Both the brain’s hippocampus and the orbitofrontal cortex are involved in inference-based behavior, according to a Northwestern Medicine study published in PLOS Biology.
The previous prevailing theory was that the hippocampus catalogued variables and relationships among them, while the orbitofrontal cortex made choices. But this study demonstrates the two regions work more closely together than previously understood, according to Thorsten Kahnt, PhD, assistant professor in the Ken and Ruth Davee Department of Neurology Division of Behavioral Neurology and senior author of the study.
“Maybe the orbitofrontal cortex is accessing information from the hippocampus, or the hippocampus is helping find information at its end,” said Kahnt, also an assistant professor of Psychiatry and Behavioral Sciences and at the Weinberg College of Arts and Sciences. “It’s not clear what they’re specifically doing, but the overarching concept is in line with the strong anatomical connections between these regions.”
The idea of inference-based behavior goes to the very underpinnings of behavioral neurology: behavior, both in humans and in animals, cannot be explained purely based on experience.
“If you could only make decisions based on what you’ve already experienced, you would be stuck,” Kahnt said. “That ability to synthesize previous experience and make a mental simulation or inference to impact future decision making is very important.”
Kahnt and his collaborators used a sensory preconditioning paradigm to study inference-based behavior. Participants repeatedly observed two pairs of visual stimuli, “A-B” and “C-D”. Stimulus “B” was then associated with a reward, while stimulus “D” was associated with no reward.
Then, the investigators presented stimuli “A” and “C”, measuring activity in the orbitofrontal cortex and in the hippocampus using functional magnetic resonance imaging (fMRI). Because “A” and “C” had never been directly associated with reward, participants had to make an inference that “A” would likely lead to a reward while “C” would not.
Kahnt found that during the initial learning phase, signatures matching the “A-B” and “C-D” pairs were found in both the hippocampus and in the orbitofrontal cortex.
“We previously believed that the hippocampus learns the associations and the orbitofrontal cortex uses these associations later on, but we found that the orbitofrontal cortex is involved from the beginning,” Kahnt said.
During the actual inference stage of the test, when stimuli “A” and “C” were administered, the investigators found only the orbitofrontal cortex was activated in a sequential fashion; stimulus “A” triggered a representation of stimulus “B,” which reactivated the reward. Furthermore, activity in the orbitofrontal cortex and hippocampus was synchronized when participants made inferences.
“This is in line with the idea that people mentally step through these different stages in order to make a prediction, and that both regions work together to achieve this” Kahnt said.
The fact that the orbitofrontal cortex is involved from the beginning was not previously known, but is in line with the interconnectivity of the two regions. Further, many psychiatric disorders involve some level of impairment in decision-making, and many are associated with orbitofrontal cortex dysfunction. Learning more about the neural architecture of decision-making may one day lead to better treatments, according to Kahnt.
“If we can figure out what is the specific functional contribution of the orbitofrontal cortex to behavior, and we can develop techniques to stimulate or activate these brain regions, we may be able to treat disorders,” Kahnt said.
This work was support by National Institute on Drug Abuse (NIDA) grant R03DA040668 and the Intramural Research Program at NIDA grant ZIA-DA000587.