In their everyday lives, animals face unusual scents and unfamiliar objects that pose difficult problems, such as: should the animal approach, avoid, or ignore strange stimuli? Choosing appropriately is crucial for survival—not only to prevent disasters but also to gain opportunities for new information or rewards. Our team, led by MCO graduate student Korleki Akiti and postdoc Iku Tsutsui-Kimura, identified a neural circuit that plays a key role in deciding whether the animals engage with or avoid novel stimuli.
Using automated analysis of behavior, we discovered that novelty exploration patterns can be categorized into three key behavioral components: initial investigatory approach, subsequent engagement, and avoidance. The brain uses size and intensity of novel stimuli to estimate threat during initial investigation and promotes transition to either engagement or avoidance accordingly.
People react to novelty in various ways. At the extremes, some suffer from fear of novelty (neophobia), while others fearlessly interact with novelty (neophilia). Both responses potentially disturb normal life. Furthermore, these behaviors are often dynamic; most people may be initially cautious toward novelty, and then gradually habituate.
Thus, to understand the neural mechanism of novelty-driven behaviors, it is important to account for both individual variability and how responses change over time. To precisely examine novelty exploration in mice, we used modern automated analyses based on machine learning techniques. These techniques included DeepLabCut for tracking multiple body parts, invented by a former MCO graduate student Mackenzie Weygandt Mathis and others, and MoSeq for identifying behavioral motifs, invented in Bob Datta’s lab. Combining these two techniques and neural ablation, we quantitatively characterized behavioral components in each of the 78 mice in the study.
We found that all the mice in our experiment exhibited an investigatory approach when they first encountered an unfamiliar Lego toy and then transitioned to either engagement or avoidance, depending on a property called physical salience—which can encompass many sensory modalities, such as the object’s size, color, loudness, smelliness, or appearing distinct from the environment. During initial investigation, animals most likely gather information, which then directs subsequent behaviors. For instance, an animal that investigates a strange object may find the object’s physical salience too intense may avoid similar objects in the future.
To understand the neural mechanism driving these sets of behaviors, we focused on a unique subset of dopamine neurons called TS dopamine neurons. In a previous paper (Menegas et al. 2018), we reported that TS dopamine neurons respond to an object’s physical salience.
We found that ablating TS dopamine neurons reduced investigatory approach and avoidance but increased engagement with novel objects, such as Lego blocks. In other words, TS dopamine neurons act as a neural mechanism to promote cautiousness or threat prediction. We also found that animals with higher TS dopamine activities tended to behave more cautiously toward a novel object. Because TS dopamine signals physical salience of stimuli, our results suggest that animals estimate potential threat using physical salience of a novel object, which determines whether the rodent engages or avoids.
Our idea is akin to the recent proposal for food value (Dayan 2022). While nutrition can be the main benefit of food, flavor may work as a positive reward prediction to shape eating even before knowing the ultimate outcome (nutrition), which generally comes when it is too late to decide whether to eat the next bite. If a tasty food ends up not being nutritious, animals can adjust to not eat it next time. Similarly, pain, injury, and death can be the ultimate outcome (threat) of novel stimuli, and physical salience may work as positive threat prediction to shape cautious behaviors. If a salient stimulus ends up being unthreatening, animals can behave more boldly when they next encounter the stimulus. Thus, we propose that animals use physical salience to form an initial threat prediction that guides avoidance behavior, which is critical to prevent potential harm.
Our study investigates the idea that being scared by a physically salient novel stimulus is a survival strategy that animals widely adopt to prevent disasters. Now that we found a neural circuit to properly estimate a potential threat, our next question is how the brain balances fear and curiosity toward novelty.