真实创伤后视觉空间想象竞争任务干预减少闯入性记忆：系统综述与元分析
Visuospatial imagery-competing task interventions for reducing intrusive memories following real-world trauma: a systematic review and meta-analysis
一项发表于 Frontiers in Psychiatry 的系统综述与元分析（PROSPERO 注册 CRD420261336501）评估了真实创伤后视觉空间想象竞争任务干预（ICTI，如记忆激活提示后玩类俄罗斯方块游戏）对闯入性记忆频率的效果。
这项元分析汇总了真实创伤后视觉空间任务干预的RCT证据，读者可了解其对闯入性记忆频率的合并效应与证据确定性。
译文尚不完整，完整内容请切换到原文。
摘要
背景：
侵入性记忆是PTSD的标志性特征，并可预测其病程，使创伤后早期成为预防的关键窗口。视空间意象竞争任务干预（ICTI）——在记忆再激活线索后进行简短的类俄罗斯方块游戏——以工作记忆干扰假说为基础，并显示出强烈的实验室效应。此前唯一一项针对真实世界创伤的RCT综合研究仅识别出三项符合条件的研究，且均被评为低确定性。尚无预注册综述专门以侵入性记忆频率为主要结局，综合视空间ICTI在真实世界创伤中的RCT证据。
目的：
评估视空间ICTI对真实世界创伤中侵入性记忆频率的合并效应；检验明确的记忆再激活线索是否调节该效应；评估其对PTSD严重程度、抑郁、焦虑、不良事件和脱落的影响；并识别差距以指导未来试验设计。
方法：
检索了七个数据库，从建库至2026年3月，纳入在暴露于真实世界DSM-5标准A创伤的成人中开展视空间ICTI的RCT。主要结局为通过前瞻性每日日记评估的侵入性记忆频率。使用RoB 2和ROBINS-I评估偏倚风险，使用GRADE评估确定性。对具有兼容日记数据的平行组RCT，采用随机效应荟萃分析合并原始均数差。在PROSPERO注册（CRD420261336501）；按照PRISMA 2020报告。
结果：
在识别的818条记录中，13项研究（9项RCT，4项非随机）符合纳入标准。五项平行组RCT（n=359）被合并用于主要结局：与对照相比，ICTI使侵入性记忆每周减少6.98次（95% CI −9.84至−4.13；I2 = 21.9%），为中等确定性证据（因偏倚风险而降级）。PTSD严重程度也支持ICTI（MD −2.42，95% CI −4.40至−0.44），因不一致性和间接性为低确定性。由于测量异质性，抑郁和焦虑无法进行荟萃分析。未发现归因于ICTI的严重不良事件；脱落率总体较低。未合并的证据大体一致，尽管一项基于VR的试验和一项在慢性PTSD中开展的交叉试验未发现获益。
结论：
视空间ICTI在真实世界创伤后对侵入性记忆频率产生具有临床可解释性的、中等确定性的降低，并对PTSD严重程度产生低确定性但有利的效应。证据基础仍限于五项合并RCT。在更大规模的独立试验开展之前，ICTI应被视为一种有前景的靶向干预，而非已确立的治疗。
系统综述注册：
https://www.crd.york.ac.uk/PROSPERO/view/CRD420261336501，标识符CRD420261336501。
引言
侵入性记忆是一种非自愿的、感觉-知觉性的再体验现象，是创伤后应激障碍（PTSD）的标志性症状，并且在现实世界创伤后，跨诊断地出现在抑郁症、焦虑症及其他应激相关状况中（）。心理创伤在全世界约50%的人口中至少在其一生中发生过一次，而暴露后数天至数周内侵入性记忆的频率和生动程度可预测PTSD的后续病程，从而将创伤后早期确定为预防性干预的关键且时间有限的窗口（，）。
视空间意象竞争任务干预（ICTIs）——最常见的是在记忆再激活提示后进行的简短类俄罗斯方块游戏——以工作记忆干扰假说为基础。据推测，侵入性记忆主要具有视觉-知觉特征，会与类俄罗斯方块任务争夺同样有限的视空间工作记忆资源；使用这些共享资源会干扰记忆巩固及随后的非自愿再激活（）。该机制得到独立汇聚证据的支持：一项眼动范式荟萃分析报告实验室研究中的效应量为d = 0.74（）；一项涵盖12,074名参与者、131种方法的大型实验室荟萃分析确定创伤提醒和俄罗斯方块为最有效的技术（Hedges’ g = 0.80）（）；另一项荟萃分析确定视空间ICTI是唯一对侵入频率显示出显著效应的认知干预类别（g = 0.46）（）。
创伤相关心理健康状况构成了巨大且日益增长的全球公共卫生负担。在高收入国家，PTSD的终生患病率为3.9%至8.8%，而在暴露于武装冲突、大规模伤亡事件和急性医疗紧急情况的人群中则上升至20–30%或更高（，）。侵入性再体验，包括闪回和不想要的创伤记忆，不仅在诊断上处于核心地位，而且始终被列为最令人痛苦和功能损害最严重的PTSD特征之一，并与重性抑郁障碍、物质使用障碍和自杀风险的共病率升高相关（）。因此，这给全球卫生系统带来了巨大的经济负担，包括更高的医疗保健利用率、更低的工作生产力和显著的间接费用（）。
目前指南推荐用于PTSD的治疗方法，包括创伤聚焦认知行为疗法（TF-CBT）、眼动脱敏与再加工（EMDR）以及延长暴露疗法（PE），已在随机对照试验中证明能有效减轻PTSD总体症状负担（，）。然而，单次心理汇报曾在创伤后即刻情境中被普遍使用，却可能使某些人的结果恶化，且大多数早期心理治疗未能预防PTSD的发生（）。这些局限性，加上创伤聚焦治疗的高强度、高脱落率（临床试验中估计为20–35%）、低资源环境下可及性受限，以及缺乏专门针对侵入性记忆巩固的干预措施，已引发了对简短、可扩展、具有机制基础的早期干预的浓厚兴趣（）：尽管选择性5-羟色胺再摄取抑制剂（SSRIs）等药物方法可提供一定的症状缓解，但它们并不直接针对侵入性意象，且在某些患者群体中存在严重的耐受性问题（）。
尽管有坚实的实验室基础，但现有唯一基于RCT的、针对真实世界创伤人群中视空间ICTI的综合分析——Astill Wright et al.（）——仅识别出三项符合条件的研究，将所有证据的GRADE评级评为低确定性，采用了这些研究并未有足够把握度检测的主要结局（PTSD严重程度），且未进行发表偏倚评估（）。自该综述以来，真实世界创伤环境中又发表了几项额外的RCT（–），其中包括发表在The Lancet Psychiatry上的一项贝叶斯自适应试验，报告了侵入性记忆的显著减少（）。然而，各试验的结果仍存在异质性，凸显了进行严格综合的必要性。
据我们所知，尚无预注册的系统综述专门针对真实世界创伤人群，以侵入性记忆频率作为主要结局，综合视空间ICTI的RCT证据。这一被广泛讨论、具有理论基础且临床前景可期的干预措施，其证据基础仍然薄弱、方法学受限，且在RCT层面完全未被综合。在能够负责任地提出临床建议之前，该领域恰恰需要对现有证据进行严格的综合。
本系统综述和荟萃分析旨在综合RCT证据，评估视空间意象竞争任务干预（ICTIs）在真实世界创伤后减少侵入性记忆的效果。具体而言，本综述旨在：（1）估计视空间ICTIs与对照条件相比对侵入性记忆频率的合并效应；（2）检验是否存在明确的记忆再激活线索对侵入性记忆频率的干预效应具有调节作用；（3）评估视空间ICTIs与对照条件相比对PTSD症状严重程度及相关心理结局（包括抑郁、焦虑、不良事件和脱落）的影响；（4）识别当前证据基础中的关键空白，并为未来随机试验设计提供建议。
方法
本系统评价和meta分析按照系统评价和meta分析首选报告条目（PRISMA 2020）指南以及系统评价方案PRISMA-P清单进行和报告（，）。本评价方案已在国际系统评价前瞻性注册库（PROSPERO）前瞻性注册，注册号为（CRD420261336501）。
纳入标准
研究设计
只有随机对照试验（RCT）符合定量meta分析的条件，包括平行组RCT和贝叶斯自适应RCT。交叉RCT也被纳入，但鉴于记忆巩固范式中固有的遗留效应风险，在敏感性分析中单独进行分析。
非随机研究（可行性研究、单组前后测研究和单病例实验设计）仅符合纳入定性叙述性综合的条件，以提供关于干预可行性、可接受性以及在RCT中代表性不足人群中效应方向的背景证据。
人群
符合条件的研究纳入18岁及以上、曾暴露于符合DSM-5标准A所定义的真实世界创伤的成年人：实际或威胁性死亡、严重伤害或性暴力（）。无论暴露是直接的、目睹的还是职业性的（例如，急诊医护人员、急救人员），均符合条件。使用实验室模拟创伤（例如，令人痛苦的电影片范式）在健康志愿者中进行的研究被排除。
干预
任何被归类为视空间意象竞争任务的干预均符合条件，定义为需要主动参与视空间工作记忆、并明确意图干扰创伤性记忆巩固或再巩固的任务。符合条件的干预包括通过智能手机、平板电脑、计算机或掌上游戏机进行的俄罗斯方块类游戏（无论是否带有心理旋转指导），以及通过虚拟现实进行的其他视空间任务。
以下内容被排除：言语认知任务（例如，文字游戏、阅读）、仅眼动干预（EMDR）、药物制剂、仅正念或放松干预，以及无法将视空间任务分离为活性成分的多成分干预。
在任务前有或没有明确记忆再激活提示的干预均符合条件；此类提示的有无被编码并作为预先指定的调节因素进行检验。
对照
任何对照均符合条件，包括：被动对照（常规照护、无干预、等待名单）；主动注意安慰剂对照（例如，书面活动日志、播客、古典音乐）；或替代性心理干预。对照类型被编码并作为预先指定的调节因素进行检验。
结局
主要结局是通过前瞻性每日日记计数每周侵入性记忆发作次数来测量的侵入性记忆（闪回、对创伤的不自主记忆）频率。为了纳入定量荟萃分析，侵入性记忆频率必须以每周计数形式报告，并提供可提取的均值和标准差，以便进行原始均值差合并。通过回顾性自我报告问卷分量表（例如PSS侵入分量表、PCL-5再体验分量表）而非每日日记计数来测量侵入症状的研究，符合定性叙述性综合的条件，但不符合主要定量荟萃分析的条件。侵入性记忆频率在每个试验自身预先指定的主要日记时间点（范围：干预后第1周至第6周）提取，以最大化试验纳入并尽可能提供对总体干预效果最全面的估计。
次要结局包括：
PTSD症状严重程度：在干预后最早可用时间点，采用经过验证的测量工具（PCL-5、IES-R、ASDS或PSS）的总分。
抑郁：在干预后最早可用时间点，采用经过验证的测量工具（HADS-D、PHQ-2、EPDS）的总分。
焦虑：在干预后最早可用时间点，采用经过验证的测量工具（HADS-A、GAD-2）的总分。
不良事件：任何报告的不良事件、痛苦反应或归因于干预的症状加重。
语言和日期限制
检索阶段未应用语言限制。检索从数据库建库至2026年3月进行，无日期限制。所有符合纳入标准的已识别记录均以英文发表。
信息来源和检索策略
我们系统检索了七个电子数据库，从建库至2026年3月：PubMed、MEDLINE、Scopus、Web of Science、Psychology Collection（ProQuest Psychology Journals）、Google Scholar和Ovid。检索阶段未施加语言或日期限制。
我们开发了一个结合干预、结局和研究设计概念的核心检索字符串，然后根据每个数据库的语法进行调整。代表性版本如下：
（“Tetris” OR “visuospatial task” OR “imagery-competing task” OR “cognitive task intervention” OR “visuospatial interference” OR “mental rotation” OR “spatial task” OR “ICTI”）AND（“intrusive memor*” OR “intrusion*” OR “flashback*” OR “re-experiencing” OR “PTSD” OR “post-traumatic stress”）AND（“randomized controlled trial” OR “randomised controlled trial” OR “RCT” OR “random*”）。为提高敏感性，我们在针对性补充检索中加入了创伤相关术语（例如“emergency department”、“road traffic accident”、“caesarean section”和“healthcare worker”）。对于Google Scholar，针对性字符串返回的前100条记录按标题和摘要进行筛选，可能符合条件的报告被手动添加到筛选库中。完整的、针对各数据库的检索策略报告于补充文件S1。
研究选择
所有记录均导入Rayyan系统综述软件，并在筛选前进行去重（）。研究选择按两个连续阶段进行。
在第一阶段，两名独立评审员（IRs）根据纳入标准对所有标题和摘要进行筛选。只有在记录明显不符合所有PICOS标准时，才会在此阶段被排除。分歧通过讨论解决；若无法达成共识，则由第三名评审员仲裁。
在第二阶段，所有通过标题和摘要筛选的记录均被获取全文。两名独立评审员分别对每篇全文记录独立应用完整的纳入标准。全文阶段排除的原因已记录，并在PRISMA 2020流程图中有报告。
数据提取
数据由两名评审员使用标准化提取表独立提取，该表在全面提取前对前三项纳入研究进行了预试验；对提取表的修订已记录。分歧通过第三方仲裁解决。在适用的情况下，联系通讯作者以获取缺失或不明确的数据。
对每项纳入研究提取以下数据：研究特征：作者、年份、国家、期刊、研究设计和样本量。人群特征：创伤类型和场所、入组时距创伤的时间、年龄、性别、临床特征以及是否存在既存PTSD诊断。干预特征：任务类型、记忆再激活线索的存在与性质、每次治疗时长和次数、总剂量、相对于创伤的时间（小时、天、周）、实施方式以及依从性/遵守率。对照特征：类型、内容和实施方式。结局数据：所有可用时间点的所有主要和次要结局，包括不良事件和失访。
偏倚风险评估
个体随机研究的偏倚风险由两名评审员使用Cochrane随机对照试验偏倚风险2（RoB 2）工具独立评估（）。RoB 2评估五个领域：随机化过程产生的偏倚、偏离预期干预导致的偏倚、缺失结局数据导致的偏倚、结局测量中的偏倚以及报告结果选择中的偏倚。每个领域被评为低风险、一些担忧或高风险，并为每项研究得出总体偏倚风险判断。
对于纳入定性叙述性综合的非随机研究，使用ROBINS-I工具评估偏倚风险（）。ROBINS-I评估七个领域：混杂、参与者选择、干预分类、偏离预期干预、缺失数据、结局测量和选择性报告。每个领域被评为低、中等、严重或极严重风险，并为每项研究得出总体判断。
分歧通过讨论解决，若未达成共识则由第三名评审员仲裁。偏倚风险评估使用交通灯图和汇总条形图进行可视化呈现。
数据综合与统计分析
定量综合
当两项或更多研究报告了相同结局且数据充分时，使用DerSimonian–Laird估计量进行随机效应荟萃分析以估计研究间方差（τ2）（）。所有分析均使用OpenMetaAnalyst（）进行。
对于主要结局，侵入性记忆频率以每周每日日记计数报告，原始均值差计算为 ICTI 组与对照组之间的均值差。选择原始均值差是因为它们以原始测量单位（即每周侵入性记忆次数）提供临床可解释的效应估计。负值表示与对照组相比，ICTI 组的侵入性记忆更少。
一项指标敏感性分析比较了 Hedges' g 和对数发病率比合并，使用数据格式兼容的试验子集，报告于补充文件 S1。进行该分析是为了检验效应量指标的选择是否影响主要结论。
主要荟萃分析
主要荟萃分析合并了五项平行组 RCT，这些 RCT 以每周每日日记计数报告侵入性记忆频率 [Iyadurai ()、Horsch ()、Kanstrup ()、Iyadurai ()、Kanstrup ()]，使用每项试验自身预先指定的主要日记时间点（范围：干预后第 1 周至第 5 周）。
三项符合条件的平行组 RCT 被排除在主要合并之外，因为它们的数据格式与原始均值差合并不兼容，并在定性综合中以叙述方式报告。交叉试验（Kehyayan ()）作为独立的敏感性分析单独分析。
异质性评估
使用 I2 统计量和 Cochran's Q 检验评估统计异质性（）。I2 值的解释如下：0–25% 表示低异质性，26–50% 表示中等异质性，51–75% 表示显著异质性，大于 75% 表示相当大的异质性。当 I2 超过 50% 时，通过预先指定的亚组分析探索潜在的异质性来源。
预先指定的亚组分析
基于再巩固理论和 ICTI 的拟议机制，视觉空间任务前是否存在明确的记忆再激活线索被预先指定为主要调节因素假设。假设使用明确记忆再激活线索的研究比没有此类线索的研究显示出更大的效应。亚组合并估计和亚组差异检验报告于结果中。
