Continuous quasi-periodic transitions between the default mode network (DMN) and the task positive network (TPN) are known to occur in the human brain and have been shown to correlate with infraslow electrical activity^1–3. Considering the functional significance of the networks comprising of these quasi-periodic patterns (QPPs)⁴, the QPPs themselves could be serving an important functional role. However, the algorithm used for detection of these patterns is dependent on user-input, the most influential of which is the specified length of the QPP. Here we investigate the effect on observed QPP templates with varying window lengths of QPPs. 8 resting state MRI scans were collected and preprocessed according to Thompson et al.⁴ and masks for DMN and TPN were obtained. Functional data was concatenated into one timeseries and the QPP-finding algorithm described in Majeed et al.² was applied repeatedly for different QPP window lengths with the same starting point. The resulting QPP templates’ correlation strength with the functional timeseries was measured for 20 window lengths ranging from 3.6 to 72 seconds. First, we compared the relative DMN and TPN activation seen within the different QPP templates observed. Next, the average correlation strength of each QPP template observed was compared to the window length assigned for that template. Finally, we compared sliding-window correlation vectors of the QPP with the functional data for the different window lengths used.Observing DMN and TPN activation over the course of QPPs with different lengths shows that an ideal QPP template depicting an alternation between DMN and TPN occurring at a ~20 second window length (Figure 1c), though similar alternation is present in most longer window lengths. Figure 2 shows that there is an overall negative relationship between the window length and correlation strength, with correlation decreasing as window length increases. This result is further evident by observing the correlation between QPP and the concatenated time courses from all subjects for different window lengths, as Figure 3 shows. There is decreasing correlation magnitude with increasing window length.Corresponding propagation of DMN and TPN signals between QPP templates of different lengths in Figure 1 shows that a similar pattern is identified by the QPP algorithm regardless of window length used. For longer window lengths (> 15 seconds), there is almost always a transition between DMN and TPN domination. This transition is most isolated from other random events at a ~20 second template. Though longer window lengths show a network switch, occurrence of random events around the network switch may be responsible for reduced QPP correlation with functional scan. Figure 2 shows that the correlation strength depicts an overall decrease as the window length increases. However, a plateau in correlation strength is observed around the 18–21 second range, which further indicates that this might be an ideal window length to observe a complete network transition while obtaining high correlation. Strong peaks seen in the longest window lengths in Figure 3 may be occurring due to close proximity to defined static starting point for QPP algorithm, which could mean that a single subject is driving the QPP template. This can be addressed by obtaining QPP templates using multiple starting points and using hierarchical clustering to obtain a uniform QPP template representative of the entire concatenated matrix. Overall reduction in correlation vector amplitude reflects the negative relationship of window length and correlation.Our results help further understanding of QPP properties, the estimated time frame of QPP occurrence in humans, and how they can be adequately identified in functional data. Windows of less than 20s typically do not contain a complete pattern; longer windows tend to contain repeated patterns, demonstrating that the pattern itself is relatively invariant to the window length chosen. Previous work looking at DMN and TPN dynamics has shown that relative activation of these networks can influence attention levels and task performance in humans⁴. QPPs may be playing a role in the underlying organization of these networks and could themselves play a role in attention and task performance.