探討小腦與腹側被蓋區迴路於酬賞型決策歷程中之角色

Abstract

在日常生活中，「決策行為」不可或缺，它讓我們面對環境變化時得以採取相應的行動。在過往的研究當中，酬賞最大化與懲罰最小化被視為決策最具體的兩大目標。因此，了解腦如何應對酬賞經驗、如何處理酬賞經驗進而影響個體的決策，是非常重要且極需解答的課題。有別於大腦，「小腦」一般被認為主要處理動作平衡，但如今也開始被「決策行為」相關研究重視。許多文獻提出小腦參與了酬賞處理、選擇決定、以及調控了中腦多巴胺神經元。然而這些研究對於小腦如何直接的處理酬賞、如何與酬賞迴路溝通、以及如何引導做出決策的因果關係仍然欠缺說明。為了回答這些重要的問題，我們建立小鼠酬賞型決策行為模型，並結合了活體電生理紀錄以及光遺傳學來了解小腦如何處理酬賞決策歷程。本研究發現，在小腦深部核(deep cerebellar nuclei, DCN)的神經元當中，31.9 %的神經細胞在處理酬賞時增加放電頻率，18.4 %的神經細胞則在小鼠選擇左側以及右側時，分別展現完全相反的放電模式。由於小腦深部核為小腦主要對外聯絡的腦區，小腦酬賞訊息極可能與中腦腹側被蓋區(ventral tegmental area, VTA)有所聯繫。準此，進一步利用逆向追蹤劑(retrograde tracer)來標定腹側被蓋區的上游後發現，小腦深部核與中腦腹側被蓋區有著直接的神經投射。此外，實驗進一步發現腹側被蓋區中有兩群神經元分別採取了增加放電頻率以及減低放電頻率的方式來處理酬賞訊息。最後，應用光遺傳學方法操弄小腦深部核至中腦腹側被蓋區迴路的神經元後發現，改變小腦深部核神經元末梢的突觸放電能夠直接的影響腹側被蓋區神經元的放電頻率；有趣的是，在小鼠進行決策行為當下，立即改變迴路中的神經元放電能改變決策行為並且影響尋求酬賞的動機。綜合以上研究結果，藉由解剖層次到行為功能層次等方法，說明了小腦深部核能夠直接的處理酬賞訊息；更重要的，小腦經由與腹側被蓋區的迴路來影響後續行為。本研究驗證小腦參與酬賞決策歷程的神經機制，亦指引出後續探討小腦與認知功能的研究方向。

Decision making has been regarded as a consecutive process of selecting an action among numerous plans in response to diversified stimuli in the environment. Evidence has shown that the choice process is guided by maximizing rewards or avoiding punishment. Accordingly, how the brain encodes rewarding experiences and uses this information to regulate following actions is the million dollar question that has been left unanswered. Increasing evidence has proposed that the cerebellum, which stands for “little brain”, contributes to non-motor cognitive processes, especially in reward encoding, choice-related decision making on single-trial level, and the modulation of midbrain dopamine neurons. However, to date, the causal relationship and underlying mechanisms of the cerebellum and reward-encoding hubs remain much unclear. To address these questions, we took advantages of mice as a model organism and trained mice to perform a two-choice reward-based decision-making task in which mice acquired switching between two choice ports to receive maximum rewards. By means of in vivo electrophysiological recording and optogenetic intervention established on the behavioral model, we investigated how the neurons in the deep cerebellar nuclei (DCN), the sole outputs of the cerebellum, derive and convey information about rewards and actions on an unbiased population level. Overall, we identified that 31.9% of recorded neurons in the DCN constantly exhibited higher activities on rewarding outcomes over unrewarded trials. Moreover, our Fluoro-Gold retrograde neural tracing data confirmed that DCN neurons directly projected axons to the midbrain ventral tegmental area (VTA), a major region that contains the reward prediction error signal of midbrain dopamine neurons. Finally, the causal relationships of the DCN-VTA circuit was determined using optogenetics. Our data indicated that the optogenetic stimuli on the DCN-to-VTA pathway is sufficient to drive neural activity in the VTA and that activation of this pathway during reward seeking alters choice behaviors and prolongs reward-seeking time. Collectively, our results suggest that the cerebellar signals in the DCN modulate reward coding and distort choice behaviors via its direct inputs to the VTA. Our findings provided novel evidence that the cerebellum represents information regarding reward which might be integrated with midbrain reward signals to determine choice actions.