A paradigm shift in translational psychiatry through rodent neuroethology

Abstract

Mental disorders are a significant cause of disability worldwide. They profoundly affect individuals’ well-being and impose a substantial financial burden on societies and governments. However, despite decades of extensive research, the effectiveness of current therapeutics for mental disorders is often not satisfactory or well tolerated by the patient. Moreover, most novel therapeutic candidates fail in clinical testing during the most expensive phases (II and III), which results in the withdrawal of pharma companies from investing in the field. It also brings into question the effectiveness of using animal models in preclinical studies to discover new therapeutic agents and predict their potential for treating mental illnesses in humans. Here, we focus on rodents as animal models and propose that they are essential for preclinical investigations of candidate therapeutic agents’ mechanisms of action and for testing their safety and efficiency. Nevertheless, we argue that there is a need for a paradigm shift in the methodologies used to measure animal behavior in laboratory settings. Specifically, behavioral readouts obtained from short, highly controlled tests in impoverished environments and social contexts as proxies for complex human behavioral disorders might be of limited face validity. Conversely, animal models that are monitored in more naturalistic environments over long periods display complex and ethologically relevant behaviors that reflect evolutionarily conserved endophenotypes of translational value. We present how semi-natural setups in which groups of mice are individually tagged, and video recorded continuously can be attainable and affordable. Moreover, novel open-source machine-learning techniques for pose estimation enable continuous and automatic tracking of individual body parts in groups of rodents over long periods. The trajectories of each individual animal can further be subjected to supervised machine learning algorithms for automatic detection of specific behaviors (e.g., chasing, biting, or fleeing) or unsupervised automatic detection of behavioral motifs (e.g., stereotypical movements that might be harder to name or label manually). Compared to studies of animals in the wild, semi-natural environments are more compatible with neural and genetic manipulation techniques. As such, they can be used to study the neurobiological mechanisms underlying naturalistic behavior. Hence, we suggest that such a paradigm possesses the best out of classical ethology and the reductive behaviorist approach and may provide a breakthrough in discovering new efficient therapies for mental illnesses.

Unsupervised discovery of family specific vocal usage in the Mongolian gerbil

“…In principle, this requires continuous, undisturbed, and longitudinal recording that take place in naturalistic contexts. This need has led to the emergence of a variety of video tools for long-term monitoring, and an associated suite of machine-learning based analyses (Datta et al, 2019; Pereira et al, 2020; Shemesh & Chen, 2023). In contrast, most studies of natural behavior do not acquire and analyze acoustic information over prolonged periods of the life cycle.…”

“…This can be particularly severe for family behaviors which occur in protected or remote environments, such as burrows in the case of fossorial rodent species like naked mole-rats and Mongolian gerbils (Brett 1986;Scheibler 2006). Some of these limitations have been addressed with laboratory environments that partially recapitulate real-world features (Shemesh & Chen 2023). However, these studies generally focused on relatively short periods of data collection that consider single animals or dyads with no prior relationship.…”

Long-term tracking of social structure in groups of rats

“…Long-term studies to examine group behavior may be an essential component to include in translational research applications, for example, in the testing of psychotherapeutic drugs to treat social anxiety [24, 25]. In particular, an important topic for future work is to establish standardized and reproducible tests and measures that are properly representative of a full range of social behavior [27, 69]. Another important application is understanding the neural basis of social interactions.…”

“…It is therefore not surprising that the translational relevance of these tests is limited [15, 16]. Incorporating environmental and social complexity into experiments can increase the generalizability of conclusions drawn from laboratory studies [17, 18]. With rats, previous work has shown that individuals within a group have a social status related to dominance and that aggression and avoidance behavior are key elements of social interactions [13, 19, 20].…”

“…It is therefore not surprising that the translational relevance of these tests is limited [24, 25]. Incorporating environmental and social complexity into experiments can increase the generalizability of conclusions drawn from laboratory studies [26, 27].…”