Is it possible to reconstruct continuous data from _TIM.mat epochs? (Waveform discrepancy in overlapping windows)
tao g
Dear HCP MEG Team,
I am planning to run a continuous dynamic network model (TDE-HMM) on the MEG Working Memory task, and I would like to clarify if it is mathematically possible to reconstruct a continuous timeline by concatenating the epoched _TIM.mat files.
Based on the trialinfo trigger samples, consecutive trials are exactly 2.5 seconds apart. Since each epoch in _TIM.mat is 4.0 seconds long ([-1.5s, 2.5s]), there should be a 1.5-second physical overlap between adjacent trials.
根据试验信息的触发样本来看，连续的试验之间恰好间隔 2.5 秒。由于_TIM.mat 中的每个时间段持续 4.0 秒（[-1.5 秒, 2.5 秒]），因此相邻试验之间应该存在 1.5 秒的物理重叠时间。
However, when I compare these physically overlapping segments (e.g., the 1.0s - 2.5s tail of Trial $N$ versus the -1.5s - 0s head of Trial $N+1$), the waveforms do not perfectly match (Pearson $r \approx 0.71$). Attempting to crop and concatenate them creates clear step artifacts at the boundaries.
然而，当我比较这些物理上重叠的片段时（例如，试验$N$的 1.0 秒到 2.5 秒区间与试验$N+1$的-1.5 秒到 0 秒区间），这两个波形并不完全匹配（Pearson 相关系数约为 0.71）。尝试对它们进行裁剪和拼接后，在边界处出现了明显的阶梯状伪影。
My questions are: 我的问题如下：
Does the tmegpreproc pipeline apply trial-wise preprocessing (such as independent detrending or demeaning) after the continuous data is epoched, thereby breaking the absolute amplitude continuity?
tmegpreproc 流程是否在处理连续数据后，会进行试验性的预处理操作（如独立去趋势处理或均值化处理），从而打破绝对幅度的连续性？For models requiring strict continuous time series, is the official recommendation to avoid concatenating _TIM.mat and instead start from the continuous preprocessed data (_c,rfDC.mat)?
对于那些需要严格连续时间序列数据的模型来说，官方建议是避免合并_TIM.mat 文件，而是直接从预处理后的连续数据(_c,rfDC.mat)开始处理。
Thank you for your time and guidance!
感谢您的时间与指导！
Best regards, 最诚挚的问候，
tao guo 陶国
Elam, Jennifer
Sent: Friday, June 5, 2026 9:30 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Subject: [hcp-users] Is it possible to reconstruct continuous data from _TIM.mat epochs? (Waveform discrepancy in overlapping windows)
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/e03bdb8f-3185-413d-8e97-6515dca833f6n%40humanconnectome.org.
The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.