest.

Many laboratory tasks have been developed to study working memory, each of which taxes to varying degrees aspects such as the overall capacity of working memory, its persistence over time, and its resistance to interference either from task-irrelevant stimuli or among the items to be retained in working memory (i.e., cross-talk). Tasks also vary in the types of information to be retained in working memory, for example, verbal or spatial information. The question of which of these task differences correspond to differences between distinct working memory systems and which correspond to different ways of using a single underlying system is a matter of debate (e.g., D’Esposito, Postle, & Rypma, 2000; Owen, 2000). For the present purpose, we ignore this question and simply ask, Do MPH and d-AMP affect performance in the wide array of tasks that have been taken to operationalize working memory? If the literature does not yield a unanimous answer to this question, then what factors might be critical in determining whether stimulant effects are manifest?

Table 4 lists the results of 27 tasks from 23 articles on the effects of d-AMP or MPH on working memory. The oldest and most commonly used type of working memory task in this literature is the Sternberg short-term memory scanning paradigm (Sternberg, 1966), in which subjects hold a set of items (typically letters or numbers) in working memory and are then presented with probe items, to which they must respond “yes” (in the set) or “no” (not in the set). The size of the set, and hence the working memory demand, is sometimes varied, and the set itself may be varied from trial to trial to maximize working memory demands or may remain fixed over a block of trials. Taken together, the studies that have used a version of this task to test the effects of MPH and d-AMP on working memory have found mixed and somewhat ambiguous results. No pattern is apparent concerning the specific version of the task or the specific drug. Four studies found no effect (Callaway, 1983; Kennedy, Odenheimer, Baltzley, Dunlap, & Wood, 1990; Mintzer & Griffiths, 2007; Tipper et al., 2005), three found faster responses with the drugs (Fitzpatrick, Klorman, Brumaghim, & Keefover, 1988; Ward et al., 1997; D. E. Wilson et al., 1971), and one found higher accuracy in some testing sessions at some dosages, but no main effect of drug (Makris et al., 2007). The meaningfulness of the increased speed of responding is uncertain, given that it could reflect speeding of general response processes rather than working memory–related processes. Aspects of the results of two studies suggest that the effects are likely due to processes other than working memory: D. E. Wilson et al. (1971) reported comparable speeding in a simple task without working memory demands, and Tipper et al. (2005) reported comparable speeding across set sizes.

Another common working memory task is the n-back task, which requires the subject to view a series of items (usually letters) and decide whether the current item is identical to the one presented n items back. This task taxes working memory because the previous items must be held in working memory to be compared with the current item. The easiest version of this is a 1-back task, which is also called a double continuous performance task (CPT) because the subject is continuously monitoring for a repeat or double. Three studies examined the effects of MPH on working memory ability as measured by the 1-back task, and all found enhancement of performance in the form of reduced errors of omission (Cooper et al., 2005; Klorman et al., 1984; Strauss et al., 1984). Fleming et al. (1995) tested the effects of d-AMP on a 5-min CPT and found a decrease in reaction time, but did not specify which version of the CPT was used.

The data from 2-back and 3-back tasks are more complex. Three studies examined performance in these more challenging tasks and found no effect of d-AMP on average performance (Mattay et al., 2000, 2003; Mintzer & Griffiths, 2007). However, in at least two of the studies, the overall null result reflected a mixture of reliably enhancing and impairing effects. Mattay et al. (2000) examined the performance of subjects with better and worse working memory capacity separately and found that subjects whose performance on placebo was low performed better on d-AMP, whereas subjects whose performance on placebo was high were unaffected by d-AMP on the 2-back and impaired on the 3-back tasks. Mattay et al. (2003) replicated this general pattern of data with subjects divided according to genotype. The specific gene of interest codes for the production of Catechol-O-methyltransferase (COMT), an enzyme that breaks down dopamine and norepinephrine. A common polymorphism determines the activity of the enzyme, with a substitution of methionine for valine at Codon 158 resulting in a less active form of COMT. The met allele is thus associated with less breakdown of dopamine and hence higher levels of synaptic dopamine than the val allele. Mattay et al. (2003) found that subjects who were homozygous for the val allele were able to perform the n-back faster with d-AMP; those homozygous for met were not helped by the drug and became significantly less accurate in the 3-back condition with d-AMP. In the case of the third study finding no overall effect, analyses of individual differences were not reported (Mintzer & Griffiths, 2007).

