Can Wearable Devices Be Used to Identify Patients at Risk Following a Traumatic Event?

Post by Kulpreet Cheema

The takeaway

Wrist-wearable devices can be used to track pain, anxiety, and sleep-related outcomes in individuals who’ve suffered from exposure to trauma in order to identify who might be at risk of persistent symptoms vs. recovery. Less movement during the day and more sleep disruption at night was associated with increased pain while changes in the number of sleep-wake transitions were associated with changes in sleep, pain, and anxiety.

What's the science?

After a traumatic event, socioeconomically disadvantaged individuals are at a greater risk of developing adverse posttraumatic neuropsychiatric sequelae (APNS). Some APNS symptoms include pain, depression, anxiety, and sleep disruption. Data pertaining to these symptoms can be collected with wrist-wearable devices, however, the utility of such devices to measure APNS symptoms following trauma exposure is unknown. This week in JAMA Psychiatry, Straus and colleagues aimed to evaluate if wrist-wearable devices can detect biomarkers to predict recovery after a traumatic experience.

How did they do it?

Data from 2021 participants, who had come to an emergency department within 72 hours of experiencing a traumatic experience, was analyzed. After their emergency room visit, participants wore a watch by Verily Life Sciences for eight weeks and self-reported ten symptoms related to APNS. Some APNS symptoms included in the self-report were pain, depressive symptoms, sleep discontinuity, and anxiety. The authors also collected accelerometry data from the watch, from which rest-activity features of sleep and average daily activity were estimated. Linear mixed models were used to derive and validate the relationships between the self-reported symptom data and the 24-hour rest-activity features.

What did they find?

The authors found nine significant rest-activity biomarkers correlated with APNS symptoms of pain, sleep, and anxiety. Reduced daily activity variance was positively associated with increased pain, meaning that the individuals who reported pain were less active during the day and more restless during the night.

In addition, individuals who reported having anxiety and sleep quality difficulties also had more sleep/wake transitions. Similarly, fewer sleep-wake transitions were associated with improved anxiety and sleep quality. This suggests a bidirectional relationship between sleep quality and anxiety. Further, it suggests that improving sleep quality might help improve anxiety in people dealing with traumatic stress.

What's the impact?

The study was the first to examine the utility of rest-activity data obtained from a wearable device in tracking APNS-related outcomes in individuals dealing with traumatic stress. The findings suggest that the data obtained from wearable devices could be used to screen for which individuals are at risk of persistence of APNS symptoms. In addition, these biomarkers can be used in the clinic to help understand the recovery process and provide appropriate treatment approaches.

Access the original scientific publication here

A Brain Region Critical in Creating Cognitive Maps

Post by Anastasia Sares

The takeaway

The lateral orbito-frontal cortex (lOFC), is a brain region in the frontal lobe, just above the eyes, that helps us interact with the world by mapping associations between different events. While previously it has been given the role of “deploying” these maps (deciding when to use them), this new study suggests that the lOFC might be involved in creating those association maps in the first place. 

What's the science?

Reinforcement learning is how we develop knowledge through our interactions with the environment. When we perform actions, we get feedback about the results of those actions, and if they result in a reward, we are more likely to repeat those actions again later. Reinforcement learning can be divided into model-free learning, where we learn directly from the consequences of our actions, and model-based learning, where we create a map of associations (in other words, a model) that can help us make decisions based on context.

This week in Nature Neuroscience, Costa and colleagues tested what happens when the lOFC is taken out of commission.

How did they do it?

The authors first trained rats to associate sound stimuli with getting food pellets, with one sound predicting banana-flavored pellets and another sound predicting bacon-flavored pellets. The rats learned that sounds were associated with a food reward, but at this phase, there was no reason for them to learn the distinction between the two sounds.

In the second phase, both types of pellets were again given to the rats, but after they ate one kind of pellet (let’s say the banana one), they would be administered lithium chloride, which made them feel nauseous. Here, the rats learned a connection between one type of pellet and nausea, but the sounds were not involved.

In the final phase, the rats were again presented with the sounds, and the authors measured how long the animal searched for a food pellet after hearing the sound. The rats had never learned a direct association between the sounds and nausea, so they would have to create this connection themselves based on information from the first two phases. Specifically, they would need to rely on an internally-created “model” or associative map linking the relevant sound to the bad pellet (even though the type of sound had not mattered before).

