Aims: Animal experimentation often demonstrates that alcohol leads to disinhibited behaviour, such as increased aggression, increased social behaviour, or increased impulsivity. However, human experimentation demonstrates that alcohol can have either disinhibiting or inhibiting effects on behaviour, depending on salient environmental cues. Our aim was to illustrate how alcohol myopia theory could be applied to the literature assessing the effects of alcohol on behaviour in animals. Methods: The effects of alcohol on animal behaviour were reviewed in several domains, including aggression, social behaviours, and impulsivity. Suggestions for testing alcohol myopia with animal research paradigms were provided. Results: Current animal research paradigms are often designed in such a way that alcohol myopia cannot be tested. To test alcohol myopia, we recommend manipulating the salience of both impelling and inhibiting environmental cues. Conclusions: Disinhibition alone cannot explain alcohol's effects on behaviour. We contend that alcohol myopia theory helps to explain some contradictory findings in the human and animal literature. We encourage animal researchers to develop research paradigms to provide tests of alcohol myopia.
(Received 24 June 2004; first review notified 22 February 2005; in revised form 30 May 2005; accepted 31 May 2005)
How does alcohol affect behaviour? One popular notion is that alcohol causes disinhibition, or a release of natural impulses by eliminating learned inhibitions (Critchlow, 1986). For example, alcohol is used as a social lubricant at parties, which is evidence of this strong belief in the disinhibiting effects of alcohol on positive social behaviours. However, alcohol is also believed to increase antisocial impulses, such as aggression or risk-taking.
Disinhibition alone, however, cannot account for the complex, and sometimes contradictory research findings. On the one hand, animal research often demonstrates that alcohol leads to disinhibited behaviour, such as increased aggression (Miczek et al., 1993), impulsivity (Poulos et al., 1998; Evenden and Ryan, 1999), and playfulness (Varlinskaya et al., 2001), consistent with the idea that alcohol is a general disinhibitor. The human literature has also been guided to a large extent by disinhibition theory. However, recent research shows that alcohol does not always lead to disinhibited behaviour, but can either inhibit or disinhibit behaviour, depending on environmental cues.
ALCOHOL MYOPIA THEORY
Why does alcohol cause disinhibited behaviour in some cases but not in others? An explanation is offered by alcohol myopia, which states that alcohol limits cognitive capacity such that intoxicated individuals tend to focus on cues in the environment that are most salient (Steele and Josephs, 1990). Alcohol myopia postulates that intoxicated individuals are unable to attend to all relevant cues simultaneously because of the limitation of cognitive capacity associated with alcohol intoxication. In other words, alcohol produces a myopic effect causing individuals to attend primarily to, and hence be more influenced by, salient environmental cues at the expense of less salient cues. Individuals who are not intoxicated, however, are not as easily influenced by salient cues because they are better able to attend to all the relevant information in the environment.
Evidence to support alcohol myopia has been found in the domain of health-relevant behaviours, such as drinking and driving or engaging in unprotected sex. MacDonald et al. (1995) found that when asked about their attitudes towards drinking and driving, both sober and intoxicated individuals reported very negative attitudes. However, when the wording of the question included an impelling cue to drink and drive (e.g. having to drive only a short distance), intoxicated individuals reported less negative attitudes than sober individuals. This pattern is consistent with alcohol myopia because intoxicated participants were influenced by the impelling cue, but sober participants were not.
Intentions towards engaging in unprotected sex have also been shown to be influenced by salient inhibiting cues such that intoxicated individuals actually reported more prudent intentions compared with sober controls (MacDonald et al., 2000). It is important to note that disinhibition theory cannot account for these findings; it cannot explain situations in which intoxicated individuals behave more prudently than sober individuals. If alcohol is a general disinhibitor, intoxicated individuals should always exhibit disinhibited behaviour.
Findings emerging from the human literature present a different picture than those in the animal literature. Although conventional wisdom might argue that alcohol will always lead to disinhibited behaviour, it is clear that this explanation is less than satisfactory. Indeed, researchers are aware that the relationship between alcohol and behaviour is complex, and cannot be accounted for by simple disinhibition (Miczek et al., 1993). We argue that alcohol myopia might resolve some of the complexities and contradictory findings.