A related task is the B–X version of the CPT, in which subjects must respond when an X appears only if it was preceded by a B. As in the 1-back task, the subject must retain the previous trial’s letter in working memory because it determines the subject’s response to the current letter. In this case, when the current letter is an X, then the subject should respond only if the previous letter was a B. Two studies examined stimulant effects in this task. Rapoport et al. (1980) found that d-AMP reduced errors of omission in the longer of two test sessions, and Klorman et al. (1984) found that MPH reduced errors of omission and response time.

Another classic approach to the assessment of working memory is the span task, in which a series of items is presented to the subject for repetition, transcription, or recognition. The longest series that can be reproduced accurately is called the forward span and is a measure of working memory capacity. The ability to reproduce the series in reverse order is tested in backward span tasks and is a more stringent test of working memory capacity and perhaps other working memory functions as well. The digit span task from the Wechsler (1981) IQ test was used in four studies of stimulant effects on working memory. One study showed that d-AMP increased digit span (de Wit et al., 2002), and three found no effects of d-AMP or MPH (Oken, Kishiyama, & Salinsky, 1995; Schmedtje, Oman, Letz, & Baker, 1988; Silber, Croft, Papafotiou, & Stough, 2006). A spatial span task, in which subjects must retain and reproduce the order in which boxes in a scattered spatial arrangement change color, was used by Elliott et al. (1997) to assess the effects of MPH on working memory. For subjects in the group receiving placebo first, MPH increased spatial span. However, for the subjects who received MPH first, there was a nonsignificant opposite trend. The group difference in drug effect is not easily explained. The authors noted that the subjects in the first group performed at an overall lower level, and so, this may be another manifestation of the trend for a larger enhancement effect for less able subjects.

Several studies have assessed the effect of MPH and d-AMP on tasks tapping various other aspects of spatial working memory. Three used the spatial working memory task from the CANTAB battery of neuropsychological tests (Sahakian & Owen, 1992). In this task, subjects search for a target at different locations on a screen. Subjects are told that locations containing a target in previous trials will not contain a target in future trials. Efficient performance therefore requires remembering and avoiding these locations in addition to remembering and avoiding locations already searched within a trial. Mehta et al. (2000) found evidence of greater accuracy with MPH, and Elliott et al. (1997) found a trend for the same. In Mehta et al.’s study, this effect depended on subjects’ working memory ability: the lower a subject’s score on placebo, the greater the improvement on MPH. In Elliott et al.’s study, MPH enhanced performance for the group of subjects who received the placebo first and made little difference for the other group. The reason for this difference is unclear, but as mentioned above, this may reflect ability differences between the groups. More recently, Clatworthy et al. (2009) undertook a positron emission tomography (PET) study of MPH effects on two tasks, one of which was the CANTAB spatial working memory task. They failed to find consistent effects of MPH on working memory performance but did find a systematic relation between the performance effect of the drug in each individual and its effect on individuals’ dopamine activity in the ventral striatum.

Two additional studies used other spatial working memory tasks. Barch and Carter (2005) required subjects to maintain one of 18 locations on the perimeter of a circle in working memory and then report the name of the letter that appeared there in a similarly arranged circle of letters. d-AMP caused a speeding of responses but no change in accuracy. Fleming et al. (1995) referred to a spatial delay response task, with no further description or citation. They reported no effect of d-AMP in the task except in the zero-delay condition (which presumably places minimal demand on working memory).

Finally, Schmedtje et al. (1988) used a delayed matching task in which one random pattern of filled squares in a 10 × 10 matrix must be held in working memory and compared with another and judged same or different. No effect was observed with a fairly low dose of d-AMP.

In sum, the evidence concerning stimulant effects of working memory is mixed, with some findings of enhancement and some null results, although no findings of overall performance impairment. A few studies showed greater enhancement for less able participants, including two studies reporting overall null results. When significant effects have been found, their sizes vary from small to large, as shown in Table 4. Taken together, these results suggest that stimulants probably do enhance working memory, at least for some individuals in some task contexts, although the effects are not so large or reliable as to be observable in all or even most working memory studies.

Do Prescription Stimulants Enhance Cognitive Control?

