以樂高程式設計介入高齡者認知健康

Abstract

認知訓練被視為具有延緩年齡相關認知衰退的潛力。然而，傳統的認知訓練主要著重在引導受試者重複練習低階認知歷程，例如注意力、記憶或抑制任務。在此臨床註冊試驗（編號 NCT05341232）中，我們欲探究針對高階認知歷 程「主動推理」對高齡者認知健康的影響。本研究透過為期12週的樂高機器人 程式課（LRPI），介入主動推理，此歷程涉及整合多元訊息來執行觀察、預測和行動的試誤過程，其中更需要協調統合各個低階認知過程。我們假設針對主動推理介入，將可能導致神經的整體活化和未特別訓練的認知功能改善。本研究中，參與者被隨機分配到實驗組或主動對照組，並使用視覺規則推理任務 （VRIT）作為功能性磁振造影的任務，以調查在推理過程中的行為和大腦活動變化。同時，前後測尚包含神經心理學測驗，用以評估實驗中未特別訓練的認知功能。迄今為止，每組8名參與者已完成介入，初步數據顯示，實驗組的訓練可能在老年人的推理能力產生潛在的效果，並且對行為和神經層面均有改善。然而，由於樣本數較小，需要進一步的研究以充分評估訓練效果。

Cognitive training has been proposed as a promising approach to address age- related cognitive decline. However, past studies have primarily focused on repetitive engagement of lower-level cognitive processes, such as attention, memory, or inhibitory tasks. To address this gap, we conducted a 12-week Lego Robot Programming Intervention (LRPI) targeting higher-level cognitive processes, specifically active inference, in a registered clinical trial (no. NCT05341232). Active inference involves integrating multiple sources of information to perform a trial-and- error process of observation, prediction, and action, and requires the coordination of lower-level cognitive processes. We hypothesized that targeting active inference would lead to general neural activation and improvements in untrained cognitive functions. Participants were randomly assigned to the experimental or active control group, and the Visual Rule Inference Task (VRIT) was used during fMRI scanning to investigate changes in brain activity and behaviors during inference processing. Pre- and post- intervention neuropsychological tests were also conducted to assess improvement in untargeted cognitive functions. To date, 8 participants in each group have completed the intervention, and preliminary data suggests that our experimental training may have promising effect on inference processing in older adults at both behavioral and neural levels. However, due to the small sample size, further research is needed to adequately evaluate the training effects.