APPLYING ALCOHOL MYOPIA TO THE ANIMAL LITERATURE
Animal research has virtually ignored alcohol myopia. Indeed, most research paradigms in the animal literature do not allow for a true test of alcohol myopia. To test alcohol myopia, both impelling and inhibiting cues should be present. Ideally, the salience of these cues is also manipulated. In many animal research paradigms, the salient environmental cues are generally impelling ones, i.e. ones that would elicit disinhibition. For example, in testing the influence of alcohol on aggression, some researchers use a shock chamber situation (Tramill et al., 1981; Davis et al., 1993). The animal is restrained in a chamber and aggression is elicited by shocking the animal. The shock can be seen as an impelling cue to act aggressively. Therefore, when an alcohol-treated animal behaves more aggressively, is it because of the disinhibiting effects of alcohol, or is it because the shock is a salient, impelling cue to act aggressively? To test both disinhibition theory and alcohol myopia, two cue conditions would need to be present in the study. For instance, one could compare conditions in which shock is either present or absent. If aggression increases regardless of the presence of shock, this would suggest that alcohol is simply producing disinhibition. If, however, the increase in aggression in alcohol-treated animals occurs only when the shock is present, that would suggest that the shock is an impelling cue to behave aggressively.
Cutler et al. (1975) examined the effects of alcohol on social behaviours in mice and found that alcohol only increased behaviours that were already stimulated by the test situation. When the mice were subjected to a new environment, alcohol increased exploratory behaviour. However, alcohol-treated mice which were not subjected to the new environment showed no differences in their exploratory behaviour from untreated mice. Similarly, in a territorial situation, when territories had not been established, alcohol-treated mice became dominant over the entire enclosure. When territorial boundaries were established, however, alcohol-treated mice were no more likely to dominate the enclosure than the non-treated mice. Although the authors view these effects in terms of disinhibition, we contend that a better explanation might be alcohol myopia. When cues designed to elicit a particular behaviour were present, alcohol-treated mice showed increases in that behaviour. When these cues were not present, alcohol-treated mice did not differ from the non-treated mice, which is exactly what alcohol myopia predicts.
Although the results of the Cutler et al. (1975) study are consistent with alcohol myopia, only impelling cues were included in the study. It is important to note that disinhibition theory and alcohol myopia make opposing predictions when inhibiting cues are present. Disinhibition theory predicts that behaviours that would normally be inhibited, such as aggression or impulsivity, will increase under the influence of alcohol. Alcohol myopia predicts that, if powerful inhibiting cues are present, alcohol could cause animals to behave even less aggressively or impulsively than they would otherwise. The finding that alcohol intoxication can lead to more inhibited behaviour has important implications for research, especially for studies on humans, as many of the behaviours associated with alcohol are potentially harmful (e.g. drinking and driving, risky sexual behaviour, and aggression). Furthermore, including tests of alcohol myopia in animal research will lead to a better understanding of how alcohol affects behaviour.
The purpose of this paper is to explain how alcohol myopia could be tested within animal research paradigms. First, we will review the animal literature and main findings for the behavioural domains of aggression, social behaviour, and impulsivity. We believe it would be useful to consider applying alcohol myopia to these particular domains for two reasons. First, the effects of alcohol on behaviour have been widely studied in these areas. Second, the behaviours in these domains are relatively complex, making them more pertinent to the human alcohol myopia literature on social decision-making. We will suggest ways to test alcohol myopia theory within each of these domains, and finally, we will explain the specific predictions that follow from alcohol myopia and compare these predictions with those from a disinhibition perspective. We will primarily discuss the findings regarding the short-term effects of alcohol at low and moderate doses, because these are most comparable with the studies conducted on humans. Furthermore, although a biphasic dose–effect curve of alcohol on behaviour has been shown such that lower doses often cause increases and higher doses cause decreases in a particular behaviour (Miczek et al., 1993), it is possible that the effects at high doses might be owing to the sedative effects of alcohol, which is less pertinent to an alcohol myopia perspective.
The effects of alcohol on aggression have been studied extensively. To date, studies involve placing the animal in a situation where aggressive behaviour would normally be elicited (e.g. a resident–intruder paradigm) and observing the effects of alcohol. For example, when resident ciclid fish confronted intruder fish, moderate doses of alcohol produced increases in attacks (Peeke et al., 1973). Similarly, in response to their mirror image, Siamese fighting fish treated with low doses of alcohol increased their aggressive displays (Raynes et al., 1968; Raynes and Ryback, 1970).
