大鼠前扣帶皮質迴及島迴皮層中同向鏡像神經元參與其同理心行為

Abstract

文獻證實齧齒類動物具有親社會行為 (prosocial behavior)，包括同理相關的動作，然而其神經基礎目前仍不清楚。人類相關研究顯示，大腦中的前皮質扣帶迴和島腦參與疼痛同理心的運作，並推測當受試者感受到疼痛刺激或是感受到他人處於疼痛狀態時，這些區域的神經元活性會改變。本論文的研究目的在於釐清這個議題，觀察當動物本身接受雷射熱痛刺激，或是看到隔壁的同伴接受雷射熱痛刺激時，其行為以及前皮質扣帶迴和島腦神經元的反應（第二章）。結果顯示，觀察者動物不僅在自己接受疼痛刺激的時候其僵化反應 (freezing behavior) 表現的百分比會上升，在觀看同伴接受疼痛刺激的時候其僵化反應表現百分比也會上升，而且觀察者動物親近 (social approach) 受難同伴的行為也會增加。在神經元的反應上，我們發現有14.7%的前皮質扣帶迴神經元和13.7%的島腦神經元會對於同伴接受疼痛刺激有反應，而這個比例顯著高於雷射熱刺激對空氣發射的控制組 (反應比例分別在前皮質扣帶迴為7.6%，島腦為5.5%)。並且，這些行為反應和神經元變化在視覺剝奪的控制組中是不存在的。有趣的是，在這些對他人疼痛有反應的神經元中，有53.9%的前皮質扣帶迴神經元和66.3%的島腦神經元，同時會對觀察者本身接受疼痛刺激有相同或相反的興奮或抑制的反應，我們稱為同向 (shared-response) 或反向 (anti-response) 鏡像神經元。在不透明隔板的控制組中，同向鏡像神應元在前皮質扣帶迴和島腦的百分比從40.8%降至30.8%。為了探討同向與反向鏡像神經元的功能性差異，我們針對神經元和社會親近或是僵化行為做相關性探討，結果發現動物在社會親近行為之前以及當下，同向與反向鏡像神經元的活性都會上升，類似的興奮性行為不會發生在雷射對空發射的控制組，另外，抑制性反應只有專一發生於反向鏡像神經元中。 為了更進一步測試同向與反向鏡像神經元在親社會行為上的意義，我們利用Bartal等人於2011年發表的同理心相關行為作業 (Bartal et al., Science 334: 1427, 2011)，讓自由活動的動物去開門解救受困的動物，探討同向與反向鏡像神經元的活性變化（第三章）。實驗結果發現，在前皮質扣帶迴、島腦和初級運動皮質中分別有48.9%、58.8% 和51.8%的神經元是同向或反向鏡像神經元。當動物在解救受困同伴之前，其同向鏡像神經元平均活性會上升，這樣的變化在空籠或是裝有玩具的籠子等控制組不會發生，並且，前皮質扣帶迴和島腦神經元 (25.9%) 活性會專一的在開門解救受困同伴前0–5秒上升，而初級皮質神經元則表現對於開門動作非專一性的上升，無論籠子內裝的是同伴、玩具還是空的，顯示初級運動皮質的神經元的確只與開門的動作相關。綜合上述結果，這些與解救行為相關的神經元平均活性，在看見他人受到疼痛刺激與自己的接受疼痛刺激都會有上升，顯示個體在觀看別人接受疼痛刺激時，同向鏡像神經元會利用活性變化調控社會親近行為，並貢獻於親社會相關的解救行為上。

Studies have shown prosocial behaviors, including empathy-like acts, in rodents, but its neural substrates remain elusive. Evidence from human studies implicates the anterior cingulate (ACC) and insular cortices (InC) in pain empathy and thus predicts that certain neurons in these regions might be engaged by a subject’s sensing itself in pain as well as by its sensing others in pain. The present study addressed this issue by recording behavioral and single unit responses of the ACC and InC while noxious radiant heat was applied to a rat or its nearby companion (Chapter 2). Results showed that rats increased freezing behavior not only when receiving noxious stimuli but also when observing a companion being hurt. The observer rats also increased social approach to the companion. We found that 14.7% neurons in the ACC and 13.7% neurons in the InC responded to the pain stimuli applied to a nearby rat. These ratios were significantly higher than those responding to air radiation (7.6% and 5.5% in the ACC and InC, respectively). Furthermore, such behavioral and neuronal changes were absent in the visual-deprived control group. Interestingly, among these vicarious pain responsive units, 53.9% and 66.3% in the ACC and InC also responded, with the same or opposite direction (denoted as shared- or anti-response, respectively), to its own pain. Percentage of shared-response neurons in the ACC and InC also decreased from 40.8% to 30.8% by blocking the companion rat from observer with an opaque divider. To investigate the functional significance of the shared- and anti-response neurons, we examined the correlation between neuronal activity and social approach or freezing behaviors. We found enhanced activities in both shared- and anti-response neurons prior to and during the social approach behavior. Such enhancement was not seen for those neurons responding to laser heat pulses directed to the air. Inhibited activities were shown specifically in anti-response neurons. Furthermore, to test the pro-social significance of the shared- and anti-response neurons, we applied a behavioral paradigm relevant to empathy, in which a freely moving rat was trained to rescue a conspecific from a restraining trap (Bartal et al., Science 334: 1427, 2011) (Chapter 3). 48.9%, 58.8% and 51.8% of neurons responding to vicarious pain in the ACC, InC and primary motor cortex (MI) respectively, were either shared- or anti-response neurons. The shared-response neurons showed changes in the ensemble activity just prior to rescue behavior and did not show any change in the control conditions in which the restrainer was empty or had a toy rat in it. In addition, the neuronal activity recorded in the rescuing behavior showed that a group (25.9%) of ACC and InC neurons changed their activity at 0 to 5s prior to the specific rescue action for a rat which we called rescuing-related neurons, whereas the activity of MI ensembles increased around the gate-opening acts no matter they subserved rescue or not. Coherently, these rescuing-related neurons also responded with their ensemble activity to both other’s pain and self’s pain. These findings suggest that observation of a conspecific in pain may engage the shared-response neurons in rat’s cortex and their activity may be related to social approach behaviors and contribute to prosocial rescuing acts.