| Deficits in reward processing may result in individuals failing to accurately assess the reward value of different options,showing excessive pursuit of specific rewards or low hedonic response.Deficits can be observed in several psychiatric disorders such as addiction,depression,and schizophrenia(e.g.,anhedonia and amotivation),and result in serious physical and mental impairments.Currently,using pharmacological neuroenhancers to repair damage to the dopamine reward loop are considered to be the main therapeutic approaches to intervene in reward processing abnormalities and their associated psychiatric disorders.However,these studies have rarely quantitatively estimated the implicit,unobservable,neural activity processes and cognitive processing components of reward processing via employing computational models.On the other hand,previous studies have advocated the use of pharmacological treatments targeting dopamine for dopamine reward loop impairments,but the clinical translation of these drugs has been greatly limited by their non-negligible side effects(increased impulsive behaviors,induced addictions,and neurotoxicity responses).Recently,a series of animal model studies have revealed that the angiotensin Ⅱ type 1 receptor(AT1R)is co-expressed with dopamine in striatal brain regions,and AT1R blockade losartan can influence responses to reward stimuli by acting on dopamine receptor transmission.Some human neuroimaging work has also revealed that losartan can directly reduce the learning rate of loss outcomes during reward learning.However,the underlying computational neural features and molecular mechanisms of how AT1R act on reward processing remain to be elucidated.In this regard,the present study aims to identify:(1)computational behavioral features and neural mechanisms involved in reward anticipation and outcome assessment,(2)neuroimaging signatures of the AT1R blockade losartan that modulate reward processing,and(3)transcriptional features of AT1R-associated reward processing.Study 1 will combine the task-based f MRI and behavioral modeling approaches,and employ a reward probability selection task that incorporates reward learning and transfer phases to examine the computational behavioral features and neural underpinnings of computational behavior involving reward processing in a total of 46 healthy subjects(23 of each gender).Specifically,the present study found that individual reward selection accuracy varied with stimulus reward probability.The further computational modeling results reflected that individuals update their choices accompanied by reward feedback faster and prefer stimulus with a higher reward expectation value.While at the level of brain mechanisms,multivoxel pattern analysis revealed that neural representation patterns in the ventral medial prefrontal,posterior cingulate cortex,anterior insula,and dorsal medial prefrontal could differentiate between the expectation of reward and loss stimuli,whereas the anterior insula,and the ventral striatum were involved in the process of individuals’ assessment of reward and loss outcomes.The present study further explored the multivoxel neural representation patterns of expected value and prediction error and found that:(1)the superior frontal gyrus responded to higher expected value,whereas the anterior insula reflected lower expected value,and(2)the ventral striatum predicted higher prediction error,whereas the anterior insula responded primarily to lower prediction error.In addition,the present study proved that individuals’ whole-brain neural representation patterns during the reward transfer phase were able to recognize loss avoidance events and were reflected in a greater tendency to avoid loss events at the behavioral level.Together,this study portrays computational neural metrics involved in reward processing in a healthy population,and provides basic neural biological targets for further effective pharmacological modulations.To further characterize the modulatory effects of neuropharmacological tools on reward processing.Study 2 combines computational modeling and multivariate decoding with neuroimaging methods to determine the effects and underlying neural mechanisms of selective blockade of the AT1R on individuals’ positive or negative learning outcomes via using the reward probability selection task on a total of 61 healthy male subjects(losartan,n=30;placebo,n=31).During the reward learning phase,relative to the placebo group,losartan improved selection accuracy for the most difficult stimulus pairs by increasing expectancy sensitivity to the rewarding stimuli.Computational modeling results showed that losartan decreased the rate of learning to negative outcomes and increased the stability of choice behavior,while preserving learning to reward outcomes.At the neural level,these behavioral results are consistent with a pattern of increased reward prediction error signals in the orbitofrontal-striatal region and enhanced neural representation of reward outcomes in the ventral striatum by losartan.During the transfer phase,when approaching maximal reward,losartan accelerated reaction times and enhanced functional connectivity between the ventral striatum and the left dorsolateral prefrontal cortex.These findings shed light on the potential of losartan in reducing the effects of negative outcomes during reward learning and revealed corresponding macroscopic neuroimaging targets.However,the distribution of expression of the AT1R itself in the brain and how it relates to the micro-transcriptional signature of reward processing by interacting with other neurotransmitter receptors remain unclearAgainst the above-mentioned question,Study 3 first utilized brain-wide transcriptomic gene expression data from the Allen Human Brain Atlas to map m RNA expression specifically associated with the AT1R and demonstrated that the receptor is densely expressed in the subcortical network formed by the thalamus,striatum,and amygdala-hippocampus.Further metacognitive decoding revealed that the AT1R gene distribution brain map was strongly associated with memory and reward/motivation processes,and to a lesser extent with negative emotional processes.In contrast,acute blockade of AT1R using losartan inhibited spontaneous neural activity in areas of high AT1R expression(especially mediodorsal(MD)and ventromedial(VM)thalamus),while increasing spontaneous neural activity in the superior parietal gyrus,a region with relatively low AT1R gene expression.At the network level,blockade of AT1R using losartan increased functional connectivity between key nodes in the cortico-basal gangliathalamo-cortical circuits,particularly the MD and VM thalamus,the caudate nucleus,and the thalamus,as well as the anterior cingulate gyrus and the adjacent medial prefrontal cortex.Based on these results,the present study found that the regulatory effects of AT1Rs on functional connectivity in specific brain regions are related to the actions of dopaminergic,corticotropic hormone-releasing hormone,opioids,and acetylcholine systems.This suggests that some of the neural and behavioral functions of AT1Rs emerge from interactions with these signaling systems.Therefore,the present study clarifies functional relevance of AT1R gene expression brain mapping and its role in the regulation of spontaneous brain activity,providing an overarching bioinformatic characterization for understanding the pathways of AT1R influence.In conclusion,this study is the first to systematically investigate the computational neural representations of reward processing and utilize effective pharmacological means to regulate reward processing from perspectives of feature description,as well as elucidate the microscopic molecular mechanism of pharmacological regulation,which is of theoretical significance and applied value for the basic research on reward processing and the clinical treatment of psychiatric disorders associated with deficits in reward processing. |
PDF Full Text Request