AI can predict political orientation, internet access boosts well-being, and more scientific research

Artificial intelligence (AI) can predict political orientation based on photos of others’ expressionless faces, according to a study in American Psychologist. Researchers asked 591 adults in the United States about their political orientation and photographed them with the same blank expression, the same black T-shirt, their hair pulled back, with no jewelry, makeup, or facial hair. The researchers fed the responses and photos into an AI model, which had been trained on thousands of naturalistic photos of liberals and conservatives. The model pulled out structural features of the faces to predict the political leanings of the expressionless individuals at a rate significantly above chance. In a second study, 1,026 online participants’ ratings of the photos from the first study were similar to the AI ratings. In a third study, the researchers fed 3,401 profile images of politicians from the United States, the United Kingdom, and Canada to the AI model. The AI model again performed above chance.
DOI: 10.1037/amp0001295opens in new window

Research in Technology, Mind, and Behavior suggests that internet use can increase measures of well-being, such as life satisfaction and sense of purpose. Researchers examined survey data from more than 2.4 million people from 168 countries, examining links between internet access and eight indicators of well-being between 2006 and 2021 from people 15 years old and older. They found that, on average, people who had access to the internet via phone, computer, or other devices scored 8% higher on measures of life satisfaction, positive experiences, contentment with their social life, and physical well-being compared with those lacking web access—even when controlling for confounding factors such as income and education. However, women ages 15 to 24 who reported having used the internet in the past week were, on average, less happy with the place where they live compared with people who didn’t use the web.
DOI: 10.1037/tmb0000127opens in new window

According to a study of school shooters over the past 50 years published in Psychology of Men & Masculinities, nearly 70% perpetrated violence against women before or during their attacks. Researchers examined details of 59 school shootings, all carried out by boys or men, in the United States, Canada, Brazil, and Europe, and profiled the shooters. Using publicly available information such as news articles and court documents, the researchers found that 44% of the school shooters perpetrated acts of violence against women, including intimate partner violence and sexual assault, before their attacks. Another 29% of the shooters exhibited behavior common to men and boys who commit violence against women during their attacks, including striving to embody traditional masculine traits and promoting the normalization of violence.
DOI: 10.1037/men0000445opens in new window

More than just a way to keep the eyes moist, blinking helps the brain process visual information, suggests research in the Proceedings of the National Academy of Sciences. Researchers tracked the eye movements of 12 participants and combined this data with an analysis of the various frequencies present in stimuli that the participants viewed. They also noted how sensitive the participants were to different types of stimuli, such as patterns at different levels of detail. The researchers found that when the participants blinked, they became better at noticing big, gradually changing patterns. The researchers suggest that the results show that when people blink, the rapid motion of the eyelids creates patterns of light that are effective at stimulating the retina, reformatting the visual input to the brain compared with when the eyes remain open and focused on a specific point.
DOI: 10.1073/pnas.2310291121opens in new window

Microdosing lysergic acid diethylamide (LSD) may extend sleep by almost half an hour per night, suggests a study in Translational Psychiatry. Researchers randomized 80 healthy male participants in the United States into LSD or placebo groups. Participants wore smartwatches to track sleep and self-administered 10 micrograms of LSD or a placebo every 3 days for 6 weeks. The researchers found that participants in the LSD group got an additional 8.13 minutes of rapid eye movement (REM) sleep and an additional 24.3 minutes of total sleep on nights following (but not on the nights of) microdosing, but they found no significant differences in deep sleep. No significant changes were observed in physical activity patterns between the LSD and placebo groups. Additionally, no noticeable interaction effects were observed in calorie intake, distance walked, total steps taken, or activity state.
DOI: 10.1038/s41398-024-02900-4opens in new window

Young people with mood disorders are less likely to acquire a driver’s license and more likely to be involved in crashes and rack up more moving violations and license suspensions than peers without mood disorders, according to a study in JAMA Network Open. Researchers compared rates of licensure and driving outcomes among 1,879 youths with a current mood disorder and 84,294 youths without in the United States from 2004 to 2017. They found that young people with mood disorders were 30% less likely to acquire a license than those without a mood disorder. Furthermore, licensed young people with mood disorders were involved in 1.2 times more crashes than those without over their first 48 months of driving. Licensed youths with mood disorders also had 1.25 times more moving violations and 1.95 times more license suspensions.
DOI: 10.1001/jamanetworkopen.2024.5543opens in new window

