Creation and Recollection of Artificial Memories

Post by Deborah Joye

What's the science?

As animals move through the world, they learn to associate certain environmental cues with events which help them survive. These memory associations have been localized to specific brain regions and underlying circuits. Specifically, the way that mice create odor associations is well-characterized and occurs similarly across individuals. If our detailed understanding of odor memory in mice is correct, it leads to an interesting question: can a false memory be created through direct stimulation of the brain? And can this memory be recalled by something in the real world, as if it actually happened? This week in Nature Neuroscience, Vetere and colleagues use association training along with direct cell-type- and region-specific brain stimulation to demonstrate that both good and bad artificial memory associations can be created in mice and recalled by a real-world cue.

How did they do it?

The authors first trained mice to form a real odor association by pairing a specific odor (acetophenone) with a mild foot shock. Since memory associations depend on presentation of the odor and the foot shock right after one another, the authors included conditions where the odor or the foot shock were presented independently or were presented 24 hours apart from one another (too far apart for memory associations to form). In the memory test, mice were put into a box with two chambers – one with the trained odor and the other with a new odor. The idea is that if a memory of the odor-foot shock pairing has been formed, the mouse will avoid the compartment containing the foot-shock-paired odor.

The authors then repeated this experimental structure multiple times. First, to test whether the odor association could be formed by direct brain stimulation, the authors genetically altered acetophenone-specific olfactory cells to be activated by a laser (called optogenetic stimulation). Mice were then exposed to the same two-chamber box containing either acetophenone or another smell. Second, to test that both the odor and the foot-shock could be created using only direct brain stimulation, the authors used optogenetics to activate both olfactory cells and cells in the brain associated with positive or negative experience (laterodorsal tegmental or lateral habenula inputs to the ventral tegmental area, respectively). Finally, since cells in the basolateral amygdala are important for memory associations, the authors tested whether this region is likewise necessary for artificial memory associations by chemogenetically silencing it. They virally expressed an inhibitory DREADD (designer receptor exclusively activated by designer drug) to turn off basolateral amygdala cells in the presence of a specific chemical. The authors then repeated the memory association experiment, using the inhibitory DREADD to block basolateral amygdala activity a subset of mice.

What did they find?

The authors found that mice formed memory associations between an odor and a foot-shock, as expected. Direct optogenetic stimulation of acetophenone-sensitive olfactory cells paired with foot shock also produced a memory association. This memory association could be recalled by mice, as evidenced by an avoidance of the compartment containing the actual acetophenone odor. Optogenetic stimulation of olfactory cells paired with stimulation of negative experience cells produced a behavioral aversion to the acetophenone odor, like mice that had been exposed to a real foot shock. In contrast, optogenetic stimulation of olfactory cells paired with stimulation of positive experience cells produced a behavioral attraction to the acetophenone odor. In all instances, artificial memory associations were only created if stimulation of the brain regions occurred close together in time, as with real memory associations. The authors also found that real and artificial memory associations engaged similar neural circuits, as shown by markers of cellular activation. Finally, when researchers chemogenetically blocked basolateral amygdala activity with DREADDs, expressions of both real and artificial memory associations were lost.

What's the impact?

This study successfully creates a fully artificial memory in mice through direct brain stimulation. The characteristics of this artificial memory were like a natural memory: time dependent, similar brain circuits, behavioral responses were specific to the trained cue, and memory expression depended on the basolateral amygdala. This is the first study to show that artificially created memories can be recalled by a real-world cue. This study presents a valuable window into how memory associations are created and integrated with real-world experience.

Vetere et al., Memory formation in the absence of experience, Nature Neuroscience (2019).Access the original scientific publication here.

Why We Make Decisions Together

Post by Anastasia Sares

What’s the science?

Collective decision-making behaviors have been demonstrated in social animals like bees, ants, and fish. Humans are also social creatures, and like these other species, we often make decisions together, even though we strongly value autonomy. What benefit is there in giving up some of our autonomy and making a decision as a part of a group? This week in Nature Human Behavior, El Zein and colleagues suggested that we decide together in order to dilute risks and negative outcomes.

What do we already know?

Previous research in this area has focused on whether collective decision-making results in a better decision overall. In some circumstances this process is helpful, but other times a group can get derailed and make a non-optimal decision. Since group decisions aren’t necessarily better in terms of accuracy, it is important to understand why we bother with them at all. After all, we like to have a choice when deciding what kind of product to buy, or what career to pursue. Some decisions are made together out of social obligation or a sense of fairness, but this may not account for all of the collective decision-making situations we observe.

