Interactions between human orbitofrontal cortex and hippocampus support model-based inference

Abstract

Internal representations of relationships between events in the external world can be utilized to infer outcomes when direct experience is lacking. This process is thought to involve the orbitofrontal cortex (OFC) and hippocampus (HPC), but there is little evidence regarding the relative role of these areas and their interactions in inference. Here, we used a sensory preconditioning task and pattern-based neuroimaging to study this question. We found that associations among value-neutral cues were acquired in both regions during preconditioning but that value-related information was only represented in the OFC at the time of the probe test. Importantly, inference was accompanied by representations of associated cues and inferred outcomes in the OFC, as well as by increased HPC-OFC connectivity. These findings suggest that the OFC and HPC represent only partially overlapping information and that interactions between the two regions support model-based inference.

Targeted stimulation of an orbitofrontal network disrupts decisions based on inferred, not experienced outcomes

“…In both groups, recognition memory was significantly above chance (SHAM: t(21)=5.01, p<0.001; STIM: t(20)=2.70, p=0.013), and there was no difference between groups (t(41)=1.34, p=0.188, permutation test, p=0.129 Figure 4A). Moreover, as in our previous study (Wang et al, 2020), recognition memory was significantly correlated with inference-based responding (r=0.51, p=0.0005, Figure 4B). These correlations were significant within each group (SHAM,r=0.38,p=0.039, one-tailed; STIM, r=0.55, p=0.01) and did not differ between groups (Z=-0.93, p=0.178).…”

“…As such, our findings suggest that the contribution of OFC to decision making may be limited to situations that require model-based planning, and that choices based on direct experience may rely on value computations in other brain areas, such as the amygdala or striatum (Paton et al, 2006;Cox and Witten, 2019). This proposal is seemingly at odds with a large number of studies across different species that has shown neural correlates of both inferred and directly experienced value in OFC (Hare et al, 2009;Schoenbaum et al, 2009;Barron et al, 2013;Stalnaker et al, 2014;Howard et al, 2015;Padoa-Schioppa and Conen, 2017;Suzuki et al, 2017;Klein-Flugge et al, 2019;Lopez-Persem et al, 2020;Wang et al, 2020). Why would OFC represent value signals if they are not required for behavior?…”

“…Eight food odors (four sweet: strawberry, caramel, gingerbread, and yellow cake; four savory: potato chip, pot roast, garlic, and pizza) were provided by Kerry (Melrose Park, IL) and International Flavors and Fragrances (New York, NY). Odors were delivered to participants' nose using a custom-built and computer-controlled olfactometer (Howard et al, 2020;Wang et al, 2020). The olfactometer was equipped with two independent mass flow controllers (Alicat, Tucson, AZ), which allow dilution of any given odorant with odorless air.…”

“…In the current study ( Fig. 1A), we applied this novel OFC-targeted brain stimulation approach in the context of a sensory preconditioning task that was designed to isolate inference-based behavior from responding that can be based on direct experience (Jones et al, 2012;Wimmer and Shohamy, 2012;Wang et al, 2020). This task consists of three phases ( Fig.…”

Prospective Representations in Rat Orbitofrontal Ensembles

“…Meanwhile, the past sensory cues, rewards, spatial directions, and behavioral choices are also reported to be reflected in the OFC neural activities (Feierstein, Quirk, Uchida, Sosulski, & Mainen, 2006;Kennerley, Behrens, & Wallis, 2011;Nogueira et al, 2017;Riceberg & Shapiro, 2017;Saez, Saez, Paton, Lau, & Salzman, 2017;Young & Shapiro, 2011;Zhou, Jia, Feng, Bao, & Luo, 2015). Such findings have been taken as evidence that the OFC, together with contributions from other interconnected brain regions such as the hippocampus, might provide a neural mechanism with which animals could mentally travel through a task model in time and recall the past events and simulate future outcomes (Behrens et al, 2018;Wang, Schoenbaum, & Kahnt, 2020;Wikenheiser & Schoenbaum, 2016; R. C. Wilson, Takahashi, Schoenbaum, & Niv, 2014).…”