Cognitive control is a broad concept that refers to guidance of cognitive processes in situations where the most natural, automatic, or available action is not necessarily the correct one. Such situations typically evoke a strong inclination to respond but require people to resist responding, or they evoke a strong inclination to carry out one type of action but require a different type of action. The sources of these inclinations that must be overridden are various and include overlearning (e.g., the overlearned tendency to read printed words in the Stroop task), priming by recent practice (e.g., the tendency to respond in the go/no-go task when the majority of the trials are go trials, or the tendency to continue sorting cards according to the previously correct dimension in the Wisconsin Card Sorting Test [WCST]; Grant & Berg, 1948) and perceptual salience (e.g., the tendency to respond to the numerous flanker stimuli as opposed to the single target stimulus in the flanker task). For the sake of inclusiveness, we also consider the results of studies of reward processing in this section, in which the response tendency to be overridden comes from the desire to have the reward immediately.

Table 5 lists the results of 16 tasks from 13 articles on the effects of d-AMP or MPH on cognitive control. One of the simplest tasks used to study cognitive control is the go/no-go task. Subjects are instructed to press a button as quickly as possible for one stimulus or class of stimuli (go) and to refrain from pressing for another stimulus or class of stimuli (no go). De Wit et al. (2002) used a version of this task to measure the effects of d-AMP on subjects’ ability to inhibit a response and found enhancement in the form of decreased false alarms (responses to no-go stimuli) and increased speed of correct go responses. They also found that subjects who made the most errors on placebo experienced the greatest enhancement from the drug.

The stop-signal task has been used in a number of laboratories to study the effects of stimulants on cognitive control. In this task, subjects are instructed to respond as quickly as possible by button press to target stimuli except on certain trials, when the target is followed by a stop signal. On those trials, they must try to avoid responding. The stop signal can follow the target stimulus almost immediately, in which case it is fairly easy for subjects to cancel their response, or it can come later, in which case subjects may fail to inhibit their response. The main dependent measure for stop-signal task performance is the stop time, which is the average go reaction time minus the interval between the target and stop signal at which subjects inhibit 50% of their responses. De Wit and colleagues have published two studies of the effects of d-AMP on this task. De Wit, Crean, and Richards (2000) reported no significant effect of the drug on stop time for their subjects overall but a significant effect on the half of the subjects who were slowest in stopping on the baseline trials. De Wit et al. (2002) found an overall improvement in stop time in addition to replicating their earlier finding that this was primarily the result of enhancement for the subjects who were initially the slowest stoppers. In contrast, Filmore, Kelly, and Martin (2005) used a different measure of cognitive control in this task, simply the number of failures to stop, and reported no effects of d-AMP.

The Stroop task tests the ability to inhibit the overlearned process of reading by presenting color names in colored ink and instructing subjects to either read the word (low need for cognitive control because this is the habitual response to printed words) or name the ink color (high need for cognitive control). Barch and Carter (2005) administered this task to normal control subjects on placebo and d-AMP and found speeding of responses with the drug. However, the speeding was roughly equivalent for the conditions with low and high cognitive control demands, suggesting that the observed facilitation may not have been specific to cognitive control.

As mentioned earlier, cognitive control is needed not only for inhibiting actions, but also for shifting from one kind of action or mental set to another. The WCST taxes cognitive control by requiring the subject to shift from sorting cards by one dimension (e.g., shape) to another (e.g., color); failures of cognitive control in this task are manifest as perseverative errors in which subjects continue sorting by the previously successful dimension. Three studies included the WCST in their investigations of the effects of d-AMP on cognition (Fleming et al., 1995; Mattay et al., 1996, 2003), and none revealed overall effects of facilitation. However, Mattay et al. (2003) subdivided their subjects according to COMT genotype and found differences in both placebo performance and effects of the drug. Subjects who were homozygous for the val allele (associated with lower prefrontal dopamine activity) made more perseverative errors on placebo than other subjects and improved significantly with d-AMP. Subjects who were homozygous for the met allele performed best on placebo and made more errors on d-AMP.

The intradimensional– extradimensional shift task from the CANTAB battery was used in two studies of MPH and measures the ability to shift the response criterion from one dimension to another, as in the WCST, as well as to measure other abilities, including reversal learning, measured by performance in the trials following an intradimensional shift. With an intradimensional shift, the learned association between values of a given stimulus dimension and reward versus no reward is reversed, and participants must learn to reverse their responses accordingly. Elliott et al. (1997) reported finding no effects of the drug on ability to shift among dimensions in the extradimensional shift condition and did not describe performance on the intradimensional shift. Rogers et al. (1999) found that accuracy improved but responses slowed with MPH on trials requiring a shift from one dimension to another, which leaves open the question of whether the drug produced net enhancement, interference, or neither on these trials once the tradeoff between speed and accuracy is taken into account. For intradimensional shifts, which require reversal learning, these authors found drug-induced impairment: significantly slower responding accompanied by a borderline-significant impairment of accuracy.