What’s new?

The authors propose that one of the main reasons that individuals make decisions collectively is because it minimizes the risk taken by any one member. It’s what animals do when they herd or flock together, making it less likely that any one member is attacked (known as the dilution effect). Humans, even when they are not in physical danger, are very averse to certain emotional risks, especially regret or responsibility for a negative outcome. Making a decision as part of a group reduces the feeling of personal responsibility and can help us to cope with the stress of difficult decisions (like parents deciding whether or not to keep an injured child on life support). It may also protect us from social backlash (like when “whistleblowers” call out bad behavior of very powerful individuals). However, when taking the group perspective and not the individual perspective, the decrease of personal responsibility comes with its own problems: at worst, no one assumes responsibility for negative outcomes, and they are not addressed at all. Think of the bystander effect, where witnesses to an emergency situation are less likely to step in and help if others are present, or the tragedy of the commons, where individuals tend to over-use common resources.

What's the bottom line?

There are a number of factors that push us toward collective decision-making: social inclusion and fairness, the idea that we are smarter together, and, as El Zein and colleagues emphasize, protection from negative consequences. In the future, it will be important to evaluate the relative contribution of these different factors in the drive to collective decision-making. This will help us better understand the behavior of the different social groups and governing bodies that permeate human society. Perhaps then we’ll know when to say, “many hands make light work” and when to say, “too many cooks spoil the broth.”

El Zein et al. Shared responsibility in collective decisions. Nature Human Behavior (2019).Access the original scientific publication here.

Maternal Genetics, Early Life Exposures and Risk for Neurodevelopmental Disorders

Post by Stephanie Williams

What's the science?

Factors that occur before and during pregnancy, including maternal alcohol intake, poor nutrition, and stressful life events, have previously been linked with a higher risk for neurodevelopmental disorders in offspring. Most studies that assess the link between lifestyle factors and neurodevelopmental disorder risk do not properly account for maternal genotype, and could therefore be confounded by the genetics of the mother. This week in Jama JAMA Psychiatry, Leppert and colleagues assess the relationship between maternal lifestyle factors and maternal polygenic risk scores for neurodevelopmental conditions.

How did they do it?                                            

The authors analyzed data collected in an ongoing longitudinal study “Children of the 90s” of a large number (N=7921) of mothers in the United Kingdom. The dataset included genetic data and information on the health and lifestyles of the children and their mothers. Mothers were asked to report their drinking and smoking habits, use of antidepressants, and nutritional supplements. They were assigned a ‘stressful life score’ based on a self-report of whether they had experienced 18 different stressful life events.  Analysis of the mother’s blood was used to determine nutritional status and toxin exposure. Obstetric records were used to assess adverse birth events like low birth weight. The authors used previously identified risk alleles to calculate a polygenic risk score, which is a score calculated from the number of variants (single nucleotide polymorphisms) for a gene that indicates a certain amount of genetic risk for a disorder or disease. The authors were interested in investigating whether the polygenic risk scores for 3 different disorders, Attention Deficit Disorder (ADHD), Autism Spectrum Disorder, and Schizophrenia were associated with the lifestyle factors mentioned above. The authors calculated an association score for each of the lifestyle-related variables and the polygenic risk score for the 3 diseases.

What did they find?

The authors identified several associations between lifestyle factors and maternal risk alleles. Specifically, the authors identified associations between genetic risk for ADHD and SCZ and higher risk of smoking, and pregnancy BMI (higher BMI for ADHD, lower BMI for schizophrenia). Genetic risk for ADHD was associated with several additional factors, including infections, use of acetaminophen during late pregnancy, lower blood levels of mercury and higher blood levels of cadmium. The only factors found to be associated with polygenic risk score for all three disorders were maternal stressful life events during pregnancy and a higher risk for experiencing severe depression. Importantly, the authors point out that they found little evidence for associations between genetic risk for autism and schizophrenia and lifestyle factors (except for BMI with schizophrenia).

What's the impact?

This work emphasizes the importance of accounting for maternal genetics when drawing conclusions about lifestyle factors that affect risk for neurodevelopmental disorders. The authors identified for the first time associations between genetic risk for ADHD and several factors, including infections, acetaminophen, and blood levels of toxins. The results of the study could inform the care and treatment of pregnant women carrying risk alleles for neurodevelopmental disorders.

Leppert, B et al. (2019) Association of maternal neurodevelopmental risk alleles with early-life exposures. JAMA Psychiatry. Access the original scientific publication here