认知与发展心理学研究室知识库

目的 早在 20 世纪初，许多临床案例报告精神分裂症患者对诸如心肌梗塞、化脓性阑尾炎等疾病缺少疼痛相关的反应，即精神分裂症患者表现出疼痛敏感性下降。目前，有关这一现象获得的研究关注较少，且其内在的神经病理机制仍不明确，可能受到遗传变异、激素水平异常、神经系统病变等多种因素的影响。鉴于疼痛敏感性的下降与精神分裂症患者的高致残率和高死亡率有显著的关联，并且这一特征有可能成为精神分裂症易感性的预测因素，因此本研究旨在探讨精神分裂症患者疼痛敏感性的变化及其内在的神经机制，提升临床对精神分裂症患者疼痛评估的重视。

方法 结合行为测量、脑电和磁共振成像技术，我们进行了一系列研究，深入探讨精神分裂症患者疼痛敏感性的变化及其相关的神经机制。在研究一中，通过比较 21 名精神分裂症患者与 21 名年龄、性别匹配的健康人的多模态感觉刺激（痛觉、触觉和听觉）诱发的脑响应特征，我们系统分析了刺激诱发电位中与精神分裂症患者疼痛敏感性相关的特征，同时为了进一步估计基线精神状态对疼痛敏感性的影响，我们还比较了刺激前 α 神经振荡的频谱功率（6-7 Hz，枕电极；8-10 Hz，中心电极），从电生理的角度解析患者疼痛敏感性变化的可能的神经机制；在研究二中，通过比较 20 名精神分裂症患者与 20 名年龄、性别匹配的健康人由疼痛刺激诱发的 BOLD 脑响应，我们探索了两组人群脑功能活动的差异，同时对比了两组人群在静息状态下的功能连接的差异，分析了精神分裂症患者的异常脑功能特征与其疼痛敏感性的关系，进一步验证脑电实验的结果；在研究三中，我们侧重考察精神分裂症患者在疼痛信息处理相关的某些脑区是否会表现出不同的大脑白质纤维弥散特性，进一步阐明精神分裂症患者疼痛信息处理功能缺陷相关的异常结构机制。

结果 精神分裂症患者仅在疼痛强度评分上显著低于健康人群，而在触觉和听觉强度评分上与健康人群没有显著的统计学差异。精神分裂症患者由激光诱发脑响应中的 N1、N2、P2 成分的幅值均显著低于健康人群，对应成分的潜伏期也显著长于健康人群；而对于触觉和听觉诱发脑响应，仅有 N2 成分的幅值低于健康人群，其他成分的幅值、潜伏期与健康人群并无显著差异。精神分裂症患者在刺激前低频和高频 α 神经振荡均较健康人群显著增高。相关分析显示，疼痛强度评分与激光诱发电位在时域的 N1、N2、P2 成分幅值和时频域的平均功率谱值，以及刺激前低频 α 振荡之间存在显著相关性。在任务态功能磁共振扫描中，精神分裂症患者在与疼痛信息处理相关的脑区的激活程度显著小于健康人群，包括中脑导水管周围灰质、丘脑、次级躯体感觉皮层、脑岛、背侧前扣带回等脑区。其中，疼痛强度评分与疼痛刺激诱发的后脑岛 BOLD 响应存在显著的正相关。基于种子点的静息态功能连接分析发现，与健康人群相比，精神分裂症患者在静息状态下的丘脑与高级皮层区的功能连接增强，包括右侧初级躯体感觉皮层、右侧次级躯体感觉皮层、左侧后脑岛；而其中脑导水管周围灰质与高级皮层区的功能连接减弱，包含辅助运动区、背侧前扣带回、背外侧前额叶。进一步的相关分析显示，疼痛强度评分与静息态丘脑-次级躯体感觉皮层的功能连接存在显著的负相关；刺激前低频 α 振荡与静息态丘脑-初级躯体感觉皮层、丘脑-次级躯体感觉皮层的功能连接存在显著的正相关；而刺激前高频 α 振荡与静息态丘脑-初级躯体感觉皮层存在显著的正相关。弥散张量成像的研究则显示，与健康人群相比，精神分裂症患者大脑白质纤维束表现出较低的分数各向异性连同较高的平均弥散率和径向弥散系数，同时出现在胼胝体的体部、膝部、内囊和钩束等脑区。

结论 精神分裂症患者的疼痛敏感性较健康人健康人群显著下降，不仅表现为跨模态的感觉信息处理的整体功能缺陷，又包含了疼痛相关的上行和下行通路的调节机制。跨模态感觉信息处理（即所有感觉模态中的较小的 N2 幅度）的功能缺陷突出反映为精神分裂症患者异常的基线精神状态（即异常的自发性神经振荡信号）。而“疼痛矩阵”中主要脑区的刺激诱发 BOLD 脑响应减少，以及静息状态上行及下行疼痛调节通路的异常功能连接，证实了精神分裂症患者疼痛敏感性的下降也存在疼痛调节相关的功能缺陷。此外，胼胝体和内囊作为主要的白质通路，涉及疼痛信息处理相关的脑区之间的相互连接，二者异常的白质纤维束的弥散特性可能是精神分裂症患者的疼痛信息处理功能异常的结构基础。本研究揭示了精神分裂症患者疼痛敏感性异常的复杂本质，也提示临床医疗应重视对精神分裂症患者疼痛的评估，以改善患者预后，提高其生活质量。