Two studies investigated the effects of MPH on reversal learning in simple two-choice tasks (Clatworthy et al., 2009; Dodds et al., 2008). In these tasks, participants begin by choosing one of two stimuli and, after repeated trials with these stimuli, learn that one is usually rewarded and the other is usually not. The rewarded and nonrewarded stimuli are then reversed, and participants must then learn to choose the new rewarded stimulus. Although each of these studies found functional neuroimaging correlates of the effects of MPH on task-related brain activity (increased blood oxygenation level-dependent signal in frontal and striatal regions associated with task performance found by Dodds et al., 2008, using fMRI and increased dopamine release in the striatum as measured by increased raclopride displacement by Clatworthy et al., 2009, using PET), neither found reliable effects on behavioral performance in these tasks. The one significant result concerning purely behavioral measures was Clatworthy et al.’s (2009) finding that participants who scored higher on a self-report personality measure of impulsivity showed more performance enhancement with MPH. MPH’s effect on performance in individuals was also related to its effects on individuals’ dopamine activity in specific regions of the caudate nucleus.

The Trail Making Test is a paper-and-pencil neuropsychological test with two parts, one of which requires shifting between stimulus categories. Part A simply requires the subject to connect circled numbers in ascending order. Part B requires the subject to connect circled numbers and letters in an interleaved ascending order (1, A, 2, B, 3, C….), a task that places heavier demands on cognitive control. Silber et al. (2006) analyzed the effect of d-AMP on Trails A and B and failed to find an effect.

The flanker task is designed to tax cognitive control by requiring subjects to respond based on the identity of a target stimulus (H or S) and not the more numerous and visually salient stimuli that flank the target (as in a display such as HHHSHHH). Servan-Schreiber, Carter, Bruno, and Cohen (1998) administered the flanker task to subjects on placebo and d-AMP. They found an overall speeding of responses but, more importantly, an increase in accuracy that was disproportionate for the incongruent conditions, that is, the conditions in which the target and flankers did not match and cognitive control was needed.

Perceptual–motor congruency was the basis of a study by Fitzpatrick et al. (1988) in which subjects had to press buttons to indicate the location of a target stimulus in a display. In the simple condition, the left-to-right positions of the buttons are used to indicate the left-to-right positions of the stimuli, a natural mapping that requires little cognitive control. In the rotation condition, the mapping between buttons and stimulus positions is shifted to the right by one and wrapped around, such that the left-most button is used to indicate the right-most position. Cognitive control is needed to resist responding with the other, more natural mapping. MPH was found to speed responses in this task, and the speeding was disproportionate for the rotation condition, consistent with enhancement of cognitive control.

Finally, two tasks measuring subjects’ ability to control their responses to monetary rewards were used by de Wit et al. (2002) to assess the effects of d-AMP. When subjects were offered the choice between waiting 10 s between button presses for high-probability rewards, which would ultimately result in more money, and pressing a button immediately for lower probability rewards, d-AMP did not affect performance. However, when subjects were offered choices between smaller rewards delivered immediately and larger rewards to be delivered at later times, the normal preference for immediate rewards was weakened by d-AMP. That is, subjects were more able to resist the impulse to choose the immediate reward in favor of the larger reward.

Taken together, the available results are mixed, with slightly more null results than overall positive findings of enhancement and evidence of impairment in one reversal learning task. As the effect sizes listed in Table 5 show, the effects when found are generally substantial. When drug effects were assessed as a function of placebo performance, genotype, or self-reported impulsivity, enhancement was found to be greatest for participants who performed most poorly on placebo, had a COMT genotype associated with poorer executive function, or reported being impulsive in their everyday lives. In sum, the effects of stimulants on cognitive control are not robust, but MPH and d-AMP appear to enhance cognitive control in some tasks for some people, especially those less likely to perform well on cognitive control tasks.

Do Prescription Stimulants Enhance Other Executive Functions?

Having reviewed the literature on enhancement of working memory and cognitive control, we are left with a small set of studies that seem relevant to executive function but do not fit easily under the previous rubrics and have little else in common with each other.

As shown in Table 6, two of these are fluency tasks, which require the generation of as large a set of unique responses as possible that meet the criteria given in the instructions. Fluency tasks are often considered tests of executive function because they require flexibility and the avoidance of perseveration and because they are often impaired along with other executive functions after prefrontal damage. In verbal fluency, subjects are asked to generate as many words that begin with a specific letter as possible. Neither Fleming et al. (1995), who administered d-AMP, nor Elliott et al. (1997), who administered MPH, found enhancement of verbal fluency. However, Elliott et al. found enhancement on a more complex nonverbal fluency task, the sequence generation task. Subjects were able to touch four squares in more unique orders with MPH than with placebo.