Using resident–intruder paradigms, low doses of alcohol have been found to increase aggressive behaviour in mice (Krsiak, 1976), increase or decrease aggression depending on whether confrontations took place in neutral or home cages, (Miczek and O'Donnell, 1980), and to have no effects on aggression (Lagerspetz and Ekqvist, 1978). It is not surprising that alcohol often causes increases in aggression in these paradigms. The finding that alcohol-treated animals behave more aggressively when they are confronted with an intruder is consistent with alcohol myopia because the intruder acts as an impelling cue to behave aggressively. Alcohol myopia would predict that as animals become intoxicated, they will be unable to attend to all the relevant cues and will focus on the most salient cues, which in these studies would be the intruder. However, it is not clear whether disinhibition or alcohol myopia is the process by which the alcohol-treated animals behaved more aggressively because disinhibition theory would predict a similar increase in aggression. Are these effects the result of the disinhibiting properties of alcohol? Or does the intruder act as an impelling cue to act aggressively, which becomes the most salient cue for the alcohol-treated animals, thereby eliciting aggressive behaviour? It is difficult to test alcohol myopia within this paradigm because an intruder generally leads to aggressive behaviour by the resident. Even comparing the resident–intruder confrontations taking place in neutral versus home cages is difficult because the rate of aggression is often so high in the home cage that a ceiling effect might occur.
In the resident–intruder paradigm, resident rats typically display a specific pattern of aggressive behaviour resulting in defeat of the intruder, who eventually displays submissive behaviour. Although alcohol myopia is difficult to test in this paradigm, some indirect support for alcohol myopia can be found in a study by van Erp and Miczek (1997). Resident rats that showed large increases in aggression after self-administration of alcohol were classified as high in alcohol-heightened aggression. The authors speculate that alcohol made these rats unresponsive to signals of submission by the intruder rat. These signals of submission by the intruder usually lead to decreased attacks by the resident. The idea that the alcohol-treated rats attended to the salient cue to act aggressively (an unfamiliar intruder) at the expense of less salient inhibiting cues (subtle signs of submission) is consistent with alcohol myopia because it suggests that these rats were not able to attend to all the relevant cues simultaneously.
Some researchers place the animal in a competitive situation and observe the effects of alcohol on aggression. In these studies, animals are categorized based on their social status, which appears to interact with alcohol. Low and moderate doses of alcohol have been shown to increase aggression in dominant rats (Miczek and Barry, 1977) and in dominant squirrel monkeys (Winslow and Miczek, 1985). However, Pettijohn (1979) found that during a competition between three dogs over a bone, low doses of alcohol increased aggression in subordinate dogs but reduced aggression in higher ranking dogs. If alcohol were simply disinhibiting aggressive behaviour, why would alcohol produce different effects depending on social status? It is possible that environmental cues function differently for animals of different ranks. Perhaps alcohol leads subordinate dogs to attend to the salient bone at the expense of the less salient social hierarchy. If disinhibition theory were true, all alcohol-treated animals should respond the same way when provoked. Instead, the meaning of the cues may vary depending on the social status of the animals; so animals will respond differently in a competitive situation depending on their social status, which is consistent with alcohol myopia.
Shock or pain has also been used to elicit aggression in alcohol-treated animals. For example, Weitz (1974) found that alcohol increased fighting behaviour in pairs of male rats when electric foot shock was employed. In contrast, Tramill et al. (1980) found that low doses of alcohol decreased aggression towards a lever when single-restrained rats were shocked. It is difficult to compare these two studies because the cues are dramatically different. While shock was employed in both studies, and may have acted as a cue to behave aggressively, the target of aggression was another male rat in the study of Weitz (1974) and a lever in the study by Tramill et al. (1980). It is quite possible that another male rat could provoke the alcohol-treated rat such that it becomes an additional cue to act aggressively. It is hard to imagine, however, how a lever in this instance could act as an impelling or an inhibiting cue. To test alcohol myopia, it would be important to have two types of cue conditions in the same study.
To sum up, alcohol studies on several types of animals have generally shown that low and moderate doses of alcohol cause increases in aggression. However, most of this research involves subjecting the animals to situations where impelling cues to act aggressively are present, such as resident–intruder confrontations, competitive situations, or shock chambers. To test alcohol myopia, it is important to include inhibiting cues, or ones that would encourage an animal to behave non-aggressively.
The effects of alcohol on social behaviours have also been studied in animals. Alcohol has been shown to increase play behaviours in monkeys (Cressman and Caddell, 1971; Crowley et al., 1974). In contrast, Krsiak and Borgesova (1973) found that alcohol decreased all social activities in rats. This research does not provide a test of alcohol myopia because the environmental cues were not measured or manipulated. For example, the studies on monkeys involved placing the alcohol-treated monkey with three other monkeys. The resulting interactions were then observed. It is impossible to control the behaviour of the other monkeys making it difficult to know whether the salient cues available to the alcohol-treated monkey were impelling cues to act sociably, impelling cues to act aggressively, or a combination of the two.