发表偏倚
通过视觉检查漏斗图不对称性来探索发表偏倚。根据 Cochrane 手册指南（，），未进行漏斗图不对称性的正式统计检验，包括 Egger 回归检验，因为纳入合并分析的研究少于 10 项。因此，对漏斗图进行描述性解释，并且未从这一小部分研究中就发表偏倚的存在与否得出明确结论。
GRADE 评估
使用 GRADE 框架评估每个结局的证据确定性（）。对于未进行荟萃分析的结局，包括抑郁和焦虑，确定性基于偏倚风险、不一致性、间接性、不精确性和发表偏倚以叙述方式分级。
结果
在七个电子数据库中总共识别出818条记录（图1）。去除186条重复记录后，对632条记录进行了标题和摘要筛选，其中608条因不相关被排除。检索到24篇报告并评估其全文合格性，其中11篇因有记录的原因被排除。13项研究符合纳入系统评价的合格标准：9项随机对照试验和4项非随机研究。5项平行组RCT被纳入侵入性记忆频率的定量荟萃分析。8项研究仅纳入定性叙述性综合：4项数据格式与原始均数差合并不兼容的RCT和4项非随机研究。
图1
该表包含2017年至2026年间发表的13项研究表1。设计范围从RCT（平行、交叉、适应性）到可行性研究和前后对照研究。招募场所包括急诊科、产科病房、重症监护病房、难民中心和住院精神科医院。样本量各不相同：最大随机样本=146，最小可行性样本=12。平均年龄介于22岁（年轻难民）至46岁（急诊创伤患者）之间。性别分布以女性为主（分娩研究中高达100%；女性比例最低的是Freedman（），77名参与者中21名为女性，约27.3%。创伤人群包括医护人员（COVID-19）、难民、机动车事故幸存者、急诊剖宫产术后女性和慢性PTSD患者。大多数研究报告的基线侵入性记忆频率以日记计数表示：例如，医护人员中位数14–15次/周，Kanstrup（）中位数15次/周，长期创伤中平均约2次/天。创伤后时间从<6小时（急性）到>20年不等。随访从1周到6个月。脱落率：Beckenstrom（ICTI组）高达35%，Hardarson（可行性研究）为33%；许多研究报告低脱落率或无差异性脱落。
表1
| 第一作者，年份 | 研究设计 | 招募场所 | 随机样本量 | 分析样本量 | 平均年龄±SD | 性别（男/女/其他） | 创伤人群 | 创伤类型 | 基线PTSD症状 | 基线侵入性记忆频率 | 入组时距创伤时间 | 随访时长 | 评估时间点 | 脱落率 | 失访原因（如报告） |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Beckenstrom等，2026（） | 去中心化、平行组、贝叶斯适应性随机对照试验（RCT） | 收治COVID-19患者的英国NHS机构；通过重症监护学会网络、社交媒体、研讨会、海报招募 | 总计：99；ICTI：40，活性对照：39，常规治疗（TAU）：20 | 主要结局（第4周）：ICTI = 26，活性对照=31，TAU = 19（ITT分析） | 41.2±10.2岁 | 12/85/2 | 在COVID-19大流行期间暴露于工作相关创伤的医护人员（NHS） | 符合DSM-5标准A的工作相关创伤事件 | PCL-5（20项版本）；基线时49%达到可能PTSD阈值（PCL-5≥33）。各组的均值和SD可在OSF（osf.io/cs6hn）获取，但主要论文中未报告。 | 纳入要求过去一周≥3次侵入性记忆。合并基线中位数10.0次/周（IQR 6.5–17.5）。按组：ICTI 11.5次/周（IQR 9.0–18.0）；活性对照9.0（IQR 6.5–16.0）；TAU 6.5（IQR 5.0–17.75）。 | 可变；从创伤当天到数月后（“创伤后6个月窗口”） | 24周 | 基线、第4周（主要终点）、第12周、第24周 | ICTI 14/40（35%），Active Control 8/39（21%），TAU 1/20（5%）。ICTI 的总体保留率低于 TAU（BF = 8.12）。 | 因不良事件退出（n=3；日记负担 n=2，无关 SAE n=1）、时间不足（n=1）、失访或日记不完整。 |
| Kanstrup et al., 2024（） | 平行组、双臂 RCT | 瑞典的医疗机构、社交媒体、医院资源和报纸 | 共 144 例；干预组=73，对照组=71 | ITT：144；主要结局完成者：118（对照组 58，干预组 60） | 41.41 ± 10.89 | 25/118/1 | COVID-19 疫情期间瑞典的医护人员 | 工作相关创伤事件 | PCL-5（8 条目）：平均 12.10 ± 6.63（量表 0-32；可能 PTSD 的临界值通常为 19） | 每周中位数 15.0（IQR 6.25–28.75）（基线日记）。 | 72.2% 在 1–3 个月内；其他在 24 小时至 >3 个月内 | 6 个月 | 基线、第1周、第5周（主要终点）、1 个月、3 个月、6 个月 | 总体 6.3%（9/144 中止；13 例失访）。干预组：8 例失访，5 例中止；对照组：5 例失访，4 例中止 | NR |
| Kehyayan et al., 2024（） | 随机交叉试验（患者内） | 德国波鸿 LWL-大学医院心身医学与心理治疗科门诊 | 共 38 例（所有患者均按交叉顺序接受两种干预） | 38（ITT；所有提交日记数据并被随机化者） | 42.2 ± 13.3 | 7/31 | PTSD（ICD-10 F43.1；89.5%）或其他严重应激反应（ICD-10 F43.8；10.5%） | I 型（单一事件）34.2%，II 型（复杂）65.8% | DSM-5 PTSD 诊断（未报告 CAPS；临床诊断）。共病抑郁 81.6%，焦虑 42.1% | 约 14.1/周（原始：2.02/天，SD 1.92） | 距创伤事件平均 13.8 年（SD 15.0） | 8 周 | 基线（2 周日记）、T1（第2周，第一次干预）、T2（第4周，第二次干预）、T3（第8周，最终日记） | 招募的 68 例中 30 例被排除（不符合标准、退出或每周 <3 次闯入）。在分析的 38 例中，所有患者均按方案完成（交叉设计） | NR |
| Hardarson et al., 2024（） | 可行性研究（重复 AB 设计，前后对照） | SAGA 队列的子研究（冰岛女性基于人群的纵向研究）；远程招募 | 共入组 = 12 例（均为女性）；无随机化（单组） | 主要结局（第5周）：n=8；3 个月随访：n=7 | 42.42 ± 12.03 | 均为女性（100%） | 有创伤暴露的女性（社区样本），伴长期闯入性记忆 | 性暴力（n=5）、目睹死亡/重伤（n=3）、身体暴力（n=3）、机动车事故（n=1） | 基线 PCL-5：平均 36.42（SD 16.81）——中重度范围 | 每周平均 25.98（SD 29.39）（基线日记） | 距目标创伤平均 20.73 年（SD 14.65） | 第二次干预后 3 个月 | 基线（第 -1 周）、第0周（干预会话）、第1–4周、第5周（主要终点）、3 个月随访 | 4/12（33%）未完成主要结局：2 例完成了第1周但失去联系，2 例因无关应激源在第1周前退出 | 无关应激源，无法联系 |
| Deforges et al., 2023（） | 双臂、平行组、多中心、双盲优效性 RCT | 瑞士两家大学医院产科（洛桑和日内瓦） | 共 146 例；干预组（Tetris）=74，注意安慰剂对照组=72 | 6 周时 PCL-5：122（干预组 64，对照组 58）；6 周时 CAPS-5：116（干预组 62，对照组 54） | 33.9 ± 4.2 | 均为女性（100%） | 妊娠 ≥34 周计划外剖宫产（UCS）后、活产、认为分娩具有创伤性的女性 | 分娩相关创伤（计划外剖宫产，DSM-5 标准 A 事件） | 基线时采集 PCL-5（20 项）；各臂均值未报告。 | 按研究方案于产后 6 小时内基线时实施 PCL-5（20 项）；发表文献中未报告各臂基线均值和标准差。第 1 周采集每日日记；发表文献中未报告各臂第 1 周均值。 | 非计划剖宫产后 6 小时内 | 6 个月 | 基线（6 小时内）、第 1 周日记、6 周（主要）、6 个月 | 第 6 周时，PCL-5：24/146（16.4%）缺失。干预组：10 例失访（3 例未接受干预，3 例在 T1-T2 脱落，4 例在 T3-T4 脱落）；对照组：14 例失访（2 例未完成任务，5 例在 T1-T2 脱落，7 例在 T3-T4 脱落） | 退出、失访（NS）；COVID-19 疫情暂停 |
| Iyadurai et al., 2023 () | 双臂、平行组、单盲 RCT，采用延迟访问对照（出于伦理原因，在第 4 周主要终点后向对照组提供积极干预） | COVID-19 疫情期间的英国 NHS 重症监护病房（ICU）；通过重症监护学会网络招募 | 共 86 例；立即干预组 = 43，延迟（对照）组 = 43 | ITT：86（主要结局：75 例具有完整/部分数据）。符合方案集：76 | 38.7 ± 9.9 | 15/69/2 | COVID-19 期间从事临床工作的 ICU 工作人员（护士、医生、专职医疗人员） | 符合 DSM-5 标准 A 的工作相关创伤事件 | 基线时 PCL-5（4 项版本）：立即组 8.71 ± 3.64，延迟组 8.00 ± 3.27。还包括 IES-R 总分 | 每周中位数 14（IQR 9–20）（基线日记） | 混合：8.3% 在 24 小时内，19.1% 在 1 个月内，16.7% 在 1–3 个月，54.8% >3 个月；>88% 报告持续暴露 | 8 周（延迟组在第 4 周接受干预，再随访 4 周） | 基线、第 4 周（主要）、第 8 周 | 第 4 周总体失访：11/86（12.8%）。立即组：7/43（16.3%）主要结局缺失；延迟组：4/43（9.3%）缺失。无因不良事件退出 | NR |
| Deforges et al., 2022 () | 原理验证性前后对照研究（单组） | 洛桑大学医院（瑞士）张贴传单，以及既往观察性研究的联系 | 共入组 = 18（无随机化；单干预组） | 主要结局（日记）：n=18 | 33.55 ± 6.35 | 全部为女性（100%） | 在洛桑大学医院分娩 ≥6 周、过去 2 周内报告 ≥4 次分娩相关侵入性记忆（CB-IMs）的女性。 | 创伤性分娩 | 干预前 PCL-5（CB-PTSD）：均值 27.89（SD 12.14）；8/17（47%）符合 CB-PTSD 诊断标准 | 每 2 周中位数 11（IQR 6）（日记 1 = 14 天）；约每周 5.5 次 | 中位数 2.01 年（IQR 2.23）；范围 7 个月至 6.9 年 | 干预后 6 周（日记 3 至第 6 周） | 基线（日记 1：14 天）、干预日、日记 2（干预后前 14 天）、日记 3（第 28-42 天，即第 5-6 周）、1 个月 PCL-5 | 干预后 0%（干预后无脱落）；1 名参与者因未遵守干预指导被排除于主要分析 | 不依从（P18 未遵循指导；其数据被排除于主要结局分析） |
| Badawi et al., 2021 () | 可行性研究（单组、前后对照、辅助治疗） | 住院精神病医院（PTSD 治疗项目，澳大利亚） | 共入组 = 23（均接受常规治疗 + 3 次辅助俄罗斯方块游戏）；无随机化 | 主要可行性结局：n=15（完成方案）。疗效分析：部分指标 n=9–10 | NR | NR | 接受住院治疗的 PTSD 成人（CAPS-5 诊断） | 混合创伤（军事战斗、急救人员执勤、职业伤害） | 治疗前 PCL-5：均值 59.0（SD 8.0），n=9；治疗前 IES-R 侵入性记忆：均值 30.56（SD 5.36），n=9 | NR | NR | 21 天（3 周） | 第0天（基线）、第1–21天（每日三次侵入性监测）、第21天（干预后） | 8/23（34.8%）在完成3次俄罗斯方块游戏前退出：5人更换治疗（无想象暴露），3人提前出院 | 治疗过程中无想象暴露，提前出院 |
| Kanstrup et al., 2021 () | 探索性开放标签试点RCT（平行组） | 急诊科（ED），瑞典斯德哥尔摩卡罗林斯卡大学医院Huddinge院区 + 附近急诊科步入式中心 | 共41人；干预组（俄罗斯方块）= 22，主动对照组（播客）= 19 | 主要结局（第1周日记）：干预组20，对照组19；第5周日记：干预组18，对照组18 | 46.15 ± 15.77 | 女性共23人（56%），男性18人（44%）；干预组14女/8男，对照组9女/10男 | 就诊于急诊科的创伤暴露患者（混合创伤） | 交通事故（对照组36.8%，干预组36.4%）、工作/家庭中的严重事故、袭击等，符合DSM-5标准A | 基线未评估 | 基线未评估 | 74%在6小时内；26%在72小时内（方案变更） | 6个月 | 基线（急诊科第1天）、第1周（主要）、第5周（次要）、3个月、6个月 | 第1周：干预组2/22缺失（9%），对照组0/19缺失。第5周：干预组4/22（18%），对照组1/19（5%）。6个月：干预组8/22（36%），对照组1/19（5%） | NR |
| Freedman et al., 2020 () | 试点RCT（平行组，治疗实施无盲法，评估者盲法） | 急诊科（ED），以色列III级创伤医疗中心 | 共77人；SnowWorld VR = 36，对照组（无干预）= 41 | 2周：55人（SnowWorld 28，对照组27）；6个月：49人（SnowWorld 28，对照组21） | SnowWorld 35.36 ± 13.48；对照组 34.95 ± 15.47 | SnowWorld：26/10；对照组：30/11 | 就诊于急诊科的成年创伤事件幸存者。 | 84.1%为机动车事故；其他：跌倒、袭击、工作事故；符合DSM-IV标准A | 基线时PSS（创伤后症状量表）（既往创伤）：SnowWorld均值2.39，对照组2.05（未达临床升高水平） | 基线未评估 | 平均<1小时（85.5%在1小时内到达并招募；除一人外均在5小时内；一人为8小时） | 6个月 | 基线（急诊科）、2周、6个月 | 2周：22/77（28.6%）参与者缺失，两组间均衡。6个月：28/77（36.4%）缺失，对照组流失率更高（48.8% vs 22.2%），可能影响长期结局的解读。 | NR |
| Iyadurai et al., 2018 () | 概念验证RCT（平行组，单盲） | 急诊科，John Radcliffe医院，英国牛津 | 共71人；干预组（俄罗斯方块）= 37，对照组（活动日志）= 34 | ITT主要结局：71（插补）。日记回收：67（干预组34，对照组33） | 39.66 ± 16.32 | 34/37 | 就诊于急诊科的机动车事故幸存者 | 机动车事故（汽车、摩托车、骑行者、行人）；符合DSM-IV标准A1 | 基线未评估 | 基线未评估 | 平均192 分钟（干预组）至211 分钟（对照组），在6小时内 | 1个月 | 基线（急诊科）、每日日记1周（主要）、1周问卷、1个月问卷 | 1周：4/71（5.6%）缺失；1个月：9/71（12.7%）缺失。未报告差异性流失 | NR |
| Horsch et al., 2017 () | 原理验证RCT（平行组，开放标签） | 产科病房，洛桑大学医院，瑞士（急诊剖宫产后） | 共56人；干预组（俄罗斯方块）= 29，对照组（常规护理）= 27 | ITT：56（主要结局）。符合方案：49（干预组25，对照组24） | 全样本均值33.39 ± 4.22；干预组34.24 ± 3.81；对照组32.37 ± 4.47 | 全部为女性（100%） | 活产紧急剖宫产（ECS）后的女性 | 紧急剖宫产（创伤性分娩，DSM-IV 标准 A） | 未评估 | 未评估 | ECS 后 6 小时内 | 1 个月 | 基线（6 小时内）、每日日记 1 周（主要）、1 周问卷、1 个月问卷 | ITT：主要结局无失访（已插补）。符合方案集：排除 7 例（3 例干预组未玩 ≥10 分钟；4 例对照组因方案违背被排除）。1 个月：部分缺失，已插补 | NR |
