Impaired or improved driving in abstinent Ecstasy users?

A recent paper set out to examine automobile driving skills in people who had previously used Ecstasy (presumptively 3,4-methylenedioxymethamphetamine; MDMA) but were currently not using. Dastrup and colleagues (2010) used a driving simulator task in which the job was to maintain a set distance behind a lead vehicle (LV) displayed on the computer screen. The job was to stay abut two car lengths (given as 18 meters) behind the LV while accelerating to 55mph. My Google U conversion calculation makes 55 mph out to be about 25 meters / sec. I would therefore estimate the closing time between the cars as about 0.4-0.5 seconds, depending on car length and how much space you assume between these lengths. Thereafter the LV changed speed as depicted in the Figure 2 from the paper.
The horizontal line sits at the 55 mph point and you can see that the speed of the LV varies up to about 59 mph and down to about 51 mph with the maximum change taking place over about 18-20 seconds. .

As with most of the literature on cognitive function in abstinent MDMA users the biggest challenge (and limitation) is that Ecstasy consumers also tend to use a wide range of other recreational drugs, many of which might have lasting effects on their own. Consistent with the better studies on the topic, this study used

four groups of drivers: abstinent MDMA users, abstinent THC users, abstinent alcohol users, and non-drug using controls

Another problem with human user studies in general is that nobody has been exposed to exactly the same amount or variety of recreational drugs. So the researchers have to devise somewhat reasonable criteria to minimize contributions from other drugs while accepting that “pure” populations are essentially impossible to find. Additional criteria and limits for the groups were given as:

MDMA poly substance users: These individuals used MDMA ≥5 times in the last three years, and may have used THC >10 times. They may have occasionally used drugs other thanMDMAand THC, but <30 times in their lifetime. The amount of any other drug use could not exceed that in which MDMA was used. For example, someone who used MDMA 10 times, but used THC 20 times, or
cocaine 30 times was excluded.
THC poly substance users: These individuals used THC ≥10 times in the last three years, but never used MDMA. Other drug use was
<30 times in their lifetime and did not exceed the number of times
they used THC.
Alcohol users: These individuals never used MDMA. They may have tried THC or other drugs, but only a few times (<6 times) in their lifetime. Alcohol use in this group was ≥6 alcoholic drinks per week. To control for tobacco use, we attempted to recruit equal number of participants who used alcohol and tobacco (<20 cigarettes per week) and who used alcohol but not tobacco (<1 cigarette per week).
Non-drug users: These individuals were similar to the alcohol users with respect to MDMA, THC or other drug use. The last time they used any drug (except alcohol or nicotine) was ≥12 months prior
to the study. The alcohol use in this group was ≤5 alcoholic drinks
per week.

If experience is any guide, some readers will want to kvetch about the impurity of the study populations. This is an inescapable feature of human subjects investigations (the design, not the kvetching). Where possible we’d like to follow up findings with controlled studies in laboratory animals performing much simpler behavioral tasks that share some critical feature such as velocity tracking or responding quickly and accurately to stop/go signals. That way the drug exposure would be both known and identical within a group. In my view, the combination of human and nonhuman studies gives us the quickest way to determine if a given substance to which humans are exposed leads to acute or lasting behavioral problems.
There were several key parameters of the subjects’ driving performance measured including coherence (are they tracking the speed changes of the LV), average following distance, delay (how long after the LV changes does the subject require to adjust speed) and gain (under/overshooting the changes of the LV- i.e. slowing too much or speeding up too much). Examples are given in Figure 3 which depicts three individuals’ performance on the task relative to the LV speed.
The only group results that were significantly different were in the delay and following distance parameters. MDMA (2.36 sec) and THC (2.2 sec) groups responded more quickly than did the ALC (3.1 sec) and control (3.4 sec) groups to changes in the LV speed. The MDMA users selected a much shorter following distance (91 ft) than did the THC (140 ft), ALC (121 ft) and control (155 ft) groups. In both cases the differences were present even when statistically adjusting for age- and sex-related differences in performance of the task.
So how to interpret theses findings? Well, responding more quickly to speed changes seems like a good thing when it comes to driving. Following more closely than you are supposed to seems like a bad thing. So are abstinent MDMA and THC users better drivers? Or is the MDMA group impaired relative to all other groups?
One way to think about the tradeoff is as stated in the Discussion of the paper:

While all participants traveled approximately 55mph (80 ft/s), the MDMA drivers showed a mean 1.04 second shorter delay than non-drug using comparison drivers (leaving 83 more feet to respond) but drove 64 feet closer to the LV. Although abstinent MDMA users drove closer to the LV, they compensated for the risk of driving close by reacting quickly to the LV velocity changes.

To make a similar comparison, the THC group had 96 more feet in which to respond but only drove, on average, 15 feet closer, than did the control subjects. Remember, these are not people who are acutely intoxicated. There are only about three obvious hypotheses once we conditionally credit the group effects as specific to the groupings. First, that the populations who self-select into these drug use conditions start off with altered risk-tolerance and attentiveness to driving. Second, it could be that the users are made that way via brain changes induced by their drug taking- some sort of toxicity. As it happens one reasonably well-populated and productive areas of research on human drug users focuses on both pre-existing and drug-induced traits of impulsivity and risk tolerance. Third, it could be that drug experienced individuals intentionally compensate for any real or perceived performance deficits they’ve noticed in their own driving by altering their attentiveness, riskiness or both.
The more attentive / risk tolerant nature of the abstinent MDMA users’ driving phenotype in this study is interesting to consider in the light of recent political efforts to ban cell phone use when driving. Those discussions have pointed to a few studies showing the detrimental effect of such distraction on one’s ability to react to changing circumstances. Drivers are also poor at cell phone use showing that the issue here has to do with the limits of your (39/40 of you anyway) brain allocating attention to multiple tasks.
The point is, of course, that real world driving is a more demanding situation than is posed by the simulator used in this study of abstinent drug users. The next step will be to examine performance in a more multi-tasking type of situation. It may be that that MDMA and THC groups still perform faster…or perhaps they are already at the limit and additional cognitive demands overcome whatever compensations they have made.
__
Dastrup, E., Lees, M., Bechara, A., Dawson, J., & Rizzo, M. (2010). Risky car following in abstinent users of MDMA Accident Analysis & Prevention, 42 (3), 867-873 DOI: 10.1016/j.aap.2009.04.015
Driving simulator user group wiki.