Varlinskaya et al. (2001) also investigated the effects of alcohol on social behaviours in rats. Alcohol-treated rats were either exposed to social stimuli (littermate of same gender) or non-social stimuli (a cotton ball) placed in the testing chamber. They found that low and moderate doses of alcohol increased social behaviours in the social condition (an impelling cue to act sociably), but failed to do so in the non-social condition (no impelling cue), which is consistent with alcohol myopia. A better test of alcohol myopia would include an inhibiting cue, or a cue to act non-sociably. For example, one could compare the responses of alcohol-treated rats' to a littermate with their responses to an unfamiliar rat. According to alcohol myopia, alcohol should increase prosocial behaviours in the presence of a littermate (an impelling cue), but should decrease prosocial behaviours in the presence of an unfamiliar rat (an inhibiting cue). However, in this paradigm, these cues might be confounded with other aggression cues, such as dominant and subordinate positions.
Studies of the effects of alcohol on social behaviours have generally shown that alcohol tends to increase social behaviours. The challenge for testing alcohol myopia in this domain is that cues are difficult to measure or manipulate. Although it makes sense to test social behaviours in response to another animal or a group of animals, it is hard to control the behaviour of the other animals. To test alcohol myopia, one would need to design a study in which cues could be manipulated in some way to produce impelling cues (or cues to behave prosocially) in one condition and relatively inhibiting cues (or cues to behave antisocially) in another.
To test alcohol's effects on impulsivity, animals are often presented with a choice between a small, immediate reward or a large, delayed reward. The index of impulsivity, then, is based on the number of times the animal chooses the immediate reward. Poulos et al. (1998) trained rats to respond to a T-maze in which one arm led to 2 food pellets and the other arm led to 12 food pellets. Delay was then introduced for the large reward. They found that alcohol dose-dependently increased impulsivity, or choice of the arm that led to the small, immediate reward. Tomie et al. (1997) had rats choose between two levers, one producing a small, immediate reward and the other producing a large, delayed reward. They found that low and moderate doses of alcohol increased the preference for the small, immediate reward (see also Evenden and Ryan, 1999).
Although these findings are consistent with disinhibition, they are also consistent with alcohol myopia. Food administered with no delay is probably more salient than delayed food, and alcohol myopia would predict that alcohol-treated animals should focus on the salient cue (the immediate reinforcer) at the expense of the less salient cue (the delayed reinforcer). It would be useful to design a study in which the large, delayed reward was made more salient. For example, one might modify the delay-of-reward paradigm to have the large reward inaccessible to the animal during the delay but still in full view, and have the small reward hidden from view. Alcohol myopia would predict that the alcohol-treated animals should show a preference for the large, delayed reward at the expense of the small, immediate reward when the large reward is a more salient cue. However, if disinhibition theory is true, animals should be unable to delay gratification, and should impulsively choose the small immediate reward in favour of the large, delayed reward.
One T-maze study of alcohol's effects on impulsivity did provide a test of alcohol myopia. P. C. Darling, T. L. Pinder, K. G. C. Hellemans, T. A. Paine and M. C. Olmstead (2003) trained rats in a T-maze in which one arm led to a small, immediate reward and one arm led to a large, delayed reward (Unpublished data). In the first study, alcohol increased impulsivity as measured by choice of the immediate reward. In the second study, a light predicted either the small immediate or the large delayed reward, but the arm paired with the light varied across trials. In this study, alcohol caused an increase in the tendency of the rats to run towards the lit arm, regardless of whether it was associated with a small or large reward. Control rats did not attend exclusively to the lit arm. We believe the best explanation of these data is that alcohol-treated rats experienced a decrease in attentional capacity and shifted their attention towards the most salient environmental stimulus, i.e. the lit arm. It is important to note that in the second study, alcohol did not cause any increase in impulsivity as measured by choice of the immediate reward. If disinhibition theory were true, alcohol should have increased the impulsivity by increasing the choice of the small, immediate reward regardless of the light.
When differential salience between cues is not an issue, studies have found that alcohol does not increase impulsivity. For example, Evenden (1999) measured ‘reflection-impulsivity’ by training rats to wait for a light indicating which lever is the correct one to press in order to receive a food pellet. The light signal went on three times. The first light provided an unreliable indication of the correct lever, the second light was more reliable, and the third light was a very reliable indication of the correct lever. Thus, a slow response to the light would enable the rat to have the highest accuracy rate and consequently obtain the highest number of food pellets. Here, the salient cue was the indicator light, which did not differ over time. Consequently, no effects of alcohol on impulsivity were found: the rats' accuracy rate remained unaffected by the administration of alcohol (see also Evenden, 1998).