| Holmes et al., 2017 () | 可行性研究（单组，前后对照） | 瑞典（难民住所、非政府组织、社交媒体、语言课程、中央车站） | 总入组 = 22（单组；无随机化） | 主要可行性结局：n=22；日记回收：n=17（77%） | 22 ± 2.4（范围 16–25） | 17/5 | 近期抵达瑞典的年轻难民（16–25 岁）（82% 在过去 6 个月内） | 在原籍国或逃亡途中遭受的创伤（叙利亚冲突）；未进一步说明 | 未评估 | 约 12.65/周（原始：1.81/天） | 距创伤事件时间 NR（冲突持续，无法确定日期）；82% 的参与者在研究入组后 6 个月内抵达瑞典。 | 1 周（日记）加上干预后即时评分 | 基线（访谈）、每日日记 1 周、干预后可接受性评分 | 5/22（23%）未交回日记：1 人过于悲伤（来自叙利亚的消息），1 人不理解说明，3 人无理由退出 | 情绪困扰、理解不足、无法联系 |
13 项纳入研究的 研究 和 参与者 特征。
如 表 2 所示，主动干预（“ICTI”）始终包括：a) 简短的记忆提醒线索（例如，列出创伤性“热点”或短暂回忆事件），随后 b) 在心理旋转指导下玩俄罗斯方块 12–25 分钟。实施方式：面对面（平板电脑/游戏机）或通过智能手机/电脑远程（i-spero、SMART-TRIAL）。治疗师参与：由研究人员指导（包括培训后的非专业人员）或在一次指导会话后自行实施。依从性：通常 >90%（例如，Beckenstrom 中为 97%，Kanstrup () 中为 96%，Freedman 中为 100%）。对照类型：主动对照（听播客、活动日志、维基百科文章）和被动对照（照常治疗、延迟干预或基线监测）。对照任务在时长和研究者接触方面相匹配，但缺乏视空间干扰。可行性研究未使用对照组或采用被试内基线。Kehyayan et al. () 的一项交叉试验比较了俄罗斯方块与言语任务（Wiki）以及从未被针对的记忆任务。
表 2
| 第一作者，年份 | 干预名称 | ICTI 类型 | 记忆提醒线索（是/否） | 理论模型 | 创伤后时机 | 会话次数 | 会话时长（分钟） | 实施形式 | 治疗师参与 | 设备/平台 | 依从性/遵守情况 | 联合干预 | 对照类型 | 对照任务描述 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Beckenstrom et al., 2026 () | 意象竞争任务干预（ICTI） | 数字化、简短、单次指导会话 + 可选自我使用 | 是（短暂回忆目标侵入性记忆图像） | 记忆再巩固、工作记忆负荷（视空间干扰） | 可变（创伤当天至数月后） | ICTI：1 次指导 + 按需自我使用（24 周内平均 6.91 次）；主动对照：相同结构（平均 5.57 次）；TAU：0 | 指导会话 1 小时；每次使用游戏/聆听 20 分钟 | 远程、基于应用程序（i-spero 平台）用于 ICTI 和主动对照；TAU：无 | 由数字导航员（不要求是心理健康专业人员）指导进行 ICTI 和主动对照；TAU：无 | i-spero（智能手机、平板电脑、电脑）用于 ICTI 和主动对照；TAU：无 | ICTI：33/34（97%）完成了指导会话；主动对照：34/35（97%） | 允许：基线时，ICTI 38%、主动对照 38%、TAU 50% 接受心理治疗/药物治疗 | 主动对照（听音乐）+ 照常治疗（TAU） | 主动对照：听莫扎特《第一号弦乐二重奏》K.423 20 分钟 + 关于莫扎特的信息播客。TAU：无任务；继续自己的照护，试验后提供访问机会。 |
| Kanstrup et al., 2024（） | 意象竞争任务干预（ICTI）——带心理旋转的俄罗斯方块 | 数字化、简短、单次指导会话 + 可选自我使用 | 是（简要描述一个侵入性记忆“热点”） | 记忆再巩固、视空间干扰（工作记忆负荷） | 可变：创伤后 24 小时至数月 | 1 次指导会话（第 1 天）+ 按需自我使用（1 个月时 80.9% 自行使用） | 任务部分约 25 分钟（20 分钟俄罗斯方块游戏） | 远程、基于应用程序（SMART-TRIAL + 外部 www.tetris.com） | 初始会话由研究人员（非专家）指导，之后自我实施 | 参与者自己的智能手机（访问俄罗斯方块网站） | 干预组：70/73（96%）完成了指导会话；对照组：68/71（96%） | 允许继续常规照护；对心理健康治疗无限制 | 主动对照（基于注意力的安慰剂） | 听哲学播客（Filosofiska Rummet，Sveriges Radio）20 分钟，关注思想和意义。 |
| Kehyayan et al., 2024（） | 基于视空间俄罗斯方块的干预（记忆提醒 + 俄罗斯方块） | 数字化、每个目标记忆单次会话（每位患者共两次会话，交叉设计） | 是（书面报告创伤记忆，然后销毁——记忆提醒线索） | 记忆再巩固（重新激活 + 更新）和视空间干扰（双任务、心理旋转） | 长期创伤（事件后平均 13.8 年） | 每个目标记忆一次会话（每位患者针对两个记忆，一个用俄罗斯方块，一个用 Wiki，随机顺序） | 总共约 35 分钟（10 分钟记忆报告 + 25 分钟俄罗斯方块） | 面对面（实验室）使用平板电脑 | 由研究人员指导（对干预非盲） | 三星 Galaxy 平板电脑（10.1 英寸）装有俄罗斯方块应用程序（N3TWORK） | 所有 38 名患者均完成了两次干预会话（分析样本内无脱落） | NR | 言语条件（Wiki）+ 从未针对的记忆（无干预） | Wiki：阅读一篇关于邮票的缩短版维基百科文章（约 30 分钟）+ 回答 8 道选择题（约 5 分钟）。无干预：无任务，仅日记监测 |
| Hardarson et al., 2024（） | 意象竞争任务干预（带心理旋转的俄罗斯方块） | 简短、远程实施、指导 + 自我使用 | 是（参与者短暂地将侵入性记忆带入脑海，不讨论内容） | 记忆再巩固（重新激活 + 更新）和视空间干扰（心理旋转） | 长期创伤（事件后平均 20.7 年） | 至少 2 次指导会话（最多 6 次）加上按需自我指导使用 | 每次会话约 20 分钟俄罗斯方块游戏 | 远程（视频平台或电话） | 由非专业研究人员（心理学学生）指导，接受远程培训/监督，然后自我实施 | 参与者自己的电脑/智能手机（www.tetris.com） | 8/12（67%）完成了≥2次引导会话。其中：7人恰好完成2次会话，1人完成4次会话。在未完成≥2次会话的4人中：1人完成1次会话，2人完成0次会话，1人从发表文献中情况不明。 | NR | 基线阶段（无干预）——被试内对照 | 无任务；仅在一周内通过每日日记监测侵入性记忆 |
| Deforges et al., 2023 () | 视空间任务（俄罗斯方块） | 单次会话，预防性 | 无 | 视空间干扰（工作记忆负荷）、记忆巩固破坏 | 产后6小时内 | 1次会话（加3 分钟训练） | 15分钟游戏（加3 分钟训练=共18 分钟） | 面对面，在参与者医院病床旁 | 由受过培训的产科临床医生（助产士/护士）实施——有引导，非自我实施 | 手持游戏设备（Nintendo DS） | 71/74（96%）接受了干预（3人未接受，原因未说明） | 常规产后护理；无限制 | 注意安慰剂对照 | 书面活动日志15 分钟：参与者简要报告活动的性质和持续时间 |
| Iyadurai et al., 2023 () | 意象竞争任务干预（ICTI）——带心理旋转的俄罗斯方块 | 数字化、简短、单次引导会话+可选自我使用 | 是（参与者从列表中简要描述了一个侵入性记忆“热点”） | 记忆再巩固（再激活+更新）和视空间干扰（工作记忆负荷） | 不定：创伤后数天至数月（也包括持续创伤） | 1次引导会话（研究者主导）+4周内按需自我使用 | 引导会话约1小时；每次使用游戏20 分钟 | 远程，通过安全网络平台（i-spero）+Microsoft Teams进行引导会话 | 引导会话由受过培训的研究者（临床心理学家和研究助理）进行；之后自我实施 | 参与者自己的智能手机、平板电脑或电脑（i-spero平台，内嵌俄罗斯方块） | 立即组：38/40（95%）坚持完成引导会话；延迟组：交叉后38/38（100%） | 允许继续常规护理（他们原本可获得的任何治疗） | 延迟获取（常规护理4周，然后相同干预） | 无主动对照任务；对照臂在前4周接受常规护理（无任务），然后交叉接受相同的ICTI |
| Deforges et al., 2022 () | 行为干预（记忆提醒+俄罗斯方块） | 单次会话治疗（长期创伤） | 是（多感官：返回产科病房、走过产房，加上按时间顺序的口头分娩叙述，然后针对最频繁侵入性记忆的具体叙述） | 记忆再巩固（再激活+更新）和视空间干扰（带心理旋转的俄罗斯方块） | 创伤后数年（中位数2年） | 1次会话（单次会话） | 叙述任务约10–12分钟，10分钟间隔（假定为不稳定化），20 分钟俄罗斯方块游戏 | 面对面：从中性办公室开始，移至产科病房检查室，然后回到办公室 | 由研究心理学家（CD）引导——非自我实施 | 手持游戏设备（Nintendo 3DS）及俄罗斯方块 | 17/18（94%）遵守干预指导（1人不遵守被排除） | 无（排除标准：正在进行的与分娩相关的心理治疗） | 基线阶段（干预前）——被试内对照 | 无任务；仅在干预前14天内通过每日日记监测侵入性记忆 |
| Badawi et al., 2021 () | 俄罗斯方块干预（视空间认知任务） | 辅助性、多次会话（3次） | 是（常规创伤聚焦治疗会谈中的想象暴露——至少15 分钟——作为再激活线索） | 记忆再巩固（通过想象暴露再激活 + 视空间干扰） | 慢性PTSD（索引创伤后数年） | 3次会谈（每次在一次包含≥15分钟想象暴露的治疗会谈之后） | 每次会谈12分钟俄罗斯方块游戏（研究前另加心理旋转练习） | 面对面，在医院病房内，在一次治疗会谈之后 | 临床医生提供平板电脑；参与者在创伤聚焦治疗会谈后自行操作。 | Lenovo平板电脑（10英寸屏幕），装有改良版俄罗斯方块（UTS Research Version，触摸屏） | 15/23（65%）完成了全部3次会谈（在开始接受想象暴露治疗的参与者中） | 常规治疗（创伤聚焦CBT配合想象暴露，每周3–4次，持续3–5周）——所有参与者相同 | 被试内基线（干预前监测期） | 无任务；参与者在任何俄罗斯方块会谈之前，通过手机应用程序监测侵入性记忆3天（基线） |
| Kanstrup et al., 2021 () | 认知任务干预（俄罗斯方块配合心理旋转） | 单次会谈，急性期实施 | 是（参与者在纸上列出“热点”——最糟糕时刻的画面——并在任务期间保持可见） | 记忆再巩固（再激活 + 更新）与视空间干扰（双任务、心理旋转） | 创伤后6小时内（74%）或72小时内（26%） | 1次指导会谈（在急诊科）+ 可选的居家自我实施强化会谈 | 总计约20分钟（至少10 分钟不间断，总计约20分钟） | 面对面，在急诊科候诊区或床上 | 由受过培训的研究助理（心理学硕士生，非临床医生）指导，此后自行实施 | 参与者自己的智能手机，装有官方俄罗斯方块应用程序（Android/iOS） | 20/22（91%）完成了干预会谈（2人因未知原因未完成） | 标准急诊科医疗护理；无限制 | 主动对照（播客） | 使用SR应用程序在自己的智能手机上收听瑞典广播哲学播客“Filosofiska Rummet”约20分钟，按照指示专注于内容。 |
| Freedman et al., 2020 () | SnowWorld（半沉浸式虚拟现实视空间任务） | 单次会谈，急性预防（破坏记忆巩固） | 否（假设急诊科情境已足以作为记忆提醒；未明确提取热点） | 视空间干扰（工作记忆负荷）——破坏非言语创伤记忆的巩固 | 创伤后约1小时内（范围最长8小时） | 1次会谈 | 平均8.06 分钟（SD 2.02）；最长10分钟 | 面对面，在急诊科内的一间小治疗室 | 由一名研究人员指导（设置VR，指示扔雪球） | 台式Asus电脑、Apple 30英寸Cinema Display屏幕、无线鼠标、耳机 | 36/36（100%）在急诊科完成了干预 | 标准急诊科医疗护理（记录了阿片类药物、缝合、骨科评估）。无限制 | 无干预对照 | 无任务；完成问卷，然后等待10分钟。仅接受常规医疗护理。 |
| Iyadurai et al., 2018 () | 基于俄罗斯方块的认知任务干预 | 单次会谈，急性预防（破坏记忆巩固） | 是（参与者向研究人员简要描述了事故中最糟糕的时刻） | 通过视空间干扰（工作记忆负荷）破坏记忆巩固 | 机动车事故后6小时内 | 1次会谈 | 约20分钟（至少10 分钟不间断） | 面对面，在急诊科一间诊室内 | 由临床心理学家（LI）指导——非自行实施 | 带俄罗斯方块游戏的手持游戏设备（Nintendo DS XL） | 36/37（97%）玩过≥10分钟（1人因工作人员中断） | 常规急诊医疗护理 | 注意力安慰剂对照（书面活动日志） | 纸笔活动日志：列出已在急诊完成的活动及持续时间，约20分钟；无记忆提醒线索。 |
| Horsch et al., 2017（） | 视空间认知任务（俄罗斯方块） | 单次、急性预防（记忆巩固中断） | 否（通过医院情境进行内隐记忆提醒；无明确热点提取） | 视空间干扰（工作记忆负荷）——中断视觉创伤记忆的巩固 | 在ECS后6小时内 | 1次 | 15分钟（最少10分钟） | 面对面，在病床上进行 | 由研究人员（非临床医生）指导，参与者在提供的设备上玩游戏 | 带俄罗斯方块游戏的手持游戏设备（Nintendo DS XL） | 26/29（89.7%）玩过≥10分钟（3人因医疗护理中断） | 常规护理（常规产科和产后护理） | 常规护理（无干预） | 无任务；仅继续常规临床护理 |
| Holmes et al., 2017（） | 基于俄罗斯方块的认知任务干预 | 单次、可行性（针对侵入性记忆） | 是（要求参与者短暂回想创伤事件，特别是脑海中闪现的最糟糕时刻） | 记忆再巩固（重新激活）+视空间干扰（俄罗斯方块游戏） | 不定（创伤后，可能数月至数年） | 1次 | 15–20分钟俄罗斯方块游戏（平均18 分钟，SD 2.4） | 面对面（个体化会话），基于智能手机玩游戏 | 由研究人员指导（口头指示）；在参与者自己的手机上自行玩游戏 | 参与者自己的智能手机，下载了俄罗斯方块应用（免费版） | 22/22（100%）完成≥15分钟的游戏 | 无（未报告其他心理治疗） | NA | 无对照条件 |
意象竞争任务的干预和对照细节。
大多数基于日记的平行组RCT显示侵入性记忆频率降低，支持ICTI，而Freedman（）和Kehyayan（）未显示显著获益（表 3）。
表 3
| 第一作者，年份 | 侵入性记忆频率（基于日记） | 侵入评估期 | 平均/中位侵入次数——干预组 | 平均侵入次数——对照组 | 相对于基线的变化（如有） | 达到减少阈值的人数 |
|---|---|---|---|---|---|---|
| Beckenstrom et al., 2026（） | 每日日记记录侵入性记忆次数（是/否+数量） | 第4周（第22–28天），控制基线周 | ICTI：中位数0.5（IQR 0.0–5.0） | 积极对照：中位数5.0（IQR 3.0–11.5）；TAU：中位数5.0（IQR 2.5–8.0） | ICTI比积极对照少72%；比TAU少70% | 到第24周：70%的ICTI参与者无侵入性记忆；积极对照13%，TAU 20% |
| Kanstrup et al., 2024（） | 每日日记（每天4次，持续7天）——工作相关创伤侵入性记忆计数 | 干预后第5周（主要终点） | 中位数=1.0（IQR 0–3）；平均值=3.4（SD 8.6） | 中位数=5.0（IQR 1–17）；平均值=12.4（SD 17.5） | 平均值较基线减少85.9%（干预组：从平均约21.8降至3.4）。IRR = 0.30（95% CI 0.17–0.53），p<0.0001 | 50%的干预参与者在第5周无侵入性记忆（完整日记数据） |