Two groups of rats participated in this experiment. The first was a control group, and the second group had been bio-engineered (by applying a custom viral agent) so that the lOFC could be temporarily “turned off” by administering a drug just before the learning session in the first phase. If the lOFC was involved in creating association maps, then turning it off for in the first learning phase should make the rats unable to learn the association between sound cues and nausea. However, if the lOFC is only involved in “deploying” these association maps, then turning it off in phase 1 should have no effect on their ability to use them later.

What did they find?

When the control rats heard the specific sound that predicted the “bad” pellets (the ones that had made them nauseous), they did not go to the food bowl as often as when they heard the other sound. This means they had clearly learned the distinction between different sounds, the foods they signaled, and the predicted result of eating those foods, and were able to put this information together in the final test to guide behavior. In contrast, rats whose lOFC was deactivated in phase 1 reduced their trips to the food bowl for both sounds at test time. These rats thus had the ability to form a rudimentary map to guide behavior, however, this map lacked the specific information that would allow them to choose the “good” pellets and avoid the “bad” ones.

The authors also put the rats through an object recognition task, which required them to distinguish between new and old objects. In this case, the rats with the deactivated lOFC performed similarly to the control rats, indicating that the lOFC is not involved in basic learning.

Finally, the authors created some mathematical models to try and reproduce the results. They found that the best explanation of the impaired animals’ behavior was an imprecise mapping from sounds to pellet flavors during the first learning phase. 

What's the impact?

Model-based reinforcement learning is a crucial function of the brain, and abnormalities in this system can lead to maladaptive behavior. For example, problems with association maps are present in mental illnesses such as schizophrenia, substance abuse, and obsessive-compulsive disorder. Therefore, understanding more about how this area of the brain works may help us better diagnose and treat these conditions.

Access the original scientific publication here.

How Habitual Checking of Social Media Changes the Adolescent Brain

Post by Christopher Chen

The takeaway

Social media use has become nearly universal among American teenagers but its possible effects on adolescent brain development remain unclear. A new study indicates that habitual checking of social media may be disrupting the normal development of brain circuits linked to reward processing and cognitive control in the young adult brain.

What's the science?

The brain undergoes drastic changes during adolescence, particularly in regions associated with motivation, reward processing, and cognitive control. Furthermore, the maturation of these regions allows for developmentally normative neural and behavioral responses to social feedback. With its use of immediate feedback in the form of “likes” or notifications as well as its widespread use in adolescent populations, exploring how habitual social media use affects social feedback-based networks in the adolescent brain is more relevant than ever. In a recent article in JAMA Pediatrics, Maza et al. investigate differences in brain activity levels in regions associated with social feedback in adolescents who habitually check social media. 

How did they do it?

The experiment looked at approximately 200 students aged 12-13 from three middle schools in rural North Carolina. First, experimenters had the students self-report how often they checked three social media sites (Facebook, Instagram, and Snapchat). Based on this data, the participants were divided into three groups based on their rate of checking social media: habitual, moderate, and non habitual. Experimenters then used functional brain imaging (fMRI) to measure brain activity of participants during a Social Incentive Delay task, a cognitive task designed to measure anticipation of social feedback. Following initial measurements, the students took part in the same experiment each year for the next two years.

Following the completion of the study, experimenters compiled and combined data from both the Social Incentive Delay task and fMRI imaging to measure activation levels of specific brain regions from each participant during the cognitive task. They then used these individual datasets to make a general linear regression model measuring the change in brain activity levels in all three groups over time.  

What did they find?

From their generalized linear regression models, experimenters found that brain activation patterns were significantly different in habitual and non habitual checkers of social media. Interestingly, these patterns were most distinct in brain regions linked to social feedback: the insular and prefrontal cortex, ventral striatum, and amygdala. Habitual social media checkers showed a decreased sensitivity to social anticipation at 12 years of age.

In habitual checkers of social media, linear regression models revealed an increase in brain activity during social anticipation across all four brain regions over time. In non habitual and moderate checkers of social media, linear regression models revealed the opposite: brain activity decreased in all four regions. These divergent results in brain activity changes in habitual and non habitual checkers of social media suggest high social media usage impacts developmental trajectories of neural circuits linked to social feedback and cognitive control.

What's the impact?

The negative functional consequences – if any – of these increases in brain activity in habitual checkers of social media are unclear. Whether the rate of social media usage directly causes or is simply correlated to these neurological changes also remains to be seen. However, this study is the first to show distinct differences in brain development in adolescents who habitually check social media, suggesting that social media is indeed changing the young adult brain.     

Access the original scientific publication here.