A JAMA Psychiatry study reports that first-degree relatives of individuals with treatment-resistant depression (TRD) are at a significant risk of also developing TRD. Researchers analyzed national health insurance data for the entire population of Taiwan (more than 26.5 million people) for 15 years. More than half a million Taiwanese received a diagnosis of major depressive disorder (MDD) in that time frame. Of those, more than 21,000 people went on to be diagnosed with TRD. The researchers found that first-degree relatives of people with TRD, including parents, siblings, and children, were more than 9 times as likely to develop TRD than those without first-degree relatives with TRD. In addition, first-degree family members of people with TRD were more likely to die by suicide and were 2 to 4 times more likely to develop other types of psychiatric conditions, such as anxiety, bipolar disorder, obsessive-compulsive disorder, MDD, and attention-deficit/hyperactivity disorder (ADHD).
DOI: 10.1001/jamapsychiatry.2024.0378opens in new window

Physical fitness among children and adolescents may protect against developing depression, anxiety, and ADHD, according to a study in JAMA Pediatrics. Researchers examined data from Taiwanese national student fitness tests for 10- and 11-year-olds between 2009 and 2019. Researchers divided participants into two cohorts—one targeting anxiety and depression (1.99 million participants) and the other ADHD (1.92 million participants)—and monitored their subsequent mental health symptoms for at least 3 years. The researchers found improved cardiorespiratory fitness, marked by a 30-second improvement in 800-m run times, was associated with less anxiety, depression, and ADHD in female participants and less anxiety and ADHD in male participants. Enhanced muscle endurance, reflected by five additional curl-ups per minute, was associated with a decreased risk of depression and ADHD in female participants and lower anxiety and ADHD risks in male participants. Improved muscular power, indicated by a 20-cm increase in jump distance, was associated with reduced risks of anxiety and ADHD in female participants and reduced anxiety, depression, and ADHD in male participants.
DOI: 10.1001/jamapediatrics.2024.0806opens in new window

A study in the Journal of Affective Disorders indicates that compared with premenopausal women, perimenopausal women are at higher risk for depression. Researchers performed a meta-analysis of seven studies involving 9,141 women in Australia, China, the Netherlands, Switzerland, and the United States. They found that perimenopausal women had a significantly higher risk (around 40%) of experiencing depressive symptoms and being diagnosed with depression compared with premenopausal women. There was no significant increase in depression risk for postmenopausal women compared with those who were premenopausal.
DOI: 10.1016/j.jad.2024.04.041opens in new window

A study in Computers in Human Behavior indicates that teens and preteens with higher social media engagement spend more in-person time with friends. Researchers analyzed data from a longitudinal study involving 812 children born in Norway in 2003 and 2004. Researchers assessed the participants’ social skills and social media usage at ages 10, 12, 14, 16, and 18, including liking, commenting, and posting content across various platforms. They found that increased social media use was associated with increased time spent with friends offline, particularly from the ages of 12 to 14. No sex or age differences were found, but socially anxious youth were at a slightly elevated risk for reduced social skills.
DOI: 10.1016/j.chb.2024.108235opens in new window

People with visual impairments are 2.5 times more likely to engage in suicidal behavior and die by suicide at nearly twice the rate compared with those without vision problems, suggests a meta-analysis in JAMA Network Open. Researchers analyzed 31 previous studies representing more than 5.7 million people looking at suicidal tendencies and visual impairment (the researchers looked for the key phrases “visual impairment,” “low vision,” and “blindness”). In 17 studies that specifically evaluated suicidal behavior, the researchers found that people with sight problems were 2.5 times more likely to have such tendencies compared with those without sight issues. From 21 studies that assessed suicidal ideation, the researchers calculated that those with vision impairment had 2.01 times the risk. Data on death by suicide taken from eight studies revealed that the risk was 1.89 times greater for people with limited vision. Overall, vision-impaired adolescents were at the greatest risk of suicidal behavior.
DOI: 10.1001/jamanetworkopen.2024.7026opens in new window