| Kehyayan et al., 2024（） | 每日日记（每24小时一次）——分别记录三个目标记忆的侵入次数 | 共8周；干预前（2周）与干预后（6周，跨越两个交叉期） | 俄罗斯方块：干预前平均2.02（SD 1.92）→干预后1.92（SD 1.88）（-5.0%变化） | Wiki：干预前 2.12（1.89）→ 干预后 2.02（2.18）（-4.7%）；无干预：干预前 2.14（2.22）→ 干预后 1.89（2.19）（-11.7%） | 相对风险（Tetris/Wiki）= 0.947（95% CI 0.85–1.05），p=0.31；无干预 vs Tetris RR = 1.060（0.98–1.15），p=0.15；无干预 vs Wiki RR = 1.004（0.93–1.09），p=0.92 | NR |
| Hardarson et al., 2024（） | 每日日记（每天4个时间点）——侵入性记忆的总计数 | 第5周（第二次干预会话后） | 均值 6.98（SD 5.73），n=8 | 基线（第-1周）：均值 25.98（SD 29.39），n=11 | 从基线到第5周减少68%（符合方案） | NR；在3个月时，从基线减少91% |
| Deforges et al., 2023（） | 分娩相关侵入性记忆的每日日记（产后第一周） | 产后前7天 | 各组的第1周日记均值和SD未在出版物中报告，仅通过Group × Day GLMM进行分析。 | 各组的第1周日记均值和SD未在出版物中报告，仅通过Group × Day GLMM进行分析。 | 各组的第1周日记均值和SD未在出版物中报告，仅通过Group × Day GLMM进行分析。 | 各组的第1周日记均值和SD未在出版物中报告，仅通过Group × Day GLMM进行分析。 |
| Iyadurai et al., 2023（） | 每日在线日记（每天一次）——工作相关创伤的侵入性记忆数量 | 第4周（第22–28天） | 立即组：中位数 1（IQR 0–3）；均值 4.03（SD 10.68） | 延迟组：中位数 10（IQR 6–17）；均值 12.46（SD 9.28） | IRR = 0.31（95% CI 0.20–0.48），p<0.001（第4周时立即组 vs 延迟组） | NR |
| Deforges et al., 2022（） | 分娩相关侵入性记忆（CB-IMs）的每日日记（14天）——参与者每天勾选一个方框并简要描述内容 | 日记2：干预后前14天（主要结局）；日记3：第28-42天（第5-6周） | 日记2中CB-IMs的中位数 = 2（IQR 4）；日记3中 = 2（IQR 3） | 基线（日记1，干预前14天）：中位数 = 11（IQR 6） | 从日记1到日记2的中位减少81.8%（IQR 39.6%）；从日记1到日记3的中位减少76.9%（IQR 29.0%） | 15/17（88%）从日记1到日记2减少≥50%；16/17（94%）从日记1到日记3减少≥50% |
| Badawi et al., 2021（） | 手机应用日记（ExpiWell）每天三次——侵入性记忆的数量 | 第0–21天（第一次Tetris会话前的基线期，然后每次会话后） | NR | NA | 部分参与者的效应量较小（例如，参与者15：从基线到第一次会话后，Tau = -0.52，p<0.001） | NR |
| Kanstrup et al., 2021（） | 纸笔每日日记（每天4个时间点），持续7天 | 第1周（干预后第1–7天）（主要时间点） | 均值 3.85（SD 8.57），n=20 | 均值 7.37（SD 7.88），n=19 | NA | NR |
| Kanstrup et al., 2021（第5周）（） | 相同的日记方法 | 第5周（干预后第28–34天）（次要时间点） | 均值 0.28（SD 0.57），n=18 | 均值 2.89（SD 6.43），n=18 | 干预组在第5周的侵入性记忆比对照组少90% | NR |
| Freedman et al., 2020（） | 无每日日记。PTSD症状量表（PSS）——侵入性症状严重程度的回顾性报告（非频率计数） | 创伤后2周和6个月 | 2周时PSS侵入性记忆：SnowWorld均值 3.54（SD 3.62），n=28；6个月时：均值 2.50（SD 3.26），n=28 | 2周时PSS侵入性记忆：对照组均值 2.70（SD 3.45），n=27；6个月时：均值 2.05（SD 2.89），n=21 | NA | NR |
| Iyadurai et al., 2018（） | MVA侵入性记忆的每日纸笔日记（每天/每时间段勾选方框） | 事故后前7天（事故当天至第7天） | 均值 8.73（SD 11.55），n=37（ITT插补） | 均值 23.26（SD 32.99），n=34（ITT插补） | NA | NR |
| Horsch et al., 2017（） | 每日日记（每天勾选）记录与 ECS 相关的侵入性创伤记忆（“闪回”） | ECS 后前 7 天（出生当天至第 7 天） | ITT：均值 4.77（SD 10.71）；符合方案：均值 3.54（SD 11.16） | ITT：均值 9.22（SD 10.69）；符合方案：均值 9.00（SD 9.32） | NA | NR |
| Holmes et al., 2017 () | 每日纸笔日记（每早/午/晚勾选）——创伤侵入性记忆计数 | 干预会话后 1 周（仅干预后；无干预前日记基线） | 每周均值 12.65（SD 9.71）；约每天 1.81（n=17） | NA | 未评估 | NR |
主要结局：侵入性记忆频率。
*在 Freedman et al., () 中：试验的主要结局是 2 周时的 PSS 总分（而非单独的侵入子量表）。此处展示侵入子量表以便与其他研究比较；标注为与其他试验中每日日记计数不同的指标。
*在 Iyadurai et al., () 中：展示原始均值；该试验发表的 d=0.67 来自对数转换分析，以处理计数数据偏态。
Beckenstrom ()：ICTI 中位数 0.5/周 vs. 活性对照 5.0 和 TAU 5.0。
Kanstrup ()：ICTI 中位数 1.0 vs. 对照 5.0；IRR = 0.30。
Iyadurai ()：ICTI 均值 8.73 vs. 对照 23.26（减少 62%）。
Horsch ()：ICTI 均值 4.77 vs. 对照 9.22（减少 48%）。
Hardarson ( () – 可行性)：第 5 周减少 68%，3 个月减少 91%。
Kehyayan ( () – 交叉设计)：无显著差异（RR = 0.95），可能由于遗留效应。
如 表 4 所示，大多数 RCT 报告 PTSD 症状显著减轻，支持 ICTI。例如，Beckenstrom ()：PCL-5 效应量（贝叶斯因子）极强（BF > 1000）vs. 对照。Kanstrup ()：1 个月时 PCL-5 OR = 0.21。Iyadurai ()：均值差 -2.29，p < 0.001。抑郁和焦虑效应较不一致；一些试验（如 Iyadurai ()）在 4 周时 PHQ-2 未发现显著差异，而其他试验显示第 4 周显著减轻但之后无。功能和生活质量：观察到积极效应，包括工作投入改善和倦怠减少（SWEBO）、WHODAS 评分更好、失眠更低（SCI-08）以及 EQ-5D-5L 生活质量更高。无严重不良事件归因于 ICTI。脱落率普遍较低且组间相当。
表 4
| 第一作者，年份 | PTSD 症状严重程度（CAPS、PCL-5、IES-R） | 抑郁症状（PHQ-9、BDI） | 焦虑症状（GAD-7、HADS） | 功能结局（工作/社会功能） | 睡眠结局 | 与侵入相关的痛苦 | 生活质量 | 急性应激症状 | 记忆的情绪影响 | 研究结论 |
|---|---|---|---|---|---|---|---|---|---|---|
| Beckenstrom et al., 2026 () | PCL-5：ICTI 比活性对照（第 4 周 BF = 1546）和 TAU（第 4 周 BF = 7732）降低更多。无 ICTI 参与者在第 12/24 周符合可能 PTSD 标准。 | 第 4 周 ICTI 相比活性对照（BF = 80.1）和 TAU（BF = 77.6）降低。后期时间点结果不一。 | 第 4 周 ICTI 相比活性对照（BF = 13.4）和 TAU（BF = 53.4）降低。第 12/24 周变化不定。 | WHODAS：所有时间点 ICTI vs TAU 有极端证据（BFs 508–2230）；vs 活性对照 BF = 138.1（第 4 周）、2.8（第 12 周）、21.0（第 24 周）。 | 第 4 周 ICTI 相比活性对照（BF = 4.3）和 TAU（BF = 23.4）失眠减轻；持续。 | ICTI 参与者相比两个对照者报告的痛苦/干扰更少，尤其是在第 24 周。 | EQ-5D-5L：第4周焦虑/抑郁方面ICTI较两个对照均获益；其他维度无差异/结果不一。 | NR | 由痛苦/干扰评分涵盖。 | EQ-5D-5L: ICTI benefit at week 4 for anxiety/depression dimension (Bayes Factor extreme). Other dimensions null/mixed. At week 12 and 24, evidence weakened. Full BF values available in the original publication. |
| Kanstrup et al., 2024 () | PCL-5 (8-item) lower in intervention at all timepoints (week 1 to 6 months): e.g., week 5 OR = 0.21 (0.10–0.42). IES-R intrusions are lower at all timepoints (OR 0.16–0.25). | NR | NR | WHODAS: better cognition and personal care at 6 months. Lower burnout (SWEBO) at 6 months (OR = 0.46). Less work impact from intrusions (OR 0.22–0.56). | The Sleep Condition Indicator (SCI-02) was better at week 1, 1 month, and 3 months (OR 2.16–3.92), but not at 6 months. | IES-R intrusions (see PTSD). Also, less disruption of concentration due to intrusions (OR ~0.40). | Self-Rated Health (SRHR) was higher (better) in the intervention at all timepoints (OR 2.61–3.13). | NR | Less impact of intrusive memories on daily functioning (OR 0.23–0.47) and occupational functioning (OR 0.22–0.56) across 1–6 months. | A brief, single-session remote ICTI (Tetris + mental rotation) reduces intrusive memories and PTSD symptoms in traumatized healthcare workers, with effects lasting up to 6 months; it is safe, scalable, and superior to active control. |
| Kehyayan et al., 2024 () | NR | NR | NR | NR | NR | NR | NR | NA | NR | In a crossover RCT of PTSD patients (~13.8 years post-trauma), neither Tetris nor verbal control reduced intrusions vs no intervention; possible carryover effects suggest the need for parallel-arm RCTs. |
| Hardarson et al., 2024 () | PCL-5: baseline 36.42 → 3-month 18.71 (d=-0.71). UMT frequency: baseline 3.25 → 3-month 1.29 (d=-2.36). UMT distress: baseline 45.42 → 3-month 24.00 (d=-0.76). | PHQ-9: baseline 11.83 → 3-month 8.29 (d=-0.46) | GAD-7: baseline 8.75 → 3-month 5.00 (d=-0.44) | SDS (Sheehan) functional impairment decreased; WHODAS 2.0 improved; impact of intrusions on daily functioning reduced | Impact on sleep: reduced at 1 month, slight increase at 3 months, but still below baseline; nightmares reduced to near zero at 3 months | UMT distress (see PTSD); intrusion-related distress reduced over time | Not assessed | NA | Vividness and distress ratings reduced; concentration disruption duration decreased. | Remote ICTI in long-standing trauma was feasible and acceptable, with large reductions in intrusive memories (68% at 5 weeks, 91% at 3 months) and improvements in PTSD, depression, anxiety, and functioning; a full RCT is needed. |
| Deforges et al., 2023 () | PCL-5 (self-report): At 6 weeks: marginal effects (total p<0.06; intrusions p<0.0525; arousal p<0.0552). At 6 months: significant total (β=-0.65, p=0.041), alterations in cognition/mood (p=0.0016), and arousal (p<0.0289). CAPS-5 (clinician): No significant group differences at 6 weeks or 6 months. | NR (measured but between-group results NR in publication) | NR (measured but between-group results NR in publication) | NR | NR | NR | NR | NR | NR | A single-session Tetris-based intervention after unplanned cesarean section reduced self-reported CB-PTSD symptoms at 6 months, but not clinician-rated PTSD, and was feasible with preventive potential. |