Raven’s Progressive Matrices (Raven, 1990) are a nonverbal intelligence test that requires subjects to complete sequences of designs with the appropriate choice from a multiple-choice set. Analogical reasoning is emphasized. Mattay et al. (1996) found that subjects scored higher on such items after d-AMP than placebo.

Two variants of the Towers of London task were used by Elliott et al. (1997) to study the effects of MPH on planning. The object of this task is for subjects to move game pieces from one position to another while adhering to rules that constrain the ways in which they can move the pieces, thus requiring subjects to plan their moves several steps ahead. Neither version of the task revealed overall effects of the drug, but one version showed impairment for the group that received the drug first, and the other version showed enhancement for the group that received the placebo first.

Kennedy et al. (1990) administered what they termed a grammatical reasoning task to subjects, in which a sentence describing the order of two letters, A and B, is presented along with the letter pair, and subjects must determine whether or not the sentence correctly describes the letter pair. They found no effect of d-AMP on performance of this task.

Schroeder, Mann-Koepke, Gualtieri, Eckerman, and Breese (1987) assessed the performance of subjects on placebo and MPH in a game that allowed subjects to switch between two different sectors seeking targets to shoot. They did not observe an effect of the drug on overall level of performance, but they did find fewer switches between sectors among subjects who took MPH, and perhaps because of this, these subjects did not develop a preference for the more fruitful sector.

Only two of the eight experiments reviewed in this section found that stimulants enhanced performance, on a nonverbal fluency task in one case and in Raven’s Progressive Matrices in the other. The small number of studies of any given type makes it difficult to draw general conclusions about the underlying executive function systems that might be influenced.

Summary Regarding Stimulant Effects on Cognition

The title question, whether prescription stimulants are smart pills, does not find a unanimous answer in the literature. The preponderance of evidence is consistent with enhanced consolidation of long-term declarative memory. For executive function, the overall pattern of evidence is much less clear. Over a third of the findings show no effect on the cognitive processes of healthy nonelderly adults. Of the rest, most show enhancement, although impairment has been reported (e.g., Rogers et al., 1999), and certain subsets of participants may experience impairment (e.g., higher performing participants and/or those homozygous for the met allele of the COMT gene performed worse on drug than placebo; Mattay et al., 2000, 2003). Whereas the overall trend is toward enhancement of executive function, the literature contains many exceptions to this trend. Furthermore, publication bias may lead to underreporting of these exceptions.

Null results are generally less likely to be published. Consistent with the operation of such a bias in the present literature, the null results found in our survey were invariably included in articles reporting the results of multiple tasks or multiple measures of a single task; published single-task studies with exclusively behavioral measures all found enhancement. This suggests that some single-task studies with null results have gone unreported. The present mixed results are consistent with those of other recent reviews that included data from normal subjects, using more limited sets of tasks or medications (Advokat, 2010; Chamberlain et al., 2010; Repantis, Schlattmann, Laisney, & Heuser, 2010).

Open Questions Concerning the Cognitive Neuroscience of Cognitive Enhancement

Given the size of the literature just reviewed, it is surprising that so many basic questions remain open. Although d-AMP and MPH appear to enhance retention of recently learned information and, in at least some individuals, also enhance working memory and cognitive control, there remains great uncertainty regarding the size and robustness of these effects and their dependence on dosage, individual differences, and specifics of the task.

In addition, while the laboratory research reviewed here is of interest concerning the effects of stimulant drugs on specific cognitive processes, it does not tell us about the effects on cognition in the real world. How do these drugs affect academic performance when used by students? How do they affect the total knowledge and understanding that students take with them from a course? How do they affect various aspects of occupational performance? Similar questions have been addressed in relation to students and workers with ADHD (Barbaresi, Katusic, Colligan, Weaver, & Jacobsen, 2007; Halmøy, Fasmer, Gillberg, & Haavik, 2009; see also Advokat, 2010) but have yet to be addressed in the context of cognitive enhancement of normal individuals.

