A new study using artificial intelligence has identified specific behavioral patterns that predict alcohol preference in adolescent mice. The results of this study indicate that reduced levels of natural reward sensitivity and sociability are strong indicators of alcohol consumption at this developmental stage. In contrast, these behavioral traits did not predict alcohol preference in adult mice. These results were published in the journal Alcohol: Clinical and experimental studies.
Adolescence represents a distinct period of brain development characterized by profound changes in neural structure. This phase often coincides with the onset of alcohol use and can lead to long-term health problems and dependence. Clinical observations show that teens exhibit a complex mix of behaviors that may increase their likelihood of drug experimentation. Traits such as risk-taking, anxiety, and reward responsiveness are often associated with drug use.
Previous animal studies have yielded mixed results regarding which specific behaviors consistently lead to increased alcohol intake. Some studies link high anxiety to drinking, while others find no association or an inverse relationship. These discrepancies may be due to the fact that early studies often analyzed single behaviors in isolation. Real-world susceptibility can include a combination of multiple interacting traits.
The researchers sought to clarify these contradictions by examining multiple behaviors simultaneously. They used machine learning algorithms to analyze combinations of traits to predict which mice would prefer alcohol. This approach provides a comprehensive view of how different aspects of personality interact to create vulnerability. By comparing adolescents and adults, the researchers also aimed to see whether predictors of alcohol use change as the brain matures.
The research team used two different strains of mice to ensure that the results applied to both genetically similar and genetically diverse populations. Samples included C57BL/6 inbred mice and Swiss outbred mice. The animals were divided into two age groups. Adolescents started at 40 days of age and adults started at 120 days of age. The adolescent group consisted of 46 mice and the adult group consisted of 79 mice.
Over three days, the mice underwent a series of behavioral tests to establish their phenotypic profile. The researchers first assessed novelty-seeking behavior using the hole board test. For this assessment, animals were placed in a playing field with holes in the floor, and the researchers counted how often the mice dipped their heads into the holes.
Anxiety level was assessed using an elevated plus maze. The device consists of two open arms and two closed arms that are raised above the floor. The researchers measured the percentage of time the mice spent with the arms open versus the time they spent with the arms closed. This indicates that anxiety is reduced.
Social behavior was assessed with the three-chamber sociability test. The test measured how much time mice spent in a chamber containing an unfamiliar mouse compared to an empty chamber. This indicator clearly demonstrated the animal’s natural tendency towards social interaction.
Coping behaviors were tested using the forced swim test. The researchers observed the animals underwater and measured the amount of time they spent actively climbing. This test is commonly used to assess how animals respond to unavoidable stress.
Finally, the researchers measured the animals’ responses to natural rewards. Mice were housed individually and given free choice between a bottle of water and a bottle of sucrose solution. The amount of sugar water consumed compared to plain water served as a measure of the animals’ sensitivity to pleasure and natural rewards.
After these behavioral assessments, mice entered a 5-day alcohol preference phase. They were housed individually and had two bottles at their disposal. One bottle contained water and the other contained a 10 percent ethanol solution. The researchers calculated ethanol preference by comparing the amount of alcohol consumed and total water intake.
The team then used a machine learning technique known as pattern regression. Train a computer model by splitting your data into a training set and a test set. The goal was to see if the model could learn enough about the relationship between behavioral profiles and subsequent alcohol consumption to predict the drinking habits of mice they had not yet analyzed.
A machine learning model successfully predicted alcohol preference based on behavioral characteristics in adolescent mice. The correlation between predicted and actual preferences was statistically significant. However, this model failed to find the predicted pattern in adult mice.
Among adolescents, two specific behaviors stood out as the strongest predictors. The first was sucrose preference, which had a positive predictive value. Mice that consumed more sugar water were significantly more likely to consume more alcohol later in the experiment. This suggests that increased sensitivity to natural rewards increases the desire for alcohol.
The second significant predictor was sociability. Analysis revealed a negative relationship between social behavior and alcohol consumption. Adolescent mice that spent less time interacting with other mice were more likely to prefer alcohol. This means that decreased sociability acts as a risk factor for increased drinking in this age group.
Other factors included in the model did not contribute significantly to the predictions. Anxiety level and novelty-seeking behavior had a lower impact on the model’s ability to predict alcohol preference. This contradicts some previous theories that suggest anxiety is a major factor in adolescent drug use.
This study provides evidence that the drivers of alcohol consumption may be fundamentally different between adolescents and adults. The failure of this model to predict adult drinking suggests that behavioral traits established in adulthood do not determine alcohol preferences in the same way as during development. This means that the adolescent brain is in a uniquely vulnerable state.
The strong association between sugar preference and alcohol intake suggests that the brain’s reward system plays a central role during adolescence. This indicates the potential involvement of the dopamine system in processing reinforcing stimuli. The authors also suggest that the orexin system, which regulates reward seeking and feeding, may be a relevant biological mechanism.
Findings on low sociability highlight the protective nature of social interactions. This raises the possibility that oxytocin, a hormone involved in social bonding, influences the reward of alcohol at this developmental stage. Adolescents may be more sensitive to the social effects of alcohol, or social isolation may increase their compensatory desire for the pharmacological effects of ethanol.
There are several limitations to consider when interpreting these results. This study used a relatively small sample size, particularly in the adolescent group. Machine learning is a powerful tool, but it typically requires large datasets for maximum robustness. The researchers used cross-validation techniques to mitigate this, but larger studies are needed to confirm.
Furthermore, because this study was conducted in mice, the results of this study do not directly translate to human behavior. Environmental conditions in mice are highly controlled and differ from the complex social environments experienced by human adolescents. Humans also face cultural and peer pressures that cannot be replicated in rodent models.
The specific strain of mouse used may also affect the results. Although the researchers used both inbred and outbred strains to increase generalizability, genetic factors still play a role. Different strains often have different baseline levels of anxiety and sociability.
Future studies are needed to validate these results with larger groups of animals and potentially different species. Scientists may also investigate whether targeting the orexin or oxytocin systems could help reduce vulnerability to alcohol in adolescents. Understanding the biological basis of these behavioral predictors may ultimately lead to better prevention strategies for teens.
The study, “Behavioral profiles predict ethanol preference in adolescent mice but not adults: A machine learning approach,” was authored by Liana CL Portugal, Bruno da Silva Gonçalves, Emily de Assis Fagundes, Maria Fernandes Freire de Sá, Cláudio Carneiro Filgueiras, Ana Carolina Dutra-Tavares, Alex C. Manhães, and Yael. Abreu-Villaza, Anderson Ribeiro-Carvalho.