| Iyadurai et al., 2023 () | PCL-5 (4-item): lower in immediate vs delayed at week 4 (mean diff -2.29, p<0.001). IES-R total lower (diff -0.8, p<0.001). | PHQ-2: no significant difference at week 4 (diff -0.52, p=0.19); but within-arm at week 8 showed a reduction. | GAD-2: lower in immediate at week 4 (diff -0.93, p=0.016). | Work engagement (SWEBO) was higher (diff 0.46, p<0.001); burnout was lower (diff -0.61, p<0.001). Sick days: no difference at week 4 (IRR 1.50, ns); at week 8, within-arm reduction. | Insomnia (SCI-08): better in immediate at week 4 (diff 5.38, p<0.001). | Distress, concentration disruption, and interference all decreased immediately (all p<0.001). | EQ-5D-5L quality of life was higher in immediate (diff 12.32, p=0.002). WHODAS is functioning better (diff = -9.78, p < 0.001). | NA | Lower impact on work functioning and other life areas (both p<0.001). | A brief remote ICTI (Tetris + mental rotation) reduced intrusive memories at 4 weeks in ICU staff, with improvements in PTSD, anxiety, sleep, functioning, and quality of life; it was safe, acceptable, and effects were replicated after crossover, supporting further trials. |
| Deforges et al., 2022 () | PCL-5 (CB-PTSD) total severity: pre 28.71 (SD 12) → post 12.29 (SD 8.84), p<0.001, Hedges’ g=1.47 (large). All 4 symptom clusters significantly reduced (p<0.01 to <0.001, all g>0.8). Of 8 participants with probable CB-PTSD pre-intervention, 0 met criteria post-intervention (p=0.013). | EPDS (Edinburgh Postnatal Depression Scale) pre-intervention: mean 10.67 (SD 4.63) – above clinical cutoff (10.5). Not reported as outcome post-intervention | Not assessed | Not assessed | Not assessed | NR | Not assessed | NA | Participants reported reduced impact on daily life (e.g., less fear of becoming pregnant again) in qualitative feedback. | This is the first test of a single-session behavioral intervention for an old, real-life single-event trauma. Its promising findings support further randomized controlled trials and suggest a potential new approach for CB-PTSD treatment. |
| Badawi et al., 2021 () | PCL-5 total: pre 59.0 (SD 8.0) → post 50.89 (17.51), Hedge’s g=0.47 (n=9). IES-R intrusions: pre 30.56 (5.36) → post 28.00 (6.38), g=0.37 | HADS depression: pre 12.25 (4.86) → post 10.67 (4.03), g=0.38 (n=8) | HADS anxiety: pre 15.00 (2.65) → post 13.11 (4.60), g=0.38 (n=9) | Not assessed | Not assessed | Not assessed | Not assessed | NA | Not assessed | A feasibility study showed that Tetris as an adjunct to trauma-focused therapy is feasible in inpatient PTSD care, with small to moderate reductions in intrusions and modest improvements in PTSD, depression, and anxiety; optimal dosing requires further study. |
| Kanstrup et al., 2021 () | IES-R total at 1 week: Intervention mean 21.3 (SD 15.4), Control 28.0 (16.9), d=0.41 (-0.26,1.07). At 1 month: Intervention 15.0 (12.6), Control 22.5 (15.2), d=0.52. At 6 months: Intervention 19.5 (21.1), Control 16.9 (16.5), d=-0.15 | HADS depression subscale at 1 week: Intervention 1.43 (1.91), Control 2.17 (2.50), d=0.33 (-0.41,1.06) | HADS anxiety subscale at 1 week: Intervention 2.14 (2.28), Control 3.11 (2.70), d=0.38 (-0.35,1.12) | Work and Social Adjustment Scale (WSAS) at 1 week: Intervention 10.0 (11.4), Control 14.9 (11.0), d=0.45 (-0.23,1.12) | Self-rated sleep (SRSR) at 1 week: Intervention 5.91 (2.35), Control 5.00 (2.98), d=-0.38 (-1.11,0.36) | IES-R intrusion subscale at 1 week: Intervention 1.15 (0.94), Control 1.60 (0.88), d=0.49 (-0.19,1.17) | Self-rated health (SRHR) at 1 week: Intervention 5.82 (1.10), Control 5.37 (1.21), d=-0.40 (-1.12,0.33) | NA | Concentration disruption due to intrusions at 1 week: Intervention 2.30 (2.64), Control 2.45 (2.52), d=0.06 (-0.67,0.79) | A brief single-session ICTI (Tetris + mental rotation) delivered acutely in the Emergency Department reduced intrusive memories at 1 week in trauma-exposed adults, with the intervention group showing 90% fewer intrusions than control at Week 5. The pilot trial supports the feasibility and exploratory efficacy of ICTI for acute trauma in an emergency care setting. |
| Freedman et al., 2020 () | PSS total score at 2 weeks: SnowWorld 10.39 (11.62), Control 7.52 (9.69), p=0.22; probable PTSD: 28.6% vs 22.2%, ns. At 6 months: SnowWorld 7.93 (10.11), Control 5.90 (6.64), p=0.64; probable PTSD: 21.4% vs 14.3%, ns. | Not assessed | Not assessed | Not assessed | Not assessed | Not assessed | Not assessed | Not assessed | Not assessed | This pilot RCT found no benefit of a single VR visuospatial task (SnowWorld) vs no intervention on PTSD symptoms at 2 weeks or 6 months; it was feasible but did not replicate prior Tetris findings, possibly due to methodological differences. |
| Iyadurai et al., 2018 () | IES-R intrusion subscale at 1 week: Intervention 7.27 (5.27), Control 10.70 (7.29), p=0.024, d=0.54. PDS total at 1 week: ns (d=0.20). At 1 month: No significant differences on IES-R or PDS. PTSD diagnosis at 1 month (PDS): Intervention 12.9%, Control 9.7%, OR = 1.40, ns. | HADS total at 1 week and 1 month: no significant differences (d=0.15 and 0.04) | HADS total (combined anxiety/depression) – no differences | Not assessed | Not assessed | IES-R intrusion subscale (see PTSD) – significant at 1 week only | Not assessed | Not assessed | Not assessed | A single Tetris session with a memory cue within 6 hours post-MVA reduced intrusive memories at 1 week (62% vs active control), with effects limited to intrusion symptoms; it was feasible, acceptable, and safe, with no 1-month benefit. |
| Horsch et al., 2017 () | PDS (Posttraumatic Diagnostic Scale) at 1 month: total score ITT: Intervention 13.5 vs Control 6.8 (p=ns). Per protocol: no significant differences. PTSD diagnosis at 1 month: ITT 12.1% vs 25.1% (ns) | HADS depression at 1 week and 1 month: no significant group differences (ITT and per protocol) | HADS anxiety at 1 week and 1 month: no significant group differences | Not assessed | Not assessed | Not assessed | Not assessed | ASDS total score at 1 week: no significant difference. ASDS avoidance, arousal: no differences | Not assessed | A single 15-min Tetris session within 6 hours post-cesarean reduced intrusive memories (~48%) in the first week vs usual care, with good feasibility and safety; no significant PTSD difference at 1 month. |
| Holmes et al., 2017 () | Not assessed | Not assessed | Not assessed | Concentration disruption: 82% reported >1 min disruption per intrusion; 68% >5 min; 32% >30 min. General concentration difficulties (0-10): mean 5.68, SD 3.59. Positive correlation with intrusion frequency (r=0.64, p=0.001). | Not assessed | Not assessed | Not assessed | NA | Not assessed | A feasibility study among young refugees (n=22) showed that diary monitoring, combined with a single Tetris session after a memory cue, was feasible, acceptable, and well-received, supporting the further development of scalable smartphone-based interventions. |
Secondary outcomes: PTSD, depression, anxiety, and functioning.