Another empirical question concerns the effects of stimulants on motivation, which can affect academic and occupational performance independent of cognitive ability. Volkow and colleagues (2004) showed that MPH increased participants’ self-rated interest in a relatively dull mathematical task. This is consistent with student reports that prescription stimulants make schoolwork seem more interesting (e.g., DeSantis et al., 2008). To what extent are the motivational effects of prescription stimulants distinct from their cognitive effects, and to what extent might they be more robust to differences in individual traits, dosage, and task? Are the motivational effects of stimulants responsible for their usefulness when taken by normal healthy individuals for cognitive enhancement?

Finally, all of the questions raised here in relation to MPH and d-AMP can also be asked about newer drugs and even about nonpharmacological methods of cognitive enhancement. An example of a newer drug with cognitive-enhancing potential is modafinil. Originally marketed as a therapy for narcolepsy, it is widely used off label for other purposes (Vastag, 2004), and a limited literature on its cognitive effects suggests some promise as a cognitive enhancer for normal healthy people (see Minzenberg & Carter, 2008, for a review).

Nondrug cognitive-enhancement methods include the high tech and the low. An example of the former is transcranial magnetic stimulation (TMS), whereby weak currents are induced in specific brain areas by magnetic fields generated outside the head. TMS is currently being explored as a therapeutic modality for neuropsychiatric conditions as diverse as depression and ADHD and is capable of enhancing the cognition of normal healthy people (e.g., Kirschen, Davis-Ratner, Jerde, Schraedley-Desmond, & Desmond, 2006). An older technique, transcranial direct current stimulation (tDCS), has become the subject of renewed research interest and has proven capable of enhancing the cognitive performance of normal healthy individuals in a variety of tasks. For example, Flöel, Rösser, Michka, Knecht, and Breitenstein (2008) reported enhancement of learning and Dockery, Hueckel-Weng, Birbaumer, and Plewnia (2009) reported enhancement of planning with tDCS.

Low-tech methods of cognitive enhancement include many components of what has traditionally been viewed as a healthy lifestyle, such as exercise, good nutrition, adequate sleep, and stress management. These low-tech methods nevertheless belong in a discussion of brain enhancement because, in addition to benefiting cognitive performance, their effects on brain function have been demonstrated (Almeida et al., 2002; Boonstra, Stins, Daffertshofer, & Beek, 2007; Hillman, Erickson, & Kramer, 2008; Lutz, Slagter, Dunne, & Davidson, 2008; Van Dongen, Maislin, Mullington, & Dinges, 2003).

Open Questions Concerning the Epidemiology of Cognitive Enhancement

It is known that American college students have embraced cognitive enhancement, and some information exists about the demographics of the students most likely to practice cognitive enhancement with prescription stimulants. Outside of this narrow segment of the population, very little is known. What happens when students graduate and enter the world of work? Do they continue using prescription stimulants for cognitive enhancement in their first jobs and beyond? How might the answer to this question depend on occupation? For those who stay on campus to pursue graduate or professional education, what happens to patterns of use? To what extent do college graduates who did not use stimulants as students begin to use them for cognitive enhancement later in their careers? To what extent do workers without college degrees use stimulants to enhance job performance? How do the answers to these questions differ for countries outside of North America, where the studies of Table 1 were carried out?

Another important epidemiological question about the use of prescription stimulants for cognitive enhancement concerns the risk of dependence. MPH and d-AMP both have high potential for abuse and addiction related to their effects on brain systems involved in motivation. On the basis of their reanalysis of NSDUH data sets from 2000 to 2002, Kroutil and colleagues (2006) estimated that almost one in 20 nonmedical users of prescription ADHD medications meets criteria for dependence or abuse. This sobering estimate is based on a survey of all nonmedical users. The immediate and long-term risks to individuals seeking cognitive enhancement remain unknown.

An entirely different set of questions concerns cognitive enhancement in younger students, including elementary school and even preschool children. Some children can function adequately in school without stimulants but perform better with them; medicating such children could be considered a form of cognitive enhancement. How often does this occur? What are the roles and motives of parents, teachers, and pediatricians in these cases? These questions have been discussed elsewhere and deserve continued attention (Diller, 1996; Singh & Keller, 2010).

Regarding other methods of cognitive enhancement, little systematic research has been done on their prevalence among healthy people for the purpose of cognitive enhancement. One exploratory survey found evidence of modafinil use by people seeking cognitive enhancement (Maher, 2008), and anecdotal reports of this can be found online (e.g., Arrington, 2008; Madrigal, 2008). Whereas TMS requires expensive equipment, tDCS can be implemented with inexpensive and widely available materials, and online chatter indicates that some are experimenting with this method.