Feola et al. (2000) tested impulsivity using a Go–Stop task, where animals must respond as quickly as possible to a ‘go’ stimulus, but must withhold their response to a ‘stop’ stimulus which follows the ‘go’ stimulus. Impulsivity was measured by the ‘stop’ time, or how long it takes for the animal to stop responding. They found that alcohol increased ‘stop’ time, suggesting that disinhibition occurred. However, the ‘go’ stimulus was a visual cue (a light) and the ‘stop’ stimulus was an auditory cue (a tone). It is difficult to say, then, whether one of these cues was more salient than the other. It would be interesting to know whether the same effect would occur if the stimuli were reversed. To include tests of alcohol myopia in a stop-go task paradigm, it would be important to manipulate the salience of the stimuli, such as the brightness of a light or the loudness of a tone. If alcohol myopia is correct, intoxicated animals' behaviour will depend more on whichever cue is more salient (i.e. alcohol-treated animals will ‘go’ or ‘stop’ faster when the stimuli are especially salient), but the behaviour of control animals will depend less on the salient cue. Alternatively, counterbalancing the stimuli signalling ‘go’ and ‘stop’ would be helpful, as this would ensure that the results were not due to the specific types of stimuli used.
To sum up, when impulsivity is measured using the delay-of-reward paradigm or with the go–stop task, alcohol often appears to increase impulsivity. However, in the Evenden studies (1998, 1999), low and moderate doses of alcohol appear to have no influence on impulsivity. In the Evenden studies, the salience of the cue (i.e. the indicator light) was held constant. Alcohol myopia would indeed predict that alcohol should have no effect on impulsivity in such a paradigm. With the exception of the study by P. C. Darling, T. L. Pinder, K. G. C. Hellemans, T. A. Paine and M. C. Olmstead (2003) (unpublished data), most studies were not designed in such a way that alcohol myopia could be tested. Therefore, it is not clear whether the increased impulsivity found in many studies is the result of disinhibition or because of the effects of alcohol myopia.
Alcohol myopia could also be tested in lever pressing for food reward paradigms. To do this, cues could be made impelling or inhibiting by letting the cues signal reward for lever pressing (e.g. presence of green light signals that reward will be delivered) or no reward for lever pressing (e.g. presence of red light signals that reward will not be delivered). Following this, both cues could be presented simultaneously (e.g. present both green and red lights). The salience (e.g. brightness) of these cues could also be manipulated. If alcohol myopia were true, when conflicting cues are present, alcohol-treated animals should respond more to whichever cue is most salient, whether it be impelling or inhibiting. If disinhibition theory were true, alcohol-treated animals should have difficulty withholding the lever-pressing behaviour and will produce more lever pressing in all conditions.
Finally, discriminative stimulus control is another area in which alcohol myopia could be tested. One discrimination procedure that could be used is the matching to sample task using compound stimuli. In this task, animals are presented with a compound sample stimulus (e.g. a red patch above horizontal lines). They are then presented with the test stimuli (e.g. red and orange patches, or horizontal and vertical lines) and are reinforced if they respond to the stimulus that matches the sample. This task is particularly well suited to testing alcohol myopia because the compound stimuli should compete for attention. The salience of the stimuli could also be varied. Alcohol myopia would make the counterintuitive prediction that alcohol treated rats' performance would be more accurate than control rats' performance when one stimulus is more salient than another. For example, if rats were presented with a bright red patch together with relatively dim horizontal lines, alcohol-treated rats should attend primarily to the red patch at the expense of the lines. When they are then presented test stimuli consisting of red and orange patches, alcohol-treated rats will respond to the correct patch because they attended primarily to the red patch in the sample phase. However, control rats should be able to attend to both stimuli simultaneously when presented with compound sample stimuli. Therefore, attention to the lines should diminish attention to the red patch. Control rats would then be less accurate when presented with the test stimuli. In contrast to alcohol myopia, disinhibition theory would not predict that alcohol would improve performance on a matching to sample task.
In most animal studies, the salience of environmental cues is not manipulated, making it difficult to test alcohol myopia. Therefore, when alcohol produces an increase in disinhibited behaviour, it is impossible to know whether this is because of the disinhibiting effects of alcohol, or whether it is owing to salient impelling environmental cues. We have provided suggestions for how animal researchers might design studies to test alcohol myopia in the behavioural domains of aggression, social behaviours, and impulsivity, among others. We encourage animal researchers to develop ways to test alcohol myopia within animal research paradigms.
We would like to thank Cella Olmstead and Lee Fabrigar for their comments on the earlier drafts of this manuscript. This project was supported by a New Investigator Award from the Canadian Institutes of Health Research awarded to T.M.