Depression outcomes
Depression was assessed in four studies using different measures: HADS-D [Iyadurai (), Kanstrup (), Horsch ()], PHQ-2 [Beckenstrom (), and Iyadurai ()]. No meta-analysis was performed due to the heterogeneity of instruments and the insufficient number of studies reporting comparable metrics. Individually, Beckenstrom () reported strong Bayesian evidence for ICTI reducing PHQ-2 scores at week 4 compared to active control (BF = 80.1) and TAU (BF = 77.6), though effects were mixed at later timepoints. Iyadurai () found no significant difference in PHQ-2 at week 4 (mean difference –0.52, 95% CI –1.23 to 0.19). Kanstrup () reported a small effect favouring ICTI at 1 week (Cohen’s d = 0.33, 95% CI –0.41 to 1.06). Horsch () found no significant group differences at 1 week or 1 month. Overall, evidence for an effect of ICTI on depression is inconsistent and limited by small numbers and heterogeneous measures.
Anxiety outcomes
Anxiety was assessed using GAD-2 [Beckenstrom (), Iyadurai ()], and HADS-A [Kanstrup (), Horsch ()]. Meta-analysis was not performed due to the heterogeneity of measures and limited comparable data. Beckenstrom () reported strong evidence for ICTI reducing GAD-2 at week 4 versus active control (BF = 13.4) and TAU (BF = 53.4), though later timepoints were mixed. Iyadurai () found a significant difference in GAD-2 favouring immediate ICTI at week 4 (mean difference –0.93, 95% CI –1.68 to –0.17). Kanstrup () reported a small effect at 1 week (Cohen’s d = 0.38, 95% CI –0.35 to 1.12). Horsch () found no significant differences at 1 week or 1 month. Collectively, the evidence tentatively favours ICTI for reducing anxiety in the short term, but larger controlled trials with consistent measures are needed.
Between-group mean difference in intrusive memory frequency (ICTI vs. Control) at primary timepoint
A random-effects meta-analysis of raw mean differences (ICTI minus control) was conducted for the five parallel-group RCTs that reported intrusive memory frequency per week. The pooled mean difference was –6.98 intrusive memories per week (95% CI –9.84 to –4.13), favouring the ICTI intervention. The effect was statistically significant as the confidence interval did not cross zero. Heterogeneity was low (I2 = 21.9%, p = 0.275), indicating that the five studies provided a consistent estimate of the treatment effect. Individual study estimates ranged from –14.53 to –3.52, with three of the five studies showing statistically significant benefits favouring ICTI. The forest plot is presented in Figure 2. Randomized studies that could not be included in the primary raw mean-difference meta-analysis because of incompatible design or outcome format are summarized in the qualitative and narrative synthesis below.
Figure 2
Moderator analysis – memory reactivation cue
A random-effects subgroup analysis was conducted to examine whether the presence of an explicit memory reactivation (MR) cue modified the intervention effect on intrusive memory frequency Figure 3. Amongst the four studies that used an explicit MR cue (Iyadurai (), Kanstrup (), Iyadurai (), and Kanstrup ()), the pooled mean difference was −7.65 intrusive memories per week (95% CI −11.01 to −4.29), favouring ICTI, with low heterogeneity (I2 = 28.1%, p = 0.243). The single study without an explicit MR cue (Horsch ()) showed a mean difference of −4.45 intrusive memories per week (95% CI −10.06 to 1.16). Although the confidence intervals overlapped, suggesting no clear difference in treatment effect between cue and no-cue studies, this comparison should be interpreted with considerable caution. Because the no-cue subgroup included only one study, a reliable formal test for subgroup differences could not be estimated. Therefore, this analysis should be considered exploratory and hypothesis-generating rather than definitive.
Figure 3
PTSD symptom severity (1–4 weeks)
The forest plot suggests that the intervention was associated with lower PTSD symptom severity compared with control, with an overall pooled mean difference of -2.42 (95% CI -4.40 to -0.44) Figure 4. Because the confidence interval does not cross zero, this overall effect is statistically significant. However, the results were not fully consistent across studies, as shown by moderate-to-substantial heterogeneity (I2 = 60.88%, p = 0.037), indicating that the study estimates varied meaningfully. The clearest benefit appears in the 1-month subgroup, where the pooled effect was also significant (MD = -2.98, 95% CI -4.74 to -1.23; I2 = 52.69%), whilst the less-than-1-month subgroup was based on a single study and showed a positive but imprecise estimate (MD = 2.87, 95% CI -2.78 to 8.52), so no firm conclusion can be drawn from it. Overall, the pattern favours the intervention, but the moderate heterogeneity suggests that the magnitude of benefit may differ across studies and follow-up timepoints.
Figure 4
It is worth noting that different PTSD instruments (PCL-5, IES-R, PDS/ASDS, PSS) were pooled as raw mean differences. This limits clinical interpretability, as the scales have different ranges; results should be interpreted with caution.
Crossover trial sensitivity analysis
A standalone sensitivity analysis of the crossover study by Kehyayan et al. () was conducted to evaluate the intervention effect within this within-subject design, yielding a pooled risk ratio of 0.95 (95% CI 0.85 to 1.05), which was not statistically significant (Z = 1.00, p = 0.32) Figure 5. This estimate indicates no clear benefit of the visuospatial intervention over the verbal control condition in this single crossover trial, as the confidence interval crosses the null value of 1.0. This null finding contrasts with the significant pooled effect observed in the five parallel-group RCTs, and may be attributable to methodological factors inherent to the crossover design, such as potential carryover effects in the absence of a washout period, or to the chronic, long-standing nature of the trauma in this population (mean 13.8 years post-trauma), which may be less responsive to a single-session intervention. Overall, this sensitivity analysis reinforces the decision to restrict the primary quantitative synthesis to parallel-group RCTs whilst highlighting the need for further parallel-arm trials in chronic populations.
Figure 5
Qualitative and narrative synthesis
Seven studies were not included in the primary raw mean-difference meta-analysis of weekly diary-recorded intrusive-memory frequency. These comprised three randomized trials with outcome formats or designs incompatible with raw mean-difference pooling, and four non-randomized feasibility or pre–post studies. These studies were therefore summarized narratively to contextualize the pooled findings and to describe evidence from different populations, intervention formats, and outcome measurement approaches.
Randomized trials not included in the primary raw mean-difference meta-analysis
Three randomized trials were not included in the primary pooled raw mean-difference analysis because their data were not directly compatible with the five diary-count RCTs included in the main forest plot.
Beckenstrom et al. () conducted a three-arm Bayesian adaptive randomized trial amongst UK NHS healthcare staff exposed to COVID-19-related occupational trauma. The study compared ICTI with an active control condition and treatment as usual. At the Week 4 primary endpoint, the ICTI group had substantially fewer intrusive memories than both comparators, with a median of 0.5 intrusive memories per week compared with 5.0 per week in the active control group and 5.0 per week in the treatment-as-usual group. The Bayesian analysis provided strong evidence favouring ICTI over both active control and treatment as usual. Although this study was not included in the raw mean-difference meta-analysis because of its three-arm Bayesian adaptive design and use of model-based estimates, its findings were directionally consistent with the pooled effect observed in the primary analysis.
Deforges et al. () evaluated a single-session Tetris-based intervention after unplanned cesarean section in a multicentre, double-blind randomized trial. The primary outcome was childbirth-related PTSD symptom severity rather than diary-recorded intrusive-memory frequency. At 6 weeks, self-reported PCL-5 total scores and intrusion subscale scores showed marginal effects favouring the intervention, whilst clinician-rated CAPS-5 outcomes did not show significant between-group differences. At 6 months, self-reported PCL-5 total scores favoured the intervention. Although Week 1 diary data were collected, per-arm means and standard deviations were not reported, and the diary outcome was analysed using a group-by-day model. Therefore, this trial could not be included in the primary raw mean-difference meta-analysis. Overall, the trial provides supportive but indirect evidence, with stronger effects on self-reported PTSD symptoms than on clinician-rated outcomes.
Freedman et al. () was a pilot randomized trial of SnowWorld, a semi-immersive virtual reality visuospatial task, in emergency department trauma survivors. This study differed from the Tetris-based ICTI trials in both intervention type and outcome measurement. It did not use a daily diary of intrusive-memory frequency; instead, outcomes were assessed using the Posttraumatic Symptom Scale. At 2 weeks, PSS intrusion scores were numerically higher in the SnowWorld group than in the control group, and the trial did not demonstrate a significant benefit on PTSD symptoms at either 2 weeks or 6 months. This study therefore represents an important exception to the generally favourable pattern of findings and suggests that the therapeutic effect observed in Tetris-based ICTI studies may not automatically generalize to all visuospatial tasks or to questionnaire-based PTSD outcomes.
Non-randomized studies
Four non-randomized studies were included in the narrative synthesis only because they lacked a concurrent randomized control group or did not provide data suitable for controlled effect-size pooling.
Hardarson et al. () evaluated a remotely delivered imagery-competing task intervention in 12 trauma-exposed women with long-standing intrusive memories, with a mean time since trauma of approximately 20.7 years. The intervention was feasible and acceptable, and participants showed large reductions in intrusive memories, including a 68% reduction at Week 5 and a 91% reduction at 3 months. PTSD, depression, anxiety, and functioning also improved over follow-up. However, because the study used a feasibility pre–post design without a concurrent control group, these improvements cannot be attributed definitively to the intervention.
Deforges et al. () tested a single-session intervention combining childbirth-related memory reactivation and Tetris gameplay in women with childbirth-related intrusive memories. Participants showed a large median reduction in childbirth-related intrusive memories after the intervention, and most achieved at least a 50% reduction. PTSD symptom severity also improved substantially. This proof-of-principle study supports the feasibility of targeting long-standing childbirth-related intrusive memories, but its uncontrolled design means that the findings should be interpreted as preliminary.
Badawi et al. () examined adjunctive Tetris gameplay after imaginal exposure sessions in an inpatient PTSD treatment setting. The study suggested that adding a visuospatial task after trauma-focused therapy was feasible, with small-to-moderate reductions in intrusion-related and PTSD outcomes. However, the study was limited by attrition, small sample size, and the absence of a separate control group, making it difficult to isolate the specific effect of Tetris from the broader inpatient trauma-focused treatment programme.
Holmes et al. () assessed feasibility and acceptability of a Tetris-based intervention amongst young refugees aged 16–25 years. The study demonstrated that participants could complete diary monitoring and smartphone-based Tetris gameplay, and intrusive memories were commonly associated with concentration disruption. This study did not test efficacy using a controlled design, but it supports the feasibility of delivering a simple, scalable intervention in a refugee population.
Integrated narrative interpretation
Overall, the qualitative and narrative evidence was broadly consistent with the primary meta-analysis, particularly amongst Tetris-based studies using memory reactivation and diary-based intrusive-memory outcomes. Beckenstrom et al. () provided strong randomized evidence favouring ICTI despite being incompatible with raw mean-difference pooling, whilst Deforges et al. () suggested possible benefits on self-reported childbirth-related PTSD symptoms but provided less direct evidence for diary-recorded intrusive-memory frequency. The non-randomized studies further supported feasibility, acceptability, and potential clinical utility across diverse settings, including remote delivery, childbirth-related trauma, inpatient PTSD treatment, and young refugees.
