In analyses based on Statistical Parametric Mapping (SPM), we sometimes encounter situations where results become statistically significant within a specific time window of the analysis. However, it is crucial to note that while the assumption of data normality holds true within this specific window, it is violated in another segment where neither parametric nor non-parametric tests demonstrate a significant difference (attached figure).

In such circumstances, the fundamental question arises: which category of tests should we employ to report the portion of the results that exhibit both normality and statistical significance? Should we opt for parametric or non-parametric tests?

In other words, should we sacrifice statistical significance due to the lack of normality across the entire time window and utilize non-parametric tests? Or, considering the normality of the data within the significant window, can we leverage parametric tests?