A study in Nature Human Behaviour suggests that people remember images better when they think they viewed them longer, and that perceived viewing duration depends on how the images are constructed. Researchers asked 52 participants in the United States to look at 252 images and estimate how much time had passed while they were looking at them. They found that when people looked at larger, uncluttered scenes—big, empty warehouses, for example—they estimated that they viewed the images for longer than they actually did. Alternatively, the participants experienced time contraction when looking at more constrained, cluttered scenes such as an image of a fully loaded refrigerator. In a second study, 48 participants were more likely to remember the images they thought they viewed for longer. The researchers’ explanation for the results is that when people see something highly detailed, their sense of time is dilated to get more information about the scene.
DOI: 10.1038/s41562-024-01863-2opens in new window

Research in Scientific Reports suggests that dreaming can help people forget mundane experiences and better process extreme events. Shortly before their bedtime, researchers asked 125 women in the United States to rate the intensity of a series of images depicting negative and neutral experiences. The next day, the participants recorded the presence and content of their dreams and their overall mood. Two hours after waking, the women repeated the emotional picture task from the night before with the addition of some new images to measure image recall and emotional reaction. Participants who reported dreaming had better recall and had muted responses to negative images compared with neutral images, a pattern not observed for those who did not remember dreaming. Additionally, the more positive the dream, the more positively the participants rated negative images the next day. The researchers suggest that the findings could lead to interventions that increase dreaming to help people work through difficult life experiences.
DOI: 10.1038/s41598-024-58170-zopens in new window

A study in The Lancet Neurology indicates that Alzheimer’s disease both starts earlier and progresses faster in people with Down syndrome. Researchers recruited 137 participants with Down syndrome and 49 participants with autosomal dominant Alzheimer’s in the United States and the United Kingdom. They then used positron-emission tomography (an imaging technique used to measure metabolic activity and other physiological processes) to map the relative development of tau tangles and amyloid plaques (protein abnormalities that precede cognitive decline in Alzheimer’s) and tracked which brain regions were affected. The researchers observed that tau tangles and amyloid plaques accumulate in the same areas of the brain and in the same sequence in both groups. However, the process happens earlier and more quickly in people with Down syndrome, and the levels of tau are greater for a given level of amyloid.
DOI: 10.1016/S1474-4422(24)00084-Xopens in new window

The effects of childhood maltreatment can stretch long into adulthood. These traumatic experiences boost the risk of poor body and brain health many years later, found a study in the Proceedings of the National Academy of Sciences. Researchers examined MRI brain scans from 21,738 participants in the United Kingdom, ages 40 to 70, and noted their body mass index, c-reactive protein levels (a blood marker of inflammation), and experiences of childhood maltreatment and adult trauma. A statistical model integrating this data showed that childhood maltreatment was correlated with increased body mass index and greater rates of trauma in adulthood. Participants with a history of maltreatment showed signs of dysfunction in their immune systems. Adding the MRI data to the model revealed that widespread abnormal increases and decreases in brain thickness and volume late in life were associated with greater body mass index, inflammation, and trauma that were borne out of childhood maltreatment.
DOI: 10.1073/pnas.2304704121opens in new window

When presented with responses to ethical questions, most people rated the answers from an AI program more favorably than those coming from fellow humans, suggests research in Scientific Reports. Researchers asked 299 participants in the United States to rate the quality of moral evaluations (but did not mention that half were generated by people and half by an AI program). They found that the participants preferred the AI’s moral reasoning over the humans’ reasoning along almost all dimensions, including virtuousness, intelligence, and trustworthiness. Next, the researchers revealed to the participants that an AI had authored half the responses. They then asked the participants to identify the source of each evaluation (human or AI), which they did correctly between 58% and 82% of the time, levels that were significantly greater than chance. The researchers suggest that the participants could distinguish human from AI responses primarily because they deemed that the best evaluations likely came from the AI.
DOI: 10.1038/s41598-024-58087-7opens in new window

AI-generated measurements of social media can create more precise assessments of depression and anxiety than population surveys, according to a study in npj Digital Medicine. Researchers used an AI system to analyze nearly 1 billion tweets from more than 2 million geolocated users in the United States living in 1,418 counties representing all 50 states. They found that the AI system’s assessment of public well-being was comparable to that of the biggest public polls but more granular in terms of time and space, with the AI determining rates on a shorter timescale and within smaller geographical regions than the polls. As a sign of its external validity, the AI system’s predictions of other community measurements commonly associated with mental health, such as mortality rates, were better than the public surveys. The AI-produced scores were also better at predicting other social, economic, and political variables.
DOI: 10.1038/s41746-024-01100-0opens in new window