However, the narrative synthesis also highlights important limitations. Evidence from uncontrolled studies cannot establish causal efficacy, and improvements may reflect regression to the mean, natural recovery, concurrent treatment, or repeated monitoring. Freedman et al. () was the clearest exception to the overall favourable pattern, showing no significant benefit of SnowWorld VR compared with control. This may reflect differences in intervention content, absence of an explicit memory reactivation cue, shorter task exposure, or use of questionnaire-based PTSD outcomes rather than prospective diary-recorded intrusive-memory frequency. Taken together, the narrative synthesis supports the plausibility and feasibility of ICTI across populations, whilst reinforcing that the strongest efficacy evidence comes from the five parallel-group diary-based RCTs included in the primary meta-analysis.
Funnel plot of primary intrusive-memory
Figure 6 presents the funnel plot of the five studies in the primary meta-analysis. Visual inspection showed no obvious asymmetry, with studies distributed reasonably around the pooled estimate. However, because fewer than 10 studies were available, formal statistical tests are underpowered and were not performed. No definitive conclusion regarding publication bias can be drawn from this small sample.
Figure 6
Risk of bias assessment
In Figure 7, the ROB 2 assessment indicates that of the 9 included RCTs, 2 were judged at low risk of bias overall, 4 had some concerns, and 3 were at high risk of bias, supporting the internal validity of the evidence base. Several studies, including Kanstrup et al. (), Deforges et al. (), and the most recent reports, showed low risk across nearly all domains and were judged as low risk overall (, ). In contrast, a few studies raised concerns or were rated as high risk, mainly due to deviations from the intended intervention, missing outcome data, or selective reporting. Notably, Kanstrup et al. (), Freedman et al. (), and Horsch et al. () were judged as high risk overall, reflecting weaknesses in one or more key domains rather than a uniform lack of quality (, , ). Therefore, the figure indicates that the evidence is generally reliable, but the findings should still be interpreted with some caution because a minority of studies contributed a higher risk of bias that could have influenced the estimated effects.
Figure 7
As shown in Figure 8, the ROBINS-I assessment indicates that the non-randomized evidence was generally limited by important methodological concerns, with all included studies judged to be at serious overall risk of bias. The main source of bias was confounding, which was rated serious across every study, suggesting that baseline differences and other uncontrolled factors may have influenced the observed effects. In addition, most studies reported moderate concerns about participant selection, deviations from intended interventions, and, in some cases, missing data, although classification of the interventions, outcome measurement, and selection of the reported results were mostly rated as low risk.
Figure 8
GRADE assessment
The GRADE assessment revealed moderate-certainty evidence that the ICTI substantially reduces intrusive memory frequency compared to control conditions (Table 5). Across five parallel-group randomized controlled trials, 398 participants were randomized, of whom 359 contributed analyzable primary-outcome data to the pooled raw mean-difference analysis. ICTI reduced intrusive memories by a mean difference of −6.98 intrusions per week (95% CI −9.84 to −4.13), favouring the intervention. Certainty of evidence for this outcome was downgraded by one level for risk of bias, because two of the five pooled trials (Kanstrup (), Horsch ()) were judged at high overall risk of bias on RoB 2, and a further two (Iyadurai (), Iyadurai ()) had some concerns; only one pooled trial (Kanstrup ()) was judged low risk. For PTSD symptom severity, the evidence was rated as low certainty due to serious inconsistency (I2 = 60.9%, p = 0.037) and indirectness arising from the pooling of different PTSD instruments (PCL-5, IES-R, PDS/ASDS, PSS) as raw mean differences. The pooled mean difference of –2.42 points (95% CI –4.40 to –0.44) suggests a modest reduction in PTSD symptoms favouring ICTI, but this finding should be interpreted with caution given the heterogeneity and measurement differences across studies. Publication bias was rated as undetected for the primary outcome, as a funnel plot of the five studies showed no obvious asymmetry; however, formal statistical testing was not performed due to the small number of studies (k < 10), and definitive conclusions about publication bias cannot be drawn. Overall, the evidence provides moderate certainty that ICTI reduces intrusive memories, reflecting residual concerns about risk of bias in the contributing trials, and the effect on PTSD symptom severity is less certain and requires further investigation with standardised outcome measures in larger trials.
Table 5
| Certainty assessment | № of patients (n = 359 Analyzed) | Effect | Certainty | Importance | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| № of studies | Study design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | ICTI | Control | Relative (95% CI) | Absolute (95% CI) | ||
| Intrusive memory frequency (follow-up: range 1 weeks to 4 weeks; assessed with: Daily diary) | ||||||||||||
| 5 | randomised trials | seriousd | not serious | not serious | not serious | nonea | 182 | 177 | MD –6.98 (–9.84 to –4.13) | 6.98 fewer intrusions per week 95% CI: 9.84 fewer to 4.13 fewer | ⨁⨁⨁◯ Moderateda | CRITICAL |
| PTSD symptom severity (follow-up: range 1 week to 4 weeks; assessed with: PCL-5, IES-R, PDS/ASDS, PSS; Scale from: 0 to 80) | ||||||||||||
| 5 | randomised trials | not serious | seriousb | seriousc | not serious | none | 285 | 286 | MD –2.42 (–4.40 to –0.44) | MD 2.42 points on scale lower (4.4 lower to 0.44 lower) | ⨁⨁◯◯ Lowb,c | CRITICAL |
GRADE summary of findings for imagery-competing task intervention vs. control.
CI, confidence interval; MD, mean difference.
Explanations.
aA funnel plot was generated for visual inspection of the five studies included in the primary meta-analysis. In accordance with the Cochrane Handbook, formal statistical tests (e.g., Egger’s regression test) were not performed because they are underpowered with fewer than 10 studies. The funnel plot showed no obvious asymmetry, but definitive conclusions regarding the presence or absence of publication bias cannot be drawn from this small set of studies. Therefore, publication bias was rated as undetected, with no downgrade applied, but this limitation is acknowledged narratively.
bI2 = 60.9%, p = 0.037.
cDifferent instruments pooled as raw MD.
dDowngraded one level for risk of bias: two of the five pooled trials [Kanstrup (), Horsch ()] were judged at high overall risk of bias on RoB 2, and two further trials [Iyadurai (), Iyadurai ()] had some concerns; only one trial [Kanstrup ()] was judged low risk.
Discussion
This systematic review and meta-analysis evaluated visuospatial imagery-competing task interventions for reducing intrusive memories after real-world trauma. Across five parallel-group randomized controlled trials that reported weekly diary-recorded intrusive-memory frequency, ICTI reduced intrusive memories compared with control conditions. The pooled raw mean difference was −6.98 intrusive memories per week, with a 95% confidence interval from −9.84 to −4.13 and p < 0.001. Heterogeneity was low, with I2 = 21.94%, and the certainty of evidence for this primary outcome was rated as moderate, downgraded one level for risk of bias because two of the five pooled trials were judged at high overall risk of bias. The PTSD symptom severity analysis also favoured ICTI, with a pooled mean difference of −2.42 points, but the certainty of evidence was low because different PTSD instruments with different score ranges were combined as raw mean differences and heterogeneity was moderate. Depression and anxiety outcomes could not be pooled because measures and reporting formats differed across studies. The qualitative and narrative synthesis generally aligned with the primary finding, especially in studies that used Tetris-based interventions and prospective diary measures of intrusive-memory frequency. Randomized studies that were not included in the primary meta-analysis also tended to support the beneficial effect.
The present findings extend previous laboratory and translational reviews by focusing on real-world trauma rather than analogue trauma exposure. Asselbergs et al. reported a moderate effect of cognitive interventions on intrusive memories in trauma film experiments, with visuospatial interference tasks showing clearer benefit than verbal tasks (). Varma et al. later examined a wider range of lab-analogue methods and found that techniques designed to reduce intrusive memories produced small to moderate effects, with trauma reminder plus Tetris showing benefit when administered immediately after, or after a delay following, analogue trauma (). Astill Wright et al. reviewed consolidation and reconsolidation therapies for PTSD and re-experiencing, and identified cognitive task memory interference with memory reactivation as one of the few protocols with evidence of benefit for intrusive memories (). Our review agreed with these prior syntheses but moved the evidence base closer to clinical translation by restricting the primary meta-analysis to real trauma populations and diary-recorded outcomes.
The pooled finding also aligned with the symptom focused framework proposed by Iyadurai et al., which treats intrusive memories as a clinically relevant target linking cognitive science and clinical application (). The real-world trials included in the primary analysis showed a consistent direction of effect across emergency department trauma, emergency caesarean section, and occupational trauma settings (, , , , ). Non-pooled randomized evidence gave a more mixed picture. The Bayesian adaptive trial by Beckenstrom et al. supported the same direction of effect, although its model-based reporting and three-arm design were not compatible with raw mean-difference pooling (). The START trial by Deforges et al. provided supportive but indirect evidence because the main extractable outcome was childbirth-related PTSD symptom severity rather than per-arm diary frequency (). By contrast, Freedman et al. found no significant benefit of SnowWorld VR, and Kehyayan et al. found no significant benefit of Tetris over verbal control in a crossover trial of chronic PTSD (, ). These findings suggest that intervention content, reactivation procedures, trauma chronicity, and outcome measurement may influence observed effects.
Several mechanisms may explain why ICTI reduced diary recorded intrusive memory frequency. Dual representation theory proposes that traumatic memories can be stored with strong sensory-perceptual representations and weaker contextual integration, which may increase involuntary sensory intrusions (). Cognitive models of PTSD similarly emphasize data-driven trauma encoding, negative appraisals, and poorly integrated autobiographical memory as factors that maintain re-experiencing symptoms (). Within this framework, a brief reminder may reactivate the trauma image, whilst Tetris or another visuospatial task taxes the same limited visuospatial working memory resources needed to maintain or restabilise vivid imagery.
Experimental work supports this account. James et al. () showed that Tetris after memory reactivation reduced intrusive memories of experimental trauma, and early cognitive-science studies proposed that visuospatial gameplay could compete with the build-up of visual flashbacks (). In the present review, the memory reactivation subgroup analysis was compatible with this interpretation, but it did not prove mechanism because the no-cue subgroup contained only one study. The diary-based primary outcome is also relevant. Daily diary methods capture prospective intrusive-memory counts and avoid some limitations of retrospective questionnaire scoring, although they still depend on participant adherence and interpretation of diary instructions (). The mechanism should therefore be viewed as plausible rather than established.
Current PTSD guidelines recommend active monitoring or trauma-focused psychological therapy after early clinically important symptoms, and established PTSD treatments such as trauma-focused CBT, cognitive processing therapy, prolonged exposure, and EMDR remain the main recommended interventions (, , ). ICTI is not yet positioned within these guidelines as a standard treatment or preventive intervention. The present findings do not challenge those recommendations. They support a narrower conclusion: a brief, targeted, low intensity intervention may reduce intrusive-memory frequency after real world trauma and may be suitable for further testing as an adjunct or early symptom-focused strategy.
The next clinical trials should use adequately powered, parallel-group designs in clearly defined trauma populations. Acute trials should recruit emergency department and maternity patients within prespecified time windows and should report weekly diary counts, PTSD diagnosis or severity at one and three months, adverse events, and functional outcomes. Delayed intervention trials should test occupational trauma and chronic intrusive memories using the same diary outcome and a standardized PTSD scale. Factorial designs could test whether memory reactivation is necessary by comparing Tetris with and without an explicit hotspot cue.
This review had several strengths. It focused on real-world trauma rather than laboratory analogue trauma, separated diary-based RCTs from studies with incompatible outcome formats, used a clinically interpretable raw mean difference, and evaluated certainty of evidence. It also distinguished the primary pooled analysis from narrative evidence, which reduced the risk of mixing questionnaire outcomes, Bayesian estimates, crossover data, and uncontrolled pre-post designs in a single effect estimate.
Several limitations remain. The primary meta-analysis included only five RCTs, which limited precision for moderator analyses, funnel plot interpretation, and assessment of publication bias. The PTSD severity analysis was exploratory because different instruments were pooled as raw mean differences. Residual heterogeneity may reflect trauma type, intervention timing, comparator intensity, delivery mode, and follow-up duration. One analytic decision differed from the initial broad plan: the primary pooled analysis was restricted to parallel-group RCTs with diary-count data rather than combining incompatible metrics. This improved interpretability but reduced the number of pooled studies.
Conclusion
ICTI reduced weekly diary-recorded intrusive-memory frequency in five parallel-group randomized trials of real world trauma. The effect was clinically interpretable and supported by moderate-certainty evidence, downgraded from high for residual risk of bias in the contributing trials. Evidence for broader PTSD symptom severity was low certainty, and evidence for depression, anxiety, and functional outcomes remained less conclusive. ICTI should therefore be considered a promising targeted intervention for intrusive memories rather than an established standalone treatment for PTSD. Larger independent trials with standardized diary and PTSD outcomes are needed before routine clinical implementation.
Statements
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.
Author contributions
MA: Conceptualization, Project administration, Supervision, Writing – original draft. SA: Validation, Visualization, Writing – original draft, Writing – review & editing. IA: Validation, Visualization, Writing – original draft, Writing – review & editing. AAA: Validation, Visualization, Writing – original draft, Writing – review & editing. AS: Investigation, Methodology, Writing – original draft, Writing – review & editing. RY: Formal analysis, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing. MF: Conceptualization, Visualization, Writing – original draft, Writing – review & editing. ANA: Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the Deanship of Graduate Studies and Scientific Research at Qassim University (QU-APC-2026).
