Scientists have found that damage to a specific brain region, the magnocellular mediodorsal thalamus, in monkeys caused behaviors similar to those seen in humans with paranoia, such as heightened sensitivity to changes and difficulty learning from outcomes. These findings, published in Cell Reports, suggest that this brain region plays a critical role in the development of paranoia, providing a potential target for future treatments.
Previous research has established that the ability to form and adjust beliefs about actions and their consequences is essential for advanced cognition. Disruptions in this ability are linked to maladaptive cognitive and behavioral states, such as paranoia.
Paranoia is a mental state characterized by intense and irrational suspicion or mistrust of others, often involving beliefs that others intend to harm or deceive. It can manifest as exaggerated feelings of persecution or conspiracies against oneself.
Prior studies have implicated various brain regions, including the prefrontal cortex and the thalamus, in the processes of belief updating and paranoia. However, the exact mechanisms and specific brain regions responsible for these disruptions remain unclear.
One approach to studying these mechanisms has been the use of probabilistic reversal learning tasks, which require individuals to adapt their choices based on changing reward contingencies. This method has been effective in identifying behavioral patterns associated with flexibility and persistence in decision-making. While these studies have provided valuable insights, they often focus solely on human subjects or a single species, limiting the ability to generalize findings across different species and neural architectures.
The authors behind the new research sought to address these limitations by adopting a cross-species approach that aligns data from monkeys with human data
The study involved a total of twenty male rhesus macaque monkeys and 1,225 online human participants, categorized based on their levels of paranoia. The monkeys were divided into groups with excitotoxic lesions in either the magnocellular mediodorsal thalamus (MDmc) or the orbitofrontal cortex (OFC), and a control group with no lesions. The lesions were created surgically, and the monkeys’ subsequent behavior was compared to that of the control group.
Both monkeys and human participants completed a probabilistic reversal learning task, which required them to choose between three options with changing reward probabilities. For the monkeys, this task was performed on a touch-sensitive monitor, with food pellets as rewards. In the human version of the task, participants received points instead of food. The task involved an initial phase where one option had the highest probability of reward, followed by a reversal phase where the reward probabilities changed, requiring participants to adjust their choices accordingly.
“Participants have to figure out what’s the best target, and when there’s a perceived change in the environment, the participant then has to find the new best target,” said Steve Chang, associate professor of psychology and of neuroscience in Yale’s Faculty of Arts and Sciences and co-senior author of the study.
Behavioral data were collected on win-switching (changing choices after a reward) and lose-staying (repeating choices after no reward) behaviors, indicative of flexibility and persistence in decision-making. Computational modeling using the hierarchical Gaussian filter (HGF) was employed to estimate belief parameters related to volatility (the tendency to expect changes) and value learning (the rate of learning about the values of each option). These parameters helped quantify how participants updated their beliefs in response to changes in reward contingencies.
“Not only did we use data in which monkeys and humans performed the same task, we also applied the same computational analysis to both datasets,” said Philip Corlett, an associate professor of psychiatry at Yale School of Medicine and co-senior author of the study. “The computational model is essentially a series of equations that we can use to try to explain the behavior, and here it serves as the common language between the human and monkey data and allows us to compare the two and see how the monkey data relates to the human data.”
The study found significant differences in behavior and belief updating between the lesion groups and the control group of monkeys. Monkeys with lesions in the MDmc exhibited increased win-switching and reduced lose-staying behaviors, indicating heightened sensitivity to changes in reward contingencies.
These monkeys also showed elevated volatility beliefs and decreased value learning rates, particularly after the reversal in reward probabilities. This pattern suggests that MDmc lesions lead to an exaggerated response to perceived changes in the environment, similar to behaviors observed in paranoid individuals.
In contrast, monkeys with OFC lesions displayed the opposite pattern: decreased win-switching, increased lose-staying, and elevated value learning rates, with no significant change in volatility beliefs. This behavior indicates a failure to adapt to changes in reward contingencies, leading to more persistent and less flexible decision-making. The computational models supported these observations, showing distinct effects of the lesions on belief updating parameters.
When comparing these findings to human participants, those with high levels of paranoia exhibited similar patterns to the MDmc-lesioned monkeys. High-paranoia individuals showed higher win-switching rates and elevated volatility beliefs, along with lower value learning rates. These parallels suggest that the MDmc plays a critical role in the processes underlying paranoia, and that disruptions in this region can lead to behaviors associated with excessive sensitivity to environmental changes and difficulties in learning from outcomes.
By demonstrating how specific brain lesions affect decision-making and belief updating in monkeys and drawing parallels to human paranoia, the research offers new insights into the neural mechanisms of paranoia.
“It allows us to ask how we can translate what we learn in simpler species — like rats, mice, maybe even invertebrates — to understand human cognition,” said Corlett, who, along with Chang, is a member of Yale’s Wu Tsai Institute, which aims to accelerate understanding of human cognition.
“It could also enable researchers to evaluate the precise mechanisms by which pharmaceutical treatments influence brain activity related to states like paranoia. “And maybe down the road we can use it to find new ways to reduce paranoia in humans,” added Chang.
The study, “Lesions to the mediodorsal thalamus, but not orbitofrontal cortex, enhance volatility beliefs linked to paranoia,” was authored by Praveen Suthaharan, Summer L. Thompson, Rosa A. Rossi-Goldthorpe, Peter H. Rudebeck, Mark E. Walton, Subhojit Chakraborty, Maryann P. Noonan, Vincent D. Costa, Elisabeth A. Murray, Christoph D. Mathys, Stephanie M. Groman, Anna S. Mitchell, Jane R. Taylor, Philip R. Corlett, and Steve W.C. Chang.