Introduction As early as the turn of the 20th century, clinicians observed patients with schizophrenia (SCZ) failing to respond to the pain of acute diseases, such as myocardial infarction, ruptured appendix, etc. So far, pain insensitivity in SCZ has received little attention and its underlying neural mechanisms has been poorly understood. Multiple facets, such as genetic aberrance, abnormality of hormone levels and lesions of the nervous system, might be resulted in the alteration of pain sensitivity. Notably, pain insensitivity is associated with increased morbidity and mortality in patients with SCZ. It is also possible that pain insensitivity might serve as a prodromal predictor of susceptibility for SCZ. Thus, our study is aimed to clarify the underlying mechanism of this phenomenon and appeals for attention to pain assessment on patients with SCZ.

Methods In the current study, we investigated altered pain sensitivity and its neural mechanisms in SCZ compared to healthy controls (HC) by combining psychophysics, electroencephalography (EEG), and functional magnetic resonance imaging (fMRI) techniques. In Experiment 1, we compared verbal ratings and EEG responses (laser-evoked potentials, LEPs; somatosensory-evoked potentials, SEPs; auditory-evoked potentials, AEPs) to nociceptive, non-nociceptive somatosensory, and auditory stimuli between 21 SCZ patients and 21 matched HC. To further estimate the impacts of mental state at baseline on between-group perception variability, we compared the spectral power of prestimulus EEG oscillations (alpha oscillations: 6-7 Hz at occipital electrodes and 8-10 Hz at central electrodes) between SCZ and HC. In Experiment 2, we examined blood-oxygen-level-dependent (BOLD) responses to nociceptive stimuli and resting-state functional connectivity (using thalamus and periaqueductal gray as the seeds) in 20 SCZ patients and 20 matched HC. In Experiment 3, we focused on the difference of white matter diffusivity in regions associated with nociceptive information processing between SCZ and HC to further clarify the structural mechanism underlying pain insensitivity in schizophrenia.

Results As compared to HC, SCZ showed significantly lower ratings to nociceptive stimuli, but similar ratings to non-nociceptive somatosensory and auditory stimuli. For laser-evoked potentials (LEPs), significantly lower N1/N2/P2 amplitudes with longer latencies of LEPs were observed in SCZ. In contrast, almost all amplitudes of somatosensory-evoked potentials (SEPs) and auditory-evoked potentials (AEPs) were not significantly lower in SCZ than HC, except N2 amplitude. For prestimulus EEG responses, the magnitudes of spectral power of prestimulus EEG oscillations at both lower and higher alpha frequencies were significantly larger in SCZ. Significant correlations were found between the ratings to nociceptive stimuli and the N1/N2/P2 amplitudes of LEPs in the time domain, the magnitudes of LEPs in the time-frequency domain, as well as the magnitudes of spectral power of prestimulus EEG oscillations at lower alpha frequencies. For task fMRI, there were lower BOLD activations to nociceptive stimuli in the key brain regions associated with nociceptive information processing, including periaqueductal gray (PAG), thalamus, primary (S1) and secondary (S2) somatosensory cortices, insula, dorsal anterior cingulate cortex (dACC) in SCZ than in HC. Particularly, the BOLD responses of the posterior insula were positively correlated with the ratings to nociceptive stimuli. For resting-state fMRI, thalamus exhibited stronger resting-state functional connectivity with right S1, right S2, and left posterior insula, whereas PAG showed weaker functional connectivity with the supplementary motor area, dACC, and dorsolateral prefrontal cortex in SCZ than in HC. Additionally, the ratings to nociceptive stimuli were negatively correlated with the resting-state functional connectivity between thalamus and S2, while the magnitudes of spectral power of prestimulus EEG oscillations at lower alpha frequency were positively correlated with the resting-state functional connectivity between thalamus and S1/S2. Similarly, positive correlations were found between the magnitudes of spectral power of prestimulus EEG oscillations at higher frequency and the resting-state functional connectivity of thalamus with S1. For diffusive tensor imaging, SCZ patients presented decreased fractional anisotropy and increased mean diffusivity, as well as increased radial diffusivity in a range of structures, including the body and genu of the corpus callosum, the internal capsule, and the uncinate bundle.

Conclusions These results demonstrated that beyond a general dysfunction of cortical sensory information processing across modalities, the pain insensitivity in schizophrenia also relied on a specific deficit of ascending and descending pathways modulating nociceptive information processing. A general deficit in cortical processing of sensory information across modalities (i.e., smaller N2 amplitude across all sensory modalities) was reflected by the abnormal mental state at baseline (i.e., abnormal prestimulus EEG oscillations in SCZ). A nociceptive-specific deficit in cortical processing of nociceptive information was demonstrated by the decreased neural activity in key regions of the “pain matrix”, and the abnormally activating function of the ascending pain pathway at resting state, together with a possible degenerated function of the descending pain pathway in SCZ. Moreover, abnormal diffusivity of white matter in the corpus callosum and internal capsule, which served as a major pathway combining brain regions associated with nociceptive information processing, may be a structural basis related to the dysfunction of nociceptive information processing in SCZ. Taken together, these findings call attention to systematic assessment on pain insensitivity, in order to improve prognosis and quality of life in SCZ.