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of this manuscript, the authors used Claude (Sonnet 5; Anthropic, San Francisco, CA, USA) for language editing and formatting. The authors reviewed all content for accuracy and take full responsibility for the work.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpsyt.2026.1925949/full#supplementary-material
References
1
American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington, VA: American Psychiatric Publishing (2013). doi: 10.1176/appi.books.9780890425596
2
KesslerRCAguilar-GaxiolaSAlonsoJBenjetCBrometEJCardosoGet al. Trauma and PTSD in the WHO world mental health surveys. Eur J Psychotraumatol. (2017) 8:1353383. doi: 10.1080/20008198.2017.1353383
3
BrewinCRGregoryJDLiptonMBurgessN. Intrusive images in psychological disorders: characteristics, neural mechanisms, and treatment implications. Psychol Rev. (2010) 117:210–32. doi: 10.1037/a0018113
4
HolmesEAJamesELCoode-BateTDeeproseC. Can playing the computer game “Tetris” reduce the build-up of flashbacks for trauma? A proposal from cognitive science. PLoS One. (2009) 4:e4153. doi: 10.1371/journal.pone.0004153
5
LeeCWCuijpersP. A meta-analysis of the contribution of eye movements in processing emotional memories. J Behav Ther Exp Psychiatry. (2013) 44:231–9. doi: 10.1016/j.jbtep.2012.11.001
6
VarmaMMZengSSinghLHolmesEAHuangJChiuMHet al. A systematic review and meta-analysis of experimental methods for modulating intrusive memories following lab-analogue trauma exposure in non-clinical populations. Nat Hum Behav. (2024) 8:1968–87. doi: 10.1038/s41562-024-01956-y
7
AsselbergsJvan BentumJRiperHCuijpersPHolmesEASijbrandijM. A systematic review and meta-analysis of the effect of cognitive interventions to prevent intrusive memories using the trauma film paradigm. J Psychiatr Res. (2023) 159:116–29. doi: 10.1016/j.jpsychires.2023.01.028
8
World Health Organization. World Mental Health Report: Transforming Mental Health for All. Geneva: World Health Organization (2022).
9
KoenenKCRatanatharathornANgLMcLaughlinKABrometEJSteinDJet al. Posttraumatic stress disorder in the World Mental Health Surveys. Psychol Med. (2017) 47:2260–74. doi: 10.1017/S0033291717000708
10
DavisLLScheinJCloutierMGagnon-SanschagrinPMaitlandJUrganusAet al. The economic burden of posttraumatic stress disorder in the United States from a societal perspective. J Clin Psychiatry. (2022) 83:21m14116. doi: 10.4088/JCP.21m14116
11
BissonJIRobertsNPAndrewMCooperRLewisC. Psychological therapies for chronic post-traumatic stress disorder in adults. Cochrane Database Syst Rev. (2013) 2013:CD003388. doi: 10.1002/14651858.CD003388.pub4
12
National Institute for Health and Care Excellence. Post-Traumatic Stress Disorder. NICE Guideline NG116. London: NICE (2018).
13
RoseSBissonJChurchillRWesselyS. Psychological debriefing for preventing post traumatic stress disorder. Cochrane Database Syst Rev. (2002) 2002:CD000560. doi: 10.1002/14651858.CD000560
14
ImelZELaskaKJakupcakMSimpsonTL. Meta-analysis of dropout in treatments for posttraumatic stress disorder. J Consult Clin Psychol. (2013) 81:394–404. doi: 10.1037/a0031474
15
Astill WrightLHorstmannLHolmesEABissonJI. Consolidation/reconsolidation therapies for the prevention and treatment of PTSD and re-experiencing: a systematic review and meta-analysis. Transl Psychiatry. (2021) 11:453. doi: 10.1038/s41398-021-01570-w
16
DeforgesCSandozVNoëlYAvignonVDesseauveDBourdinJet al. Single-session visuospatial task procedure to prevent childbirth-related posttraumatic stress disorder: a multicentre double-blind randomised controlled trial. Mol Psychiatry. (2023) 28:3842–50. doi: 10.1038/s41380-023-02275-w
17
IyaduraiLHighfieldJKanstrupMMarkhamARamineniVGuoBet al. Reducing intrusive memories after trauma via an imagery-competing task intervention in COVID-19 intensive care staff: a randomised controlled trial. Transl Psychiatry. (2023) 13:290. doi: 10.1038/s41398-023-02578-0
18
RamineniVMillrothPIyaduraiLJakiTKingslakeJHighfieldJet al. Treating intrusive memories after trauma in healthcare workers: a Bayesian adaptive randomized trial developing an imagery-competing task intervention. Mol Psychiatry. (2023) 28:2985–94. doi: 10.1038/s41380-023-02062-7
19
KanstrupMSinghLLeehrEJGöranssonKEPihlgrenSAIyaduraiLet al. A guided single session intervention to reduce intrusive memories of work-related trauma: a randomised controlled trial with healthcare workers in the COVID-19 pandemic. BMC Med. (2024) 22:403. doi: 10.1186/s12916-024-03569-8
20
KehyayanAThielJPUnterbergKSaljaVMeyer-WehrmannSHolmesEAet al. The effect of a visuospatial interference intervention on posttraumatic intrusions: a cross-over randomized controlled trial. Eur J Psychotraumatol. (2024) 15:2331402. doi: 10.1080/20008066.2024.2331402
21
BeckenstromACBonsallMBMarkhamASladeORamineniVIyaduraiLet al. A digital imagery-competing task intervention for stopping intrusive memories in trauma-exposed health-care staff during the COVID-19 pandemic in the UK: a Bayesian adaptive randomised clinical trial. Lancet Psychiatry. (2026) 13:233–47. doi: 10.1016/S2215-0366(25)00397-9
22
PageMJMcKenzieJEBossuytPMBoutronIHoffmannTCMulrowCDet al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. (2021) 372:n71. doi: 10.1136/bmj.n71
23
MoherDShamseerLClarkeMGhersiDLiberatiAPetticrewMet al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev. (2015) 4:1. doi: 10.1186/2046-4053-4-1
24
OuzzaniMHammadyHFedorowiczZElmagarmidA. Rayyan—a web and mobile app for systematic reviews. Syst Rev. (2016) 5:210. doi: 10.1186/s13643-016-0384-4
25
SterneJACSavovićJPageMJElbersRGBlencoweNSBoutronIet al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. (2019) 366:l4898. doi: 10.1136/bmj.l4898
26
SterneJACHernánMAReevesBCSavovićJBerkmanNDViswanathanMet al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. (2016) 355:i4919. doi: 10.1136/bmj.i4919
27
DerSimonianRLairdN. Meta-analysis in clinical trials. Control Clin Trials. (1986) 7:177–88. doi: 10.1016/0197-2456(86)90046-2
28
WallaceBCDahabrehIJTrikalinosTALauJTrowPSchmidCH. Closing the gap between methodologists and end-users: R as a computational back-end. J Stat Softw. (2012) 49:1–15. doi: 10.18637/jss.v049.i05
29
HigginsJPTThompsonSGDeeksJJAltmanDG. Measuring inconsistency in meta-analyses. BMJ. (2003) 327:557–60. doi: 10.1136/bmj.327.7414.557
30
EggerMDavey SmithGSchneiderMMinderC. Bias in meta-analysis detected by a simple, graphical test. BMJ. (1997) 315:629–34. doi: 10.1136/bmj.315.7109.629
31
HigginsJPTThomasJChandlerJCumpstonMLiTPageMJet al eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 6.5. Cochrane (2024).
32
GuyattGHOxmanADVistGEKunzRFalck-YtterYAlonso-CoelloPet al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. (2008) 336:924–6. doi: 10.1136/bmj.39489.470347.AD
33
HardarsonJPGambleBThorarinsdottirKStephensenESKanstrupMGudmundssonTet al. Developing a brief cognitive task intervention to reduce long-standing intrusive memories of trauma: a feasibility study with remote delivery for women in Iceland. Clin Psychol Eur. (2024) 6:e11237. doi: 10.32872/cpe.11237
34
DeforgesCFortDStuijfzandSHolmesEAHorschA. Reducing childbirth-related intrusive memories and PTSD symptoms via a single-session behavioural intervention including a visuospatial task: a proof-of-principle study. J Affect Disord. (2022) 303:64–73. doi: 10.1016/j.jad.2022.01.108
35
BadawiASteelZMahoneyCBerleD. Feasibility of an adjunctive cognitive task in the treatment of posttraumatic stress disorder. Discov Psychol. (2021) 1:11. doi: 10.1007/s44202-021-00011-5
36
KanstrupMSinghLGöranssonKEWidoffJTaylorRSGambleBet al. Reducing intrusive memories after trauma via a brief cognitive task intervention in the hospital emergency department: an exploratory pilot randomised controlled trial. Transl Psychiatry. (2021) 11:30. doi: 10.1038/s41398-020-01124-6
37
FreedmanSAEitanRWeinigerCF. Interrupting traumatic memories in the emergency department: a randomized controlled pilot study. Eur J Psychotraumatol. (2020) 11:1750170. doi: 10.1080/20008198.2020.1750170
38
IyaduraiLBlackwellSEMeiser-StedmanRWatsonPCBonsallMBGeddesJRet al. Preventing intrusive memories after trauma via a brief intervention involving Tetris computer game play in the emergency department: a proof-of-concept randomized controlled trial. Mol Psychiatry. (2018) 23:674–82. doi: 10.1038/mp.2017.23
39
HorschAVialYFavrodCHarariMMBlackwellSEWatsonPet al. Reducing intrusive traumatic memories after emergency caesarean section: a proof-of-principle randomized controlled study. Behav Res Ther. (2017) 94:36–47. doi: 10.1016/j.brat.2017.03.018
40
HolmesEAGhaderiAErikssonEOlofsdotter LauriKKukackaOMMamishMet al. I can’t concentrate”: a feasibility study with young refugees in Sweden on developing science-driven interventions for intrusive memories related to trauma. Behav Cognit Psychother. (2017) 45:97–109. doi: 10.1017/S135246581600062X
41
IyaduraiLVisserRMLau-ZhuAPorcheretKHorschAHolmesEAet al. Intrusive memories of trauma: a target for research bridging cognitive science and its clinical application. Clin Psychol Rev. (2019) 69:67–82. doi: 10.1016/j.cpr.2018.08.005
42
BrewinCRHolmesEA. Psychological theories of posttraumatic stress disorder. Clin Psychol Rev. (2003) 23:339–76. doi: 10.1016/S0272-7358(03)00033-3
43
EhlersAClarkDM. A cognitive model of posttraumatic stress disorder. Behav Res Ther. (2000) 38:319–45. doi: 10.1016/S0005-7967(99)00123-0
44
JamesELBonsallMBHoppittLTunbridgeEMGeddesJRMiltonALet al. Computer game play reduces intrusive memories of experimental trauma via reconsolidation-update mechanisms. Psychol Sci. (2015) 26:1201–15. doi: 10.1177/0956797615583071
45
SinghLAhmed PihlgrenSHolmesEAMouldsML. Using a daily diary for monitoring intrusive memories of trauma: a translational data synthesis study exploring convergent validity. Int J Methods Psychiatr Res. (2023) 32:e1936. doi: 10.1002/mpr.1936
46
Department of Veterans Affairs, Department of Defense. VA/Dod Clinical Practice Guideline for the Management of Posttraumatic Stress Disorder and Acute Stress Disorder. Washington, DC: Department of Veterans Affairs and Department of Defense (2023).
47
BissonJIBerlinerLCloitreMForbesDJensenTKLewisCet al. The International Society for Traumatic Stress Studies new guidelines for the prevention and treatment of posttraumatic stress disorder: methodology and development process. J Trauma Stress. (2019) 32:475–83. doi: 10.1002/jts.22421
Keywords
imagery-competing task intervention, intrusive memories, memory reactivation, meta-analysis, post-traumatic stress disorder, systematic review, Tetris
Citation
Allahim MA, Alsallal SA, Almohaimeed IA, Alfouzan AA, Alrashed AS, Abualjamal RY, Alharbi MF and Alqifari AN (2026) Visuospatial imagery-competing task interventions for reducing intrusive memories following real-world trauma: a systematic review and meta-analysis. Front. Psychiatry 17:1925949. doi: 10.3389/fpsyt.2026.1925949
Received
02 July 2026
Revised
29 August 2026
Accepted
23 September 2026
Published
13 October 2026
Volume
17 - 2026
Updates
Copyright
© 2026 Allahim, Alsallal, Almohaimeed, Alfouzan, Alrashed, Abualjamal, Alharbi and Alqifari.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Moayed A. Allahim, moayedallahem@gmail.com
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.
来源：Frontiers in Psychiatry · frontiersin.org
猜你喜欢
- 基于语音识别MDD患者自杀风险：模型开发与4周随访验证Frontiers in Psychiatry · 1 小时前
- 混合方法研究：澳大利亚四家急诊科员工面对职场暴力的4Fs应激反应Frontiers in Psychiatry · 7 天前
- Frontiers in Psychiatry 发表氯胺酮精神病学应用系统综述Frontiers in Psychiatry · 12 天前
- Frontiers in Psychology研究：AI伴侣拟人化感知与问题性情感依赖的关联Frontiers in Psychology · 2 小时前
- Frontiers in Psychiatry 网络元分析：传统中式健身功法对大学生心理与体质的比较效果Frontiers in Psychiatry · 4 天前